.u A ' ‘ , ~ ‘ . -.-_-. . ~.‘H;l.‘.,. J .- __~ . ,‘ ‘ A u I .1. . H.“ .. '.-,.. ~ - . .- . I f d4. \.‘n1.1“"I‘i.‘\§x.u\vl_\"} ,‘ _ , ,.,., ‘ ,,l ‘ , l A ‘ , , ‘ H .Hu. . II '1‘. _. I ..,A, . . I . V lulu Mu . I v _ .‘ . . FOOD DISTRIBUTION IN A UI‘I'IN‘AMERICAN CITY (Cali, CoIombia) Thesis for the Degree cf Ph. D. MICHIGAN STATE UNIVERSI‘IY COLIN B. GUTHRIE 19'32 LIBRARY Michigan Sta” University 21 3 ~ , IIIII thesis entitled Fm)? DISIRIBuT/no 1w KI “(W {iMEIQ'C-M CITY ' I I I This is to certify that the I I I I (CALI/ COLOM5¢ R) presented by Comm/D)- éUWflIE has been accepted towards fulfillment of the requirements for 1)” - D degree i113 USI “r E55 A?) M I [(1 Major professor Date IL M [6‘11 ABSTRACT Food Distribution in a Latin-American City by Colin B. Guthrie This dissertation is a study in a rather Specialized area in the field of economic development, being a sub-topic of the very important subject of the integration of urban and rural economies of a society in an intermediate stage of economic develOpment. It is a study dealing with fbod distribution in the city of Cali, Colombia, the political and economic capital of the Departamento del Valle del Cauca. the Cauca Valley Department or Province. The purpose of the study was to evaluate the perfbrmance of the fbod distribution system of the city and to make recommendations that would tend towards remedying any deficiencies that might be detected. This study uses direct observation and primary data. Preliminary observation and interviews suggested global hypotheses which in turn defined specific hypotheses. Surveys were mounted to collect data to test the specific hypotheses. and analysis of these data led to the conclusions and recommendations described in the dissertation . SI 2 -- Colin B. Guthrie The basis of the study is an examination of the fOod-consumption habits of the papulation of Cali, and of the purchasing habits and preferences of the Cali housewife, the food-purchasing agent. Food consumption was investigated by means of a market basket survey, while preferences were deduced by comparing attitudes towards types of fOOd retailers with purchasing patterns. This indicated that locational convenience (store accessibility) was the factor which dominated retailer selection. Next in importance were quality and variety of store offerings, with prices in fourth place. The relatively low importance of prices showed that if low prices were to be made available to the consumer in Cali this would have to be done by locating food stores in the residential areas, because consumers were not prepared to travel to where low-price stores were located. Other consumer data showed that half the families in Cali spent two-thirds or more of their income on fbod, while two-thirds spent half or more of their income in this way. However. the Lorenz curve fbr the city was so deep that only about forty percent of the city's income was Spent upon foodstuffs. Study of the performance of the principal fbod-distribution institutions, the wholesalers and retailers, indicated that they did not take excess profits (some meat dealers constituted an exception to this generalization). nor were there any internal inefficiencies that would suggest they incurred avoidable costs. Therefore lower food prices could not be obtained by putting pressure upon individual . 3 -- Colin B. Guthrie operators. Improved performance could only be obtained by better adaptation of institutions to the requirements of other components of the system. It was found that the existing system was well-adapted to food distribution based upon public markets. Market stalls, which were narrowly specialised by product, were well served by the wholesalers who were also product-Specialised. But Cali is now too large for the consumer to be well served by a system of this type. Transportation costs of consumers who bought food outside their own neighborhoods (principally in the public markets) were equivalent to about 3% of total food expenditures. Also, it could be shown that food stores were potentially more efficient than market stalls in the sense of having lower unit transaction costs and observation suggested that a public market large enough to be viable would need more customers than could be fbund living within reasonable walking distance. That is, public markets could not provide the locational convenience required by the housewife. It followed that an improved food distribution system would have to be based upon fbod retailers located in the residential areas. Data from the market basket survey showed that if the service areas of these outlets approximated reasonable walking distance (three city blocks) there would be sufficient patronage for each store for it to achieve substantial economies of scale. This implies a new type of food retailer. since there are at present in Cali no large food stores located in middle-income or low-income neighborhoods. 4 -- Colin B. Guthrie These stores would have to be supplied by full-service full- line wholesalers. which do not at present exist in Cali. Therefore wholesalers of this nature will have to be developed. Such wholesalers. however. cannot replace the existing product-Specialized wholesalers whose primary marketing function was deduced to be that of establishing market-clearing prices and thus compensating for the absence of commodity exchanges in Colombia. Although the pr0posed system has one more level of distribution than the existing system its aggregate Operating costs should be lower, and locational analysis suggests that competitive pressures will be strong enough to ensure that these savings will be passed along to the consumer. A surprising result of the analysis is that social dislocation attributable to the new system will be negligible, as far as throwing pepple out of work is concerned. While the new system will employ fewer peOple than the existing one, normal attrition in the existing system is far greater than would be the displacement caused by any foreseeable rate of introduction of the new system. FOOD DISTRIBUTION in a LATIN-AMERICAN CITY (Cali, Colombia) by 0', Colin 3? Guthrie A,THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Marketing and Transportation 1972 n ”A: N o4 v- : _:. "3 l l i I I“ db ACKNOWLEDGEMENTS It is gratifying to have this Opportunity to extend my thanks to those without whose help this work would never have been begun, or, if begun, finished. Dr. Donald A. Taylor made it possible for me to join the research team, and his experience was an invaluable asset to PIMUR at the research design stage of the project. Dr. Donald S. Henley provided invaluable help at the evaluation stage of the project, and I am deeply indebted for the help he has given me as chairman of my thesis committee. Nelson Suarez and Kelly Harrison, co-leaders of the project. were never-failing in their resourcefulness, finding ways round every obstacle. To put into the consumer surveys the resources that were demanded required considerable trust in the research design, and I am grateful to them for giving me that trust. Perhaps the biggest single contribution was that made by the insight, intelligence, and outstanding organizational skill of Guillermo Nolta. To Guillermo must go most of the credit for any success achieved by the consumer surveys. I am also grateful for the contributions of Ruben Cruz Cruz and Miguel Arce, contributions limited only by the scope offered to the contributors and not by their ability or enthusiasm. A great part of the present dissertation depends upon the work done by David Lloyd-Clare and Hugo Deque in the research upon retailing and wholesaling. My indebtedness to them is both inmense and obvious. Finally, I want to record my gratitude to my wife.Kay, whose contributions to this dissertation have been all-pervasive, and but for ii iii whom the dissertation would probably not have been started and certainly would never have been finished. Although this dissertation could not have been written without the help acknowledged herein, it goes without saying that any errors are solely my responsibility. ACKNOWLEDGEIENTS ..... . .................................. LIST OF TABLES ................................................... LIST OF FIGURES ... ............................................... Chapter I. INTRODUCTION ..... ...................................... Statement of Purpose .................................. Utility of Study ...................................... Structure of Research ...... ............................ General Framework . .......... .. ............ .. .......... Specific Issues ... ...... ..... ........... .............. Social Justification ..... ..... ............... . ........ The Setting - Cali Colombia ...... ....... . ........ . ..... Historical and Geographical .... ..... ............... 2. Cali: Tapographical .. ...... ..... ..... ............. 3. Food Retailers ......................... ..... ..... 4. Facilitating and Ancillary Institutions ............ Sources of Finance ....................... ......... Transportation .......... ........... ............... Training Facilities ........................ ....... Packaging ...... ................................. Government Intervention in Food Distribution ...... 5.5wmmy. ..H.u.u.u.u.u.u ............... Review of the Literature .. ..... . ...... ............ Outline of theDissertation ........................ 1. II. THE CONSUMER .......................................... 1. Introduction ...... ..... .... ......... . ............. A Note Upon Classification Criteria ............... 2. Demographic Characteristics .............. . ......... Education .... ..................................... Households . .......... . ............................ Age ....... ........................................ TABLE OF CONTENTS iv Page ii vii X and Chapter Page 3. Employment and Income Distribution ... .............. 61 4. Food Expenditures and Income ............ . .......... 69 Source of Data ...... ............................ 69 A Note upon the Mathematics of Regression Analysis. 74 5. Expenditures Upon Selected Food Products ........... 75 6. Transportation COStS 0.0 I. .00... O. O. O... .0 O. I. .00 .00 83 Summary ........ ........................................ 85 III. CONSUMER-RETAILER INTERFACE ..... ...................... 87 1. Store Patronization Patterns ....................... 97 2. Shopping Frequency ................................. 106 3. ShOpping TranSportation Costs ................ . ..... 111 4. Miscellaneous Services ...................... .... 113 5. Factors Affecting Choice of Retailer ............... 116 6. Conclusions ........................................ 128 IV THE URBAN FOOD DISTRIBUTION SYSTEM ...... . ............. 129 Introduction .......................................... 129 Retailers ..... ............... . ................. . ...... 133 Minor Retailer Types ............ . ........ ............ 133 Non-Profit Oriented Retailers ........................ 135 Non-Systemic Retailers ......... ........ ......... ..... 144 Self-Service Stores .................................. 147 Chain Stores .......................................u. 150 Public Markets ....................................... 151 Centrally Located Traditional Stores ....... .......... 158 Non- -Centra11y Located Traditional Stores ............. 162 Meat Distribution ....... ....... .................. 166 Retailers - Evaluation of Perfbrmance ................. 172 The Wholesaler - Retailer Interface ................... 178 The Wholesale Sector .................................. 183 Wholesalers of Grains and Processed Staples .......... 184 WholeSaler-Retailers ..... . ........... ..... ..... ..... 191 Fruits and Vegetables Wholesalers .. .................. 192 Wholesale Distribution of Branded Foods .............. 196 Summary ...;..... ............ ............ .............. 197 V CONCLUSIONstNO RECOMMENDATIONS ....................... 199 Evaluation of Existing Food Distribution Institutions . 202 An Improved Food Distribution System .................. 208 Locational Analysis of Proposed Retailers ....... ...... 210 Economic Analysis of Proposed System Components ....... 217 Organization ...... ................................... 224 Social Consequences .......... ... ..................... 230 The Contribution of Government ..... ................... 233 Evolution of the Pr0poses System ...................... 236 Sumary . . . . ........................................... 339 vi Chapter Page Appendix 1 POPULATION ESTIMATES O 0 00000 O O O O 0000000 O 000000000000000 242 II CONSUMER SAMPLING AND SURVEY METHODOLOGY .......... .... 248 III EXPERIENCE WITH THE ATTITUDE-MEASURING SECTION OF THE QUESTIONNAIRE ........................................ 254 IV COMPARISON OF MARKET STALLS AND STORES ................ ‘ 259 BIBLIOGRAPHY ..................................................... 263 Table I II. II II II II. II. II II II II II II II .1.1. .2.1. 1.1. .1.2. .2.1. .2.2. 2.3. 3.1. .3.2. .4.1. .5.1. .5.2. .5.3. .5.4. .6.1. LIST OF TABLES Comparison of the Percentage Contributions of Selected Regions to Total National Manufacturing Activity in Colombia, 1966 ..................... Distribution of Population by Socio-Economic Level .......................................... Distribution of Per-Capita Income with each Socio- Economic Level ................................. Population and Households by Socio-Economic Level and Per-Capita Level ........................... Distribution of Education Levels ............... Family Position of Most Educated Member ........ Household Size as a Function of Socio-Economic -Level and Per-Capita Level ..................... Mean and Median Per-Capita Income for Each Per- Capita Income Level ............................ Income by Socio-Economic Level ................. Per-Capita Incomes and Food Expenditures for Four Per-Capita Income Levels ....................... Monthly Per-Capita Food Expenditures by Per-Capita Income Levels .................................. Monthly Family Food Expenditures by Socio-Economic Leve .......................................... Income-Elasticities of Demand for Food Products, by Per-Capita Income Levels .................... Projections of Monthly Consumption of Selected Food Products in Cali for 1969, 1979 and 1989 .. Eonsumer Transportation Costs by Socio-Economic evel ........ . ................................. vii Page 22 27 54 54 56 57 59 67 67 71 77 78 79 83 .u. u‘. I”. Table 111.1. . 111.1. . 111.1. . 111.2. . 111.2. . 111.2. . 111.3. 111.4. . 111.5. . 111.5. . 1V.1 IV.2. IV.3. 1V.4. 1V.5. IV.6. 1V.7. viii Percent of Families in Each Socio-Economic Level Patronizing Different Types of Retail Stores ..... ................................... Percent of Food Expenditures in Each Socio- Economic Level Spent in Different Types of Retail Store .................................. Percentage of Families Interviewed Patronizing Principal Outlet Types, by Product and Socio- Economic Level ................................ Percent of Consumers making Shopping Trips .... Distribution of ShOpping Trips and Sales Volume by Day of the Week ............................ Average No. of Shopping Trips Per Family for Selected Retailer Types ....................... Transportation Costs of Non-Local Shopping Per Family Per Week by Socio-Economic Level ....... Use of Consumer Credit in Food Purchasing ..... Relative Importances of Detailer Attributes, by Method of Estimation and Product Group ........ Indices of Attribute Importances .............. Patronization of Caja Grocery Stores, by Socio- Economic Level ................................ Retail Distribution of Raw and Pasteurized Milk .......................................... Distribution of Market Stalls by Speciality and Market ....................................... Characteristics of Large Graneros and Wholesaler- Retailer ...................................... Characteristics of Operators of Large Graneros and Wholesaler-Retailer ....................... Characteristics of Small Graneros and Tiendas . Characteristics of Operators of Small Graneros and Tiendas ................................... Page 97 97 99 106 108 109 112 114 122 123 139 145 154 159 161 163 165 Table IV.8. 1V.9. IV.10. IV.11. IV.12. IV.13. 1V.14. 1V.15. 1V.16. V.1. v.2. V.3. v.4. V.5. A.11.1. ix Average Margins Taken by Meat Dealers ......... Relative Importance of the Main Retailer Types in the Distribution of the Principal Foodstuffs Retailers' Gross Margins by Outlet Type ....... Financial Comparison of Outlet Types .......... Relative Prices of Foodstuffs by Retailer Type Percent of Retailers Purchasing from Regular Suppliers Extent and Terms of Credit Received by Retailers Margins needed and obtained by Wholesalers ..... General Characteristics of Wholesalers of Fruits and Vegetables ................................. Distribution of Vendors in Satellite Markets by Product Type ................................... Potential Sales of Retailers Serving l8-Block Areas, by Socio-Economic Level of Location ..... Projected Income Statements of Proposed Retailers, at Varying Levels of Monthly Sales ............. Projected Investment Requirements of Proposed Retailers, at Varying Levels of Monthly Sales .. Projected Income Statement of Proposed Wholesaler Comparison of Selected Characteristics of Partic- ipants in the PIMUR Consumer Survey and Market Basket Survey .................................. Page 169 172 174 175 176 179 180 185 192 204 213 218 218 222 251 LIST OF FIGURES Figure Page 1.1. Map of Colombia ......................................... 18 1.2. The Department of Valle and the Geographic Location of the Cauca Valley ............................ 20 1.3. Socio-economic levels of barrios and locations of public markets, Cali, 1969 ........................... 26 11.3.1. Lorenz Curve, Cali, 1969 ................................ 63 II.3.2. Distribution of family income, Cali, 1969 ............... 64 II.3.3. Distribution of per-capita income, Cali, 1969 ........... 65 II.4.1. Food Expenditure Versus Income, Cali, 1969 .............. 72 V.1. Illustrating Service Area of Proposed Retailers ......... 212 I. :6 ‘1 l .4 ‘- CHAPTER I Introduction Statement of Purpose In September 1968 a research team made up of people from Michigan State University and from various Colombian institutions began a study of the food distribution system in the city of Cali, Colombia. The purpose of the study was: .IaIIto designwa food distribution system for Cali which would make ‘L:_‘food available to inhabitants, particularly to those in middle- . ’iaeoae and low-income categories at lower prices than the existing system::and to make recommendations to the Colombian authorities concerninggcourses of action which would lead to the realization of such an improved system, -'E):£6 trace the effects upon the economy of the region of a hypo- thetical reductionufhgfood prices. The research fiésm, working under the acronym of PIMUR, finished its studies in late 1969, and its final report is now available in English1 as well as Spanish. The objective of the present dissertation is to present a marketing analysis of the data collected 1Market Coordination in theQevelopment of the Cagca Valley Re ion--Colombia, Research Report No. 5, Latin American Studies Center, Michigan State University, East Lansing, 1970. by PIMUR in the area of urban food distribution, in order to complement the action-oriented PIMUR final report. We shall try to show how marketing principles and concepts were used in structuring the research, in analyzing the data collected, and in designing an improved food distribution system. In so doing, we hope to illustrate the usefulness of these principles and concepts in understanding economic systems. especially economic systems in underdeveloped economies. In particular, I we hope to show that marketing principles and concepts are powerful I enough (or fundamental enough) to be transferred without modification from the context of the developed U. 5. economy to that of underdeveloped economies. While marketing techniques and practices should be tailored Ffl“__,,ii.WLL to fit each society, the principles and concepts underlying them are the same everywhere. Utility of Study, It is hoped that this study will contribute to the under- standing of the marketing process in a particular ambience and, by extension, marketing processes in general. Without such an under- standing marketing effort may be poorly directed and marketing efficiency impaired; in particular, legislation and other governmental action affecting the market can be ill-aimed and confusing, as has happened in the past. If this study should cast any light upon these processes its utility will be amply established. Unlike the PIMUR report this dissertation is not directly \flfluw- __ HM.~—.ww intended to foment marketing reforms aimed at reducing food prices. m N'W” ‘3 MW But if it can provide any help or guidance to people engaged in research of the PIMUR type then again its utility is not in question. For more than a quarter of the population of Cali the diet is nutritionally inadequate. and there is no reason to believe that the same is not true for most South American urban populations‘. If this study can be of assistance to anyone working to help ameliorate this inadequacy then it will serve a useful purpose. Malnutrition is not an economic good. In addition to improved nutrition the reduction of food prices can have consequences of interest to the development economist and the economic planner. Persistent underdevelopment is a complex phenomenon, and it is not proposed to enter into a detailed discussion of it here, but one of the most obvious aspects of it is a very high rate of urban unemployment. This in turn leads to a low average income (low in effect as well as in statistical averaging, as a result of the extended house- hold culture which exists in Latin America). Low average income means that after providing for food and shelter there is very little effective demand for manufactured goods. Since the manufacture of such goods accounts directly or indirectly for the vast bulk of urban employment this weak demand serves to perpetuate the low-employment conditions. If demand could be increased for low-priced manufactured goods, manufactured goods whose production function is linear in labor (with a high constant of proportionality), this would, subject to certain conditions of entrepreneurial readiness to respond, give rise to increased employment. The resulting increase in income would in turn further increase demand for the products of the factories and at least maintain the increased level of employment or even promote further increases. 1Henley found some evidence of this in La Paz, Bolivia Market Processes in La Paz, Bolivia, Research Report No. 3, Latin- r can tu es enter, Michigan State University, East Lansing A . I?‘ 1:” I... .15 Several conditions have to be simultaneously satisfied for this happy state of affairs to be realized. Entrepreneurs have to be alert to the opportunities available. and both willing and able to respond. This implies a certain willingness to take risks, access to financing. and ready availability of increased supplies of raw materials. The factories must be located in the community in which the increased demand is to be generated, or the feedback linkage will be broken, and the goods to be demanded must have a high labor content or the feedback linkage will still be disrupted. In all these, the economic planners can contribute to setting the scene, by judiciously combing sub- sidization, taxation and allocation of raw materials and credit, and by providing timely reports of economic conditions. The most ticklish problem in this program is stimulating demand. Merely pumping money into the economy is unlikely to be fruitful, ' because it leads to inflation and is frowned upon by the international financial institutions whose goodwill is very important to most under- developed countries. Also it is difficult to introduce the new money into the system where it will best serve the present purpose, i.e. in the hands of low-income consumers. Reducing the price of food, however, is a very attractive way of priming the demand pump. About two-thirds of the inhabitants of Cali spend over half their income upon food], so that a reduction in the price of food by 10% is equivalent to an increase in real income of 5% or more for two-thirds of the population. The total increase in 1See Chapter II a" I": o a .‘u effective demand can thus be quite subStantial. 0n the other hand the absolute increase in effective real income per family or per person is quite small. so that the new demand will tend to be directed towards low-ticket goods like tableware, kitchenware, and clothing, which in general are goods with a high labor content. But the proportional increase in demand for these and other goods can be quite dramatic. A 5% increase in real incomes in an environment where basic necessities (food and housing) occupy 90% or more of incomes is equivalent to doubling (or more) the discretionary income of two-thirds of the population. A prerequisite of efforts to reduce food prices to the consumer will be introduction of a marketing mentality into the food distribution sector. As a general rule food distribution in Latin American cities gelled into its present format many years ago. when the «WWI- _. cities were much smaller than they are now. With few exceptions, the W- ”fin-I, 3'... food distribution systems consist of non-integrated multi- level channels, .... ._._-—.-._.-.__ ...-..‘hn-h. *—-—— Hm- ...?“ I; . with atomistic conditions obtaining at the retail level and to a great extent even at the wholesale level. This situation extends also to the country assemblers and rural producers. This has created a situation in which each participant in the food distribution system merely repeats each day the time-hallowed activities, no participant is in a situation to have an overview of the whole system, and no participant is in a position to introduce changes by his own efforts]. Introduction of changes of the type suggested by PIMUR will compel these participants to re-evaluate their activities in the light of consumer 1A notable exception to this rule in Colombia is IDEMA, a Department of the Ministry of Agriculture, whose responsibility is the maintenance of orderly markets for agricultural products, and whose authority is very broad-ranging. requirements. functions to be performed, and the nature of other participants in the distribution system. In short, if they are to benefit from the changes they will have to adopt a marketing mentality, and in all likelihood this will spread to other sectors of the economy as well. In particular, an understanding of this mentality must spread to those in the public sector responsible for overseeing economic affairs. Otherwise marketing activities will be starved of resources and hampered by ill-considered legislation. The introduction of a marketing mentality is. we believe. a vitally important element in economic development. Structure of Research . ammlhmwwk The purpose of the research is to acquire an understanding of the food distribution system of Cali sufficient to evaluate the “N efficiency of the existing system and to design an improved system in -~L_._fl_ ,flfi, compatible with the resources available and the conditions obtaining {in the city. The_criterion of_1MPerement shall be that 05 selling foods cheaper to low-income consumers than does the existing system ”...“.H \fi- while still providing the various services required by consumers. I The fbod distribution system is actually a system, and must be analyzed as such. It is a system of interlinked components, and its overall behavior is a function of the characteristics of these components and the nature of the interlinkages. The first step is to define what components will constitute \ .w-unv-‘IW‘ '"' ' the system and where these components are interlinked. In theory, the ...... LAUGH?“— system is practically boundless, because for almost any activity anywhere in the world a chain of causation can be discovered which will eventually impinge upon food distribution in Cali. Fortunately, the immediate system. that which can have perceptible effects upon food distribution in Cali, is a good deal more compact. We shall take as the direct system the suppliers (country assemblers, food processors, etc.), wholesalers, retailers, and consumers. We are interested in the flows of food through these ...-“v..- *r“ -‘I"_ w.- M" ' “ _-.........,.—.. ”m” ".aa.:‘h components (or, equivalently, the flows of money in the opposite “mat-qt WV“. ’I—an‘ “PH- direction) and how they are related to prices at the linkages. These WM ...-1’ relationships are functions of the characteristics of the wholesalers, retailers. and consumers. Our final objective is to discoverm mys of altering these characteristics which will reduce the price at the - ...... -‘mfiI-rns .3" ‘v consumer-retailer linkage while maintaining or increasing the flows of food. The consumers. clearly. it is not in our terms of reference to change. nor is the present writer qualified to make recommendations concerning the suppliers because our investigations were confined to #M'vn—‘hd—A ...—w “h...— lint-“M Cali and the suppliers operate on a national scale. So we are left with /‘ I ‘m—‘fl-m 33on1y wholesalers and retailers as the components whose characteristics /’ I it is feasible to change. The linkages between these components are fairly straight- forward and can be diagrammed as follows: suppliers wholegalers " wholesaler/retailers retailers A r consumers U.- 10 .J *2. where the lines indicate that flows of foodstuffs can occur in a downward or horizontal direction. The diagram indicates that: a) suppliers send foodstuffs to wholesalers and wholesalers/retailers; b) wholesalers receive from suppliers and send to wholesaler/retailers and retailers: c) wholesaler/retailers receive from suppliers and wholesalers and send to retailers and consumers: d) retailers receive from wholesalers and wholesaler/retailers and send to consumers; e) consumers receive from wholesaler/retailers and retailers. We have introduced here a previously unmentioned component, the wholesaler/retailer, which as its name implies combines the functions of wholesaling and retailing, because this particular component turns AQ—‘xn- h out to be of considerable interest to this study, as will be seen W w 1......LW-m4f‘rmm -me WWI-.fi‘ll later when we discuss prospective food distribution systems. The diagram should not be taken as a system graph; it is only intended to depict the paths along which food flows can occur, and is by no means complete. For instance, we have defined institutional food M consumption (hotels, hospitals, etc.) out of the system. and the diagram does not make allowance for foodstuffs to leave the system as spoilage. The system depicted in the diagram is a generalized one, and for some products the flow sequence shown is not followed. A sub- stantial proportion of the beef consumed in the city goes direct from / _/ suppliers to retailers (the so-called "forranea" beef), and most of thef l/ll raw (unpasteurized) milk goes direct from suppliers to consumers. ( However. these specialized distribution channels can be brought formally within the system as diagrammed if we consider some of the various stages shown in the diagram to be functionally incorporated into others. It is immaterial how the incorporation is considered to move. Not depicted in the diagram are the :I:I11§¥§"§:+v1ces that are involved in food distribution but which must be considered as exogenous factors, in that this dissertation does nogkgfihhgrt to make recommendations as to how their performance could be improved. This omission does not imply that such improvement is not important (PIMUR investigated all these areas and made appropriate recommendations); it merely reflects the fact that thorough analyses of these factors would each constitute a study as large as, or larger than, the present one. The factors concerned are: i) banking and financial services ii) transportation services iii) governmental intervention, direct (e.g., the IDEMA food stores), indirect (as in price controls), and legislative These factors affect the food distribution sector, and changes in their characteristics will have noticeable effects upon food distribution. But these effects will not be spectacular, except possibly for the hypothetical case of particularly ill-judged legislative action. This is because they are not linked into the food distribution system in a way that permits feedback. Strictly speaking, feedback k does exist, but#1nwfact,mthe feedback linkages are rather tenuous. M'- Different considerations apply to another sector: the urban manufacturing sector. Earlier we traced the way in which a reduction in food prices could hopefully lead, via demand for manufactured goods. 10 to reduced levels of unemployment and higher per capita incomes. This would lead to a much greater demand for food than that indicated by the straightforward consideration of income elasticities. Since the price-elasticity of supply of foodstuffs is likely to be very low in the short run (especially for beef) the effect of this demand-amplifi- cation would be that of strong negative feedback, tending to restore food prices to their original level. However, Cali only accounts for about 4% of the population of Colombia, so any reasonable increase in wfl_flh .- v- _ ... H”, " '~*w—p .... the city's demand for food will not substantially affect the prices of ”...-a...- '- umu~mnw ~--—— “...—...... ...—Fm... foods at the supplier level, which are determined by national demand “.3, mm“ «rm UWI‘W rmcIL‘AMI1:_—q"mw_esim Nev-14h and supplngonditions To the extent that economies of scale are 5:511:51e in food distribution there exists potential for positive feedback, with higher demand bringing about lower prices, but this will choke off fairly rapidly. Perhaps the most important point here is that conditions in Cali do not seem to be suitable for developing the "boot- strap" increases in demand and employment described earlier. Only a small proportion of the manufacturing in Cali is concerned with low- ticket items with high labor content. It appears that the beneficiaries of any increase in demand for low priced manufactured goods in Cali will be the other Colombian industrial cities, principally Medellin and Bogota. We therefore feel justified in omitting this sector from our —-- p.- w 0-,.1‘ h M— at“... "wt H.“.Mm “...... I ...... r " H‘hw w' m L‘...‘ diagram;fi Even if this increased demand for low priced consumer goods does result eventually or directly in increased demand for goods of 1An exception to this is constituted by many fruits and vegetables, grown locally exclusively for the Cali market, but these me up only a small proportion of consumer food expenditures. 11 the type manufactured in Cali this will not help much, for these goods generally have a low labor content. We end up. therefore. with the simple system diagrammed above. This system contains no feedback loops of any importancel_so ' "‘ “'a'KV-z. . r... W questions of stability do not arise. For the present study this is """' ‘41-‘30“ fortunate. for it means that errors in parameter estimations resulting from random sampling variations will not result in amplified errors in estimates of the performance of the system. Specific Issues To repeat. the objective of the study is to design a food distribution system that will enable the consumer to buy food at lower prices than existed as of the time of the field work, viz. February 1969. We have defined our system so that this statement must be taken to mean that we will try to reduce the total mark-up between suppliers and _-._ v—r;__.._ ‘0‘ retailers. as well as expenses other than direct food expenditures ____. ___ ”Jh-Mflwmefivw: ”My.“ ~— wt (e. 9. travel) incurred by consumers in the course of making food purchases. ....“me ......“ .. . ”...-va u—m, «n . main“ in“ m.- M Imwhwmw "‘ "' " I'm an ., h ”In .4. ... . This in turn implies that the present mark-up is greater than it need be, given the functions that must be performed by the system. The words “need be" in the previous sentence have important connotations. associated with the concept of efficiency. This concept is well understood in the physical sciences, particularly thermo- dynamics. but it has proved most elusive in marketing. If the present study were a problem in thermodynamics it would be relatively simple to measure the efficiency of the system relative to its theoretically possible efficiency, and then decide whether the discrepancy was sufficiently large to warrant the effort and risk of trying to reduce 12 it. A piecemeal analysis would be required only if the latter decision was in the affirmative. in which case we would try to determine whether this piston. that throttle valve. or yonder boiler was in some manner malfunctioning. But the present study does not enjoy the benefits available to thermodynamicists, and has toAstart_with a_piecemea1fil approach. This has its dangers. We might find. for example, that the interlinkage between wholesalers and retailers, the way that transactions were carried out between them, were excessively clumsy and costly. we could then recommend a streamlined transaction mode, hoping to reduce the final price of the goods handled. only to find that the changes in operating modes required of both parties to the transaction would result in higher operating costs and a higher final price. This is why we spent some time defining the system under discussion and justifying the omission of various possible system components. He are now at least fairly certain that we have missed no important components of the system. and that providing we follow the effects of changes through the various components listed we shall arrive at approximately the right result. The research itself can be cast in the form of seeking to w «u.-—.a¢nfi-‘~14—~M_~W-M 21m umw- ' 2“ fir’flu , prove a set of hypotheses The global hypothesis, which follows from ! ¢ “‘2’?! LB. {tLfiJD the statement of purpose of the research. is that the mark-up attached “gun-“HM - an- 4. ”gm “ism Ami ""- " iv . ‘.v a" gif} to foodstuffs on their passage through the system is excessive. But .. '5m'~ —*-“‘- —-- w-vr- nun—r “ 'W‘" “WW-m as we have already noted. this hypothesis is not directly testable, .11" :7pbecause marketing as a science has not yet‘HEValoped the appropriate :‘theory. We can only show that the mark-up is excessive by showing that we can design a system which provides food at a lower final cost. 13 So we must break this hypothesis down into others which are directly testable. l. The first such hypothesis is that excess profits are being taken by one or more components of the system. In this context excess profits will be considered to be net profits which are equivalent to a return on investment greater than that available from investing in other businesses of similar riskiness plus, in the case of owner-operators, an imputed salary equal to that paid to hired managers of similar operation. If this hypothesis turns out to be correct then we must seek to ww' Mao-I‘m” ~ ~ 4““ not“... ‘C—FJM -—4 4— “rm-1e- rung-N...“ *. discover the mechanism which prevents market forces from eliminating WW» umflffl 1_..“0 ~.‘ Hon-0M op..- H—‘W 'v— these excess profits. The most probable form of mechanism would be barriers to entry. of which the following are examples: MN mm”, W?- a) forcible. such as a Mafia- -type operation b) legal. such as licensing requirements c) difficulty of obtaining capital d) a "closed shop" understanding. according to which suppliers refuse to deal with prospective new wholesalers, and so on down the line e) it may be difficult and time-consuming to acquire the necessary expertise and establish contact with suppliers, customers, and credit and transportation agencies f) practices such as tied sales, particularly tying credit to sales 9) peculiarities of the market such that it is obvious to prospective entrants that their entrance will result in profitless operation for all market participants This list would seem to include all possible barriers, and each item leads to individual sub-hypotheses associated with the main hypothesis. 2. Ci {Lie ) f, 3- i l 1 l 3. f d, 14 Some or all of the components of the system have unnecessarilx.h19h,. '13-‘- I “"Hfm. operating costs and thus are forced to charge high mark-ups. The --"«*.‘T:-‘-, WWII-N Mum .— test of unnecessarily high operating costs (and thus margins) is whether we can design components that would return normal profits (as defined above) with lower margins while maintaining the levels of service of the existing components. What we are looking for here are deficiencies withlg_the existing components. such as poor managerial practices which directly increase costs or managerial #- 4 ——-' Win-1mm.- can. inadequacies which keep the scale of operation at uneconomically low levels. The interlinkages between the several components are such as to impogeihighécost modes of operation and/or prevent the members of the '—"‘"I'-.n-- components from growing to sizes where economies of scale are 4.- .."v— “ 'm‘flmgw - .... 1 “" ""V" ""'"““’ ‘J‘ n- - "hW-r‘-1 Ir- 2* ‘ “'4' obtainable. One or more of the following contribute to such a situation: a) Consumer requirements are such that the only stores which can 7;; satisfy them are small units and/or units with high operating costs. Such requirements are refusal to travel more than a {JAGfi‘ {fixv very short distance to a store (thus limitingwthe_potentiglg o‘dquw ' _sales volumewofflstoresl, insistence on cost-adding services. such as credit and home delivery. and persistence in making multiple small purchases. b) The nature of the transactions between components are such as to impose small operating scales or high operating costs upon one or both transacting parties. That is, transactions are fully or largely negiotated rather than routinized for one or more of the following reasons; inadequate information on prices requiring extensive search on the part of the buyer, inadequate information on demand requiring extensive search on the part of the seller and/or leading to scale-restricting, risk-averting practices. absence of reliable grading standards and/or absence of mutual trust, necessitating purchase by inspection, poor assortments held by sellers. necessitating many small purchases by buyers, 15 tradition-bound practices such that innovations would bring no direct reward to the innovator or rewards incommensurate with the risks involved. Again. the lists of possible causes of dysfunctional practices is as complete as we can make them. and lead to a set of sub-hypotheses associated with the main hypothesis. From this stage on the research follows standard practice. The sub-hypotheses are converted into questionnaire items, and then |-fl‘.‘.fi. follows the procedures of sample design, questionnaire pre-test and 'dfiim ..b, “-’i hang/w re-design, field work. and data processing and analysis. The following .—.—— F'm- ' 5" “wand-W45”. ~1mmu'wwt-‘Ww chapters of this dissertation will be concerned with presenting the results of this analysis and with drawing the indicated conclusions. Social Justification One cannot lightly undertake to make changes in institutions afi important as food distribution. Food distribution is an extremely complex set of economic activities. highly inter-related, and the effects of ill-Judged action could be extremely serious to the population of the region concerned. However, the recommendations arising from the research upon which this dissertation is based were carefully formulated and, we believe, free from danger on this score. At worst, they will merely fail in their objective of reducing food prices. Socio-cultural systems are also very complex and are very vulnerable to shocks, especially because the effects can be very slow in appearing. We should therefore make an attempt to foresee the effects of our actions and to clarify the issues involved. even though we might not be entitled to make the final judgement of values. That, in all equity, must be left to the people affected, or their represen- tatives. 16 Our direct objective is to lower the price of food to low- WWI tau: -mq'... .Iwgwm,hr‘~~ "1:“? "'.I"'<.!"l!!"-"'--...,3 7' income dwellers in an urban community. There would seem little to be "..-“..- —-—_——— ‘ ——— i‘flv— ww— -"" “r. :- apprehended on this core. We are not aware of any studies that show social cohesiveness in normal communities to be directly dependent upon the existence of malnutrition in major segments of the population. As for the secondary objective. the encouragement of industrialization, the horrors of nineteenth-century industrialization have been very amply documented. and we must consider whether they are likely to be repeated in modern Colombia. The worst aspects appear to fall into two categories, the exploitation of labor. particularly child labor. and the shock of dislocation on changing a structured rural environment for a disorganized urban one. As far as the first is concerned, experience indicates that the situation can be remedied by legislation, and Colombian labor legis- lation is both liberal and adequately enforced. For the second. we can observe that rural-urban dislocation is occurring without industrializ- ation in Latin America, and that in this case industrialization can only help, by providing employment to people who otherwise would be totally disenfranchised. And although the Western industrialized countries have yet to find satisfactory answers to the problems of affluence the underdeveloped countries have a long way to go before these problems will start worrying them. A more serious problem is posed by those employed in food distribution itself. The research in Cali showed that there were almost nine thousand retail outletsfioperating in 19§21 far more than M an efficient food distribution system would need. Rationalisation ' “VI.- 1x41- HIM—Mi... "‘ of the system would certainly threaten the livelihoods of fifteen to l7 twenty thousand families. Would such rationalisation produce net social benefits? There is. of course. no way in which this question can be conclusively answered. We believe that there would be net social benefits. In the first place. the burden of supporting an inefficient food distribution system falls particularly severely upon the poor, those least able to afford it. In the second place, there are reasons, given below, for believing that the misfortunes of the retailers will not be anything like as severe as the figures above would indicate. l. A substantial proportion of the small retail stores do a good deal of their business selling beverages for consumption on or off the premises. This is a high-margin line, and provides most of their profit. It should not be reduced. and might even be increased. by rationalisation of food retailing. 2. The normal attrition rate amongst these operators is very high. It was found in Cali that between making up lists of stores and returning three months later to carry out interviews one-quarter of the stores listed had gone out of business. Any rates of introduction of new stores, short of a crash program, should be slow enough for the normal attrition to remove those who would be injured by low-priced competition. 3. Market stall vendors, congregated into six public markets, account for almost four thousand of the retail outlets. As will be shown later. these markets serve almost exclusively their own neighborhoods. Thus by Judicious zoning in licensing new retail stores the shock to stall operators from new types of competition can be cushioned. 18 Caribbean Sea =arranquilla \ l \ Panama Venezuela ‘ ‘ Bucaramanga O Agdellin Bwota Pacific 0"." Buenaventura OCali ‘ 5 Ecuador Peru + N Map of Colombia Fig.1.! Brazil 19 4. The two thousand stall operators who specialize in fruits and vegetables will in all probability remain competitive for some time to come. given the difficulty of routinizing the marketing of these products. The Setting - CaliA#Colombia l. Historical and Geographical Three mountain ranges dominate Colombia's geography, and have dominated most of its history. They start at a node on Colombia's southern border with Ecuador and run north, separating only slightly from each other as they go. From east to west the three ranges constitute the barriers separating the plains and the Amazon basin from the valley of the Magdalena, the Magdalena from the valley of the Cauca, and the Cauca from the swamps and forests of the Pacific coastal area. These two valleys, together with the area around the short Sinu River in the north-west corner of the country, account for the great bulk of Colombia's food production and also of its population. Cali is located in the Cauca valley. The Cauca River originates high up in the Andes, near the node where the three ranges join, and arrives at the level valley floor some twenty miles south of Cali. This level valley stretches to Cartago, a town about l00 miles north of Cali, where the river descends through rapids to the swamps of the Atlantic region and eventually joins the Magdalena to flow on to the Caribbean. The City of Cali is the capital of the Department of Valle del Cauca, the political entity which consists of the valley and the land on either side roughly up to the crests of the two defining mountain ranges, together with a westward sweep to the Pacific between the Maya and San Juan rivers. The valley proper which is roughly one hundred 20 c4 0 ”#AL.E:&\‘~‘r<> ./ 's. ' L /{I2au '3, 4 [SARA / 1:: In. {Oiled .- I/ibnuugn a:fi?:' ‘-. / om. U0 .3, f'J :3 c I Mosulff “ E \ Id Jag" g l , CHO ca \ “I. ' 3 .:. ’ ., - I. K-\_1'8’m. "2' 0" O ' ‘ mi? . I \‘m’u/K.‘ ., b;‘.- d 0” .I 'l I ~\..../ 7 a . A ‘1. . 00000 Y 00 ,me IIII'“ “'- .I' In?" “'3" 17:13. I! I y) L. . PACIFIC OCEAN 1 .. . -.m . I, I I v j I. V .~ «II In ' § 1 \ I Can I v anew-r — on / Q _ j k :1“ . M- I A \ "I - ,. - / .‘\ \J .K . l srwaou "- v=r C A U C A «+0. Railroad Cocoa Vouoy VALL EY. THE DEPARTMENT OF VALLE AND THE GEOGRAPHIC LOCATION OF THE OAUOA FELZ 21 and twenty miles long, has an average width of ten miles. So we have a long narrow valley running north and south. with the Cauca River as its dominant feature. The valley floor is extremely flat, so that a great deal of land area adjacent to the river is subject to flooding and is permanently marshy. North of Cali the river runs along the western side of the valley. so that all the cities between Cali and Cartago were sited east of the river. the swamps making it difficult to obtain access to the west. Cali alone is sited to the west of the river about five miles south of the place where the river first swings into the hills. This location cut Cali off from the main part of the valley (even today the town of Palmira. twenty miles to the east, is a much more important ‘ marketing center for agricultural products than Cali is), but gave it access to the south and, more important. to Buenaventura. Colombia's only port on the Pacific. Cali's growth began with the opening of the Panama Canal, which enabled coffee to become an important export crop. Coffee came up the Cauca by boat to Cali. where it was transshipped to Buenaventura, and imported goods flowed in the other direction. In the early Thirties the railroads displaced the river boats, but Cali was still the inland terminus of the link with the Pacific. The same was true even when all-weather roads replaced rail with road transport. However, Cali's monopoly is about to cease. A new and much better road from Buenaventura to Buga, some sixty miles north of Cali was approaching completion in l969. It should be of interest to economic geographers to watch the development of the two cities when this road is finally opened. (C. The locational monopoly has served Cali well. The import- substitution policy introduced in the Thirties encouraged manufacturers, and Cali was well situated for factories dependent upon imported raw materials and semi-finished goods. In the late Fifties a number of foreign firms established manufacturing facilities in the city. and in 1969 Cali was the third industrial city in Colombia. Table I.l.l. presents comparative data for Bogata. Medellin and Cali. the country's three main cities. Strictly the data apply to the departments in which the three cities are situated, but only a negligibly small amount of manufacturing is done outside the main urban areas. Table I.l.l. Comparison of the Percentage Contributions of Selected Regions to Total National Manufacturing Activity in Colombia, l966 Region Antioquia Bogata, D.F. Valle Per-cent of Establishments l6.7 24.l 15.5 Per-cent of Personal Employed 25.4 26.0 l7.5 Per-cent of Total Remuneration 27.0 25.0 19.4 Per-cent of Gross Production 20.2 22.4 20.4 Per-cent of Value Added 23.3 23.3 2l.l SOURCE: Boletin Mensual de Estadistica. DANE, #2l2 (October 1968) It can be seen from the figures in Table l.l.l. that Cali accounts for one-fifth of the country's manufactures, and one-sixth of its manufacturing employment. 23 The population of Cali has grown rapidly in recent years. The census of 1964 showed a population of 639,000, and comparison of this with the result of the census of 1951 showed that the growth rate between these dates was equivalent to a compounded rate of 7i%. Clearly a substantial immigration flow from other parts of Colombia has to be hypothesized to account for such a rate of growth. This population increase has been accompanied by a geographical extension of the city east toward the River Cauca. We estimate the population of Cali, as of February 1969, to have been 8l3,000. Details of the calculations involved in this estimate are given in Appendix l to this chapter. At an average household size of 6.27 persons1 , there were l30,000 households in Cali. Details of income distribution, education levels, and other demographic statistics are given in Chapter II, where we discuss the inhabitant of Cali in his capacity as a consumer of foodstuffs. As can be imagined, this rapid expansion has brought growing pains with it. For instance, the city has encountered great difficulty in providing an adequate level of public services to the new urban areas. Another problem which is the crux of this dissertation, is that the city is far too large to be served by its original food distribution system. The rapidity of growth has been such that alternative systems have not had time to evolve, although the social costs of the existing system's deficiencies are affecting a great many people. IPIMUR. Consumer Survey 24 Originally, Cali was built on the low foothills of the Western Range. and expansion eastward was inhibited by periodic flooding of the ‘ river. In l954 the autonomous Corporation, CVC. was formed whose function, broadly speaking. was the provision of infrastructure or external economies in the departments of Cauca, Valle, and Caldas. Cauca and Caldas quickly drapped out of the scheme, but Valle has benefited substantially from its activities. Amongst the first of these was the construction of a levee along the west bank of the Cauca, which has permitted drainage of the land between Cali and the river, and by l969 urbanization of this area had been extensive. The limiting factor was the provision of utilities, principally water and sewers, the cost of which was becoming prohibitive. Planeacion Municipal, the city's plan- ning department, has decided upon a policy of discouraging lateral expansion and encouraging vertical expansion in the form of high-rise apartment blocks near the city center where the provision of utilities is cheaper. 2. Cali: Topographical It is worthwhile describing the layout of Cali at some length, because one of the criteria used for classifying the consumers, that of socio-economic level, is very closely associated with well-defined regions of the city. . The center of the city (in the sense of being the heart of the business district) is the Plaza Cayzedo. The Rio Cali, a small river which passes only a few blocks to the north of the Plaza, formed until recent times the northern boundary of the city. About five miles to the north again, the Cauca sweeps in almost to the edge of 25 the hills. Until recently the east was bounded by swamp lands. Until around 1930 Cali expanded slowly; by this time it extended perhaps half a mile west of the Plaza Cayzedo. a mile east, and a mile south. The Rio Cali formed its northern border. The wealthy then started building north of the Rio Cali. This was the era associated with the completion of the railroad along the valley, and with the establishment of manufactures resulting from the "substitution of imports" policy. Up until this time the Plaxa Cayzedo had been the city's principal food market, but now the vendors were moved down to the Calvario district, which was already the center of the wholesale food trade. The distance involved in this move was not great, being only six blocks south and one block east. Here they were established in the building which, after expansion in 1938, became the Galeria Central or Central Market as it was in 1969. At this time the central market area was still the south-eastern boundary of the city's business district. The next spurt of Cali's growth came in the late Forties and early Fifties when the suddenly increasing migration to the cities noticed all over South America was intensified in Colombia by the after- math of the "Bogotazo" in 1948]. This is the period of growth that is still continuing. It was accompanied, naturally, by a corresponding areal growth of the city which can be most conveniently described in terms of socio-economic classes. 1The "Bogotazo" is the name given to the riots and other convulsions following the assassination of the left-wing political leader Jorge Eliecer Gaitan in Bogota in 1948. The near-anarchy which followed this event has dominated Colombian politics ever since. Order was rapidly restored in the cities and large towns, but sporadic civil war and terrorism continued in the countryside for several years, degenerating into banditry. Many pe0ple fled the countryside at this time, seeking the relative security of the cities. 26 Qua-guxuaenxiugsu-nan_~o 331.381.... wan-IE3.“ .3232 299°.e-ugaaxaea 83922:: 3.8%: saueaex: 82.83 833.833 .xxssaaa: azuoma 2L). ’2.” Q 27 The neighborhoods of the city, the barrios, are classified by the municipal government into nine socio-economic levels, which correSpond roughly to the social class of its residents as shown in Table 1.2.1. Table I.2.l. Distribution of P0pulation by Socio-Economic Level, Cali, February 1969 Socio-Economic Level Class Percent of Population (a) I Upper-Upper 4.8 II Lower-Upper 4.2 III Upper-Middle 7.7 IV Lower-Middle 12.4 V Upper-Lower 50.0 1 VI Lower-Lower 20.9 VII, VIII & IX Industrial, Penal Military -- (a) These figures are those obtained from the PIMUR Consumer Survey With the expansion of the city, the upper classes started building on the land to the north-west and up the hills north of the Rio Cali. This movement has now ceased because the hills are built-upon up to the level of the municipal water treatment plant, so that houses built further up would not be able to obtain water. At present upper- class areas are being developed in the extreme north and south of the city on lower land. Middle-class housing is growing both to the north 28 and to the south-west, a growth pattern that has continued for some time. Some years ago this growth encircled, then absorbed, the village of San Fernando, about four miles south of the city center. Working-class housing is spreading southwards on the east of the middle-class areas described above. Prior to this the construction of the levees mentioned in the previous section made possible urban- ization of the land east of the city center down to the Cauca, some five miles from the Plaza Cayzedo. This urbanization consisted entirely of working class housing. The tugurios or slums have develOped on land that no one else wants, either because of its inaccessibility, its propensity to flooding, or the absence of sewage and water services. There are four principal slum areas in Cali:] 1. Siloe, on a steep hillside on the south-west of the city. 2. Terron Colorado and Bella Vista, built on the hills west of the city center a bit further out and higher up than the upper-class districts of Arboleda and Normandia. 3. La Union, in thesnuth-east of the city (actually just over the city's official political border) 4. La Isla, along the Rio Cali about two miles of the city center. These are officially classified as tugurios. It will be noted that, except for the last named, they are all on the periphery of the city. It is also true that the poorer working-class areas are amongst those further from the city center, and that, except for some 1For a description of living conditions in these slums see Robert L. Heilbroner, "The Great Ascent", Harper & Row, New York 1963. Chapter II. 29 destitutes who scrape out an existence in the lawless area to the south of the Galeria Central, the poor in Cali live far away from the center.I ———~‘___‘ This will be seen to be an important factor in our evaluation of the performance of the food distribution system. Figure 1.3. illustrates the foregoing. The only additional comment is that the old city is now classified as almost entirely middle- and lower-middle class. The map also indicates the principal thoroughfares of the city. The most important shapping area is the city-center. This is the only area that serves the whole city, the others are only intended to serve their more-or-less inmediate neighborhood. In particular, the Galeria Central area is the hub of Cali's food distribution, both retail and wholesale. Up until the time that the field work for this thesis was performed in 1968-1969, the Galeria Central was the hub of Cali's food distribution, both wholesale and retail. In 1970 the Galeria Central building was torn down, as part of the program of beautification of the city in preparation for the Pan-American Games in 1971. Tenth Avenue, which runs alongside of the Galeria Central, was enlarged to a four- lane arterial street, and this entailed the tearing down of the market buildings. North of the Rio Cali is the Centro Comercio del Norte (Commercial Center of the North), a shopping center modeled on North American lines, and catering to the high-income families who live nearby. In the rest of the city there are no well-defined shopping areas, al- though the satellite markets, especially Alameda in the south, have 3O attracted some merchants. Other concentrations of retail stores are found along some of the principal roads marked on the map. 3. Food Retailers In the Introduction a diagrammatic representation of Cali's food distribution system was described. This identified five major components, which were suppliers (external to the city), wholesalers, wholesaler-retailers, retailers, and consumers. These will be described in detail in Chapter IV. However, in Chapter II and III we shall be discussing the interface between the consumer and the retailer, so it will be useful at this stage to describe the various types of food retail outlets operating in the city, in order to define the terms that will be used in these discussions. There are three major types of food retailers in Cali. These are the public markets, the traditional food stores, and the self- service stores. There are also many other types, some important and some unimportant, which display a wide range of operating characteristics. Together, these form a fourth, "miscellaneous", group. The Public Markets. The public market system consists of the Galeria Central (including the two Calvario markets) and five satellite markets]. Their locations are noted in Figure 1.3. Each market consists of a large building housing many small market stalls. The market stalls specialize by product, and are segregated inside the buildings according to the four main product groups of meats, non-perishables, fruits and k 1A sixth satellite, Alfonso Lopez, was under construction in 1969, and has since come into service. In 1970 the Galeria Central was torn down to make way for road improvements. 31 vegetables, and dry goods. In the streets around the markets numbers of street vendors sell fruits and vegetables. Some of these vendors are officially associated with the markets, in that they pay rentals to the city, while other are unofficial and, in fact, prohibited by law. The Galeria Central, by dint of its size and central location, is the most conspicuous food retailing institution in Cali, and has come to epitomize food retailing in that city. Just as, in the U.S.A., when one thinks of food retailers one thinks of A&P, so in Cali one thinks of the Galeria Central. It is reputed to be, of all food out- lets in Cali, the one which offers the lowest prices, the greatest variety of products, and the widest range of qualities. It is also Spectacularly dirty, extremely crowded, and infested with thieves and pickpockets. The satellite markets can best be described as being like the Galeria Central but more moderate in all their characteristics. Traditional Stores. These are personal-serviCe stores, with the customer on one side of a counter and the shapkeeper and the stock-in-trade on the other. There are two distinct types, the small neighborhood stores and the large stores around the Galeria Central. In terms of numbers and diSpersion round the city the 319393, the small neighborhood store, is Cali's most important food retail outlet. Tiendas can range in size from a roadside stall to a store with 2,000 square feet of floor Space, and can show any of a wide range of special- izations. Many of them are essentially small bars or sidewalk cafes. Those larger tiendas which sell only foodstuffs without specializing in 32 any one line are called graneros (literally, grain-dealers) but the distinction between the two types is not precisely drawn. As often as not the term applied depends on which title the Operator cares to put on his signboard. We shall follow PIMUR's practice of defining as tiendas neighborhood stores with sales of up to C01. $20,000 per month, and as small graneros stores with sales of from Col. $20,001 to Col.$50,000 per month. The main offering of these stores is locational convenience. They make fOOd available all over the city. Individually they do not offer a full line of foodstuffs (only 30% of tiendas sell meat), but collectively they do because they are so numerous and dispersed that almost every house in Cali is within easy reach of several such stores. The variety they offer within any one line, even collectively, is rather poor, especially in the case of fresh fruits and vegetables. The large traditional stores are located in the immediate vicinity of the Galeria Central. A typical store of this type is open- fronted (no front wall, no display windows), 20 to 30 feet wide and 30 to 60 feet deep. They sell only non-perishable staples and processed foods. Although they do not sell meat directly they rent space in their stores to meat stalls (bancos de carne). The distinction between the large graneros and the wholesaler- retailers, the two types of centrally-located traditional stores, is that the former sell only to final consumers while the latter sell also to other retailers. The wholesaler-retailers are, on average, larger than the large graneros, but otherwise there is little difference batween them. 33 Self-Service Stores. These are of four types, supermarkets, chain stores, cOOperatives, and grocery stores of the Cajas de Compensacion Familiar (Family Allowances Funds). Supermarkets in Cali are located in or near upper-class districts. They are large stores, by Cali standards, departmentalized, and with a self-service format. Of all food retailers in Cali they offer the widest range of foods (with the exception of the public markets). In terms of depth of offering and merchandising technique they are best in luxury goods (canned fruits and juices), good in staples, fair in meats and rather poor in produce. Their prices are competitive with those of the public markets. They differ from supermarkets in the U.S.A. in a crucial aspect. They are not mass-merchandisers, Operating on a philoSOphy of low margins and quick turnover, and measuring success at least partly in terms of market share. Instead they are large neighborhood stores catering to the upper classes. Although they have adopted the format of the North American supermarket they have not adopted its philOSOphy of Operation. Chain stores are large stores of the five-and-ten type, and are mostly found near the Galeria Central. They aim at the lower- middle and working classes, with aggressive merchandising and competitive pricing, and if the crowds flocking through their stores are any indication they do this very successfully. Foodstuffs do not form a large part of their business, and they offer only non-perishables, mostly processed foods. It has become almost standard practice amongst them to use rice as a loss-leader in their frequent promotions. 34 These chains would well repay further study. Of all organiza- tions in Colombia they appear to be the ones which have best solved the managerial problems of organization, training, delegation, and control, as well as having apparently consciously appraised their place as marketers. Dissemination of this problem-solving ability would be a valuable form of economic aid. Cajas (de Compensacion Familiar, Family Subsidy Funds). There are two grocery stores operated by these institutions in Cali. The Cajas are described in more detail in Chapter IV. The two stores are open only to the employees of firms affiliated to the respective Cajas. ‘The stores are large, with a self-service format. Their product offering is not complete, only one store handling fresh meat, and they concentrate upon non-perishables and canned goods. One of these stores had been Opened only a few months before PIMUR's field work started. Cooperatives. In February 1969 there were seventeen consumer cooperatives operating in Cali. They were officially under the supervision of the Superintendency of Cooperatives in Bogota, and enjoyed various privileges which are detailed in Chapter IV. They are open to the public, but draw their memberships from amongst employees of individual firms or institutions, so that one finds, for example, the cooperative of the employees of Carvajal & Co., of the employees of the Pacific Railroad, of the employees of the Government of Valle. They have mostly located in a nondescript manner, neither in the city center nor in the barrios, so that their members find them difficult of access. Some have located near to where their members work, which helps in this respect, but which is only a partial solution 35 to the location problem. In general, they offer only a limited range of goods (dry goods and non-perishables), their prices are not very competitive, and their stock control is poor. There are some exceptions to this generalization. Miscellaneous Retail Store Types. These comprise the following types: Truckers (Camioneros) Specialty Stores Stores operated by IDEMA Puestos de Salud (Health Post Stores) Markets in nearby towns Comisariatos Farmers' Market Ambulant Vendors Of these, only the specialty stores and the truckers are of any significance in Cali's food distribution. The truckers are, strictly, people who sell raw (unpasteurized) milk door to door from a truck or pickup, although the name can be applied to people delivering pasteurized milk. The truckers handling raw milk account for about half of Cali's milk supplies, the rest being pasteurized milk sold either through tiendas or delivered directly. The most important group of specialty stores are those stores which deal in fresh meat, catering especially to upper-class clients. Other stores specialise in poultry and eggs, and there is at least one large establishment specialising in fish and sea-food. Another group, large in numbers if not in volume of sales, is the so-called expendios de leche, which handle raw milk. These 36 stores, which are basically one—room operations, are found in barrios of S.E.L.V. and VI, and appear to act as retail outlets for truckers in areas where demand is too scattered or irregular to warrant door- to-door delivery service. The stores operated by IDEMA (Instituto De Mercadeo Agricola, Agricultural Marketing Institute, a department of the Ministry of Agriculture) supplied only a negligible proportion of Cali's food consumption at the time of PIMUR's field work. However, because of the financial resources behind IDEMA, it must be considered to be of at least potential importance as a food retailer. In February of 1969 IDEMA was operating three stores in Cali, and had just started on a program of sending truck-trailers equipped as stores into the lower- class barrios to sell staples. The Puestos de Salud are public health clinics located in the lower-class barrios. They enter into food retailing only in that some of them sell staples at cost to needy cases. Comisariatos (Commissaries) are food outlets operated for the use of armed forces personnel only, but this name is commonly used to refer to the stores associated with the Puestos de Salud. The Farmers' Market is not a separate institution but an arrangement whereby a fenced-off area outside Siloe market is reserved on Saturday mornings for the use of small farmers to sell direct to consumers. This arrangement came into effect shortly before PIMUR's field work started. The Ambulant Vendors travel round the residential areas selling mainly produce from barrows. They did not appear to be very numerous. They should not be confused with the street vendors associated with the public markets. 37 4. Facilitating and Ancillary Institutions In this section we will look briefly at the more important factors that set the tone of the business ambience within which food distribution is performed. These are sources of finance, transporation, personnel training facilities, and the government's part in setting the rules of the game. Sources of Finance Finance in Colombia in 1969 was dominated by the government's fight against inflation, and few people thought the fight would ease within the foreseeable future. The brunt of the fight fell upon credit, which was extremely tight. Interest rates ran at 25% per year for "triple A" industrial concerns, and at five per cent per month for personal loans secured upon lands and houses. Banks had been forced to cut down drastically on conventional banking activities. Conditions set by the Bank of the Republic had in effect established a reserve ratio of 50%. Interest rates on deposit accounts were limited by law to 4% in a period when inflation was running at about 11% per year, thus severely limiting deposit funds. And the interest charged on bank loans were limited to 14%. Instead of the standard banking services the banks put their efforts into the administra- tion of loans to firms eligible for special treatment. These firms, in agriculture-related industries, and the so-called "development industries", were eligible for loans at advantageous interest rates (8% to 15% per .year), and banks making such approved loans could discount them with the Central bank . 38 Since none of these loans were available in commerce or in the marketing sector, food distribution was effectively cut off from any financial services from the banks. Thus the food distribution sector was essentially self-financed in respect of short-term or circulating capital. The same is also true of long-term or fixed capital, but for different reasons. The open money market in Colombia is still very restricted by the small number of firms and investors and the total lack of any reliable information concerning firms. For this reason capital- raising is done by the entrepreneur from his own resources and those of his family and friends. The government-sponsored credit lines described above constitute an attempt to get round this difficulty since they are long-term loans. Transportation Three types of transportation are of interest to the food distribution sector; inter-city goods transportation, intra-city transportation and intra-city passenger transportation. Colombia possesses two railroad systems, one Operating along the Cauca valley, with a branch to Buenaventura, and the other along the Magdalena. Although freight rates by rail are slightly lower than by road, this does not compensate for the long delays in rail shipments, typically two weeks as compared with two days by road. In a country Where credit is very hard to obtain and costs upward of 2% per month and in which wholesaler's gross margins are under 4%, a delay of two weeks “is prohibitive. 39 Road transport in Colombia seems to be quite effective, at least as far as fbod haulage is concerned. PIMUR heard few complaints about inter-city tranSport from peOple in the food sector. This is not to say that road transport is efficient, in the sense of having low freight rates, few delays, etc.. Truckers are almost all small Operators, vehicles are mostly old (PIMUR reckoned that two-thirds of the trucks hauling foods were over ten years old), shipping is disorganized, and pilfering and even hijacking are common. But the goods seemed to arrive in adequate volume without the transpOrtation adding greatly to the final price. The tranSportation charge on a stem of plantain brought 130 miles from Quindio to Cali, for instance, amounted to only 8% of the price to the consumer. Intra-city transport is another matter. The stem of plantain that cost Col./$1.20 to truck the 130 miles from Quindio cost Col./$O.60 to haul 15 miles from the Galeria Central to Alemeda Market, and 25% of tienda operators said that arranging transportation was their primary difficulty in making purchases. This is not surprising. There are seven thousand food retailers in Cali, and every day they all, or nearly all, come to the Galeria Central area to make their purchases. Then they have to arrange trans- portation of their goods to their respective places of business, so that every day between the hours of six and eight o'clock in the morning seven thousand shipments are made. For the rest of the day there is little activity of this type. Thus the use of these -ransport resources is very inefficient; taxis are the only participants in this cross- town transportation activity that can reckon on being adequately occupied throughout the day . 40 The situation described is a clear illustration of the concept of suboptimization. Every participant is Operating under his own least- cost conditions, but the final cost to the consumer is probably some 3% to 5%1 greater than would be the case with a whole-system organization, because the adaptation between participants is so poor. Passenger transportation facilities are very good, considering the rapid growth of the city, which amongst other things has left road construction well behind. There is an extensive network of bus routes radiating from the center of the city, and services are frequent. Fares, at the equivalent of U.S. $.03, are not expensive. Out at the edge of the city the routes are more widely spread and become rather circuitous. This, together with poor road surfaces, means that the trip into the city center can take upwards of two hours. The low incomes of people who live out at the periphery of the city also mean that bus fares, cheap as they are, become a matter of some consideration. One feature of the bus system in Cali is that virtually every bus route in the city passes within two blocks of the Galeria Central. As we shall see in Chapter IV, this is probably a major factor in accounting for the importance of the Galeria Central as a food retailing institution. There are many taxis in Cali, and again fares are very low by standards obtaining in the U.S.. A fairly common practice among middle- class and lower—middle class housewives in Cali is to travel to the Galeria Central by bus and return home by taxi. This gets around one of the major drawbacks of the bus, that it is not designed for trans- porting loaded shopping-baskets. \* 1The system proposed in Chapter VI offers cost savings of this 0 r‘der. 41 Training Facilities We are interested here in establishments providing training in retail management. While there are many institutions in Cali which offer preparation for careers in business, ranging from the Universidad Del Valle, with its M.B.A. program, to small one-room enterprises which teach typing, we could not find any which gave courses of instruction in the management of small or medium-sized retail stores. The logical place to set up courses of this type would be SENA, the National Apprenticeship Service. SENA trains young people in such fields as motor mechanics, machine-tool operation, electronics, farm machinery, and commerce, and also offers refresher courses for practitioners in these areas. SENA appears to be willing to add programs of instruction according to demand; for example, in 1969 the Cali branch had recently started courses in the operation of fishing gear and vessels in conjunction with a drive to start a commercial fishing industry in Buenaventura. Packaging One of the keys to the operation of self-service stores is extensive packaging, and this practice can increase the rate at which customers can be served in personal-service stores. In Cali, even in the self-service stores, packaging practices are still rather primitive. One of PIMUR's investigations, therefore, was concerned with looking into the state of the packaging art in Colombia, and in particular with determining the extent to which suppliers of packaging materials were prepared to accommodate the requirements of their customers. 42 The conclusions reached were that: a) the industry is in general vigorous and actively seeking applications for new packaging techniques and materials. b) with the exception of fibre for sacking, glass, and some paper pulp made from sugar-cane, all packaging materials are imported and in consequence rather expensive. c) the high cost of packaging materials tends to confine their use in food distribution to luxury foods for which consumers are prepared to pay high prices. d) the industry was in a position to contribute to reducing spoilage of fruits and vegetables by means of improved containers], but marketing these containers would be difficult in view of the existing structures of the production and marketing of fruits and vegetables. In summary, the packaging industry is in a position to satisfy the needs of the food distribUtion sector, present and future. Its main weakness would seem to be a dependence upon imported technology, with its consequent dependence upon high-cost imported packaging materials. Government Intervention in Food Distribution This intervention takes three forms: legislation, direct intervention, and indirect intervention. 2 Legislation concerning food distribution is voluminous , if not very effectively enforced nor well-directed. There are, for example, the 1For example, cardboard cartons had been designed for transportation of fresh pineapple. 2There is a good discussion of this topic in PIMUR Final Report, p.315. 43 populist laws against speculation, which, insofar as they are effective at all, only prevent the populace from receiving the benefits of speculative activities. Perhaps the most significant legislative interventions have been those which created the Price Control Office, and Decree 2420 of September 1968 which gave far-reaching powers to IDEMA, the Agricultural Marketing Institute, which is a department of the Ministry of Agriculture. These have resulted in direct intervention in food marketing, as described below. The Price Control Office, the Spperintendencia de Precios, is responsible for the enforcement of price controls. In theory, all prices in Colombia are controlled, but in practice control is applied only to "articles of basic necessity". The prices on manufactured goods are set by Planeacion Nacional (the National Planning Department) on the basis of discussions with, and data supplied by, the manufacturers, and as a result they are fairly realistic. Thus there is little temptation to transgress the controls. For most foods the data are supplied by the Ministry of Agriculture, but there are not sufficient data available, and the delays involved in the price-setting process make it too slow to cope with seasonal variations in supply. In any case, the lack of standards for most foods, and variations in qualities make enforcement difficult, and rigged scales and other devices can be used to nullify the effects of the price controls. As compared with the Price Control Office, which attempts to regulate food marketing by decree, IDEMA attempts to promote orderly marketing of foods by acting as a participant in the agricultural marketing 44 system. IDEMA was given very wide powers by Decree 2420 of 1968, but in 1969 its activities impinged on food distribution in three main areas: 1. it had a monopoly on all imports and exports of foodstuffs other than coffee. I 2. it operates a price support system for the production of some staples such as corn and beans. In this activity it has the handicaps of inadequate data upon which to base its price-setting decisions, and insufficient funds with which to support these prices. The obverse side of this activity, selling the products thus purchased, is also intended to affect food distribution in a socially-desirable manner. For instance, consumer cooperatives are supposed to receive preferential treatment in purchasing staples from IDEMA. 3. it maintains a program of trying to keep food retail prices down by operating retail stores and by announcing in local newspapers the prices at which these stores are currently selling staples. This program is described more fully in Chapter IV. Indirect governmental intervention in food distribution has consisted principally of encouraging the formation of "philanthropic" institutions, such as the consumer cooperatives and the Caja stores. There are, of course, many other ways in which the government's actions have affected food distribution. We have already mentioned the effects of credit restrictions, and the high import duties on tinplate and plastic sheeting have confined their use in food packaging to items of luxury nature. These effects, however, are more incidental than intended or deliberate. f. I r' I' 45 5. Summary This has been a very brief desceiption of the city of Cali, and of the various types of food retailers which operate in the city. The purpose was to provide the background for the detailed description of the operation of the food distribution system which forms the substance of the following chapters of this thesis. Review of the Literature This dissertation is a study in macro-marketing, an analysis of an entire marketing system in an underdeveloped country. Very few such studies have been carried out. In fact, the only important studies of this type have been, with one exception, those carried out by the Latin American Market Planning center at Michigan State University. While there have been many good studies of food marketing by those interested in agricultural economics and agricultural marketing they are of only limited interest to us here because of their rather different approach. Although generalizations are dangerous it seems to be fair to categorize these studies as studies in commodity marketing, interested principally fin flows and prices. We are concerned with an advanced and extended _ -_-.—-—— —-»—~ form of institutional marketing, in which the main interest is in the —4__.._._-__~ market participants and_the relationships between them. The two types a” _ k-'—*_ of study complement each other, but the research methods are so different that there is little prospect of methodological cross-fertilization. The pioneering study in our field of interest was the study of food distirbution in Puerto Rico carried out by Galbraith and Holton I 46 in the early 1950's]. This study was basically in the standard micro- economic form. Cost data were analyzed to determine apprOpriate scales of operation for wholesalers and retailers, and certain restrictive trading practices were identified, chiefly in the area of food importing. However, they went beyond simple economic analysis in that they took notice of the vertical relationships between market participants, and considered the feasibility of introducing new formats. For example, they weighed the possibility of introducing the supermarket format, and came to the conclusion that this type of store would be incompatible with consumer buying practices. In 1965 the first of the Michigan State University studies took place, also in Puerto Rico. This was followed by studies in Brazil (in Recife), in Bolivia (La Paz), and most recently in Colombia (Cali)2. These were all multi-disciplinary studies, using primarily m..— a— . '— 1John K. Galbraith and Richard H. Holton, "Marketing Efficiency in Puerto Rico", Harvard University Press, Cambridge, Mass., 1955* 2C.C. Slater et. al., "Food Marketing in the Economic Develop- ment of Puerto Rico", Research Report No. 4, Latin AmeriCan Studies Center, MiChigan State University, East Lansing, 1970 "Market Processes in the Recife Area of Northeast Brazil", Research Report No. 2, Latin AmeriCan Studies Center, MiChigan State University, East Lansing, 1969 "Market Processes in La Paz, Bolivia", Research Report No. 3, Latin American Studies Center, MiChigan’State University, East Lansing, 1969 H.M. Riley et. al., "Market Coordination in the Development of the Cauca Valley Region - Colombia", Research Report No. 5T,—Latin Amefitan Studies Center,“Michigan State University, East Lansing, 1970 éThis document will be referred to subsequently as the "PIMUR Final eport" ..__ l 47 the skills of marketers and agricultural economists and drawing also __—...—-——— upon communications and social sciences. They were extended in scope, considering food marketing from the farm to the kitchen, as well as the ancillary functions such as transportation and financing. The Recife ! study also contained an analysis of the entire economy of the region, while the Cali study contained special analyses of various sectors of the Colombian national economy. tut As well as the primary reports, these studies also generated a number of dissertations and articles dealing with various associated topics, so that there is now a fair-sized body of relevent literature available from the Latin American Studies Center at Michigan State University. There has been very little other relevant writing done on this subject, however. One useful article has been written by Currie]. Based upon the experience of his years of work in Colombia, it offers a clear overview of food marketing in that country, and contains several interesting observations. Bennett2 studied the effects of governmental intervention in food marketing in Santiago, Chile, and his description of the food distribution system of that city shows some interesting difference between it and those of the cities studied by the Michigan State group. 1Lauchlin Currie, in Reed Moyer and Stanley Hollander (eds.), "Markets and Marketing in Developipg Countries", Richard D. Irwin, Homewood, Ill. ,1968. pp. 1201124 2Peter 0. Bennett, "The Role of Government in the Promotion of Efficiency in Retail Marketing of Food Products in Greater Santiago, Chile", Proceedings of the American Marketing Association, p. 105, Fall 1965 48 Research Report No. 5 of the Latin American Studies Center at Michigan State University, which reports the results of the Cali study, is the basic document for this dissertation. It makes available in English the final report of PIMUR, the group which carried out the study (PIMUR is an acronym of the Spanish for Integrated Urban-Rural Marketing Project), and will be referred to hereinafter as the "PIMUR Final Report". PIMUR fielded sixteen separate (but related) studies, two of which, the Wholesale/Retail Study and the Consumer Study, provided the basic data used in this dissertation. Reports on these two studies have been published (in Spanish) as PIMUR Technical Reports Nos. 6 and 7. Outline of the Dissertation Chapter II, The Consumer, describes the inhabitants of the city. A feature of this chapter is the presentation of the results of a survey of food consumption practices. Chapter III is devoted to a discussion of the food purchasing behavior of the Cali housewife. Both observation of purchase practices and analysis of attitudinal data indicate that locational convenience I I (accessibility) is by far the most important characteristic that the Cali lihousewife requires of her food suppliers. Chapter IV consists of a review and evaluation of the various institutions involved in food distribution in Cali. Emphasis is placed upon trying to understand why each of these institutions has taken the a form it now bears, in order to determine whether any of the existing ‘l-fih. .... —— —.._—.—.-.. my-” institutions could contribute towards, or form a significant part of, an improved food distribution system. 49 In Chapter V the work of the previous chapters is brought together in designing a food distribution system that will provide food at lower prices than obtain at present to low-income people. It will be found that totally new (to Cali) types of retailers and wholesalers will be needed. CHAPTER II The Consumer 1. Introduction In the introduction to this study, we depicted the food distr- ibution system of Cali as a set of interlinked components constituting a path along which foodstuffs flow from a source component, the supplier, to a sink component in the form of the consumer. Within the terms of reference to this study these two terminal components must be regarded as exogenous, as given, in the sense that we cannot expect to make recommendations which would require that they be changed. The consumer affects the adjacent component of the system, the retailer, in two ways. One, the volume of food purchased, defines the potential market for the adjacent component, and the other, where and under what conditions the purchases are made, determines which retailers will survive to benefit from this market. The second effect will be considered in Chapter III. In this chapter we will be concerned with J fi--_ the potential market. More specifically we will be seeking information v.___ on: 1. Domestic food consumption, in order to know the volume that flows through the system 2. How this consumption will vary in response to changes in the general level of food prices and changes in real income 3. “Intensity" of demand measured as food expenditure per city block. This information will be used in determining the potential sales for retails in various parts of the city. 50 51 The data sources of this information are the PIMUR Consumer Survey and Market Basket study. These are described in detail in the appendix to this chapter, The Consumer Survey sample comprised 629 families, of which 521 declared their income, and the Market Basket sample 240 families of which 190 declared their income. The first sample was a simple random sample from the entire population of Cali, while the second was made up of Consumer Survey respondents who offered to participate in the extra study. Statistical analysis showed no evidence of sampling bias in these two samples.1 A Note Upon Classification Criteria In this and other parts of this study we will classify consumers according to two criteria, per-capita income level and run-bdrm _ ..., .. ... ...-...-..I \- u-__ —-.—- .....r.._ —.4.-...-. . om-va-u—IM -MMJ— wM \ socio-economic level. ”Per-capita income is commonly used as a variable in econometric wdrk. This is a construct computed as family income divided by family size, or N.N.P. divided by population, or similar methods, and suffers from various defects. It would be preferable, in relating expenditures to incomes, to use income per spending unit as the independent variable. This is particularly true for such goods as are consumed by the family as a unit, such as houses and stoves, where the family is the spending unit and family income should be the independent variable. However, comnodities like food, which are consumed by family members rather than by families as a whole, can satisfactorily be related to per-capita income. lSee appendix to this chapter for a description of sampling methodology and tests for sample bias. 52 A good deal of precautionary work went into the selection of number of per-capita income levels and location of break points. Too many categories can obscure overall trends in a welter of local irregu- larities, too few can conceal important counter-trend movements. And even if the right number of categories is selected, unfortunate selection of break points can introduce the same effects. Ideally, break points should be located at places where changes in trends are found. In order to forestall poor location, tabulations of shopping patterns were made using ten categories of per-capita and family income, and carefully inspected for evidence of local patterning. No significant patterning of this type was found, so that we felt free to choose the number and locations of break-points to facilitate statistical analysis, in the absence of other factors. On this basis three breaktpoints were selected -—"* We“ to divide the samples approx1mately iInto four equal parts as follows: -.. ,‘WH‘A _7" “wt-FJQP' uniwm-J‘a rye—1"“?! Haw-m1.) NM .. 7'. IW pt-IhA-A wk” ; Income Level 1 g _3_ 4 Per-capita income range - (Col $/month) O - 125 126f 240 241 - 500 500 & over Mean Per-capita income H , vr~. (Col $/month) 86I 176 ; 352 , 1157 . A.“ \. This classification, of course, applies only to those families which revealed their income to our interviewers; The other classification we shall use is socio-economic level of the barrio (district) in which the respondent lived. The municipal government of Cali recognizes nine socio-economic groups, of which the last three (penal, military and industrial areas) are not of interest to us. Barrios are put in one of the six residential groupings according to their score on a rather complicated points system which takes into 53 account such things as: The type of housing and type of construction Level of provision of public services (water, power, garbage collection) Closeness of arterial roads Income of residents and their occupation. In chapter I it was shown how barrios of the various levels tend to form distinct regions of the city, and it was also noted that some of the prominent types of retailers are geographically associated with certain of these regions. We will look into the consequences of this in chapter III. In this chapter we shall use the socio- economic level (or w.- ‘1’ J'“' wu" v n. “"T‘Wifl‘v‘t' I! I halyuiuAWW‘vflww" S.E.L.) as the classificatory criterion chiefly when we consider the m Win-m mm..- a...- H""-’"¢iw u- M «I‘- -. - . .. v «up... ...”...3 ,. .9... . ,. , 1 potential market_jacingwretai]ers. For this purpose it has the advantage vaw'1w that S.E.L. is easily observable, while average income can be determined only through a survey, which is expensive. Since S.E.L. is geographical (defining a barrio) it is very closely connected with problems of store location in nature. It can also be argued that S.E.L., being to some extent under the control of the individual (he can choose to live anywhere he likes within his means) and being a multi-dimensional concept, is more indicative of life-style than income level, which is uni-dimensional and is essentially a given factor for each individual, assuming that he always seeks to maximize his income. We do not propose to delve further into this, except to note that analysis of the Market Basket data indicates that the two criteria are of roughly equal power. The coefficient of variation of expenditure on the main food items (the ratio of standard 54 deviation of expenditure to mean expenditure) is approximately the same whether classification is by per capita income level or socio-economic level. To complete this section Table II.l.l. shows the relationship between income level and socio-economic level. It can be seen that the two sets of criteria are strongly associated with each other. Table II. 1.1. Distribution of Per—Capita Income with each Socio- Economic Level. Cali, February 1969 Per-Capita Income Col.$/Month I II III IV V VI Percentual Distribution 0 - 125 0 0 8.0 4.7 27.5 50.1 126 - 240 0 4.0 4.4 35.8 35.2 32.9 241 - 500 0 10.7 28.0 45.0 28.8 15.6 over 500 100.0 85.3 59 6 14.5 8.5 1.4 Source: PIMUR, Consumer Survey Table II. 1.2. shows the way that these two classifications divide the population of Cali. Table II. 1.2. Population and Households by Socio-Economic Level and Per-Capita Level. Cali, February 1969 Socio-Economic Level I II III IV V \\///’VI\\ Percent of Population 4.8 4.2 7.7 12.4 50.0 3 20.9 ; Percent of Households 4.6 4.9 8.3 11.9 49.9 ’\ 20.4% Per-Capita Income Level 1 2 3 4 Percent of Population 25.5 29.1 25.3 20.1 Percent of Households 20.3 27.8 28.7 23.2 Source: PIMUR, Consumer Survey 55 2. Demographic Characteristics The population of Cali is. of Indian, Spanish and Negro blood, and of all mixtures in between. As can be imagined from its rapid growth, about half of the population is not native to the city, coming instead from other parts of Colombia. The Consumer Survey showed that 46% of the population was born in Valle, 24% in the group of Departments north of Valle (the so-called "Paisa Group"). 14% from Narino and Cauca south of Valle, and the remainder (16%) from other parts of Colombia. Of those born in Valle, a substantial proportion was born outside Cali. Only 33% of heads of the households interviewed had lived all their lives in Cali. This later figure underestimates the proportion of Cali-born heads of households, since it excludes those who have lived for a while elsewhere, but the error is probably small. It was believed that the immigrants to Cali, thought to be poor and untrained, would be concentrated in the lower socio-economic levels. This belief was not borne out by the Consumer Survey, which showed the proportion of immigrants to be practically the same at each level. Consistent with this is the observation that only 13% of the heads of households were born in rural areas. This is probably an underestimatel, but it appears that the migration flow is not a simple flow from the countryside to the cities. “Rather it would seem, there is also a substantial flow from towns to cities. While this contradicts the simple model of Cali's growth described in the appendix to chapter I, 1Two factors could contribute to this: first a reluctance to be considered a country bumpkin could result in misrepresentation, and second. many of those declared as born in a city were probably born in small villages which for census purposes were considered urban areas, and in which customs are probably more rural than urban. 56 it does not seriously invalidate it. The towns and small cities have population pressures upon employment similar to, pOpulation pressure upon land in the rural areas, so that the numerical results of the simple model will not be greatly, if at all, in error. Education Table II. 2.1.presents figures upon education level attained by three members of each family interviewed; the household head, the housewife, and the member of the house with the highest education level. Clearly these categories overlap somewhat. In one-sixth (16.6%) of instances the head of the household was also the housewife, while table II. 2.2.shows the breakdown of “most educated family member" by Table II. 2.1. Distribution of Education Levels. Cali,February 1969 Household Family Member Education Level heads Housewives with most3 Education % % % None 1 3.9 4.5 0.2 Primary: Incomplete 31.0 40.3 16.8 Complete 1 “27.0 29.5 22.0 Secondary: Incomplete 22.6 21.1 38.7 2Complete 9.0 3.0 10.9 University 6.5 1.6 11.6 Notes: 1. "Incomplete" means not having completed the official number of years at each level (five for primary, six for secondary) 2. This category includes all university-level education, completed or not. 3. These classifications are not "pure". See table II. 2.2 and text. Source: PIfiUR, Consumer Survey position in family. The result is that column I of table II. 2.1 indicates levels slightly lower than those applicable to male household heads due to the dilution by less educated housewives; similar considerations apply 57 even more strongly to column 3. However, the data as presented are more suitable for our purpose than more strictly segregated data would be. Table II. 2.2. Family Position of Most Educated Member. Cali, February 1969 Family Position Proportion of Most Educated Members % Household head 44.4 Housewife 14.7 Son/Daughter 34.8 Nephew/Niece 1.8 Other 4.3 Source: PIMUR COnsumer Survey The figures of Tables II. 2.1.and II. 2.2.are interesting. First, they indicate that in most cases there is a family member with more education than the head of the household. In general this will be a junior member, i.e. a younger brother of the household head or one of the older children, thus the indications are that the level of education in Cali is increasing. Second, although 45% of housewives and 35% of household heads are functionally illiterate (taking this to describe those who have not completed their primary education) only one household in six does not have at least one member who is functionally literate. This is a point of some importance in considering the process of diffusion of inbrmation and innovations. Consumer perceptions of new types of food outlets should be relatively free of misconceptions and prejudices caused by ignorance and perpetuated by illiteracy. Perhaps more important, literacy makes feasible the dispersion via newspapers and fliers of factual information concerning prices and qualities. 58 Advertising by radio cannot convey such information effectively in a way that facilitates comparisons of the offers of competitors. Literacy thus makes it possible to extend to neighborhood stores the market discipline of price competition, which at present is only effective in wr .—. the public markets and amongst the stores of the Galeria Central area. “(Fun—‘- ...___. _fi " Third, preliminary results of studies of changes in value systems amongst sections of the population of Colombia by Dr. Weber and his 1 indicate that in Cali increased colleagues at the Universidad de Valle levels of education are associated with a "Westernization" of value systems. While "Westernness" is not so well defined a concept that all its implications and connotations are clear, when applied to food purchasing patterns it would certainly imply a high valuation of tangible factors such as hygiene, packaging and quality, and a low valuation of intangible psychological satisfactions of following custom. We can say, therefore, that the figures in the third column of table II 2.1 provide strong evidence that the Cali consumer is becoming psychologicallymprepared to r—vsfi—nup-nfl—“f t'l-fl " patronize StQI§§_di£fenentnjn format from the traditional tiendas and ‘W “Wm, ‘I‘-~\,.‘ public markets. ‘3 ""95 Households Since the concern of this study is with food consumption and purchasing, we took a rather specialised view of the Cali family, regarding it as a food purchasing unit rather than a social or kinship group. The words family and household are used interchangeably in this restricted sense. Thus we are_§onsidering maids‘ang.Iheit,§hil§r9n.m- M where applicable, as family or household members «AI-Iv 0W ...:M- ‘—~*"""""“ The households interviewed in the Consumer Survey ranged in size from one person to sixteen, with a modal value of five, and a mean value of 6.27. 1 Personal Communication 59 Table II. 2.3.shows how family size varied with socio-economic level Table II 2.3. Household size as a Function of Socio-Economic Level and Per-Capita Income Level. Cali, February 1969 Socio-Economic Level I II III IV V VI Number of Persons 6.55 5.35 5.79 6.48 6.27 6.43 Per-Capita Income Level 1 2 3 4 Number of Persons 7.95 6.62 5.37 5.48 Source: *PIMUP Consumer Survey and level of per-capita income. One reason why so many large households occur in Income Level 1 is that a given income divided amongst a large number of family members results in a low per-capita income, so that members of large families tend to be placed in low per-capita income groups. The variation of household size with S.E.L. reflects life-style more accurately, we believe, than does the variation with income level. Families of S.E.L. I tend to have more servants (wealthier) and are more traditional in style (inherited wealth) than those in S.E.L. 11. Those in S.E.L. II are typically professional families, with university-trained members and with a life-style more like that found in middle-class families in the U.S.. As we go down the social scale, we return to the traditional Latin life-style, with its emphasis on large "extended" families. However, this conclusion is the result of casual rather than systematic observation, and reflects only the present writer's interpretation. Age Cali is a city of the young. The Consumer Survey found that 14% of the population was less than five years old, and 43% less than sixteen years old. More directly important to us here is the age of the Cali housewife, who does the food purchasing. The average age of the housewife was 38% years, and that of the heads of households 42% years. Fewer than ten per-cent of housewives are under 24, 70% are 30 or over, 60 and 30% are over 43 years old. This age distribution and the average level of education amongst housewives would lead one to suspect that the Cali housewife is rather set in her ways, and not likely to welcome drastic changes in food retailing methods, nor to introduce changes in her family's diet. Now is an appropriate time to evaluate the foregoing in respect of food consumption patterns. To what extent will the factors described above need to be taken into account in analysing the results of the Market Basket Survey? We have already noted that the age and education of the Cali housewife are such as to make spontaneous changes in her shapping patterns unlikely,_§o that the Market Basket results are safe from invalidation on this score. Against this must be set the potential effects of improving educational levels of the population, but this will only be apparent in the long run and there will be ample time for the food distribution system to adapt to any changes from this source. At present, in only 10% of families do the children participate in decisions concerning which foodstuffs should be purchasedl. Family size and composition are other variables which can ____‘___ ’ A—-—— .1- uhwv ‘m‘FM‘L-fli‘h ‘... .M effect food consumption expressed on ma.per-capita basis. However, these I v-m‘I-u ”WM-W W PHI. .f “WI”, again are factors which will change only very slowly, although they may change faster amongst some sectors of the population (see table II. 2.3. and associated discussion). We will be content tolcheck these factors (using family size as our measure) in the Market Basket sample in order to ensure that in this respect it is representative of the population. 1PIMUR, Consumer Survey 61 At the outset we anticipated that area of origin would be an important factor in the analysis of the Market Basket data. It was believed that an ever-increasing stream of migrants was coming into the city, unskilled, poorly-educated migrants with fOOd consumption habits which, given the geography of Colombia, would be quite different from those of Cali. The results of the Consumer Survey lead us to believe that area of origin can safely be neglected. The rate of inward flow is much smaller than anticipated (a yearly rate equivalent of 2% of Cali's 1969 population, instead of 4.5%) and the majority of migrants were previously city-dwellers, and thus more cosmopolitan in their customs than country-folk. Furthermore, the process has been going on so long (since the Panama Canal opened in 1914), that by now the customs in Cali must closely approximate the average of the people exporting regions of Colombia. 3. Employment and Income Distribution Secondary data upon employment and income in Cali were not good in 1968 - 1969. The most comprehensive source was I.C.S.S. (the Colombian Social Security Institute), but although all employers were supposed by law to be registered with the institute, the large proportion of small and cottage industries hampered the enforcement of the law. In 1969 teeth were put into the regulations when it was announced that salaries and wages would only be allowed as tax deductions upon presentation of the corresponding I.C.S.S. receipts, but this of course has no effect 'upon businesses that don't appear on the tax rolls. A study made in 19681 showed that the main sources of lawful 1Centro de Estudios Sobre Desarollo Economico (CEDE), Encuestas Urbanas Sobre Empleo y Desempleo, Apendice Estadistico, Bogota, July 1968. 62 employment in Cali were manufacturing (transformation industries) with 34% of the labor force, service industries with 22% and commerce with 18%. Unemployment figures are also unsatisfactory in Colombia r— ----r There appears to be no on- going measuring of unemployment comparable to that carried out by the Bureau of Labor Statistics in the U.S. Such figures as are available are not based upon the number of people seeking employment, but rather upon the total number of employed and 1 the total number of able-bodied people. On this basis PIMUR estimated the unemployment level to be between 20% and 25%. In seven per-cent ...... ———-_-.._... ma- “"wd‘r ,..~.a._l .. ...», #1.),me 5 I I of the families interviewed no family member had a job, although only in one case did the family declare an income of zero. One question asked in the Consumer Survey was “May we know the total monthly income of the family?" (literal translation). Of the 629 families interviewed 521 told PIMUR their income, a surprisingly high response rate. Figure II.3.1.the Lorenz curve, and figure II.3.2. the cumulative family income curve, summarise the responses to this question. The Lorenz curve shows the deep sag typical of the under- develOped countries. _Fifty per-cent of families account for only 17% of all income, compared with 15% in the Recife area of Northeast Brazil (1967), 23% in La Paz, Bolivia (1967) and 29% in the U.S.A. (1964)2. The Lorenz curve for Cali is very similar to that for Recife. As well as being distributed in a highly non-uniform manner, incomes in Cali us». ...—u VM W ..-“... .. _ IPIMUR, Technical Report #7, Cali 1969. p.30 2Lorenz curves for these four economies are compared in PIMUR Technical report #7, p.59 63 1 Figure 11.3.1. Lorenz Curve. Call 1969 100 ‘04 Percent of income Famiry income.pesos per month 64 Fig. 11 3.2. Distribution oi lamiiy income.Cai see 3000‘ 1000‘ F---—r-——— r f i 1 o 20 40 Jo Percent of families (Cumulative) Per-capita income. pesos per month 65 Fig.n.3.3. Distribution oiper-capita income Cali 1969 Percent of people (Cumlative) 66 are very low. As Figure II.3.2. shows, the median family income was COL/$1,500 per month (Colombian Pesos), equivalent to U.S.$l,040 pgg_ year}. In l969 foodstuffs in Cali cost roughly half as much as in the U.S.A.,2 but even allowing for this U.S.$l,O4O is not much to support a family of six or seven fbr a year. The mean family income was Col/$2,500 a month, but so skewed was the distribution that fewer than 30% of families made as much as this, so that 70% of families subsisted on less than the equivalent of U.S.Sl,750 per year. Tables II.3.l.and II.3.2. show how the two classification methods we use are related to income. The figures for the lowest per-capita income group (table 11.3.1.) are particularly vivid; one quarter of the population has to subsist on Col.$4/day (roughly U.S. $0.25), and one-eighth on Col.$3/day. Before leaving this section we should consider the validity of these results. As we noted before, the data were obtained as answers to the question "May we know the total income of this family?", and interviewers were instructed to stress the word “total". The question was phrased in this manner to reduce embarrassment to the respondent in case of refusal, embarrassment which might have led her to refuse to answer the following part of the questionnaire. We anticipated a rather higher rate of refusal than the 18% actually encountered. 1This uses the official exchange rate of U.S.$l - Col.$l6.9O in February 1969. Using the black-market rate of $17.50 would make the comparison even more striking. 2Personal observation 67 Table 11.3.1. Mean and Median Per-Capita Income for Each Per-Capita Income Level. Cali, February, 1969 Per-Capita Income Level 1 2 3 4 Per-Capita Income Range 0-125 126-240 241-500 over 500 (Col.'$/month) Mean Per-Capita Income 86 176 352 1157 (Col.$/month) Median Per-Ca ita Income 92 167 333 893 (Col.$/month Percent of Population 25.5 29.1 25.3 20.1 Source: PIMUR, Consumer Survey Table II.3.2. Income by Socio-Economic Level. Cali, February 1969 Socio-Economic Level I II III IV V VI Mean Household Income 11,185 5,568 4,440 2,490 1,603 1,013 (Col.$/month) Mean Per-Capita Income 1,687 1,009 710 400 251 157 (Col.$/month) Percent of Population 4.8 4.2 7.7 12.4 50.0 20.9 Source: PIMUR, Consumer Survey In any survey work where a large proportion of refusals arise it is important to check whether bias is likely to be introduced from this source. Accordingly we compared the non-declarers with the total sample in respect of distribution of socio-economic levels, education, family size and proportion of Cali-born head of households. For none of these was the difference significant even at the 10% level. We feel, therefore, that the refusals were not systematic but random and that no bias will be introduced by using or neglecting the data coming from those who would not, or could not reveal their incomes. Although we asked for total incomes, we believe that the answers received pertained to only direct cash incomes and, perhaps, rental incomes. Other sources of income were probably not included. 68 On a national accounts basis the difference is significant, but to have tried to obtain information upon all sources of income would probably have been fruitless, because the interviewees were housewives who would probably not be acquainted with details of things such as dividends, social security payments and separation payments. Furthermore, to have attempted to obtain these details would have prolonged the interview intolerably. So the figures given here are not directly comparable with such figures as, for instance, per-capita gross national product. In any case, we argue that the figures PIMUR obtained are those buying on available cash, not upon employer's payments into security;nm m - "_fl 7 _ , _,, “n.1,..-“— .... ' “Wm—..-"..- 7“"- funds, nor upon future separation payments.‘ Admittedly, the existence #Mm'" of these payments reduces the need to "put something aside for a rainy day", and thus may increase the proportion of cash income that the family decides appropriate to spend on food, but this is a secondary effect. In any case, since a greater part of these payments is. pr0portional to direct cash income, the effect will only be that of slightly altering the coefficients in the regression equation relating food expenditures to true total income. Exceptions to this are the self-employed and those people, usually the lower end of the income scale, whose employers are not registered with ICSS. It is quite possible that many of the replies given to the question concerning income, were false, in greater or lesser degree. Two reasons for this come quickly to mind - desire for concealment of wealth and desire for aggrandizement, and distortions from either reason could have been introduced in what the earner told the housewife, or what the housewife told the interviewer, or both. 69 These two motivations would have opposite effects upon the responses, those seeking concealment understating income and those seeking aggrandizement overstating it, and so would tend to offset each other on the average, but they would be additive in introducing variance in the correlation of food expenditures and income. To some extent gross misstatements could be detected by comparing income with occupation (none were found), but minor ones could not. The only secondary data available against which PIMUR's data could be compared were those from ICSS, but they were not strictly comparable. For one thing the PIMUR data included the lumped earnings in those families with multiple wage earners, and for another the ICSS data did not include employees of many small businesses (usually low paid) nor the self-employed, principally concentrated at the extremes of the income range (e.g. lawyers and bootblacks). Nevertheless, a comparison was made of the ICSS data with the middle portion of the PIMUR data, and showed that the two sets of data were consistent with each other. Therefore we feel that if the biases described above are present, they at least roughly offset each other and were not introduced systematically. The comment above concerning the regression equation still stands, of course. 4. Food Expenditures and Income Source of Data. The data presented in this section comes from the PIMUR Market Basket Survey. In this survey the food purchases of sample of Cali families were recorded for an entire week: The sample consisted of those respondents to the Consumer Survey who agreed to participate in the further study. It was originally hoped that around 400 families would participate; 330 agreed to do so, but the drop-out rate was rather 70 high, and in the end only 240 families produced usable results. Of these only 190 had revealed their income. Although smaller than expected this number is sufficient for the greater part of the purposes for which the Market Basket Survey was intended. The principal casualty was the possibility of determining whether and to what extent the poor pay more for their food than the rich, because the two-way classification required (by income level and by retailer type) would make the individual samples too small to be usable. The loss is probably not great for there were some uncontrolled variables which in all likelihood would have made the results very questionable. Chief amongst these were quality of goods and honesty of weights and measures. The composition of the_Market Basket sample is discussed in the Appendix to this chapter, where it is shown that, with only minor reservations,rit can_be_takenflasflrepresentativewofuthg_pppulgtj9n f i,“ The first attempt at relating food expenditure to income used . simple first-order regression analysis for each per-capita income level. The results of this were satisfactory and consistent from level to level, but the correlation coefficients were disappointingly low. The shape of the line segments when graphed, suggested that a second order regression would better fit the situation, and this was tried using a program obtained from the Computer Center at Michigan State University. The results given by this were very satisfactory. The "best fit" equation derived in this way was v = 53.50 + 0.3424x - 0.0000872x2 where Y = monthly per-capita food expenditure X = monthly per-capita income 71 The correlation coefficient was 0.78, and its square 0.61. This can be interpreted as meaning that of the observed variance around the mean expenditure 61% can be attributed to differences in incomes. This is a high correlation coefficient; one seldom finds a human activity of this type so dependent upon one factor. The regression equation is plotted in Figure II.4.1. Also plotted on this figure is a line through the origin of slope 0.5 (the dotted line). Points lying on this line denote food expenditures equal to 50% of income. As can be seen, it intersects the regression curve at an income of Col.$280 per month, and figure 11.3.3 shows that 60% of the population of Cali spend 50% or more of their income upon food. Similarly, it can be shown that 36% Spend two-thirds (67%) or more of their income in this way. These figures are indicative of either very low productivity in the economy of the city or grossly inequitable distribution of income, or both. They certainly support the comments made in the Introduction about the importance of studies of the PIMUR type. Table II.4.1. summarizes these points. Table II.4.1. Per-Capita Incomes and Food Expenditures for Four Per- Capita Income Levels. Cali, February 1969 Income Level 1 2 3 4 Overall Mean Per-Capita Income (Col.$/month) 90.59 174.76 341.26 1058.93 377.61 Food Expenditures (Col.$/month) 80.98 109.83 163.98 288.22 152.79 Mean Per Capita Food Expenditure as of % of Income 85.6% 62.8% 48.1% 27.2% 40.5% Source: PIMUR, Market BaSket Survey The mean per-capita incomes in Table 11.4.1 differ slightly from those of Table 11.3.1. in consequence of sampling variations. Overall, 40.5% of income in Cali is Spent on fbod, compared with some Per Month) '0 O O 1 Per-Capita Food Expenditure (Pesos 72 Fig.ll.4.1. Food Expenditure Venue Income.Ca|i.1969 2504 5? ‘IOO-i ' 01 O V 51—160 260 360 460 500 600 700 800 Per-Capita Income (Pesos Per Month) 73 twenty percent in U.S.A. in 1967-1968. The comparison is not favorable to Colombia. Yet the tabulated figure is, in reality, over-optimistic because of the concentration of income in the upper income-levels. A more representative figure would be that for the individual at the median income of Col./$206 per month, whose food expenditures amount to Col./$121 per month, or 59% of his income. The usual application of equations of this type described above is prediction. In the present case, the effect of changes in real income upon food consumption can be predicated, subject to the usual assumptions of cross-sectional analysis. As it happens we shall not use this equation for such a purpose, prefering to use the first-order regressions for each per-capita income level. In justification for this we can point out that: 1) use of the straight-line first-order segments vastly simplifies the computations. Using the second order curve would necessitate tedious numerical integration, unwarranted by the accuracy of the available data. 2) the straight-line segments over their respective ranges of applica- bility, correspond closely to the second-order curve, and where they differ it would seem reasonable to believe that with a total of eight arbitrary coefficients they should fit the data better than the second- order curve with only three coefficients. The results of these calculations are presented in Table II. 5.4.(p.80) A further use which can be made of the regression equation is that of predicting the effect of blanket changes in food prices upon food expenditures. Strictly speaking, such prediction is a matter of determining price-elasticity of demand, and for this we do not have sufficient or apprOpriate data. However, the Hicks-Slutsky analysis 74 shows that the total effect of a change in prices is made up of both income and price effects. Here we are considering a generic group of goods for which the substitution (price) effect is small, a group which accounts for a large proportion of total expenditure and for which the income effect is therefore large. So we can argue that a total reduct- ion in food prices is equivalent to an increase in real income, and that the effect upon food consumption will closely correspond to that pred- icted by the regression equation for such an increase in income. Applying this, for example, to the case of an individual with an income of Col.$100 per month, we find that a 10% reduction in food prices will result in a 7% reduction in food expenditure. Furthermore, this will increase the money available for non-food expenditure by no less than 45%. For an individual with an income of Col /$200 per month the corresponding reduction in food expenditures would be just under 7%, and the funds available for non-food expenditures would increase by 10%. A Note upon the Mathematics of Regression Analysis There are certain conditions which must be fulfilled if the full range of analysis, including estimates of confidence ranges and tests of significance of coefficients, is to be applicable in studies of regression. The principal such condition is that of homoscedasticity, that the variance around the regression line be constant. This condition is violated in the present case, the data being heteroscedastic to the extent that the standard deviation of observations around the regression line is closely proportional to income (the independent variable). Johnston1 shows how this situation can be handled, at least for the case 1Johnston, J."Econometric Methods", New York, McGraw-Hill 1963, p.210 75 of linear regression, by a simple transformation of variables, but the resulting equation, involving the reciprocal of the independent variable, is awkward to use. The main effect of violating the homoscedasticity equation is to make the analysis too rigorous - coefficients of significant importance could be indicated as non-significant by the formal analysis. In the present case the formal analysis shows the coefficients to be highly significant, so there is no need to go through the transformation. Another condition that the formal analysis specifies is that of pre-determined values of the independent variable. This however, is a pure formality, specified only because without it the concept of variance around the regression line cannot be formally defined. The main point here is that if we take as our criterion of "best fit“ that of minimizing the squared deviations around a line a least squares regression provides the "best fit", irrespeCtive of whether the formal mathematical conditions are satisfied. 5. Expenditures Upon Selected Food Products For the purpose of the Market Basket Survey 86 products were identified, including six "others", one for each of meats, grains, dairy products, vegetables, fruits and processed foods. These six groups were considered separately because they have distinct distribution channels, and each group formed the subject of a commodity-flow studyl. 1These studies are described in PIMUR Einal Report 76 All told these "other" groups accounted for only 2.37% of food expenditures, indicating that the 80 specifically indicated foods were well chosen. Several of these products were sub-classified for the benefit of other researchers in the PIMUR program. For example, fourteen cuts of beef and three kinds of potatoes were recognized. Clearly amongst so many foodstuffs, there must be a sizeable number whose importance in the typical Cali family's budget is small. Also in discussing so many items one tends to find onedsview of the forest obscured by trees. Analysis of the Market Basket data showed that there were twenty-one products which each accounted for one percent or more of food expenditures in Cali, and which together account for 77.7% of all food sales. We will confine our discussion to these products, together with six "others", one for each of the groups listed above. Table 11.5.1.1ists weekly food expenditures of these selected products, together with their relative importances in the food budget, by per-capita income level, while Table II 5.2 does the same but with breakdown by socio-economic level. In this latter table levels I, II, and 111 have been aggregated because for each of these levels the number of families sampled was too small to provide separately useful data. Various points of interest arise in these tables. Lard, for example, is an inferior good which is rapidly replaced by cooking oil, as income rises. Surprisingly there is no such substitution between unpasteurised (raw) and pasteurised milk (bottled), indicating at least 77 00>00m 000000 000002 02200 0000zom 0.000 00.000 0.000 00.000 0.000 00.000 0.000 00.000 0.000 00.00 00000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 000000000 00000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 00000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000 000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000 000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 000000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 000 0000000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 000000 00000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 .000 00000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 000000 00000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 00000 0.0 00.00 0.0 00.00 0.0 00.00 0.00 00.00 0.00 00.0 0000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 00000 00000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 0.00000 0000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 03000 000: 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 0000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 0000: 00000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 0000000 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0.0 00.0 0000 0.00 00.00 0.00 00.00 0.00 00.00 0.00 00.00 0.00 00.00 0000 0.0 00.0 0.0 00.00 0.0 00.0 0.0 00.0 0.0 00.0 0000 0 0.000 0 0.000 0 0.000 0 0.000 0 0.000 :mgm>o w m N H _.m>m._ mEoucH Agucoe\mvmpm>00 mEouc0 0000001000 00 .0003000000xm 0000 000000-000 0000002 .000 .H.m.00 mpnmh 78 xm>02m 000000 000002 02200 0000000 000 u H 0.000 00.000 0.000 00.000 0.000 00.0000 0.000 00.0000 00000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.000 . 000000000 00000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 00000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00 00 . 00000 000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.0 0000 0.0 00.0 0.0 00.0 0.0 00.00 0.0 00.00 00000 000000 0.0 00 00 0.0 00.00 0.0 00.00 0.0 00.00 000000000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 000000 0.0 00.00 0.0 00.00 0.0 00 00 0.0 00 00 00000 0.0 00.00 0.0 00 00 0.0 00.00 0.0 00.00 000 0000000 0.0 00.0 0.0 00.00 0.0 00 00 0.0 00.000 000000 00000 0.0 00 00 0.0 00.00 0.0 00.00 0.0 00.00 00000000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 .000 00000 .0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 00000000 0.0 00 00 .0.0 00.00 0.0 00.00 0.0 00.00 00000000 0.0 00.00 0.0 00 00 0.0 00.00 0.0 00.00 000000 00000 0.0 00.0 0.0 00.0 0.0 00.00 0.0 00.00 0000000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 00000 0.00 00.00 0.0 00.00 0.0 00.000 0.0 00.00 0000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 00000 00000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 0.00000 0000 0.0 00.00 0.0 00.00 0.0 00 00 0.0 00.000 00000 0000 0.0 00 00 0.0 00.00 0.0 00.00 0.0 00.00 0000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 00000 00000 0.0 00.0 0.0 00.00 0.0 00.00 0.0 00.00 0000000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 0000 0.00 00.000 0.00 00.000 0.00 00.000 0.00 00.000 0000 0.0 00.00 0.0 00.00 0.0 00.00 0.0 00.00 0000 0 0.000 0 m.0¢u 0 0.000 0 0.000 00>m0 00Eocoum-o0uom 0000 00020000 .0000 .Fm>m0 00500oum-o0uom 00 0000000000xm 0000 000500 0000002 .m.m.00 m0300 . -, . L . m Q7K 4 ‘ M, pH N N ‘K r . nus .NI!V..-.JTV.,AN -- swab‘u‘vy‘hf : nVPL -\:~.UA~ ..‘C r \A; u A u J‘JvaLQ “dead ‘ Lat u ‘ usa...l-s;U.. . agingd . ..v p to)”. A «UZFVU: n .5.» p . . Gnu )LCJL.~.».. v» 79 00>00m 000000 000002 03200 0000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 000000000 00000 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0 00000 0000.0 0000.0- 0000.0 0000.0- 0000.0 0000.0 0000.0 00000 000 0000.0 0000.0- 0000.0 0000.0- 0000.0 0000.0- 0000.0- 0000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 00000 000000 0000.0- 0000.0- 0000.0- 0000.0 0000.0 0000.0- 0000.0 000000000. 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0- 0000.0 000000 0000.0 0000.0 0000.0- 0000.0 0000.0- 0000.0 0000.0 00000 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0 - 000 0000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0- 000000 00000 0000.0 0000.0 0000.0 . 0000.0 .0000.0 0000.0 0000.0 00000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 .000 00000 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0 0000.0 00000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 00000000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 000000 00000 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0 0000000 0000.0- 0000.0 0000.0 _ 0000.0 0000.0 0000.0 0000.0 00000 0000.0 0000.0- 0000.0 0000.0 0000.0 0000.0 0000.0 0000 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0- 00000 00000 0000.0 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0- 0.00000 0000 0000.0- 0000.0 0000.0- 0000.0 0000.0 0000.0 0000.0 00000 0000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0- 00000 00000 0000.0- 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000000 0000.0- 0000.0 0000.0 0000.0 0000.0- 0000.0 0000.0 0000 0000.0- 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000 0000.0- 0000.0 0000.0 0000.0 0000.0 0000.0 0000.0 0000 00000 000 00000 000 00000 000 00000 0 0 - 0 0 0 - 0 0 0 - 0 0 0000000 00>m0 050000 000000-000 0000 00000000.000u .00m>m0 050000 000000-000 00 .00000000 0000 000 000500 00 000000000000-000000 .0.0.00 00000 80 ' Tab1e 11.5.4, Projections of Month1y Consumption of Se1ected Food Products in Ca1i for 1969, 1979, and 1989 Under Two Assumed Rates of Growth of Income Month1y Consumption ( 000's of 1969 pesos) Year 1969 1979 1989 Income Growth Rate (% per year) - 1 1.5 1 1.5 Product: Pork 3687 5862 5994 9316 9771 Beef 26280 41550 42368 65666 68481 Fish 2183 3580 3717 5867 6337 Chicken 2740 ‘ 4358 4457 6928 7267 Other Meats 4008 6525 6751 10617 11393 Eggs 3703 5819 5915 9138 9469 Mi1k éRaw) 5247 8492 8760 13738 14662 Mi1k Past.) 4924 7720 7840 12098 12508 Other Dairy 4177 6485 6552 10058 10287 Rice 10159 15781 15948 24488 25060 Beans 2626 4159 4244 6584 6877 Lenti1s 1588 2578 2572 3992 4176 Other Grains 2665 4267 - 4378 6829 7212 Potatoes 3516 5570 5686 8822 9219 Tomatoes 2069 3295 3372 5245 5509 Other Veg. 4926 7895 8105 12648 13371 P1antain 4123 6522 6652 10313 10761 Other Fruits 5876’ 9291 9474 14684 15314 Cooking 0i1 3437 5466 5590 8689 9116 Sugar 2786 4267 4281 6523 6570 Coffee 2122 3368 3441 5344 5594 Choco1ate 1404 2324 2424 3843 4186 Canned Foods 1930 3280 3462 5554 6181 Lard 2250 3634 3746 5869 6251 Pan Sugar 3542 5488 5538 8492 8664 Break 5527 8610 8713 13399 13754 Other Processed -6444 10638 11080 17540 19059 Tota1 123941 196766 201060 312281 327051 Source: PIMUR Market Basket Survey Note: Popu1ations for 1969, 1979, and 1989 taken as 813,000, 1,238,000 and 1,880,000. 81 that the price difference between the two is not a factor inhibiting the use of pasteurised mi1k in Ca1i. The 1uxury nature of canned foods is c1ear1y shown, and the same is true of fruit. This 1ast rather dispe1s the romantic notion of tropica1 countries as p1aces where exotic fruits are to be had a1most for the asking. Yet these fruits do grow a1most wi1d. This on1y goes to show the high marketing costs of matching the scattered distribution of the fruit with the concentrated demand of a 1arge city. Beef is by far the most important item in the food budget, accounting for 21.5% of food expenditures. Beef, mi1k and rice together make up 38% of this food budget. No other item 1isted is of great individua1 importance in the fami1y food budget in Ca1i. This means that efforts expended in changing their particu1ar distribution channe1s or production conditions wou1d not have much effect upon tota1 Ca1i food expenditures, nor on the cost of 1iving in the city. Attempts to reduce the cost of 1iving by reducing food prices shou1d be aimed at changing the operating conditions of 1arge segments of the city's food distribution system, so that the benefits of more efficient distribution methods can be ref1ected in 1ower consumer prices for a wide range of products. The cost, in terms of pub1ic funds, of setting up a new type of country assemb1er of mangos, for examp1e, wi11 be of the same order of magnitude as that of setting up a new type of urban food who1esa1er or retai1er. But two types of who1esa1er and two types of retai1er shou1d be enough to reconstitute the entire urban food distribution system, whi1e forty or more new institutions wou1d be required mere1y to improve the rura1 assemb1y practice for fruits and vegetab1es. 82 Tab1e II 5.1.wi11 be of use 1ater, when we make estimates of the potentia1 demand for food in various sectors of the city. ' As part of the ana1ysis, first order regressions were ca1cu1ated of expenditures upon each item as functions of income, for each per-capita income 1eve1. From these in turn were ca1cu1ated point income e1asticities of demand for each product at the centroid of each income 1eve1. as we11 as arch e1asticities between adjacent pairs of income 1eve1s. These figures are presented in Tab1e II.5.3.principa11y as reference materia1]. A1so ca1cu1ated and presented in Tab1e II 5.4 were tota1 month1y consumption figures for the various foodstuffs for the years 1969, 1979, and 1989. These figures are based upon income inoreases of 1% per year at a11 1eve1s. and popu1ations of 813,000 for 1969, 1,238,000 for 1979, and 1,888,000 for 1989. The popu1ation figures were computed 'according to the mode1 described in Appendix I.of Chapter II. Imp1icit in the ca1cu1ations were the fo11owing assumptions: a) that the distribution of income wi11 not change as the pOpu1ation grows b) that tastes wi11 not change in the intervening period c) that a11 prices remain unchanged The e1asticities 1isted in Tab1e II. 5. 3-vary in dramatic fashion from 1arge to sma11 and from positive to negative, whi1e the consumption figures of Tab1e II.5.1.from which the e1asticities were ‘ca1cu1ated show much smoother and more regu1ar patterns. The reason 1They are s1ight1y different from those in the PIMUR Fina1 Report, because the data used have been c1eaned up, a process which had not been comp1eted at the time the PIMUR report went to press. The differences are unimportant. . ~ 83 for this 1ies in the definition of e1asticity as the ratio of proportiona1 change in the independent variab1e. ComputationaTTy, this comes out as : e1asticity = sTope of regression 1ine X income + consumption (in the present case of income e1asticity) When consumption is on1y a sma11 proportion of income, the resu1t is to mu1tip1y the regression 1ine s1ope by a 1arge factor (100 in the case of a product accounting for 1% of income). With such sma11 consumption items, sma11 abso1ute errors of measurement or sma11 samp1ing errors resu1t in very 1arge errors when the resu1ts are expressed as e1asticities. For this reason the consumption projections of Tab1e II 5.4 were ca1cu1ated direct1y from the regression 1ine coefficients, a1though the changes in income are so sma11 that the errors resu1ting from using e1asticities wou1d be sma11 a150. 6. Transportation Costs Many consumers in Ca1i do not do their food shopping 10ca11y, but rather at stores which are not within wa1king distance from their ' homes. The transportation costs invo1ved in this shou1d be counted as food costs if we are to estab1ish a basis for comparing the performance of the existing system with that of any a1ternative system. Tab1e II. 6.1.presents some figures concerning this aspect of food purchasing Tab1e 11.6.1. Consumer Transportation Costs by Socio-Economic Leve1. Ca1i, February 1969 Socio-Economic Leve1 I-III IV V VI Overa11 Proportion of Fami1ies using Transportation % 84% 43% 51% 57% 55% Neek1y Transporation cost per Fami1y Co1./$ 5.13 3.75 4.56 7.20 5.11 Transportation Costs as ercent of Food costs 1.27% 1.32% 2.48% 4.85% 2.38% Source: PIMUR Consumer and Market Basket Surveys 84 The tab1e shows that the cost of transportation adds 2.38% to the tota1 domestic food expenditures of the Ca1i popu1ation, and that this percentage is doubTed for the fami1ies of the Towest socio-economic 1eve1. Nhi1e these tranSportation expenses might to some extent be associated with other activities than food purchasing, for examp1e trave1 to and from work, the context and phrasing of the questionnaire makes this un1ike1y1. A point to be observed in the tab1e is that the figures for S.E.L. I - 111 are unrea1istica11y 10w, principa11y because peop1e who did their trave11ing by private automobi1e natura11y dec1ared their transportation costs as zero. This method of transportation was the principa1 one found in S.E.L. I, and was common amongst fami1ies of S.E.L. II and 111, whi1e in the remaining socio-economic groups it was virtua11y absent. The second row of tab1e 11.6.1. expresses transportation costs averaged over a11 fami1ies in each S.E.L. But not a11 the fami1ies in each S.E.L. shop non-1oca11y (e.g. 51% of fami1ies in S.E.L. V), so that the transportation costs of each fami1y that shops non-1oca11y goes up in inverse proportion. That is, the average fami1y of S.E.L. V. that shops non-10ca11y spends 4.56 e .51 = Co1./$8.96 on transportation each week, which is about equiva1ent to about 5% of the food bi11 of the average fami1y of this socio-economic 1eve1. 1The question was phrased "The 1ast time you did your main food shapping what transportation did you use?". Trave1 on food was an accepted, and ear1ier in the interview the distinction had been made severa1 times between main shopping and other shopping. 85 Summary The purpose of this chapter was to study the food consumption habits of the popu1ation of Ca1i, and such demographic variab1es as affect food consumption. It was found that expenditure upon foods was c1ose1y re1ated to income and to the socio-economic 1eve1 of the barrio — .n... . "a... tug... MWmmmmwn—n ,mmmmumm’ m‘lhil lav-uh 4 of residence. This is of importance in considering the prospects for M neighborhood food retai1ers. It was a150 shown that the effects upon food consumption of migration into the city cou1d safe1y be considered negTigib1e. An important point is that food_expenditures are concentrated fi— _upon re1ative1xnfew products, Beef a1one accounts for 21.5% of a11 food expenditures, beef, mi1k, and rice together account for 38%. Twenty- one products absorb 77.7% of the money spent upon food, and eighty products absorb 97.6%. Even taking into consideration the various grades of each product, this indicates that a food retai1er wishing to cater for substantia11y a11 of his c1ients' food requirements wi11 not find difficu1ties arising through having to maintain a very wide range of stock. Simi1ar considerations app1y to who1esa1ers. 1 The importance of this study was amp1y confirmed by the data showing the very 1ow income of a1most a11 inhabitants of Ca1i, and the very 1arge proportion of income that is spent upon food. The secu1ar1y-improving 1eve1 of education of the popu1ation { suggests that new forms of retai1er who offer rea1 advantages over existing forms wi11 not encounter prob1ems of acceptance through their 1 potentia1 customers' being tradition-bound. A 1arge proportion of consumers prefer to buy food e1$ewhere than at the 1oca1 store, thus incurring transportation expenses which I I“ .. ... _ -‘--‘ -. dll' 86 must be_addeq_to_food expenditures. “These transportation expenses are not inconsiderab1e, and are greater, both in proportion to food expenditures and abso1ute1y, for the poor who 1ive on the periphery of}: the city. 4: Apart from meats and rice, the Ca1i housewife spreads her food expenditure over a variety of products so that no one product accounts for a significant proportion of these expenditures. Improve- 1 ments in food distribution, therefore, must come through making food- ' ~-—--——-~ distributing institutions more efficient rather than attempting ”‘7' ’ ...f -14.- _ _._ -.-. , ‘u‘wfim-,fi .9 piecemea1 improvements based on a commodity-f1ow approach. awn-mmf- /7 _ ...-“f I-l‘“_~‘n..u' .... CHAPTER III Consumer-Retai1er Interface The purpose of this chapter is to determine those specifications of a food distribution system for Ca1i which are functions of the purchasing behavior of the consumer. Chapter II dea1t with the tota1 vo1ume of foods to be hand1ed, with price sensitivity, and with the food consumption of fami1ies of the various socio-economic 1eve1s. We are concerned here with the food purchasing process, the interface between consumers and retai1ers. The purpose of this chapter is to test the hypothesis that high-cost modes of Operation and/or uneconomica11y sma11 sca1es of :operation‘are-forced upon retai1ers by the requirements of consumers. This is crucia1 to the who1e study. If the hypothesis is true we wi11. be unab1e to reduce food prices by introducing changes at the retai1 1eve1. If the hypothesis is unsupported, possibi1ities for improvement at the retai1 1eve1 exist. To test the hypothesis we first endeavor to identify the factors that are determinants of the purchaSe process, and then discover what are the requirements and preferences of the consumer in reSpect of these factors. We can diSprove the hypothesis if we can show that these 87 88 requirements and preferences can be satisfied by retai1 out1ets which are compatib1e with a distribution system more efficient (i.e. 1ower- cost) than the existing one. In addition, we sha11 be we11 on our way to the design of an improved system. The testing of the hypothesis thus fa11s into two stages, ana1ysis and synthesis. The present chapter comprises the ana1ytic stage,-and Chapter V the synthetic.‘ The importance of this part of the study can hard1y be over- estimated. Our u1timate objective is to show how the Ca1i consumer can be supp1ied with food at Tower prices than present1y obtain. It is not difficu1tIto-’desiflgnwa‘food distribution system that can do this, but un1ess the consumer is ab1e and wi11ing to patronise the retai1ers of this pr0posed system the exercise wi11 be point1ess and, if the recommendations are put into effect, even damaging. This is not a pure1y academic point. As we sha11 see in Chapter IV the stores that have been brought into existence for the specific purpose of se11ing foods at 10w prices (c00peratives, Cajas de Compensacion, IDEMA) have not had any significant effect upon food retai1ing in Ca1i. The resources that have been invested in these stores wou1d a1most certainTy have provided more socia1 benefits had they been used in other ways, such as in pension funds. There are two main 1ines of attack upon the prob1em of finding the chief determinants of consumer food-purchasing behavior. The first consists of comparing observed purchasing patterns with ...-0" observab1e demographic characteristics such as socio-economic 1eve11 TIn genera1, socio~economic 1eve1 is difficu1t to measure. In Ca1i, however, it becomes an attractive c1assificatory criterion because the barrios are officia11y categorized into six socio-economic eve1s. «:1 ,- r . E.- 89 and 1eve1s of income and education.‘ This method has the advantage that the*data are-re1ative1y easy to co11ect and are not 11ke1y to be inva1idated by such factors as poor respondent comprehension or interviewer bias. The disadvantage is that the resu1ts of this ana1ysis on1y indicate the types of stores patronised by the various categories of reSpondents. Any interpretation of the resu1ts in terms of why_these ' patronisation patterns exist must necessari1y depend upon, and be co1ored by, the ana1yst's perceptions of the types of stores concerned. This perception may not correspond with the perceptions of the respondents. The second approach e1iminates this difficu1ty. It consists of re1ating observed purchase patterns direct1y to consumer perceptions of the avai1ab1e stores. But it brings with it the difficu1ties invo1ved in measuring perceptions. These difficu1ties are twofo1d. The dimensions of the perception have to be identified, and a sca1ing instrument has to be devised to measure the intensity of perception a1ong each dimension. \Both approaches are used in this study.‘ Data requirements for the first approach were easi1y satisfied by straightforward questions in the Consumer Survey schedu1e. Devising the questions for the second approach was more comp1icated. The dimensions of perception apprOpriate here are those retai1er attributes which affect purchasing behavior. The attributes considered were chosen on the basis of compromise between comprehensive- ness and respondent exhaustion. That is, out of the virtua11y Open- ended set of retai1er attributes which cou1d conceivab1y inf1uence buyer behavior a sub-set was se1ected such that: 90 a) the 1ist was not 1onger than the respondent cou1d reasonab1y be expected to answer whi1e sti11 paying attention to the questions, and b) each attribute in the sub-set was be1ieved to have more inf1uence upon purchasing behavior than any attribute omitted. The se1ection of the attributes in the sub-set, whi1e in the fina1 ana1ysis subjective, was guided by accepted marketing theory, by 1oca11y-he1d be1iefs that were encountered in the pre1iminary investigations, and by topics suggested in other treatises on food distribution in Latin America.1 The attributes fina11y se1ected were, in order of their appearance in the interview schedu1es: Qua1ity of foodstuffs offered Congenia1ity of c1iente1e Provision of credit Provision of home de1ivery Hours of service C1ean1iness and tidiness of the estab1ishment Honesty in respect of weights and measures Price reductions and discounts Cost of transportation 0. Enab1ing the customer to se1ect the items to be purchased 11. Courtesy and attention 12. Variety of foodstuffs offered 13. Convenience of 1ocation 14. Prices of stap1es 15. Prices of non-stap1es axoooxamm-bwm—o Qua1ity, variety, 1ocationa1 convenience, and prices are important attributes of food retai1ers in a11 societies. C1ean1iness and tidiness, and honesty in respect of weights and measures are usua11y taken for granted in deve10ped countries, but in Ca1i there 1 See, for examp1e, Ga1braith, John Kenneth, and Ho1ton, Richard H. Marketing Efficiency in Puerto Rico. Cambridge, Mass.: Harvard University Press, 1955. p.69. 91 are great variations in the c1ean1iness of stores between types and between stores of the same type, and retai1ers' sca1es are wide1y reputed to be "doctored". Service and attention, and congenia1ity of c11ente1e, which are re1ative1y unimportant in the 1arge impersona1 supermarkets of the deve10ped countries, might be of major important in the face-to-face conditions of a market sta11 or sma11 neighborhood store. Se1f-se1ection (as compared with se1f-service) and price concessions cou1d we11 be important factors in the housewife's image of herse1f in her ro1e of the fami1y purchasing agent. The question of hours of business was suggested by the observation that c1osing hours for food out1ets in Ca1i vary from midnight (for some tiendas) to mid- day (for the greater part of the sta11s in the Ga1eria Centra1).Was this ref1ected in purchasing or customer-preference patterns? The be1ief that credit was very important to users of the neighborhood stores was frequent1y encountered in the pre1iminary investigations. This, as we11 as home de1ivery, were cited as important factors in food purchasing in Perto Rico by Ga1braith and Ho1ton1. It was considered desirab1e to test the importance of these factors. Deve10ping a measure of the intensity with which retai1ers were perceived to possess these attributes was rather comp1icated. The most obvious method wou1d have been to ask respondents to rate each retai1er type on each attribute using, say, a five- or seven-point 1o .cit. But Hen1ey and Farace found in La Paz that these factors, and the "socia1" factors such as the function of the 1oca1 store as that of a meeting p1ace were of 1itt1e importance. (Dona1d S.Hen1ey and R.Vincent Farace, "Consumer Buying and Communication Patterns in Bo1ivian Urban Food Retai1ing: A Pre1iminary Report". Proceedings of American Marketing_Association (Washington D.C., Winter 1967) ‘ 92 semantic differentia1 sca1e. But this wou1d have required the fi11ing- in of 300 questionnaire items (twenty retai1er types times fifteen attributes). The next option wou1d have been to confine the retai1er types considered to those patronized by the respondent, but this wou1d give no indication about why she rejected the other retai1er types. In any case, the Latin-American "mentepositiva"("positive thinking", a tendency to see everything as either b1ack or whi1e, with no shades of gray).I makes the semantic differentia1 a doubtfu1 too1 in these societies. There is too great a risk that many of the responses wi11 be based upon b1essing or damning by association. That is to say, if a retai1er is found to be good (or bad) in respect of one important attribute, the respondent, through rationa1ization or other menta1 process, is 1ike1y to consider him good (or bad) in respect of some, or even a11, other attributes. Whi1e the resu1ting attribute-intensity profi1e is undoubted1y the one upon which the consumer bases her purchasing decisions there is no way in which the researcher can determine which was the key attribute in her case. An a1ternative approach wou1d have been to abandon the pseudo- interva1 measure of the semantic differentia1 sca1e and to ask respondents to rank retai1ers in order of possession of attributes. Here again, 300 questionnaire items wou1d have to be answered, items which are, if anything, more difficu1t to answer than those of a semantic differentia1 sca1e. Therefore this approach Was rejected, on the grounds that respondent exhaustion wou1d have vitiated the answers. 1The origin of this phrase is not known. The author 1earnt it from his Co1ombian co11eagues. 93 In the event, a reduced-ranking technique was used. Respond- ents were asked which retai1er type was, in their opinion, best and which was worst in reSpect of each of the fifteen attributes. "Don‘t know", "A11 the same", and "Litt1e difference between types" were a1so acceptab1e responses. 0rigina11y, respondents were asked about three categories, best, second best, and worst, but this was changed after the pre-tests, when the second best category e1icited a 1arge proportion of "Don't know" responses. Interviewers reported that reSpondents appeared to have some difficu1ty in understanding the concept of "second best". This may be another manifestation of the mente positiva, with peop1e identifying on1y the outstanding1y good or bad performers, or it may mere1y mean that in Ca1i housewives make their retai1er-se1ection decisions on a narrow information base. Whatever the reason, it can be argued that the three-point sca1e used (best, worst, and other) contains the information which forms the data input to the housewife's decision- making process. The responses to the questions described above, together with information upon where respondents bought their food and other demographic data (income, barrio of residence, etc.), wi11 provide the bu1k of the source materia1 for this section. Other materia1 is derived from the Market Basket Study. In this chapter we wi11 use, with few exceptions, socio-economic 1eve1 as the criterion of c1assification for consumers. Pre1iminary ana1ysis of the data revea1ed that patterning in shopping behavior was more apparent under this c1assification than under c1assification by income, either per fami1y or per capita]. 1See Chapter 11, pages 52 and 53 for a description of how socio-economic 1eve1 (S.E.L.) is determined by the Ca1i municipa1 Government. 94 We wi11 use severa1 different c1assifications of retai1er types, according to the point to be i11ustrated. Ii wi11 be usefu1 at this point to review the various types of store that can be distinguished in Ca1i, and to see how the types can be meaningfu11y aggregated. The most obvious aggregation is that of pub1ic markets, comprising the Ga1eria Centra1 (inc1uding the two Ca1varios markets and the associated street vendors) and the sate11ite markets. Next W“— “ H come the neighborhood stores - the tiendas and sma11 graneros7 The . third group,ise1f-service stores,\is rather more comp1icated in that it contains both food and nonefood stores and profit-oriented and non- profit oriented stores.j This group is made up of the supermarkets and chain stores (profitforientedflsub-group) and the c00peratives and stores of the Cajas de Compensacion (non-profit oriented sub-group). In addition to a se1f-service format these storesflresembTe“each‘other in that they are 1arge, departmenta1ized, and new (in the sense of being recent arriva1s on the food retai1ing scene) enough that they have not yet deve10ped a standard or traditiona1 mode of operation. None of the remaining types of food out1et can be meaningfu11y c1assified with any other type. Fortunate1y, on1y two are of importance, the truckers hand1ing mi1k, and the specia1ty stores, of which those dea1ing in beef and pou1try are the on1y ones with a significant tota1 vo1ume of sa1es. The remaining types are of neg1igib1e importance in food retai1ing. For the most part we sha11 aggregate truckers, specia1ty stores, and the rest into the category "others", but for some _._...-.-._- —-______ _,... ...—.— -" "--— purposes we wi11 find it usefu1 to iso1ate individua1 types. The most obvious examp1e wi11 be when we discuss purchase frequency. Many pe0p1e 95 buy mi1k dai1y from truckers, and to inc1ude truckers in "others" in this context wou1d give an entire1y fa1se picture of purchase frequ- encies for this group. Another c1assification criterion which shou1d be kept in mind whi1e reading this chapter is that of type of 1ocation. Three categories of 1ocation can be distinguished, centra1, neighborhood, and random.‘ The centra1 stores comprise the Ga1eria Centra1, the ___.._. ...-v”, traditiona1 stores in the vicinity of the Ga1eria Centra1 (the 1arge graneros and the who1esa1er/retai1ers, which wi11 be discussed in Chapter IV), and many of the chain stores which are 1ocated in this area. Neighborhood stores are the sma11 graneros, the tiendas, supermarkets (which are 10cated in or near barrios of S.E.L. I and II), and the truckers. Sate11ite markets a1$o fa11 into this group. Random1y 1ocated stores are those whose 1ocation bears no apparent re1ation to where their customers 1ive nor to main transportation routes. The main examp1es of this type are the cooperatives and the Caja stores, but many of the stores which were c1assified under "others" above a1so fa11 into this category. These random1y-1ocated stores tend to concentrate in three areas, if the word "concentrate" can be app1ied in the context of randomness. They are found in the Versa11es barrio, north of the River Ca1i, a1ong the Ca11e 15, and on the periphery of the city center. (See Figure 1-3) This 1atter Tocation is presumab1y dictated by rea1-estate prices, and can be rough1y defined as being rather beyond wa1king-distance of the centra1 area, the area around the Ga1eria Centra1. 96 ];_Store Patronization Patterns Tab1es III.1.1 and 111.1.2 on page 97 sunmarize the patronization patterns discovered in PIMUR's Consumer Survey. Tab1e 111.1.1. is based upon the answers to the questions "Where do you make your principa1 purchases of X", where X was in turn grains, meat canned foods, fruits and vegetab1es. mi1k, eggs, and chicken, and Tab1e 111.1.2 is based upon answers to the subsequent question "How much did you spend (at each type of store patronized) during the 1ast two weeks?1 Various trends, dependent upon socio-economic 1eve1, are apparent in the patronization patterns. Neighborhood stores are used principa11y by the poorer c1asses, se1f-service stores by the upper —.-—-1_____._—--—-— 1 _l-H- c1asses, and the pub1ic markets by the midd1e c1asses. Many exp1anatory mechanisms cou1d be postu1ated to account for these trends, but the best, in the sense of being at the same time the simp1estand the most power- fu1, is to postu1ate that retai1er se1ection is based upon 1ocationa1 convenience for the customer. On this basis we can exp1ain the trends noted above by noting that supermarkets, which constitute an important component of the se1f-service group of stores, are 1ocated in or near upper-c1ass barrios, that a1most a11 of the midd1e-c1ass barrios are 1ocated within easy reach of pub1ic markets, and that the great bu1k of 1ower-c1ass barrios are poor1y served by other than neighborhood stores, tiendas and sma11 graneros. This behaviora1 assumption can exp1ain many more trends than the three mentioned above. Consider the shopping patterns exhibited by pe0p1e of S.E.L. I. First, notethat to these pe0p1e the cost of 1It shou1d be noted that by confining the respondent's attention to principa1 purchases (those associated with p1anned sh0pping trips) these questions tend to diminish the importance of the neighborhood stores, by e1iminating routine purchases of bread and mi1k and such unp1anned purchases as they obtain by virtue of their convenience of 1ocation. 97 Tab1e III.1.1: Percent of Fami1ies in Each Socio-Economic Leve1 Patronizing Different Types of Retai1 Stores Socio- . Economic Neighborhood Centra1 Sate11ite Se1f- Other Leve1 Stores Market Markets Service Stores 1 31.0 41.4 6.9 100.0 48.3 11 64.5 77.4 29.0 77.4 45.2 111 46.2 94.2 36.5 44.2 48.1 IV 76.0 92.0 16.0 49.3 34.7 V 80.6, 64.6 28.7 30.6 20.1 VI (80.5) .71.9‘ 13.3 21.9 13.3 Overa11 74.1 71.4 23.7 37.7 25.3 Source: PIMUR, Consumer Survey Tab1e 111.1.2: Percent of Food Expenditures in Each Socio-Economic Leve1 Spent in Different Types of Retai1 Store Socio- Economic Neighborhood Centra1 Sate11ite Se1f— Other Leve1 Stores Market Markets Service Stores 1 14.4 19.8 3.0 . 32.7w” 18.3 11 24.4 25.9 15.2 17.4 6.7 III 11.6 30.2 25.2 7.0 14.4 IV 31.8 ..31.9> 8.7 13.8 ' 6.8 V .40.6’ 21.8 20.3 7.9 5.1 VI t44.8‘ 26.5 12.0 7.8 6.4 Overa11 33.4 24.8 16.3 11.5 7.7 Source: PIMUR, Consumer Survey 98 taxis is a negTigib1e item of expense, and that many of them have private automobi1es. They a1so have househo1d he1p. To them, 1ocationa1 convenience is not a matter of time or distance; their choice of retai1er is bounded on1y by their know1edge of the retai1ers avai1ab1e, which shou1d be extensive, given their high 1eve1 of education and socia1 and physica1 mobi1ity. On this basis our exp1anation of their use of se1f-service stores above becomes fa11acious, and we must attribute this phenomenon to other retai1er characteristics. Tab1e 111.1.3a shows that peop1e in S.E.L. I use se1f-service stores principa11y as supp1iers of grains and canned goods, the Ga1eria Centra1 for fresh fruits and vegetab1es, whi1e for meat they go to a neighborhood butcher’or one of the specia1ist meat stores inc1uded in "others". In the se1f-service group, they divide their patronage between supermarkets and the Caja stores, a point which is not made c1ear in Tab1e III.1.3a. This purchasing pattern cou1d be taken as an idea1ized one, the one which wou1d be di5p1ayed by the who1e popu1ation if it had the mobi1ity of the peop1e in S.E.L. 1, and assuming that it had the same sca1e of va1ues, a dangerous but not a1together imp1ausib1e assumption. But there is one point about this purchasing pattern which merits comment. Why do so many more S.E.L. I pe0p1e buy fruits and vegetab1es in Ga1eria Centra1 than in the sate11ite markets, A1ameda in particu1ar? Upper-c1ass peop1e in Ca1i are fastidious about C1ean1iness, hea1th- conscious, and most go shopping in private automobi1es. They are a1so sensitive about persona1 safety to the point of paranoia, if the bars and bo1ts and night-watchmen that surround their houses are any indication. Yet they go to the Ga1eria Centra1, which is dirty, ' 99 Tab1e III.1.3. Percentage of Fami1ies Interviewed Patronizing Principa1 Out1et Types, by Product and Socio- Economic Leve1 Product Group Socio- Economic Out1et Canned Fruits and Leve1 Type Grains Meat Foods Vegetab1es a I Centra1 Market 11 28 4 34 Sate11ite Markets 3 O 4 5 Tiendas 11 16 8 14 Se1f-service 68 28 76 41 Others 7 28 8 6 b 11 Centra1 Market 23 32 16 13 Sate11ite Markets 16 23 6 30 Tiendas 19 23 23 32 Se1f-service 42 16 45 16 Others 0 6 3 10 c 111 Centra1 Market 29 4O 21 50 Sate11ite Markets 15 31 5 34 Tiendas 23 1O 12 4 Se1f-service 27 8 23 6 Others 6 11 4 6 d IV Centra1 Market 31 39 9 52 Sate11ite Markets 3 11 O 15 Tiendas 35 39 20 28 Se1f-service 24 7 23 1 Others 7 4 5 4 e V Centra1 Market 28 25 8 35 Sate11ite Markets 14 22 4 27 Tiendas 37 49 12 35 Se1f-service 17 1 11 0 Others 4 3 2 3 f VI Centra1 Market 36 27 2 4O Sate11ite Markets 8 11 2 10 Tiendas 39 54 8 43 Se1f—service 1O 2 6 1 Others 7 6 O 6 Source: PIMUR, Consumer Survey 100 unhygienic, infested with purse-snatchers and pickpockets, and where there is no parking space. A1ameda market is better in a11 these respects and in addition specia1izes in se11ing fruits and vegetab1es to upper-income customers. Two-thirds of the sta11s in A1ameda market se11 fruits and vegetab1es versus Tess than ha1f in the Ga1eria Centra1 (see Tab1e IV.3.), and some hand1e commodities such as 1ettuce f1own in from Bogota. Yet seven times as many consumers of S.E.L. I buy fruits and vegetab1es in the Centra1 Market as buy them in sate11ite markets. As Figure 1.3. shows, A1ameda market is further away from the barrios of S.E.L. 1 than is the Ga1eria Centra1, and the same is true of the other sate11ite markets. One can on1y surmise that even for the physica11y mobi1e peop1e of S.E.L. I distance from home is sti11 an important factor in retai1er se1ection, a1though the mechanism behind this phenomenon is not c1ear. The S.E.L. 11 shopping pattern (Tab1e III.1.3b) shows a pronounced swing away from se1f—service stores and towards tiendas and sate11ite markets, in this case exc1usive1y A1ameda market. This . appears to be a consequence of the ease of access to A1ameda from the 1arge S.E.L. II barrios of the San Fernando area. The increase in tienda patronization is a1$o associated with San Fernando. At the time of the fie1d work there was on1y one supermarket in this area, and it was suffering from poor management. A1though supermarkets are aimed at peop1e in the upper socio-economic 1eve1s, and a1though peop1e in S.E.L. II are high1y mobi1e, yet the effects of distance were strong enough to provide an Opening for the neighborhood stores to compete. 101 Geographica11y, S.E.L. IV barrios are more centra1 and have Tess convenient access to sateTTite markets, whi1e many S.E.L. III barrios have easy access to A1ameda market. This is ref1ected in the patronization patterns shown in Tab1es III.1.3c and III.1.3d, in the spTitting of custom between the centra1 and sateTTite markets. Another interesting trend deve10ps in the midd1e c1asses in the way they repTace the seTf-service store patronization of the upper c1asses. It appears that the trade which in the upper c1asses goes to the se1f- service stores is SpTit in the midd1e c1asses into two streams. The tiendas get the sa1es of non-perishabTes (grains and canned foods) and the pub1ic markets get the sa1es of perishabTes, meat and fruits vegetab1es. This runs counter to the overa11 trend in shopping patterns for this group, which emphasizes Tocationa1 convenience in that the seTf-service stores are repTaced by the more accessibTe neighborhood stores and pub1ic markets. The counter-trend indicates that in purchasing perishabTes there are some factors that are more important than maximization of Tocationa1 convenience. What these factors are wi11 be discussed Tater in this chapter. The importance of ease of access continues to be dispTayed in the shopping patterns of pe0p1e in socio-economic 1eve1$ V and V1. Usage of "other" stores fa11$ off substantia11y, and tienda usage increases. Patronization of the sate11ite markets ref1ects the 1ocation of these out1ets. Three of the five sate11ite markets are 1ocated in barrios of S.E.L. V, one of which is on the bus route serving the 1arge S.E.L. VI barrio of La Union. One sate11ite market was intended to serve the S.E.L. VI barrio of SiToe, a1though its ' 102. Tocation is Tess than idea1 for this purpose. The fifth sate11ite market is, as we have noted, situated in the midd1e-c1ass barrio of A1ameda. We have been ab1e to exp1ain the difference in food purchasing patterns of the various socio-economic 1eve1$ in Ca1i in terms of the Tocation of the purchaser re1ative to her preferred stores. This 1eads_toflthe conc1usion that Tocationa1 convenience is the most ”_IJ—HH‘ —--. _— important factor in the consumer's se1ection of retai1ers. Yet the _._+_—_- - Ga1eria Centra1 is patronized by seventy percent of the popu1ation and se115 one-quarter of a11 the food consumed in the city. How do we reconciTe these two contradictory observations? If we cannot reconciTe them we wi11 be forced to conc1ude that 1arge food-retai1ing comp1exes 1ike the Ga1eria Centra1 can exert an attraction strong enough to overcome the desire for Tocationa1 convenience on the part of the consumer. This cou1d make very great differences in our recommendations. We be1ieve that the contradiction is more apparent than rea1. Whi1e the most obvious characteristics of the Ga1eria Centra1 comp1ex, its great size and range and depth of offered goods, and its proximity to downtown non-food stores, gives it an attraction for shoppers that the "neighborhood" store cannot h0pe to emuTate, there are two other factors which are a1most certainTy more important than thos mentioned above in exp1aining why so many pe0p1e patronize the Ga1eria Centra1. The Ga1eria Centra1 j§_the neighborhood store for a .....- ,significant proportion of CaTi's popu1ation. The ILMA study of 1965 1 1ILMA, Estudio de1 Sistema de Mercados Propiedad de EMCALI, (ILMA, Bogota, 1965) p.11-11. 103 found that 15% of the patrons of the Ga1eria Centra1 made their journeys there on foot. AppTying this to the proportion of peop1e that use the Ga1eria Centra1 (71.4%, according to PIMUR) we find that between ten and eTeven percent of the popu1ation of CaTi 1ives within wa1king. distance of the Ga1eria Centra1, and uses it as their nearest food store. .— M This makes the centra1 market comp1ex the convenience store for about \———— 90,066”Bébp1e. The reason that so many peop1e wa1k to the Ga1eria _CentraT is that there are practica11y no tiendas or sma11 graneros, or indeed fbod stores of any kind, within about ten b10cks of the centra1 market comp1ex. This is douthess the resu1t of the Ga1eria CentraT's estab1ishment in its present Tocation in 1931 as the city's centra1 market. The effect of this wou1d have been to inhibit the estab1ishment of neighborhood stores within competing distance, an effect which sti11 endures. The other thing that contributes to the importance of the Ga1eria Centra1 is the fact that it is the natura1 focaT point of the -— city's mass transport system. Every bus route in the city passes within tWo b10cks of the Ga1eria Centra1. Furthermore, the bus fare system in CaTi is a f1at-rate one. So the Ga1eria Centra1 is not on1y the most accessibTe out1et for one-tenth of the city's popu1ation, it is a1so the second most accessibTe out1et for anyone 1iving within easy reach of a bus route. Since Ca1i has a very 1arge number of bus routes this 1atter category includes practica11y everybody in the city. Therefore we wou1d expect to find that anyone who considered their nearest retai1er to be deficient in one or more respects wou1d be 1ike1y to use the Ga1eria Centra1 to make up the deficiency. In other words, we wou1d expect to find the Ga1eria Centra1 used 1arge1y as a specia1ty store by a 1arge number of peop1e. Consideration of the patronization 104 and sa1es figures for the centra1 market bears out this conc1usion. Setting aside the ten or e1even percent of the popu1ation who use it as a neighborhood store, and who presumab1y obtain about 90% of their food needs there, we find that around sixty percent of the popu1ation use it to obtain just over one-quarter (27%) of their food needs. The specia1ty store nature of the patronization is corroborated by Tab1e III.1.3, which shows that fruits and vegetab1es are the items which principa11y attract peop1e to the Ga1eria Centra1, with meat as a secondary attraction. As we noted ear1ier, a feature of the Ga1eria Centra1 area is the exceTTent assortment of fruits and vegetab1es avai1ab1e there. Two factors, then, are of dominant importance in giving the Ga1eria Centra1 its striking position in food retai1ing in Ca1i. These are its originaTNestabTishment as the city's centra1 market at a time when the centra1 market accounted for the bu1k of retai1 food sa1es in the city, and its situation at a focaT point of mass transportation. These two factors have provided the voTume of business which has enabTed the Ga1eria Centra1 to offer the outstanding breadth and depth of assortment which is its most dishinguishing feature as a retai1 center. Since neither condition can be reproduced in Ca1i in the future we can remove from further consideration the idea of designing a food distribution system for the city around 1arge comp1exes 1ike the Ga1eria Centra1. This ana1ysis a1so exp1ains why attempts to do this (the estab1ishment of the sate11ite markets) have not been very successfuT, and provides at 1east a partiaT exp1anation for the poor perfbrmance of the non-profit oriented stores in Ca1i]. 1 135 144 These stores are discussed in detaiT in Chapter IV, pages 105 If we accept this ana1ysis we are driven to the conc1usion that any food distribution system we might recommend must be based upon food retai1 out1ets featuring Tocationa1 convenience as the most --.-......— .1 a. prominent component of their marketing mix. For stores serving the Imidd1e and Tower c1asses (90% of the pOpuTation) this mean Tocating actua11y in the residentiaT barrios. It a1so means that these retai1ers can count upon obtaining very 1itt1e trade other than that avai1ab1e from peop1e 1iving with wa1king distance of the stores. Even so 1arge and "obvious" a retai1 out1et as the Floresta sate11ite market derived, in 1965, 86% of its patronage from peop1e 1iving within wa1king distance]. SpeciaTty stores wi11 continue to exist, and even pr05per if they Tocate strategicaTTy on streets served by severa1 bus routes. But for basic foods the neighborhood store has a great competitive advantage over a11 other out1ets, and these basic fbods account for the great bu1k of food expenditures in Ca1i. If our reconmended out1ets cannot compete on equa1 terms with the neighborhood stores in-respect of Tocationa1 convenience they wi11 not be ab1e to carry out their primary purpose, which is to make ‘ food avai1ab1e to the popu1ation of Ca1i at Tower prices than the existing food distribution system can offer. 1ILMA, op. cit., P.111-15 106 2. Shoppinngrequency Peop1e who have waited in 1ine in a supermarket wondering why on1y one check-out counter out of six is in Operation wi11 probabTy agree that the rate of fTOw of customers is a factor that shou1d be taken into consideration in the Tayout and operation of a store. Tab1e 111.2.1 presents the data from which rate of customer fTOw can be estimated for food stores in Ca1i. Tab1e 111.2.1. Percent of Consumers making Shopping Trips, by Socio- Economic Leve1 and Day of the Week. Ca1i, February 1969. Socio- Economic Percent of Consumers making Shopping Trips Leve1 Mon. Tues. Wed. Thur. Fri. Sat. Sun. I-III Tiendas 68% 65% 65% 71% 74% 58% 61% Others 54 61 54 58 ,‘74 100 62 IV Tiendas 74 74 74 74 78 74 65 Others 78 87 52 74 110 100 74 V Tiendas 72 74 75 74 72 75 75 Others 49 52 48 48 42 68 53 v1 Tiendas 81 81 73 '31 73 77 31 Others 38 44 4O 4O 44 63 63 Source: PIMUR, Market Basket Study- The figures as they stand have to be treated with care. They probabTy understate the number of visits made to tiendas. This is a consequence of the data coTTection and processing methods used, which were such that a maximum of one trip to each ofififteen types of out1ets cou1d be recorded. Thus a housewife cou1d visit tiendas two or more times in one day but on1y one trip wou1d be recorded. SimiTarTy, she cou1d visit, say, three types of out1ets in one shopping trip and three 107 trips wou1d be recorded. This accounts for the rather odd statement that, in Tab1e 111.2.1, 110% of S.E.L. IV consumers make trips to "other" stores. StrictTy Speaking, this datum indicates that the average S.E.L. 1V housewife visits 1.1 stores other than tiendas on Fridays. In using Tab1e 111.2.1, therefore, the ana1yst shou1d be aware that the data for "other" stores ref1ect visits rather than shopping trips, whi1e the data for tiendas do not ref1ect mu1tip1e visits, to pick up items that were overTOoked on what might be described as ”primary" visits. The figures given in Tab1e 111.2.1. are actua11y much more usefu1 for our purposes than the caveats noted above wou1d indicate. This is because the tienda data ref1ect the frequency of visits to convenient, but not u1tra-convenient, stores. They indicate the fTOw through a store which offers Tocationa1 convenience, whi1e reducing the effects of the fTOw caused by treating the store as if it were a Tarder. This is preciseTy the infOrmation required fOr the design of our hypotheticaT store. ProbabTy the best estimate that can be made from Tab1e 111.2.1. is that our hypotheticaT store wiTT have to be prepared to serve three- quarters of its regu1ar CTients every day, and a11 of them on Saturday and Sunday. If there are no changes in purchasing habits a11 of this wi11 have to be done during the forenoon._ Market Basket Study data indicated that 95% of a11 food purchases were done during the morning. Whi1e it wOqu be desirab1e to spread the consumer fTOw throughout the day rather than concentrate it during the forenoon hours, we do not know the extent to which this might be possibTe. ‘The custom of buying food before midday is consistent with the practice, based on technicaT \... few-fl" - _. A - _ 108 considerations, of distributing fresh meat and produce to retai1 out1ets during the earTy morning hours. The housewife wi11 natura11y" time her shopping to take advantage of the better conditions of .1 freshness and se1ection avaiTab1e in the morning. 1 If this were the on1y causative factor behind the custom of shopping in the morning, the retai1er cou1d fairTy easi1y spread the customer fTOw over the whoTe day by such devices as using refrigeration to keep goods fresh, re-stocking his she1ves two or more times per day, and offering sma11 premiums to customers patronising the store at sTack periods. But if it were the case, as is quite probabTe, that it is a Tong-standing tradition in Ca1i to eat the main mea1 at midday there is nothing the retai1er can do to change the custom of buying fOod in the morning. It is our Opinion that the first exp1anation is the correct one, but this is pure1y an Opinion. Tab1e 111.2.2 shows the distribution of shopping trips and sa1es voTume by the day of the week. This tab1e indicates that Friday and, particu1ar1y, Saturday are the days on which those housewives who do not buy food on a day-to-day basis do their main shopping. This is verified by the data of Tab1e 111.2.3, which shows the breakdown of trips between se1ected retai1er types according to the day of the week. Tab1e 111.2.2 Distribution of Shopping Trips and Sa1es VoTume by Day of the Week. Ca1i, February 1969. Day Mon. Tues. Wed. Thurs. Fri. Sat. Sun. Percent of Trips 13 13 13 13 15 T7 16 Percent of Sa1es 10 T4 11 TO 17 25 13 Source: PIMUR, Market Basket Study 109 Tab1e 111.2.3. Average Number of Shopping Trips Per FamiTy, for Se1ected RetaiTer Types. Ca1i, February 1969. RetaiTer Type Mon. Tues. Wed. Thurs. Fri. Sat. Sun. Tiendas .73 .74 .73 .75 .73 .73 .74 Truckers .20 .20 .23 .25 .21 .17 .18 Centra1 Market .14 .14 .12 .13 .13 .24 .16 Sate11ite Markets .12 .15 .16 .15 .14 .16 .18 Others .23 .24 .18 .20 .24 .29 .21 Source: PIMUR, Market Basket Survey Here it can be seen that the stores showing rises in customer fTOw on Saturday are the Ga1eria Centra1 and "others", the stores associated with trips away from the neighborhood. A practice which has an effect upon the time patterns of consumer purchasing behavior is that used by many empToyers in Ca1i in paying sa1aries and wages. Payments are made on a1ternate Fridays, which accounts in part for the purchasing surges on Saturday. A150, the pay-checks are distributed after the banks.c105e on Friday, and in CaTi they do not open on Saturday. So the provision of check-cashing or equiva1ent servige can be a matter of some importance to retai1ers .. .uflr-w—a—u. -..—..— if they are to obtain their share of the Saturday trade. The foregoing discussion has indicated that fOod buying in Ca1i is predominantTy a matter of mu1tip1e sma11 purchases. This mi1itates against efficient food retai1ing, since a given instaTTation can process on1y a given number of purchases.' If the purchases are 1 sma11, the tota1 sa1es vo1ume wi11 be sma11, and the margin required to 1 cover the fixed costs wi11 be 1arge. We are therefore faced with 1 contradictory requirements. Good service impTies neighborhood stores, 110 but neighborhood stores impTy inefficient retai1ing, in the sense of high margins. "nh*”“#“JCOnsideration of the nature of the purchases indicates a possibTe way out of this impasse. Three types of purchase can be distinguished: ;3) \SmaTT purchases, of few items, in which 1itt1e or no se1ection is invoTved. The archetype of such purchases are those invoTving bread and/or mi1k. I. _l b) Medium-sized purchases of severa1 items, with some se1ection invoTved. Representative.of such purchases are those of meats and produce. .c1 Large purchases, invoTving many items. These are the "main" 1 shopping trips, when grains and non-perishabTes are bought in quantities sufficient to 1ast severa1 days. The se1ection process is more comp1ex, invoTving aTTOcation of funds between products as we11 as se1ection amongst varieties of any one product. The first type of purchasing is usua11y of dai1y occurrence, the second varies from dai1y to week1y, most typicaTTy occurring two or three times a week. The third type is a week1y phenomenon, a1though in the Market Basket Study many housewives were found to buy grains and non-perishab1es in amounts intended to 1ast for two weeks or more. The three types of purchasing can be fbund singTy or in combination, in the sense that in a shopping trip of the second type products characteristic of the first type wi11 often aTSO be bought, whi1e in trips of the third type products of the first two types wi11 a1$o be bought. The point to be made is that the three types of purchases need different se11ing and check-out faci1ities, and the objective in 111 designing the store shou1d be that of maximising the number of customers that can be processed in the day rather than that of maximising the exposure of customers to the merchandise. We wi11 return to this in Chapter V when we set out our recommendations for an improved food distribution system. 3. Shopping Transportation Costs The previous section considered shopping trips from the point of view of Toad upon the stores. Here we consider them as a Toad upon the consumer. In Section 6 of Chapter II Tab1e 11.6.1 presented data upon transportation costs incurred in shopping, expressing them in terms of food expenditures of the popu1ation. Tab1e 111.3.1 presents the data in rather more detaiT and in terms of the Toad upon those peop1e who shOp non-TocaTTy. For this reason shopping frequencies differ from those quoted in the previous section. The tab1e brings out some interesting points. The increase of frequency as one goes down the socio-economic sca1e is presumab1y re1ated with reducing degree of possession of refrigerators. The cost per trip faTTs from S.E.L. I through S.E.L. V, as buses repTace taxis as transportation. The rise noted for S.E.L. V1 is rather surprising. PresumabTy it ref1ects the Tocation of most of the barrios of this 1eve1 on the outskirts of the city, but bus fares in Ca1i are of the f1at-rate type, so that Tocation shou1d not affect the cost of trave1. However, the quaTity of bus service at the edges of the ciiy is re1ative1y poor, in the sense that the routes are more wide1y spaced than they are nearer the city center. ATso, the bus routes serving the two Targest barrios of S.E.L. V1, SiToe and La Union, run aTong the boundaries of these 112 Tab1e 111.3.1. Transportation Costs of Non-LocaT Shopping Per Fami1y Per Week, by Socio-Economic Leve1 . Ca1i February 1969. Socio-Economic Leve1 I II III IV V VI Average number of trips 1.08 1.06 2.03 2.94 3.48 3.96 Average cost/trip (COT.$) 4.30 4.65 3.66 3.00 2.57 3.17 Average cost/week (CoT.$) 4.67 7.44 7.31 8.80 8.93 12.56 Average f80d exp. week 616 403 393 296 205 159 (COT.$) Transportation cost as percent of food exp. (%) 0.8 1.9 1.9 3.0 4.4 8.0 Source: PIMUR, Consumer Survey barrios - they do not enter the barrios. Thus many housewives, returning on the bus from the city center with a heavy basket or baskets of food- stuffs, are faced with extra expense because they have to pay someone to heTp them carry the basket(s) to their homes. Another practice, that of trave11ing to the city center by bus and returning by taxi, is presumab1y aTso caused by the rather inadequate service the buses can offer a housewife carrying heavy baskets. In spite of the greater expense of this mode of trave1 23% of those housewives of S.E.L. VI who shopped non-TocaTTy were found to use it.3 1This tab1e app1ies on1y to those conSumers who did their principa1 fbod purchasing non-TocaTTy. 2These are average week1y food expenditures for consumers buying non-TocaTTy. For this reason they are sTightTy different from the averages obtained in the Market Basket Study, but the differences are sma11 (Tess than 2%). 3PIMUR, Consumer Survey 113 Tab1e 111.3.1. makes it c1ear that the costs of shopping for food away from the neighborhood become increasingTy heavier, both absoTuteTy and as a percentage of food expenditures, as one goes down the socio-economic sca1e. This indicates that there is a considerabTe degree of dissatisfaction with the service provided by the neighborhood stores. It aTso indicates that the estab1ishment of satisfactory stores in the neighborhoods wi11 bring substantiaT benefits in the form of savings of transportation costs, and that these savings have the very desirab1e characteristic that the greatest benefits wiTT go to those who most need them, peop1e in the Towest socio-economic 1eve15. 4. MisceTTaneous Services It is wide1y be1ieved in Ca1i that consumer credit is an important factor in food retaiTing, especiaTTy in the poorer barrios. A” ...- IIEL an...- 4. ~.-.. For instance, the manager of ACOPI, one of the two Cajas de Compensacion in Ca1i that operate food stores, ton PIMUR interviewers that he had noted that his store was being patronized principa11y by the higher-paid empToyees of the affiTiated firms. He attributed this to his inabiTity to give credit, and expressed his be1ief that the tiendas in the barrios used the extension of credit as a device to attract and hon customers. Ga1braith and HoTton'l stated that consumer credit was important in food retai1ing in Puerto Rico, and Nason2 found it to be moderateTy important in the marketing mix of neighborhood stores in Recife, BraziT. Against this, Hen1ey3 found 1itt1e evidence that credit was a .. din... ‘- --—. “mm—WP?”— 1 2Robert W. Nason, "Urban Market Processes in Recife, BraziT", unpubTished Ph.D. dissertation, Michigan State University, 1968. p.98. op.cit., pp. 20-21 3DonaTd S. Hen1ey and R.Vincent Farace, "Consumer Buying and Communication Patterns in BoTivian Urban Food Retai1ing", American Marketing Association, AnnuaT Meeting, Washington D.C., 1967. 114 significant factor in food retaiTing in La Paz, BoTivia, and PIMUR found the same thing in Ca1i. Tab1e 111.4.1. summarizes the data obtained from the PIMUR Market Basket Study on the use of consumer credit in food purchasing. Note that the c1assification of consumers in this tab1e is’according to income and not according to socio-economic 1eve1. (There'are two reasons for changing the c1assification here. First, there were few peop1e in the upper socio-economic 1eve1 amongst the Market Basket Study participants, which, in conjunction with the very Timited use of credit, made it difficu1t to obtain a statisticaTTy significant number of responses from credit users. Second, if credit use is important to consumers as a necessity rather than mere1y as a convenience, as the be1iefs mentioned above wou1d indicate, then we wou1d expect to find a strong negative association between income 1eve1 and credit use. As Tab1e III.4.1. shows, the association is very weak, and is positive. The users of credit are, if anything, those who 1east need it, indicating that credit is a convenience rather than a necessity. .3. ...-W «— ~ Tab1e 111.4.1 Use of Consumer Credit in Food Purchasing. Ca1i, February 1969 Credit as Percent Per Capita Income (CoT.$/month) of Tota1 0-125 126-240 241-500 over 500 No. of Purchases 8.7 8.5 4.4 3.1 Food Expenditures 9.3 9.7 10.0 15.3 Source: PIMUR, Market Basket Study The sma11 amount of credit used aTso indicates that it is not an important convenience. This is verified by the fact, shown in the Market Basket Study, that most of the purchases on credit were purchases H— of mi1k from_truckers. Here the convenience of the vendor is as 115 important as that of the buyer. (Some mi1k de1iverers wiTT on1y accept payment on speciaT days when a cTerk makes the rounds with the driver of the vehicTe.) ResuTts from the PIMUR WhoTesaTe/RetaiT Study confirm that credit is not important in food retai1ing in Ca1i. Of the tiendas interviwed on1y 40% gave credit, and these on on1y 22% of their sa1es; the corresponding figures for retai1ers in the pub1ic markets were 30% and 20%, respectiveTy. OnTy the COOperatives do much se11ing on credit, reporting that 80% of their sa1es were made in this mOde. In this case there is no evidence that credit avai1abiTity is of significant importance to the customers. Rather, it appears that giving credit by the COOperatives is a ideoTogicaTTy-based poTicy. The cooperatives have been encouraged in this by having been given the priviTege of being ab1e to garnishee the sa1aries of their debtors. The conc1usion is that consumer credit is not necessary in '1‘:- WAFW the marketing mix of the food retai1er in Ca1i. RetaiTers who refuse _pflvvw a—n—II'I-M" , "'“'"*~gfiw""-"""' M” " to give credit wou1d Tose on1y a negTigibTe amount of business. On the i Other hand, there are no doubt many fami1ies in the poorer districts whose income is irreguTar or comes from day-to-day casuaT Tabor, and for these peop1e credit at the TocaT food store cou1d be important. The Market Basket Study showed that amongst the Towest income consumers 12.8% of the purchases at the tienda were on credit, whi1e in the second-Towest income group 11.4% of purchases at'the tienda were on credit. For both of these groups the tienda accounted for about ha1f of a11 food bought, so that on1y about six percent of the fami1ies in the two Tower income groups appear to need credit faciTities.1 I TThat is, of the ha1f of the peop1e in these income groups who use the tienda about theve percent use credit. This corresponds to about six percent of a11 the peop1e in these income groups. 116 Home de1ivery is very TittTe used. Mi1k de1iveries accounted for a1most aTT the home de1iveries noted in the Market Basket Study, and in the Consumer Survey Tess than one percent of housewives had used home de1ivery in their 1ast purchase of foods other than bread or mi1k. Carry-out service is virtua11y unknown in Ca1i, at 1east as a service provided by retai1ers. Around the pub1ic markets there are usua11y sma11 boys who wiTT carry the housewife's basket for her for a sma11 fee (usua11y one peso at the time of the fie1d work). ATmost a11 food purchasing in Ca1i is carried out during the morning hours. Fewer than five percent of the purchases recorded in the Market Basket Study were made after midday. As far as service is concerned, the retai1er of food in Ca1i need be Open for business on1y untiT noon, as is the practice of a1most a11 staTThoners in the pub1ic markets. @deed, as we noted ear1ier in discussing shOpping frequency, the prob1em facing food retai1ers in this respect is to persuade their customers to spread their purchases over the whoTe working day, in order to make effective use of the fixed faci1ities (space, dispTay 1. faciTities, check-out, faciTities)) I! 5. Factors Affecting Choice of RetaiTer The previous sections of this chapter have concentrated on the topics of where and in what fashion the Ca1i housewife buys food. OnTy in one instance, where we noted the substantiaT custom-attracting power of Tocationa1 convenience, have we touched upon the reasons why_she““w *u— .. “III--1 I‘m-WI “m aTTocates her custom as she does. Yet this is an extremeTy important .-_r-I—-.- ' ~.1-~—.. up ‘ matter. The most economicaT and efficient food distribution system ‘/ wou1d be a faiTure if it did not take into account the requirements of the consumer. 117 UnfortunateTy, it is extremeTy difficu1t to ascertain why peop1e do things in one way rather than in another. Motivation research is not by any means a mature science1 , and survey instruments are just not avaiTab1e "off the hook".* This means that an investigator who wishes to obtain information using methods of the type associated with motivation research has to design his own specific instrument and is then faced with prob1ems of determining its reTiabiTity and vaTidity. VaTidation of instruments deaTing with non-standard situations is virtua11y impossibTe. NevertheTess, in view of the importance of this tOpic it was feTt that an attempt shou1d be made to probe into the more important aspects of food retai1er se1ection on the part of the Ca1i housewife. There are two objectives for this study. The first is to identify those retai1er characteristics or attributes which the housewife seeks in her food supp1iers. The second is to test the hypothesis that consumer requirements force retai1ers to adopt operating formats which entaiT high operating costs and/or uneconomica11y sma11 sca1es of operation. It can be seen that this hypothesis must be shown to be faTSe befOre the present study can be continued. A mode1 of the consumer's decision-making process was needed to guide the coTTection of the data for testing the hypothesis, and a very simpTe mode1 was assumed. First, the consumer was assumed to be a rationaT shopper, patronizing the store(s) which best fquiTTed her requirements. Second, the dimensions of these requirements were taken to be the store attributes described ear1ier in this chapter (see page 1Luck, 0.0., WaTes, H.G., and TayTor, D.A., "Marketing Research", EngTewood CTiffs, Prentice-HaTT, 2nd. ed., 1961. Chap. 18. 118 Each consumer was regarded as attaching a 1eve1 of importance or saTience to each attribute, and a1though these TeveTs cou1d vary from consumer to consumer it was assumed that the profiTes were sufficientTy simiTar to make it meaningfuT to ca1cu1ate average TeveTs i.e. there were centra1 tendencies in the importances attached to each attribute. FinaTTy, each consumer was assumed to have evaTuated a11 avai1ab1e food retai1ers in respect of the degree to which they possessed each attribute, and to shop at the retai1er for whom some monotonicaTTy increasing function of the products of saTience and degree of possession of a11 attributes was greatest. The assumption of evaTuating all_food retai1ers is principa11y ' made to simpTify the mathematics and data-coTTecting requirements. A1though the assumption is patentTy faTse, this does not significantTy vp ‘..‘.-ar-‘.-~ ... _‘~ affect the vaTidity of the ana1ysis. In fact, the assumption is over- ; .-.: ...-g rigorggs; A11 that is needed is that the housewifeTS information be. pagequate to permit her to identify the store best suited to her requirements and a few (say, two or three) of its nearest competitors. This is a much more reaTistic assumption than that of evaTuating gll_ food retai1ers, yet the resu1ts wi11 be the same in each case, and the more rigorous assumption 1eads to easier data-coTTection and ana1ysis. If this mode1 is a reasonabTe representation of the consumer's retai1er se1ection process the data that we have to measure are where the consumer shops and how she evaTuated the retai1ers in respect of the attributes. If we compare the two sets of data we shou1d find that . attributes of high saTience are those for which preferred retai1ers were high1y rated. 119 We have described earTier (pages 90 - 93) the instrument which was used to coTTect the attribute rating data. Respondents were asked which of a11 the retai1er types in Ca1i were best, and which were worst, in respect of each of the fifteen attributes se1ected as being the primary dimensions of retai1er attractiveness. As was described on page 93 , the evidence suggests that the three-point sca1e thus ( obtained (best, not mentioned, and worst) contains, in effect, a11 the information that goes into making the retai1er se1ection. .It is effectiveTy equiva1ent to a sca1e of good, neutraT, and bad. The ana1ysis consists essentiaTTy of determining for each attribute the proportion of peOpTe who shopped at the stores they said were best. SuppTementary ca1cu1ations wou1d be those to determine the proportions of peop1e shopped at the stores they said were worst, a high proportion in these cases indicating unimportant attributes. These proportions can be ca1cu1ated in two ways, with the denominator either inc1uding or exc1uding those respondents who answered "Don't know". Arguments can be presented fOr both methods. Tab1e III.5.1.echudes the "Don't know" responses whi1e the resu1ts tabuTated in Tab1e 111.5.2. inc1udes them. It does, however, exc1ude the 67 respondents (10.5% of the sampTe) who answered "Don't know" for a11 attributes, both for best and fOr worst. In order to show how much difference there is in the denominators under the two bases of ca1cu1at- ion the numbers of “Don't know" answers received are_1isted in the first two coTumns of Tab1e 111.5.1. I It is of interest to note the distribution of "Don't know" answers. OriginaTTy it was thought that because respondents were being asked for their opinions in these matters, rather than their experiences 120 (and the interviewers were instructed to emphasize this point) there wou1d be few such answers. As can be seen, there were many, and the number of them appears to vary inverseTy as the importance of the attribute. Thus, there are re1ative1y few "Don't know" responses for such primary variab1es as prices, quaTity and Tocation, and very many for attributes which are known to be of TittTe interest, particu1ar1y home de1ivery and wiTTingness to give credit. A150, the number of such responses to the questions asking for the worst far outnumber in every case those to questions asking for the best. A1though unanticipated, this pattern is consistent with our assumed mode1. If home deTivery is unimportant to you, you wi11 not ask your retai1er whether he provides this service, and simiTarTy for credit. And once a retaiTer is beTOw an acceptab1e TeveT in respect of an attribute you are not going to go to any great Tengths to estimate just how bad he is in this reSpect. The ten percent of respondents who repTied "Don't know" to a11 the questions in this section, and the many more who gave this response to a1most a11 of them, are rather more difficu1t to fit into the assumed mode1. Various reasons can be hypothesized to account for this phenomenon. These peop1e might evaTuate stores on the basis of an entire1y different set of attributes from those in the questionnaire, but it is unTikeTy that the set of attributes chosen cou1d be entire1y, or a1most entire1y, mispTaced. They might be pure1y passive shoppers, not evaTuating stores at a11. For instance, they might shOp on1y where their husbands teTT them, or on the basis of Tong-ingrained habits, or simpTy be passive, as the resu1t of-a very Tow 1eve1 of education or Tack of aquaintance with the stores avai1ab1e, as might be the case with 121 recent migrants to Ca1i. Yet again, they might have been too timid to express an opinion or afraid of making "wrong" answers and appearing stupid. Each exp1anation wou1d predict that such peOpTe wou1d be found predominantTy amongst the poorest and 1east-educated members of the popu1ation, and indeed the proportion of such responses amongst peop1e of S.E.L. VI was, at 20%, significantTy higher (at the .OT 1eve1 of probabiTity) than the overa11 proportion of 10.5%. Tab1e 111.5.1. presents, for each of the three principa1 food product groups, the proportion of respondents who bought that product at the store which they rated as best in reSpect of the indicated attribute, expressed as a proportion of a11 those respondents who gave a "best" rating. That is, the "Don't know" responses are exc1uded. ATternate coTumns present the same ca1cu1ations for "worsts". Tab1e 111.5.2. inc1udes the "Don't know" responses, and aTso combines the best and worst evaTuations to give a sort Of "Index of Attribute Importance", ca1cu1ated as Index of Importance = Egg?!- where B - number buying at retai1er rated best, W . number buying at retai1er rated worst, and 562 - sampTe size, exc1uding the "a11 don't know" respondents. It shou1d be observed, incidentaTTy, that the number of "Don't know" reSponses 1isted in Tab1e III.5.1. aTso exc1udes the "a11 don't know" responses. These two tabTes reveaT TittTe that was not anticipated in our ear1ier discussion. That is, Tocationa1 ggnyegj Inn--\- u“ .. ence 399....11.S....... —‘- .- coroTTary, Tow transportation cost, are the most important attributes, .uaosmaogzu =3ocx u.:oa= uwsmzmcm ocz mucmucoammg we Ahoy songs: we» mmmp vmucooms smxocx u.coo= ea Lungs: mg» mew mmczmwm mmmchp mm>g=m gmszmcou assz "mussom 122 ON. m_. AN. m_. mp. om. New on mafiamom-=oz to moowca NN. mm. mm. mN. mp. mm. mm_ mm mopaaom co mooeca o_. as. «F. as. o_. _e. sow mm =o_omoou ocowcoscou mp. Nm. mm. Pm. ~_. Fm. mew mm sooom new sooesa> mp. ON. mu. NN. NF. mm. om, Fa smootsou can mow>com om. NP. om. n_. my. mm. mum _o_ c_omo=o macaw oooeom emu as. am. Me. as. o_. mm. mop mm omou cowoaocoamcac» so; NF. mm. _N. mm. up. mm. sop um monsoomea new metamtam 5,. mm. mm. mm. mp. mm. mo_ mm motsmaoz new apnoea: wN. me. am. m_. o_. om. o_F mm otoom to amazepemoFu mp. mm. m_. mm. m_. em. «on map amazemsm to mesa: 2. 2. 2. S. 2. 2... one 8... .558 go: so. mm. mo. _m. mo. mm. mam mew oweocu mo>ew mp. mm. we. mm. ~_. mm. opm mm opoocoaeu Seamaoea 4P. mmmwwv Fm. mN. mp. . mm. PPN 4N sowpaso :82 tom :83 one. “to: :8 one: :8 mmpnmummm> use muwzgm new: mcmoeo mzocx u.:oo= mannrguu< F: umsoz van «mom umuma mempwmumm um mcwxzm mpaomm mo =o_ueoaogm mo guess: sea. seascnoa .cpou .aaotw bosoots can cowoaewomm to oogooz s5 .moospwtooa sopwaoom so moocaotoasH oswoa_om .F.m.HHH apps» ..i‘ja(. 123 whi1e home deTivery is the 1east important. Credit faciTities appear -no;; important than previou51y indicated, but a gTance at the "Don' t know" coTumns of Tab1e 111.5.1 resoTves the contradiction. ReTativeTy few peOpTe even know which retai1ers aregood or bad as far as giving credit is concerned, but a re1ative1y high prOportion of these peop1e are concerned about credit. Tab1e 111.5.2. Indices of Attribute Importances. Ca1i, February 1969 Ti Attribute Index of Importance . Fruits and Grains Meat Vegetab1es QuaTity of Goods .19 .13 s22‘ PTeasant CTienteTe .17 .10 .09 Gives Credit .13 .15 .12 Home OeTivery .07 .01 .OT vj>HOUFS of Business .18 .16. .T3 C1ean1iness of Store .103 -.07 -.08 Weights and Measures {22 .13 .17 Bargains and Discounts .12 .061 .09 Low Transportation Cost t28i> £31) “267 Can Se1ect Goods OneseTf .09 -.04 -501 Service and Courtesy .18 .06 .06 Variety and Stock guz ’13 Convenient Location Q31) 1,343) Prices of StapTes .17 .11 Prices of Non-StapTes .15 - .01 .02 Source: PIMUR, Consumer Survey "Can se1ect goods oneseTf" and "CTeanTiness of store“ are identified by the Indices as rather unimportant, but the remaining attributes a11 bunch together, and cannot be definiteTy arranged in order of importance. The instrument and the ana1ysis are not W-WH‘ "“"”‘”'""""'" * "‘ ~ . sufficientTy precise to distinguish between them. “‘hh Ma- 1See text for a definition of this Index. 124 There is an interesting aspect of this ana1ysis concerned with the attribute "Hours of business", the number of hours per day that retai1ers are open to customers. There are substantiaT differences between retaiTer types in this characteristic, with many staTTs in the Ga1eria Centra1 cTosing at midday whiTe many tiendas stay Open untiT Tate in the evening and the supermarkets and Caja stores keep regu1ar business hours. Yet, as we noticed earTier, virtua11y a11 food purchasing is done during the morning. This means that in spite of the wide differences between retai1ers in this respect, hours of business wi11 not be perceived by consumers as a variabTe which enters into the retaiTer-se1ection process, since a11 retai1ers are Open during the morning. This conc1usion is verified by the observation that no Tess than one-sixth of responses to the Consumer Survey indicated that the Ga1eria Centra1 was best in respect of hours of business, and Tess than one-quarter gave the tienda this rating. Now the Indices described above give some idea of the ranking of the various attributes in order of importance, but some degree of absoTute measurement is wanted. In particu1ar, we want to know what attributes warrant some effort on the part of the retai1er to achieve, and which do not. By knowing this we can avoid potentia1 waste of time and money. We want, in fact, to identify those attributes which are actua11y sought by customers and those which customers are wiMTing" O sacrifice in favor of 'sought' attributes. If we recognize these two c1asses of attributes we wou1d 1ike to find one or more attributes on or near the border between the two, because by identifying such attribute(s) we can refer to the Indices and sort attributes into 125 'sought' and 'sacrificed' categories. It foTTows at once thatfihours of business is such a border- Tine attribute. Customers can neither seek nor sacrifice this attribute, so that, as far as retaiTer se1ection is concerned, it is an attribute of neutraT importance. Therefore attributes whose Indices are greater- 1 ~t§§h that of hours of business can be considered to be activeTy sought I by consumers, whi1e those with Indices of Tesser magnitude are those 5 which customers are wi11ing to forego. On this basis, it can be seen that in meat purchasing on1y Tocationa1 convenience is activeTy sought, that in purchasing grains honesty in respect of weights and measures aTso faTTs into this category, whi1e in buying fresh fruits and vegetab1es quaTity and variety and stock are added to the 1ist of sought attributes. Again, there are severa1 attributes which have indices approximateTy equaT to that of hours of business, and can therefore be considered as effectiveTy neutraT as far as the retaiTer-se1ection process is concerned. Further insight into the retaiTer-se1ection process can be derived from comparing the attribute profi1es for the three product groups. First, however, we must note some characteristics of the products themseTves. Meats are characterised principa11y by their perishabiTity, the sheTf-Tife of meats being measured in hours. This is because meat distribution in CaTi is a "hot-meat" system, in which w “an. _, “1’ refrigeration is not used. Fresh fruits and vegetab1es are character- ...-.....- ised by the great variety of individua1 products and by their fairTy high perishabiTity. Soft fruits, once at the retai1er, have a sheTf Tife of one or two days, whi1e p1antain and potatoes can be kept for 126 a week or so. Grains and processed foods can be kept effectiveTy indefiniteTy, for periods up to severa1 months. These characteristics exp1ain some of the phenomena observed in food purchasing, as summarized in the indices of Tab1es 111.5.1 and 111.5.2. Nearness to the point of consumption is an over-riding consideration in meat purchasing, (nearness in time, that is,) whi1e freshness (i.e., quaTity) becomes on1y important,and of importance comparabTe with that of variety of offering, in the case of fruits and vegetab1es. We wou1d therefore expect to find product characteristics over-riding retaiTer characteristics in the case of meats, and this is the case, Tocationa1 convenience being the most important attribute and the on1y one above the neutraT point. SimiTarTy, we are not surprised to find quaTity, interpreted as freshness, and variety of offering becoming sought rather than sacrificed attributes in the purchasing of fruits and vegetab1es. When it comes to purchasing grains the imperishabiTity of the product re1ieves the buyer of time constraints. This aTso heTps to re1ieve her of p1ace constraints, since she can save up the money required for transport. Therefore we can expect retai1er character- istics to dominate product characteristics in the purchasing of grains. That is to say, if we compare the attribute profiTe associated with grains purchasing with those of purchases of meats and fruits and vegetab1es the differences in the case of grains, if any, wiTT be indicative of preferred retai1er characteristics. When the comparison is made, we find that the attribute profi1e of the "preferred" retai1er is suggestive of the 1arge, modern store. 127 The various attributes which increased significantTy in the case of grains purchases are those in which the stores of this type made their best showings in terms of consumer perceptions. The chain stores, whose marketing mix is aimed at the pOpu1ar market, received particu1ar1y favorabTe ratings in these respects. A1though bargains and discounts might be taken to suggest market staTTs and haggTing, the responses to the Consumer Survey indicated that consumers interpreted this attribute principa11y as meaning the Tow-prices poTicies of the Caja groceries and the chain stores. Again, the attribute of service and attention was one in which the persona1-service store types, the tienda and the market staTTs, did badTy. Consumers apparentTy prefer the impersona1 poTiteness of the personneT of 1arge stores to dea1ing direct1y with the store owner. HI I This is an important conc1usion. It means that 1arge, modern stores, with their attendant opportunities for Obtaining economies of sca1e, are not inconsistent with the requirements of CaTi's consumers. There is another important point that is brought out by the additionaT ana1ysis. Prices, as we have seen, are surprisingTy Tow on the sca1e w” of attribute requirement. This eTiminates, at 1east partiaTTy, what ”was reputed to be the main attraction of the Ga1eria Centra1. The main attraction is now seen to be the quaTity and variety of the goods on offer, and this attraction is sufficient to overcome, at 1east in the eyes of many Ca1i housewives, the Tocationa1 inconvenience of the city center. The importance of this conc1usion is c1ear. The retai1 out1ets must be Tocated in the residentiaT areas, and they must offer an adequate se1ection of fOodstuffs - they must be fuTT-Tine fOod stores.’ 128 This impTies stores considerabTy Targer than the existing sma11 graneros, and is consistent with the point made earTier about the need for obtain- ing economies of sca1e impTying 1arge stores. 6. ConcTusions The formaT purpose of this chapter was to test the hypothesis‘ that the requirements of the Ca1i consumer force the Ca1i food retai1er into high-cost modes of Operation and/or an uneconomica11y sma11 sca1e of Operation. 1 In testing this hypothesis it was found that: a) except for the provision of Tocationa1 convenience, the consumer does not require any services other than those direct1y associated with the retai1 function of resoTving discrepancies of quantity and assortment. In particu1ar, such cost-inducing services as provision of credit, home de1ivery, and Tong hours of business are not necessary. b) there is no evidence that the consumer prefers sma11 stores. The hypothesis is therefore disproven, except insofar as the ——-——--H fl-fil-I- _— provision of Tocationa1 convenience, which impTies the dispersion of food retai1 out1ets amongst the residentiaT areas of the city, is incompatibTe with the operation of stores 1arge enough to achieve economies of sca1e compared with the existing tiendas and sma11 graneros. Later, in Chapter V, it wi11 be shown that size and dispersion are not incompatibTe. CHAPTER IV The Urban Food Distribution System Introduction In Chapter 111 there was deve10ped the primary specification for any new food distribution system for the city of Ca1i. This specification is that such a system must be based upon retai1ers situated in the residentiaT areas of the city, in order to provide Tocationa1 convenience for the housewife. CTearTy, these retai1ers must offer a 1 fuTT Tine of foodstuffs , and probabTy a se1ection of househon suppTies, if the advantages of Tocationa1 convenience are to be fuTTy rea1ised. Our objective is to design a system that wiTT make food avaiTab1e w. _na-fi cram-..-- ....- .-—- ._,_ “— .... ...-..., ”hr-aur- ‘ .- -«-——- —-_. to the consumer at Tower _cost than obtains at present, and this impTies retaiT “"155-W1EO-12£9§,§919§ voTumes, in order to reduce the impact of fixed costs upon margihs This requirement is consistent with the evidence, noted in Chapter III, that the housewife is not averse to shopping in 1arge, modern stores. w..- The purpose of this chapter is to Took for information on two tOpics. The first is to find whether any of the existing types of food retai1ers in CaTi can reasonab1y be expected to deveTop spontaneousTy into the type of retai1er required by the projected system. The second topic is that of determining why the existing retai1ers and other 1By this we mean that the retai1er shou1d offer some selection in a11 of the main groups of foodstuffs (meats, grains, fruits and vegetab1es, and processed stap1es). 129 130 institutions of the food distributionsystem in Ca1i assumed their present formats and modes of operation. These two topics are very cToseTy re1ated. In fact, they are two facets of the same topic. SpecificaTTy, when and if we find a ‘retaiTer type which appears to be suitabTe for deveToping into the type we require we wi11 immediateTy be faced with the question of why this deveTopment has not aTready taken p1ace. We wi11 expect to find that there exist barriers to deveTOpment. In fact, if we do not find ......_.... _“l—M “aw—gm sughmbarriers this indicates that our understanding of the existing W’— food distribution system and its environment is incompTete. Even if we do not find any retai1er types capabTe of deveTopment we wiTT sti11 need to have a good understanding of the circumstances under which the present food distribution system operates. Otherwise we wiTT run the risk of designing for the new system a types of retai1ers and whoTesaTers which are incompatibTe with other institutions in the food marketing channeT. An important concept which wiTT be used in the ana1ysis is that of capacitwaer change, and its negative, that of an evoTutionary* dead end. An institutiOn wiTT be considered capabTe of change, or to ”have evOTutionary capacity, if it can make sma11 changes in its format or operating methOds which resu1t in its being better adapted to its immediate environment. The immediate environment comprises the institution's customers, supp1iers of goods, supp1iers of services (e.g., transportation), and government. The componentsflfltli:_bentteghe(_Taptaiizjongarfie:I__fi~ a) increasing rewards to the owner or operator, b) deTivering better service to customers, and c) receiving better service from supp1iers. 131 "SmaTT" changes are those which preserve the nature of the institution, those Tess in magnitude than a "quantum jump". Thus, if a tienda operator decided to add, say, potatoes to his product 1ine this wou1d be a sma11 change, whi1e going to supermarket format (i.e., seTf-service and departmentaTization) wou1d be a 1arge change. SimiTarTy, a market sta11 cou1d not evoTve direct1y into a neighborhood store, since this wou1d entire1y change the nature of the institution. These considerations app1y on1y to the institution, not to the operator. There is no reason why a sta11 operator shou1d not Teave his sta11 and buy a tienda, for examp1e. But he cou1d not, by a series of sma11 changes, convert his sta11 into a tienda. Note particu1ar1y that change is considered on1y with respect to the immediate environment. Whi1e the operator might optimize the position of his institution within the marketing channeT he cannot change the structure of the channeT as a whoTe. It is quite possibTe to find an inefficient (sub-optimaT) channeT made up of efficient (OptimaT) institutions. We are dea1ing here with a system composed of many interTinked components. The comp1exity of the ana1ysis of such systems grows very rapidTy as the number of components invoTved increases. To simpTify ”Willi- an Wan- .. . the ana1ysis it is important to identify those components which have M 1itt1e or no effect upon the behavior of the system as a whoTe. We ...-W- can cTassify such components into two categories. There are the minor W components, consisting of a few sma11 units and hand1ing on1y a sma11 .._____..._.L_— proportion of the food consumed in the city, and the "non- -systemicf “M...._—- Cid-'- components. These 1atter are institutions which are on1y TooseTy Win”- interTinked with the rest of the system, the "main stream" system, such 132 that changes in the "non-systemic" components need not induce compens- tory changes in the "main stream" components with which they are Tinked. This is a suitabTe point to discuss one of the very few generaTizations that can be made about the Operators of the_food__md_ distribution system in CaTi. A1most withgut exgeption they spend a ta...— ...... ———-—_. great deaT of time and effort on buying and TittTe on se11ing. They I ......- 2......— are active buyers and passive seTTers. This app1ies at a11 TeveTs of - distribution. SuppTiers to the system sit back and wait for orders I from whoTesaTers, whoTesaTers sit back and wait for retai1ers to come to them, and retai1ers behave in the same way towards consumers. Zffi;;;e are two main exceptions. One is that group of retaiTers which we have denoted as chain stores. These retai1ers are vigorous mer- chandisers. The other exception is provided by processors of branded (i.e., non-stapTe) processed foods, who tend to maintain 1arge and active sa1es forces;/) We do not know what are the background causes of this behavior. It cou1d be connected with the sOciaT va1ues inherited from the warrior caste that coTonised the region many years ago, it cou1d be the resu1t of a defensive preoccupation with traditiona1 skiTTs, or it cou1d stem from an "economics of scarcity" phiTosophy, a be1ief that demand aTways presses upon suppTy. ETucidation of the source wou1d be an interesting and worthwhiTe research prob1em. For the moment it is enough to note that the phenomenon exists. We belieye‘ that it is an important point to our thesis, and we wi11 return to it ”cu-...“.— in Chapter V;-s - M 133 RetaiTers The many types of retai1ers Operating in CaTi at the time of PIMUR's fie1d work are described here under the foTTowing headings (in order of discussion): 1. Minor retai1er types, those having on1y negTigibTe importance in terms of numbers and of tota1 proportion of food sa1es. . Non-Profit Oriented types, those Operated by IDEMA, the Cajas de Compensacion, the Puestos de Sa1ud, and the consumer cooperatives. . Non-Systemic retaiTers, as described above. These are mi1k deaTers and distributors of pouTtry and eggs. . SeTf-Service stores, supermarkets and chain stores. . PubTic Markets, with their associated street vendors. . CentraTTy-Tocated traditiona1 stores, the whoTesaTer-retaiTers and 1arge graneros. . Non-CentraTTy Tocated traditiona1 stores, the sma11 graneros and tiendas. . Meat distributors, at aTT TeveTs of distribution. Minor RetaiTer Types a) Markets in other towns. AnaTysis of the Consumer Survey data showed that these were of negTigibTe importance, and the requirement of consumers for Tocationa1 convenience indicates that they wiTT not become more important in the future. The towns invoTved are Yumbo, some 10 kms. north of CaTi, and Jamundi, about 15 kms. to the south. A few peop1e, mostTy of socio-economic TeveTs II and III, trave1 to these towns to buy meat. This practice is probabTy confined to peop1e who have famiTy connections in these towns. 134 b) AmbuTant Vendors se11 house-to-house from a barrow, dea1ing mostTy d) in fruits and eggs. A1though they are patronized by six percent of househons their method of Operation makes them difficu1t to inc1ude in a survey, so that TittTe is known about them. They are not often seen in the streets, and are of very sma11 sca1e. They account for Tess than one percent of the food sa1es in the city‘. Commissaries: (Comisariatos) are, strictTy speaking, stores operated by and for the use of the armed forces, and are thus outside our terms of reference. There is on1y one in CaTi. However, the name is aTso commonTy used to denote the stores Operated by the PubTic HeaTth Service, the Puestos de Sa1ud, which are described beTow. Farmers' Markets (Mercados Campesinos). This name was given to the practice of setting aside an area in SiToe Market each Saturday for the exc1usive use of sma11 farmers (campesinos). The idea was that this wou1d eTiminate the gouging middTemen and give a fairer price to both producer and consumer. This practice was started towards the end of 1968, when PIMUR's research program was being formuTated, and so was inc1uded in the program. Two students in the MBA program at Universidad de1 VaTTe, under the guidance of Dr. Hugo Duque, aTso carried out some research into its Operation. It did not show much sign of deveToping into an important out1et for produce. OnTy two percent of the housewives interviewed in the Consumer Survey had visited it. 1PIMUR, Consumer Survey 135 Non-Profit Oriented RetaiTers a) b) PubTic HeaTth Service Stores (Puestos de Sa1ud). There were ten of these registered with the Secretariat of PubTic HeaTth in CaTi in earTy 1969. They are operated principa11y by charitabTe organ- izations in conjunction with sma11 neighborhood cTinics, and are intended to serve the destitute and the needy. These stores se11 a Timited range of stap1es at cost. A1though they are presumab1y of great importance to those who use them (8% of fami1ies in S.E.L. VI and 4% of those in S.E.L. V1) they are not of any significance to the food distribution system of CaTi, and are c1ear1y not candidates for a p1ace in our proposed system. In fact, their operators and everybody eTSe wou1d Tike to see the need for their services e1iminated. Stores operated by IDEMA. IDEMA (Instituto Qg_Mercadeo AgricoTa, AgricuTturaT Marketing Institute) is a department of the Ministry of AgricuTture charged with the promotion of orderTy marketing of agricuTturaT products. In order to maintain a downward pressure on prices IDEMA runs stores in various cities in which there are son grains and some other non-perishabTes, and has a poTicy of announcing in the TocaT newspapers the prices at which these products are currentTy being son. However, the advertising is on1y done sporadicaTTy, and on a very unobtrusive sca1e. OnTy 16% of the housewives interviewed by PIMUR's Consumer Survey knew of this service. 1PIMUR, Consumer Survey 136 At the time of the fie1d work in Ca1i IDEMA operated three stores in the city. The main store was Tocated on the Carrera 15, about tweTve b10cks south of the PTaza Caycedo, the city center. Another was Tocated in the Centro Comercio de1 Norte, in an upper-c1ass neighborhood, whiTe the third was Tocated in the Tower-c1ass barrio of SiToe. EarTy in 1969 IDEMA began a program of sending mobi1e stores into the poorer areas of the city, the S.E.L. VI barrios. Sixteen Tocations are served by these traiTer-stores, each Tocation being visited for a few hours once each week. The timing of the introduction of this service was such that PIMUR cou1d make no estimate of its effectiveness, but the cost of the traiTers was very high, and the week1y visiting scheduTe does not correspond we11 to the practice of the typicaT S.E.L. VI housewife of buying food on a day-by-day basis. A cost-effectiveness ana1ysis of the traiTer- stores wou1d probabTy show them to be rather poor performers. OnTy 2.5% of the famiTies interviewed by PIMUR's Consumer Survey had patronized an IDEMA store in the two weeks preceding the interview. Yet 84% of them knew about IDEMA. The reasons given for not patronizing were, in order of frequency, no IDEMA store near the house, Tong waits for service, poor quaTity of the goods son, and Tack of variety of goods son. IDEMA's food distribution activities are carried out by a section of a department of a Ministry, and this makes its potentia1 performance as a food distributor comp1ete1y unpredictabTe. Its performance wiTT be a function of the abiTity, be1iefs, and prestige of the director of the section (who is TiabTe to be transferred to another section at any time) over-ridden by the corresponding characteristics 137 of the department head (aTso TiabTe to transfer) further over-ridden by those of the current Minister, who is 1ike1y at any time to find himseTf given another portfoTio. ATT of these posts are her by poTiticaT appointees who are changed frequent1y. In addition to this, there is the eTement of unpredictabiTity associated with changes in the re1ative prestige of government departments. In 1969 the Ministry of AgricuTture was very prestigious, and we11 supp1ied with funds. This was because current theories of economic deveTopment attributed great importance to agricuTture. This emphasis cou1d change. Another possibi1ity is that the food retaiTing program cou1d deveTop a momentum of its own, and become a more or Tess autonomous bureaucracy. Whi1e this wou1d make its behavior much more predictabTe it is to be feared that there cou1d be a tota1 Toss of dynamicism, in the manner of bureaucracies everywhere. An important point to be made in respect to IDEMA's food retai1ing activities is that the rewards system, under either a1ternative, wiTT be incompatibTe with the deveTopment of an efficient marketing system for the distribution of economic goods, which requires a consumer orientation. POTiticaT appointees seek either ideoTogicaT ends or high1y-visib1e short-term resu1ts (demagogic ends), whi1e the ends of bureaucrats are served by strict adherence to ruTes and reguTations. A possibTe justification for IDEMA's continuing a food retai1ing Operation might be that of maintaining a countervaiTing force against tendencies to monopoTy practices on the part of private food retai1 organizations. However, it wou1d possibTy be more c) 138 efficient to repTace this retai1 section with a grOUp of officiaTS charged with poTicing the market behavior of firms in the private sector and using the threat Of TegaT action to enforce good behavior. Grocery stores of the Cajas de Compensacion FamiTiar (Fami1y ATTowance Funds). The Cajas appear to be an uniqueTy CoTombian institution. According to the CoTombian Labor Laws each firm above a certain size (with a capitaT of COIJ’$100,000 or more) must pay a sum equiva1ent to 4% of its payroTT into a Caja]. The Caja in turn pays a fami1y subsidy, according to the number of dependents, to each empToyee of the member firms who earns CoTui$1500 per month or Tess. Any funds surpTus to the requirements of the fami1y subsidy must be used in ways which benefit the empToyees of the 2 member firms. Two of the eight Cajas Operating in Ca1i have used some of the surpTus funds to start food stores. .The stores seTT at Tow prices to empToyees of affiTiated firms. One store had been open for about two years when the PIMUR fie1d work was done, the other on1y a few months. Neither store is Tocated in the city center, a1though the newer one is accessibTe from a point where most of the bus routes of the city pass. This "nondescript" Tocation poTicy probabTy Ties behind the patronization pattern shown in Tab1e IV-T. As can be seen intensity of use is 1The sum invoTved is actua11y 7%, but 3% is used to sponsor the NationaT Apprenticeship System. 2There are actua11y three such stores, but one is estab1ished in the form of a cooperative. 139 Tab1e IV-T. Patronization of Caja Grocery Stores, by Socio-Economic Leve1, CaTi, February 1969. Socio-Economic Leve1 1 II III IV V VI Overa11 Percent of Fami1ies 28 20 15 13 4 2 6.7 Using Stores in Each S.E.L. Source: PIMUR, Consumer Survey greatest at high S.E.L. and faTTs throughout the socia1 sca1e. This corresponds with the experience of the manager of the newer store, that operated by ACOPII. In the same budeing ACOPI operates a grocery store and a Tow-price pharmacy, and the manager had noticed that whi1e the customers of the pharmacy were predominantTy bTue- coTTar workers the customers of the grocery were a1most entire1y drawn from the senior ranks, managers and technoTogists. A probabTe exp1anation for this is that drug purchases are unp1anned, of an emergency nature, usua11y of high unit price, and, most important, necessitate a trip into the center, since there are no pharmacies out in the residentiaT areas. Food purchases are routinized, and even if the housewife finds the service offered by the neighborhood store unsatisfactory a singTe bus trip, with no changes, wi11 take her to the Ga1eria Centra12. IACOPI is the acronym of the CoTombian Association of SmaTT Businesses. ‘ 2The author is indebted to David LToyd-CTare, then of PIMUR, fOr this suggestion. 140 A pecuTiarity of the Cajas, at 1east to those accustomed to the industriaT scene in North America, is that whiTe their objective ...-..u- _._..-- is to provide benefits to the workers they are Operated by associations —_..-~"' which represent manufacturers and businessmen. This becomes Tess incongruous when viewed in the context of the paternaTistic tradition of businesses in CoTombia, and it is certainTy true that they are in generaT carrying out the intentions of the Labor Laws in providing benefits to the Tower-paid workers. However, the fact that the Cajas are operated by manufacturer's associations gives rise to a basic defect in their potentia1 as food distributors. The associations are in competition with each other for affiTiates, and they compete by providing ever more impressive services to the empToyees of their affiTiates. Whi1e this is basicaTTy a desirab1e thing it has two drawbacks. The first is that in the struggTe to provide extra services the managers of the Cajas may weTT find that they have undertaken projects which overtax their resources or their manageriaT capacities. FaiTure of a Caja through over-commitment cou1d have serious socia1 repercus- sions. Another basic difficu1ty associated with the Cajas 1ies in their market orientation. The peOpTe who make the decisions concerning affiTiation are the manager of industriaT enterprises, not the workers; so that the Cajas wi11 offer services which wiTT impress managers. Perhaps this exp1ains the success that the Caja grocery stores in Ca1i have had in attracting customers from amongst the sa1aried empToyees of their affiTiated firms. d) 141 A11 things considered, the Caja grocery stores must be rejected as candidates for a significant p1ace in a food distribution system for Ca1i. The capitaT resources present1y tied up in this activity wou1d probabTy be better empToyed in a pension fund, which wou1d benefit both the workers and provide deveTopment funds for the nationaT economy. The purpose of the Cajas exposes them to the dangers of economicaTTy-unsound phiTanthrOpic ventures, whi1e their connections with associations representing the owners and managers renders them poTiticaTTy vuTnerabTe. And as we have seen, they may have a rather dysfunctionaT idea of what the market is that they shou1d serve. Consumer Cooperatives. In February 1969 there were seventeen consumer cooperatives se11ing foodstuffs in CaTi. Like a11 cooperatives in CoTombia they are officia11y under the supervision and controT of the Superintendency of COOperatives in Bogota, and enjoy the encouragement and support of the centra1 government. This support takes the form of tax exemptions, preferentiaT prices for transportation, preferentiaT access to IDEMA's stOcks of foodstuffs, and various other M_&_ _M items. Their poWer to garnishee the sa1aries or wages of members who s‘n-——. — owe them money makes them,of aTT the retai1er types in Ca1i, the one that does the greatest proportion of sa1es on credit. Each co- operative is associated with one firm or institution, and draws its members on1y from that firm. PresumabTy this resu1ts from a too-TiteraT reading of the history of the COOperative movement, but whatever the origin of the practice, it does an immense amount of For a fuTTer description of the priviTeges accorded the cooperatives see the PIMUR FinaT Report, pp. 320-321. 142 damage to the cooperative stores. Since the homes of the members are scattered aTT over the city the stores are inaccessibTe for their target customers, and this is virtua11y a guarantee of faiTure. A few cooperatives have Tocated cTose to the factories of the associated empToyers, so that at 1east the man of the househon can do some buying, but this is on1y a very partiaT soTution of the Tocationa1 prob1em. 1 PIMUR's Consumer Survey found that 7.3% of the fami1ies interviewed had made a purchase at a COOperative during the two weeks preceding the interview, that 6.2% regu1ar1y made purchases at a cooperative, and that 12.4% were members of a cooperative. CTearTy, ToyaTty is not a strong sentiment amongst the members, at 1east not strong enough to counteract the sundry disadvantages of patronizing these stores. In addition to the Tocationa1 inconvenience there are the drawbacks of a narrow range of goods offered, consisting of non- perishabTe stap1es and canned goods, poor stock controT resuhing in stock-outs, and non-competitive prices without compensatory dividend payments. In addition to the poor custom resu1ting from the above drawbacks the cooperatives must aTso cope with the internaT handicaps of poor1y-trained personneT, Tack of centraTized buying, and possibTy a mispTaced be1ief that adherence to the principTes of cooperativism is an adequate substitute for acquaintance with the principTes of merchandising. 1See PIMUR TechnicaT Report #7, p. 67, for a summary of the EeeeTts of interviews with the managers of severa1 cooperatives in a . 143 We wou1d not write off comp1ete1y the possibiTity that cooperatives cou1d pTay a significant part in food distribution in CaTi in the future. But this cou1d on1y happen if, a) the cooperatives changed their format from a factory-based to a barrio-based membership criterion, and b) they were to aTign under a nationaT purchasing organization such as, for examp1e, the Cooperative WhoTesaTe Societies found in Great Britain, and c) their formation were based upon a rea1ization that food distribution in CaTi is an affair of thin margins and keen competition, and not upon the beTief that the spread between farm prices and consumer prices is made up of nothing but profit for the midd1e- man. However, even if the cooperatives were to reform according to the above recommendations, we agree with the recommendations of the 1 PIMUR FinaT Report that their speciaT priviTeges be revoked without bothering with a review. If cooperatiyes cannot compete with private enterprises without beingfsubsidized it is difficu1t to see, on “economic grounds, why they shou1d exist.. It may be possibTe to make '§*Zase for measures faciTitating the estab1ishment (but certainTy not the continued operation) of farm producer cooperatives, on the grounds that the food processors cannot negotiate efficientTy with 1arge numbers of sma11 producers, but this argument does not extend to consumer cooperatives. CaTi in the twentieth century is IPIMUR FinaT Report. pp. 320-321 and p. 323. 144 not remoteTy Tike nineteenth-century RochdaTe. In summary, we feeT that we fan dismiss the non-profit oriented ; . sector of CaTi's food distribution system from further consideration. IDEMA shou1d certainTy retire from food retaiTing, and content itseTf with ensuring that background suppTies of foodstuffs are adequate for anticipated demand (at the storage, or pre-whoTesaTer, 1eve1), particu1ar1y on occasions when the system is TikeTy to be severeTy strained, such as the Pope's visit to Bogota or the proposed Pan-American Games in Ca1i. Further excursions of the Cajas into food retaiTing shou1d be ‘- discouraged, and the cooperatives shou1d be Teft to fend for themseTves. o—ugq—n-u—v—-—- ——-——- ,{1 without speciaT priviTeges. We wou1d, however, recommend that some form . of technicaT assistance be made avai1ab1e to the organizations which Operate the PubTic HeaTth posts. Many of the cases of iTTness which come to them are undoubted1y caused by maTnutrition, and any advice 1 which can improve their techniques of buying and distributing foodstuffs; 1 wiTT be advice we11 worth giving. J Non-Systemic RetaiTers . The c1assification("non-systemic" is not one in common use in the marketing Titerature, but it, or its equiva1ent, is TikeTy to appear frequent1y in the future as interest in macro-marketing grows. It is intended to signify the property of not being Tinked, or on1y very TightTy Tinked, with the system under ana1ysis. The importance of the concept ' Ties in the fact that non-systemic institutions can be ana1ysed in- dependentTy of the main system, and that changes in such institutions wiTT have on1y “simp1e” effects upon the main system, effects which can be taken into account without having to foTTow their repercussions 145 through a11 the components of the main system. Being ab1e to identify non-systemic institutions can greatTy ease the task of the ana1yst. Two sectors of food distribution in CaTi can be described as hon-systemic, the distribution of mi1k and the distribution of eggs and pouTtry. These two inc1ude the majority of the so-caTTed SpeciaTty stores, with the main exceptions being the few specia1ty meat stores, which wi11 be discussed Tater. Whi1e neither sector is comp1ete1y independent of the main system, the connections are simpTe, in the sense described above. a) Mi1k Distributors deaT in two products, pasteurized miTk and raw (unpasteurized) mi1k, and the two are handTed by distinct channeTs. Raw mi1k is distributed by truckers (camioneros) who buy from dairy farmers and se11 mostTy door-to—door. In 1969 there were 190 such truckers who had been Ticensed by the city to deaT in raw miTk. Pasteurized miTk passes through two pasteurizing p1ants in the city, and is then deTivered to food stores and to homes. Tab1e IV-2.gives detaiTs on the retai1 distribution of both types of mi1k. We can note here, anticipating Tater discussion, that sa1es of raw miTk Tab1e IV-2. Retai1 Distribution of Raw and Pasteurized MiTk, Ca1i, T969 Percent of Sa1es Type of Out1et Raw Mi1k Pasteurized MiTk Food RetaiTers 13 57 Home OeTivery 62 31 SpeciaTized Mi1k Stores (Expendios) 15 -- Institutions 10 12 TOTAL 100 100 Source: PIMUR Mi1k Study b) 146 accounts for 2% of food sa1es of the neighborhood stores (tiendas and sma11 graneros) and pasteurized mi1k for 7.4% of these sa1es. A1though pasteurized mi1k is an important item in the sa1es of the neighborhood stores, the interTinkage between the main food distribution system and mi1k distribution is sti11 weak, this means that changes in one system wiTT have TittTe effect upon the other, and cou1d be negTected un1ess extreme changes are contempTated in either system. The on1y point of contact between the two systems is the de1ivery of mi1k to the food stores, and the detaiTs of how this is done are not particu1ar1y criticaT. We can therefore assume that any recommendations that might be made concerning the main system wiTT have TittTe or no effect upon this matter. The SpeciaTized miTk stores (expendios de Teche), mentioned in Tab1e 1V-2.as out1ets of raw mi1k, are very sma11 operations, using one room of a famiTy dweTTing and are Tocated on1y in the poorer barrios. They wou1d appear to be a means whereby the truckers can obtain distribution in areas where purchasers are too few and purchases too irreguTar to warrant regu1ar de1ivery service. In 1969 there were 121 of these estabTishments honing Ticenses to se11 raw mi1k, and there may we11 have been many more Operating without Ticenses. LittTe is known about the expendios, their margin structure, whether they are independent deaTers or tied to particu1ar supp1iers, nor even why the truckers deaT with them instead of with tiendas. PouTtry and Egg_0istributors. The on1y important point of contact between pouTty and egg distributors and the main food distribution system is in egg distribution, where tiendas and sma11 graneros account 147 , 1 for 57% of retaiT sa1es. SpeciaTized retaiTers and whoTesaTer- retaiTers account for 31% of retai1 egg sa1es and 69% of broiTer retaiT sa1es in CaTiI. The whoTesaTers in the channeTs for pouTtry and eggs are aTT speciaTists, and many are owned by producers or producers' associations. The main point about this sector is that broiTer and egg production on a commercisiwseaTeiis a re1ative1y new industry in CoTombia,- T __flfl____flf,_fl-- .. . _ and the move towards forward integration in marketing these products “is newer stiTTz. Therefore we can anticipate that the industry -Wi11 be fTexibTe enough (not tradition-bound) to adapt to changes in the main system, and we can drop the pouTtry and egg marketing institutions from further consideration. SeTf-Service Stores This category comprises the supermarkets and the chain stores. The Caja groceries and the severa1 cooperatives that use a seTf-service format have been discussed ear1ier. The reason for distinguishing this group of stores it that they are profit-oriented stores which are 1arge enough to pay substantiaT sa1aries for trained fuTT-time managers and which are not cast in the traditiona1 mon of CoTombian food retaiTing. If there were any significant innovations deveToping in food retai1ing in Ca1i, it wou1d be most TikeTy to be found amongst this set of retai1ers. ‘PIMUR FinaT Report, Tab1es 3.12 and 3.13 21b., p. 141, p.144 130 institutions of the food distribution system in Ca1i assumed their present formats and modes of operation. These two topics are very cToseTy reTated. In fact, they are two facets of the same topic. SpecificaTTy, when and if we find a retaiTer type which appears to be suitabTe for deveToping into the type we require we wiTT immediateTy be faced with the question of why this deveTopment has not aTready taken p1ace. We wiTT expect to find that there exist barriers to deveTopment. In fact, if we do not find ..fl-I-I- ....— I“ -——""" _——n- SHFD_PBFFIEF5.FPIS indicates that our understanding of the existing ” food distribution system and its environment is incompTete. Even if M we do not find any retai1er types capabTe of deveTopment we wiTT sti11 need to have a good understanding of the circumstances under which the present food distribution system operates. Otherwise we wiTT run the risk of designing for the new system a types of retai1ers and whoTesaTers which are incompatibTe with other institutions in the food marketing channeT. An important concept which wiTT be used in the ana1ysis is that of capacity for change, and its negative, that of an evoTutionary my“- .— dead end. An institution wi11 be considered capabTe of change, or to ‘11. have ev01utionary capacity, if it can make smaTT changes in its format or operating methOds which_resu1t in its being better adapted to its immediate environment. The immediate environment comprises the institution's customers, supp1iers of goods, supp1iers of services (e.g., transportation), and government. The components of better *adaptat_ion are: -..... W—u—n—u—r a) increasing rewards to the owner or operator, b) deTivering better service to customers, and c) receiving better service from suppTiers. 131 "SmaTT" changes are those which preserve the nature of the institution, those Tess in magnitude than a "quantum jump". Thus, if a tienda operator decided to add, say, potatoes to his product Tine this wou1d be a smaTT change, whiTe going to supermarket format (i.e., seTf-service and departmentaTization) wou1d be a 1arge change. SimiTarTy, a market sta11 cou1d not evoTve direct1y into a neighborhood store, since this wou1d entire1y change the nature of the institution. These considerations app1y on1y to the institution, not to the operator. There is no reason why a sta11 Operator shou1d not Teave his sta11 and buy a tienda, for examp1e. But he cou1d not, by a series of sma11 changes, convert his sta11 into a tienda. Note particu1ar1y that change is considered on1y with respect to the immediate environment. Whi1e the operator might optimize the position of his institution within the marketing channeT he cannot change the structure of the channeT as a whoTe. It is quite possibTe to find an inefficient (sub-optimaT) channeT made up of efficient (optimaT) institutions. We are dea1ing here with a system composed of many interTinked components. The comp1exity of the ana1ysis of such systems grows very rapidTy as the number of components invoTved increases. To simpTify ' ‘va-n—u- _ the ana1ysis it is important to identify_those components which have Want.- "" F D I. ll. TittTe or no effect upon the behavior of the system as a whoTe. We W‘vu‘fi- ' can cTassify such components into two categories. There are the minor m.- 1“ components, consisting of a few sma11 units and handTing on1y a sma11 \______.._L_.- proportion of the food consumed in the city, and the "nonesystemief '-"" components;‘ These 1atter are institutions which are on1y TooseTy I n In“ interTinked with the rest of the system, the "main stream" system, such 132 that changes in the "non-systemic" components need not induce compens- tory changes in the "main stream" components with which they are Tinked. This is a suitabTe point to discuss one of the very few generaTizations that can be made about the Operators of the food, distribution system in Ca1i. ATmost without exception they spend a --—-—._ great deaT of timefiand effort on buying and TittTe on se11ing.' They are activeabuyers and passive seTTers. This app1ies at aTT TeveTs of distribution. SuppTiers to the system sit back and wait for orders from whoTesaTers, whoTesaTers sit back and wait for retai1ers to come to them, and retai1ers behave in the same way towards consumers. Z3;;;erare two main exceptions. One is that group of retai1ers which we have denoted as chain stores. These retai1ers are vigorous mer- chandisers. The other exception is provided by processors of branded (i.e., non—stapTe) processed foods, who tend to maintain 1arge and active sa1es foreesL/) We do not know what are the background causes of this behavior. It cou1d be connected with the socia1 va1ues inherited from the warrior caste that coTonised the region many years ago, it cou1d be the resu1t of a defensive preoccupation with traditiona1 skiTTs, or it cou1d stem from an "economics of scarcity" phiTOSOphy, a be1ief that demand aTways presses upon suppTy. ETucidation of the source wou1d be an interesting and worthwhiTe research prob1em. For the moment it is enough to note that the phenomenon exists. We be11eye_ fl"— that it is an important point to our thesis, and we wiTT return to it -u—_———--—' in Chapter Y:-m« ___—__-—-’-.I 133 RetaiTers The many types of retai1ers Operating in CaTi at the time of PIMUR's fie1d work are described here under the foTTowing headings (in order of discussion): 1. Minor retai1er types, those having on1y negTigibTe importance in terms of numbers and of tota1 proportion of food sa1es. 2. Non-Profit Oriented types, those operated by IDEMA, the Cajas de Compensacion, the Puestos de Sa1ud, and the consumer cooperatives. 3. Non-Systemic retaiTers, as described above. These are mi1k deaTers and distributors of pouTtry and eggs. 4. SeTf-Service stores, supermarkets and chain stores. 5. PubTic Markets, with their associated street vendors. 6. CentraTTy-Tocated traditiona1 stores, the whoTesaTer-retaiTers and 1arge graneros. 7. Non-CentraTTy Tocated traditiona1 stores, the sma11 graneros and tiendas. 8. Meat distributors, at a11 TeveTs of distribution. Minor RetaiTer Types a) Markets in other towns. AnaTysis of the Consumer Survey data showed that these were of negTigibTe importance, and the requirement of consumers for Tocationa1 convenience indicates that they wiTT not become more important in the future. The towns invoTved are Yumbo, some 10 kms. north of CaTi, and Jamundi, about 15 kms. to the south. A few peop1e, mostTy of socio-economic TeveTs II and III, trave1 to these towns to buy meat. This practice is probabTy confined to peop1e who have fami1y connections in these towns. 134 b) AmbuTant Vendors se11 house-tO-house from a barrow, deaTing mostTy d) in fruits and eggs. A1though they are patronized by six percent of househons their method of operation makes them difficu1t to inc1ude in a survey, so that TittTe is known about them. They are not often seen in the streets, and are of very smaTT sca1e. They account for Tess than one percent of the food sa1es in the city‘. Commissaries: (Comisariatos) are, strictTy speaking, stores operated by and for the use of the armed forces, and are thus outside our terms of reference. There is on1y one in CaTi. However, the name is aTso commonTy used to denote the stores operated by the PubTic HeaTth Service, the Puestos de Sa1ud, which are described beTow. Farmers' Markets (Mercados Campesinos). This name was given to the practice of setting aside an area in SiToe Market each Saturday for the exc1usive use of smaTT farmers (campesinos). The idea was that this wou1d eTiminate the gouging middTemen and give a fairer price to both producer and consumer. This practice was started towards the end of 1968, when PIMUR's research program was being formuTated, and so was inc1uded in the program. Two students in the MBA program at Universidad de1 VaTTe, under the guidance of Dr. Hugo Duque, aTso carried out some research into its Operation. It did not show much sign of deveToping into an important out1et for produce. OnTy two percent of the housewives interviewed in the Consumer Survey had visited it. 1PIMUR, Consumer Survey 135 Non-Profit Oriented RetaiTers a) PubTic HeaTth Service Stores (Puestos de Sa1ud). There were ten b) of these registered with the Secretariat of PubTic HeaTth in Ca1i in earTy 1969. They are operated principa11y by charitabTe organ- izations in conjunction with sma11 neighborhood cTinics, and are intended to serve the destitute and the needy. These stores se11 a Timited range of stap1es at cost. A1though they are presumab1y of great importance to those who use them (8% of fami1ies in S.E.L. v1 and 4% of those in S.E.L. v1) they are not of any significance to the food distribution system of Ca1i, and are c1ear1y not candidates for a p1ace in our prOposed system. In fact, their Operators and everybody eTse wou1d 1ike to see the need for their services eTiminated. Stores operated by IDEMA. IDEMA (Instituto Qe_Mercadeo AgricoTa, AgricuTturaT Marketing Institute) is a department of the Ministry of AgricuTture charged with the promotion of orderTy marketing of agricuTturaT products. In order to maintain a downward pressure on prices IDEMA runs stores in various cities in which there are son grains and some other non-perishabTes, and has a poTicy of announcing in the TocaT newspapers the prices at which these products are currentTy being son. However, the advertising is on1y done sporadicaTTy, and on a very unobtrusive sca1e. OnTy 16% of the housewives interviewed by PIMUR's Consumer Survey knew of this service. 1PIMUR, Consumer Survey 136 At the time of the fie1d work in Ca1i IDEMA operated three stores in the city. The main store was Tocated on the Carrera 15, about tweTve bTocks south of the PTaza Caycedo, the city center. Another was Tocated in the Centro Comercio de1 Norte, in an upper-c1ass neighborhood, whi1e the third was Tocated in the Tower-c1ass barrio of SiToe. EarTy in 1969 IDEMA began a program of sending mobi1e stores into the poorer areas of the city, the S.E.L. VI barrios. Sixteen Tocations are served by these traiTer-stores, each Tocation being visited for a few hours once each week. The timing of the introduction of this service was such that PIMUR cou1d make no estimate of its effectiveness, but the cost of the traiTers was very high, and the week1y visiting scheduTe does not correspond we11 to the practice of the typicaT S.E.L. VI housewife of buying food on a day-by-day basis. A cost-effectiveness ana1ysis of the traiTer- stores wou1d probabTy show them to be rather poor performers. OnTy 2.5% of the famiTies interviewed by PIMUR's Consumer Survey had patronized an IDEMA store in the two weeks preceding the interview. Yet 84% of them knew about IDEMA. The reasons given for not patronizing were, in order of frequency, no IDEMA store near the house, Tong waits for service, poor quaTity of the goods son, and Tack of variety of goods son. IDEMA's food distribution activities are carried out by a section of a department of a Ministry, and this makes its potentia1 performance as a food distributor comp1ete1y unpredictabTe. Its performance wiTT be a function of the abiTity, be1iefs, and prestige of the director of the section (who is TiabTe to be transferred to another section at any time) over-ridden by the corresponding characteristics 137 of the department head (aTso TiabTe to transfer) further over-ridden by those of the current Minister, who is 1ike1y at any time to find himseTf given another portfoTio. A11 of these posts are her by poTiticaT appointees who are changed frequent1y. In addition to this, there is the eTement of unpredictabiTity associated with changes in the re1ative prestige of government departments. In 1969 the Ministry of AgricuTture was very prestigious, and we11 SUppTied with funds. This was because current theories of economic deveTOpment attributed great importance to agricuTture. This emphasis cou1d change. Another possibiTity is that the food retaiTing program cou1d deveTop a momentum of its own, and become a more or Tess autonomous bureaucracy. Whi1e this wou1d make its behavior much more predictabTe it is to be feared that there cou1d be a tota1 Toss of dynamicism, in the manner of bureaucracies everywhere. An important point to be made in respect to IDEMA's food retaiTing activities is that the rewards system, under either a1ternative, wiTT be incompatibTe with the deveTopment of an efficient marketing system for the distribution of economic goods, which requires a consumer orientation. POTiticaT appointees seek either ideoTogicaT ends or high1y—visib1e short-term resu1ts (demagogic ends), whi1e the ends of bureaucrats are served by strict adherence to ruTes and reguTations. A possibTe justification for IDEMA's continuing a food retaiTing Operation might be that of maintaining a countervaiTing force against tendencies to monopoTy practices on the part of private food retai1 organizations. However, it wou1d possibTy be more C) 138 efficient to repTace this retai1 section with a group of officiaTs charged with poTicing the market behavior of firms in the private sector and using the threat Of TegaT action to enforce good behavior. Grocery stores of the Cajas de Compensacion FamiTiar (Fami1y ATTowance Funds). The Cajas appear to be an uniqueTy CoTombian institution. According to the CoTombian Labor Laws each firm above a certain size (with a capitaT of CoT/'$TO0,000 or more) must pay a sum equiva1ent to 4% of its payroTT into a Caja]. The Caja in turn pays a fami1y subsidy, according to the number of dependents, to each empToyee of the member firms who earns COT./$1500 per month or Tess. Any funds surpTus to the requirements of the fami1y subsidy must be used in ways which benefit the empToyees of the 2 member firms. Two of the eight Cajas operating in CaTi have used some of the surpTus funds to start food stores. .The stores se11 at Tow prices to empToyees of affiTiated firms. One store had been open for about two years when the PIMUR fie1d work was done, the other on1y a few months. Neither store is Tocated in the city center, a1though the newer one is accessibTe from a point where most of the bus routes of the city pass. This "nondescript" Tocation poTicy probabTy Ties behind the patronization pattern shown in Tab1e IV-T. As can be seen intensity of use is 1The sum invoTved is actua11y 7%, but 3% is used to sponsor the NationaT Apprenticeship System. 2There are actua11y three such stores, but one is estab1ished in the form of a cooperative. 139 Tab1e IV-T. Patronization of Caja Grocery Stores, by Socio-Economic Leve1, CaTi, February 1969. Socio-Economic Leve1 1 II III IV V VI Overa11 Percent of FamiTies 28 20 15 T3 4 2 6.7 Using Stores in Each S.E.L. Source: PIMUR, Consumer Survey greatest at high S.E.L. and faTTs throughout the socia1 sca1e. This correSponds with the experience of the manager of the newer store, that operated by ACOPI‘. In the same budeing ACOPI operates a grocery store and a Tow-price pharmacy, and the manager had noticed that whi1e the customers of the pharmacy were predominantTy bTue- coTTar workers the customers of the grocery were a1most entire1y drawn from the senior ranks, managers and technoTogists. A probabTe exp1anation for this is that drug purchases are unp1anned, of an emergency nature, usua11y of high unit price, and, most important, necessitate a trip into the center, since there are no pharmacies out in the residentiaT areas. Food purchases are routinized, and even if the housewife finds the service offered by the neighborhood store unsatisfactory a singTe bus trip, with no changes, wi11 take her to the Ga1eria CentraTz. 1ACOPI is the acronym of the CoTombian Association of SmaTT Businesses. ‘ 2The author is indebted to David LTqyd-CTare, then of PIMUR, fOr this suggestion. 140 A pecuTiarity of the Cajas, at 1east to those accustomed to the industriaT scene in North America, is that whiTe their objective _—--—.-.« 4* ...-u" i§,t° provide benefits to the workers they are operated by associations ”Which represent manufacturers and businessmen. This becomes Tess incongruous when viewed in the context of the paternaTistic tradition of businesses in CoTombia, and it is certainTy true that they are in generaT carrying out the intentions of the Labor Laws in providing benefits to the Tower-paid workers. However, the fact that the Cajas are operated by manufacturer's associations gives rise to a basic defect in their potentia1 as food distributors; The associations are in competition with each other for affiTiates, and they compete by providing ever more impressive services to the empToyees of their affiTiates. Whi1e this is basicaTTy a desirab1e thing it has two drawbacks. The first is that in the struggTe to provide extra services the managers of the Cajas may weTT find that they have undertaken projects which overtax their resources or their manageriaT capacities. FaiTure of a Caja through over-commitment cou1d have serious socia1 repercus- sions. Another basic difficu1ty associated with the Cajas Ties in their market orientation. The peop1e who make the decisions concerning affiTiation are the manager of industriaT enterprises, not the workers; so that the Cajas wiTT offer services which wiTT impress managers. Perhaps this exp1ains the success that the Caja grocery stores in Ca1i have had in attracting customers from amongst the sa1aried empToyees of their affiTiated firms. d) 141 A11 things considered, the Caja grocery stores must be rejected as candidates for a significant p1ace in a food distribution system for CaTi. The capitaT resources present1y tied up in this activity wou1d probabTy be better empToyed in a pension fund, which wou1d benefit both the workers and provide deveTOpment funds for the nationaT economy. The purpose of the Cajas exposes them to the dangers of economicaTTy-unsound phiTanthrOpic ventures, whi1e their connections with associations representing the owners and managers renders them poTiticaTTy vuTnerabTe. And as we have seen, they may have a rather dysfunctionaT idea of what the market is that they shou1d serve. Consumer COOperatives. In February 1969 there were seventeen consumer cooperatives seTTing foodstuffs in CaTi. Like a11 cooperatives in CoTombia they are officia11y under the supervision and controT of the Superintendency of Cooperatives in Bogota, and enjoy the encouragement and support of the centra1 government. This support takes the form of tax exemptions, preferentiaT prices for transportation, . ...-..- preferentiaT access to IDEMA's stocks of foodstuffs, and various other “qw- .__ -h_ M — hm -_-— — —- items.: Their power to garnishee the sa1aries or wages of members who OWe them money makes them,of a11 the retai1er types in Ca1i, the one that does the greatest proportion of sa1es on credit. Each co- Operative is associated with one firm or institution, and draws its members on1y from that firm. PresumabTy this resu1ts from a too-TiteraT reading of the history of the cooperative movement, but whatever the origin of the practice, it does an immense amount of For a fuTTer description of the priviTeges accorded the cooperatives see the PIMUR FinaT Report, pp. 320-321. 142 damage to the cooperative stores. Since the homes of the members are scattered aTT over the city the stores are inaccessibTe for their target customers, and this is virtua11y a guarantee of faiTure. A few cooperatives have Tocated cTose to the factories of the associated empToyers, so that at 1east the man Of the househon can do some buying, but this is on1y a very partiaT soTution of the Tocationa1 prob1em. \ PIMUR's Consumer Survey found that 7.3% of the fami1ies interviewed had made a purchase at a cooperative during the two weeks preceding the interview, that 6.2% regu1ar1y made purchases at a cooperative, and that 12.4% were members of a cooperative. CTearTy, ToyaTty is not a strong sentiment amongst the members, at 1east not strong enough to counteract the sundry disadvantages of patronizing these stores. In addition to the Tocationa1 inconvenience there are the drawbacks of a narrow range of goods Offered, consisting of non- perishabTe stap1es and canned goods, poor stock controT resuldng in stock-outs, and non-competitive prices without compensatory dividend payments. In addition to the poor custom resu1ting from the above drawbacks the cooperatives must aTso cope with the internaT handicaps of poor1y-trained personneT, Tack of centraTized buying, and possibTy a mispTaced be1ief that adherence to the principTes of cooperativism is an adequate substitute for acquaintance with the principTes of merchandising. 1See PIMUR TechnicaT Report #7, p. 67, for a summary of the EeseTts of interviews with the managers of severa1 cooperatives in a . 143 We wou1d not write Off comp1ete1y the possibiTity that cooperatives cou1d pTay a significant part in food distribution in CaTi in the future. But this cou1d on1y happen if, a) the cooperatives changed their format from a factory-based to a barrio-based membership criterion, and b) they were to aTign under a nationaT purchasing organization such as, for examp1e, the Cooperative WhoTesaTe Societies found in Great Britain, and c) their formation were based upon a rea1ization that food distribution in CaTi is an affair of thin margins and keen competition, and not upon the be1ief that the spread between farm prices and consumer prices is made up of nothing but profit for the midd1e- man. However, even if the cooperatives were to reform according to the above recommendations, we agree with the recommendations of the PIMUR FinaT Report1 that their speciaT priviTeges be revoked without bothering with a review. If cooperatives cannot compete with private enterprises without being subsidized it is difficu1t to see, on econOmic grounds, why they shou1d exist.’ It may be possibTe to make -;*;ase for measures faciTitating the estabTishment (but certainTy not the continued Operation) of farm producer cooperatives, on the grounds that the food processors cannot negotiate efficientTy with 1arge numbers of sma11 producers, but this argument does not extend to consumer cooperatives. CaTi in the twentieth century is 1PIMUR FinaT Report. pp. 320-321 and p. 323. ./ h/ 144 not remoteTy 1ike nineteenth-century RochdaTe. In summary, we feeT that we can dismiss the non-profit oriented _ sector of CaTi's food distribution system from further consideration. IDEMA shou1d certainTy retire from food retaiTing, and content itseTf with ensuring that background suppTies of foodstuffs are adequate for anticipated demand (at the storage, or pre-whoTesaTer, TeveT), particu1ar1y on occasions when the system is 1ike1y to be severeTy strained, such as the Pope's visit to Bogota or the proposed Pan-American Games in Ca1i. Further excursions of the Cajas into food retaiTing shou1d be discouraged, ‘and the COOperatives shou1d be Teft to fend for themseTves. without speciaT priviTeges. We_!guId_,_however, recommend that some form of technicaT assistance be made avai1ab1e to the organizations which operate the PubTic HeaTth posts. Many of the cases of iTTness which come to them are undoubted1y caused by maTnutrition, and any advice i e which can improve their techniques of buying and distributing foodstuffsi 1 wi11 be advice we11 worth giving. Non-Systemic RetaiTers The c1assification("non-systemic" is not one in common use in the marketing Titerature, but it, or its equiva1ent, is TikeTy to appear frequent1y in the future as interest in macro-marketing grows. It is intended to signify the property of not being Tinked, or on1y very TightTy Tinked, with the system under ana1ysis. The importance of the concept ' Ties in the fact that non-systemic institutions can be ana1ysed in- dependentTy of the main system, and that changes in such institutions wiTT have on1y "simpTe" effects upon the main system, effects which can be taken into account without having to foTTow their repercussions 145 through a11 the components of the main system. Being ab1e to identify non-systemic institutions can greatTy ease the task of the ana1yst. TWO sectors of food distribution in Ca1i can be described as non-systemic, the distribution of mi1k and the distribution of eggs and pouTtry. These two inc1ude the majority of the so-caTTed specia1ty stores, with the main exceptions being the few specia1ty meat stores, which wiTT be discussed Tater. WhiTe neither sector is comp1ete1y independent of the main system, the connections are simpTe, in the sense described above. a) MiTk Distributors deaT in two products, pasteurized mi1k and raw (unpasteurized) milk, and the two are handTed by distinct channeTs. Raw mi1k is distributed by truckers (camioneros) who buy from dairy farmers and se11 mostTy door-to-door. In 1969 there were 190 such truckers who had been Ticensed by the city to deaT in raw mi1k. Pasteurized mi1k passes through two pasteurizing p1ants in the city, and is then deTivered to food stores and to homes. Tab1e IV-2.gives detaiTs on the retai1 distribution of both types of mi1k. We can note here, anticipating Tater discussion, that sa1es of raw mi1k Tab1e IV-2. RetaiT Distribution of Raw and Pasteurized Mi1k, CaTi, T969 Percent of Sa1es Type of Out1et Raw Mi1k Pasteurized Mi1k Food RetaiTers T3 57 Home OeTivery 62 31 SpeciaTized Mi1k Stores (Expendios) 15 -- Institutions 10 12 TOTAL 100 100 Source: PIMUR Mi1k Study b) 146 accounts for 2% of food sa1es of the neighborhood stores (tiendas and sma11 graneros) and pasteurized miTk for 7.4% of these sa1es. A1though pasteurized mi1k is an important item in the sa1es of the neighborhood stores, the interTinkage between the main food distribution system and mi1k distribution is sti11 weak, this means that changes in one system wiTT have TittTe effect upon the other, and cou1d be negTected un1ess extreme changes are contempTated in either system. The on1y point of contact between the two systems is the deTivery of mi1k to the food stores, and the detaiTs of how this is done are not particu1ar1y criticaT. We can therefore assume that any recommendations that might be made concerning the main system wiTT have TittTe or no effect upon this matter. The speciaTized miTk stores (expendios de Teche), mentioned in Tab1e IV-2.as out1ets of raw mi1k, are very sma11 operations, using one room of a famiTy dweTTing and are Tocated on1y in the poorer barrios. They wou1d appear to be a means whereby the truckers can obtain distribution in areas where purchasers are too few and purchases too irreguTar to warrant regu1ar de1ivery service. In 1969 there were 121 of these estabTishments honing Ticenses to se11 raw mi1k, and there may we11 have been many more operating without Ticenses. LittTe is known about the expendios, their margin structure, whether they are independent deaTers or tied to particu1ar supp1iers, nor even why the truckers deaT with them instead of with tiendas. PouTtry and Egg Distributors. The on1y important point of contact between pouTty and egg distributors and the main food distribution system is in egg distribution, where tiendas and smaTT graneros account 147 ‘ ' for 57% of retai1 sa1es. SpeciaTized retai1ers and whoTesaTer- retaiTers account for 31% of retai1 egg sa1es and 69% of broiTer retai1 sa1es in Ca1i]. The whoTesaTers in the channeTs for pouTtry and eggs are aTT speciaTists, and many are owned by producers or producers' associations. The main point about this sector is that broiTer and egg production on a commercia1isca1e is a re1ative1y new industry in CoTombia, __wfi___fi,#*" and the move towards forward integration in marketing these products “is newer stiTTZ. Therefore we can anticipate that the industry 'WiTT be fTexibTe enough (not tradition-bound) to adapt to changes in the main system, and we can drop the pouTtry and egg marketing institutions from further consideration. SeTf-Service Stores This category comprises the supermarkets and the chain stores. The Caja groceries and the severa1 cooperatives that use a seTf-service format have been discussed ear1ier. The reason for distinguishing this group of stores it that they are profit-oriented stores which are 1arge enough to pay substantiaT sa1aries for trained fuTT-time managers and which are not cast in the traditiona1 mon of CoTombian food retai1ing. If there were any significant innovations deveToping in food retai1ing in Ca1i, it wou1d be most TikeTy to be found amongst this set of retai1ers. IPIMUR FinaT Report, Tab1es 3.12 and 3.13 2ib., p. 141, p.144 148 Supermarkets were defined by PIMUR as profit-oriented food stores satisfying three criteria: a) more than 30 sq. meters of seTTing and storage space b) seTf-service format c) departmentaTization by product Tines (which aTso impTied offering a fuTT Tine of foodstuffs) ' Nineteen such stores, of which six were branch stores, were operating in Ca1i in earTy 1969. A11 were Tocated in or near upper- cTass neighborhoods. WhiTe 7.6% of respondents to the Consumer Survey used them, a1most aTT their patronage (83%) came from peop1e in socio- economic TeveTs I, II, and III. Their tota1 sa1es account for cTose to five percent of tota1 retaiT food sa1es in Ca1i]. When one notes that upper-c1ass househons spend much more than Tower-c1ass househons upon food (per famiTy) the above figures indicate that the supermarkets are on1y capturing about one-fourth of their patrons' food expenditures. This checks with the figures presented in Tab1e III.1.3., which show that the seTf-service stores in generaT are of importance on1y in grains and canned foods. For meats and fruits and vegetab1es peop1e in the upper socio-economic TeveTs patronise the pub1ic markets and specia1ty stores. This is not reaTTy surprising. The supermarkets carry a good se1ection of canned, bottTed, and packaged foods, and of wines and Tiquors. In produce they are poor, the produce section usua11y consisting of a set of grimy bins at the back of the store honing a 1PIMUR, Consumer Survey 149 very Timited and rather sad-Tooking se1ection of fruits and vegetab1es. The meat departments are a bit better, but are stiTT not good. The supermarkets, in fact, might be described as being production- oriented or product-oriented, in the sense that they devote most of their effort to doing the things that they find easy to do, such as hand1ing non-perishabTe articTes. They are certainTy not marketing- oriented. Of the firms interviewed one advertised in newspapers, three on radio, and two used both media. Other advertising was Timited to a few firms which used posters. Price-based advertising or promotion was conspicuous by its absence. The supermarkets in CaTi, in short, are not mass-merchandisers but rather neighborhood stores of a 1arge and showy format. At present they exhibit no characteristics that wou1d Tead us to be1ieve that they wiTT natura11y become the dominant mode of food retai1ing in Ca1i in the near future. In fact, they have as a cTass been rather unsuc- cessfuT in Ca1i. Whi1e poor management (in the sense of day-to- day administration) has probabTy contributed to this, the main barrier to deveTopment is a1most certainTy their Tack of a marketing orientation. ATso, as we sha11 see Tater, the whoTesaTe sector is not compatib1e with efficient supermarket operation. There is a thriving chain of supermarkets in Bogota, which indicates that supermarkets are not incompatibTe with the CoTombian ambience, but the secret of its success seems to Tie in the quaTity of its management. Most of the managers are members of the fami1y which owns the chain, and they receive very eTaborate training. Whi1e there is no obvious reason why supermarkets, properTy managed, shou1d not pTay a part in a new food distribution system, there is no evidence that supermarkets 150 are present1y generating changes that cou1d bring about such new systems. A1though the expectation of finding constructive innovation in the supermarket sector has been disappointed, there is one point about them that is instructive. They attract to their stores a 1arge proportion of those 1iving near. The PIMUR Consumer Survey indicated that most (72%) housewives of S.E.L. I patronized supermarkets, as did many housewives of S.E.L. II and III (34% and 25% respectiveTy). Yet, as was noted above, they do this without any significant amount of advertising or pubTicity. It wou1d appear that their visibiTity is an important factor. They are 1arge stores by CaTi standards, with distinctive formats and 1arge neon signs. By reason of this visibiTity they provide their own reminder advertising. To summarise, whim-— the supermarkets_in Ca1i provide some interesting and usefu1 insights, U... ‘3‘... Han-‘- but do "9.13.3.1: present give any evidence of deveToping into the type of store needed for our proposed fOod distribution system. Chain Stores 2' m I Mo—vvv .. -\‘ . . These are 1arge stores of the type known as "five-and-tens" in the U.S.A.. They se11 a wide variety of househon goods, inc1uding a good se1ection of canned and bottTed foods and grains. There were five firms, of which at 1east two were nationaT in sc0pe, Operating ,14 stores in CaTi in eariy 1969. Most of the stores were situated near the Ga1eria Centra1. This group is not important_1n,foodlretaiTing,_eecounting for on1y about 2% of the tota1 sa1es of this sector. They are most m” wide1y knOwn fOr their practice of using rice as a Teader in their ___.-.—..—.- 151 frequent promotions. The principal importance of these stores lies in their potential as the "new waveh in Colombian retailing. Prices, products, promotion, and place all give the impression of being aimed, with few exceptions, at a Specific target market, the working-class and lower- middle class consumer. As far as we know, they are the only organizations of Colombian origin that give evidence of having successfully solved problems of organization, training, delegation, and control. _They would well repay intensive study by those interested in economic development. |[ .._--'-' ....— bissemination of their problem-solving ability and methods would be more ‘ valuable than any other form of economic aid so far devised. It is interesting to note that in l958 Peter Drucker1 described a type of enterprise which, he said, would be an extremely effective catalyst of economic development. His description corresponds precisely to these chain stores. Public Markets Cali's public market system comprises a central market, the Galeria Central, and five satellite markets, the plazas satelites, dotted around the cityz. They are administered by EMSIRVA, an agency of_the municipal government. 1Peter F. Drucker, "Marketing and Economic Development", Journal of Marketing, Vol. 22, no. 3., January l958. ZSince the PIMUR field work was done in early l969 a new satellite market has been Opened at Alfonso Lopez, on the eastern edge of the city, and the public markets of the Galeria Central area have been demolished. 152 Until 1929 the hub of Cali's food retailing was an Open- air market in the Plaza Cayzedo, the square which is now in the heart of the city's business district. Taking advantage of the opportunity offered by the centennial celebrations in 1929, the city government built a market building in Calvario barrio, and moved the stalls down there. The move was not large in terms of distance, being only six blocks south and one block east of the original location. Calvario barrio was then, and still is, the district where the food wholesalers Operated. In those days food was broughtin on mules and horses, now it comes in by truck and bus, but otherwise there has been little change in the areajmm-. “ , While the name Ga1eria Central Officially applies only to the building constructed in l929 (and since expanded) in practice it is commonly applied also to Calvario I and Calvario II, markets which occupy the two blocks east of the Galeria Central prOper. Also associated with the public markets in this area are swarms of street vendors (vendedores ambulantes) who sell primarily fruits and vegetables in the streets immediately around the market buildings. Some of these are officially recognized by the authorities, and pay a day-by-day rental to EMSIRVA, but about 75% of them are Operating illegally. The police attempt to curtail the activities of the unofficial vendors, but with little effeet.rr II|___..,.._r-----—---—-—_._ _.... _ -_ ' In addition to the food retailers associated with the public markets there are large numbers of traditional food retailers located in the vicinity of the Galeria Central. Some of these, the wholesaler- retailers, also sell foodstuffs at wholesale. Altogether, the Ga1eria Central area accounts for about one-quarter of Cali's retail sales of ”fin—'- ...r J c I I residents of the upper—middle class housing developments nearby. l53 fOod, the public market and the traditional retailers sharing this volume roughly equally. A major problem with the Galeria Central area is the extreme congestion in the streets, compounded of the heavy flow of shOppers, the activities of the street vendors, and the movement Of trucks and buses associated with the presence of food wholesalers in the vicinity. This congestion imposes a fairly heavy social Cost in wasted time, and makes refuse removal difficult. It also attracts numbers of pickpockets and purse-snatchers.-rr “ The satellite markets each consist of one large building and a fenced-in area in which street vendors can rent space. They have not attracted the concentrations of customers, traditional retailers, illegal street vendors, and wholesalers which characterise the Galeria Central (although Alameda market was showing signs of moving this way in l969) and as they are Of more modern design EMSIRVA is better enabled to keep up with the removal of refuse. Of the satellite markets, Cristobal ColongrPorvenir, and Floresta are located in working-class barrios, Alameda is in a middle-class barrio, and Siloe, which was intended to h— ‘Ih-u-n- serve the slum area of Siloe, now finds itself used increasingly by the Each public-market building houses several hundred stalls (most of the stalls in Porvenir and Siloe markets are not occupied), and each stall Specialises in the sale of one line of goods. The principal specialties are meats, grains and processed foods, fruits and vegetables, and dry goods (low-priced housewares and kitchen utensils). Stalls are segregated by specialty, all the meat stalls being located together in one part of the building, etc.. Table IV.3. I34 gives the breakdown of stalls by Specialty and market. Table IV.3. Distribution of Market Stalls by Specialty and Market, Cali, February 1969. ‘Galeria Specialty Central Alameda C.Colon Floresta Siloe Porvenir Total Fruits and Vegetables 628 243 37 65 26 66 1065 Meats and Fish 268 44 42 38 15 31 438 Grains and Staples 126 27 33 29 10 18 243 Dairy 65 7 10 -- 4 -- 86 Poultry and Eggs 4O 4 -- 2 -- 6 52 Bread and Pastas 24 2 3 4 -- -- 33 Others , 334 47 46 23 8 23 481 Total Stall Operators 1485 374 171 161 63 144 2398 Street Vendors 925 90 219 57 31 35 1357 TOTAL 2410 454 390 218 94 179 3755 % of Total 64% 12% 10% 6% 3% 5% 100% Source: EMSIRVA and PIMUR, Wholesale-Retai1 Survey The public markets handle 20% of Cali's retai1 food sales, with the Ga1eria Centra1 accounting for 12% and the satellite markets for 8%.__Of the total public market sales 81% are done by fixed stalls and 19% by street vendors]. wk! IPIMUR Final Report, p.47 155 The Ga1eria Centra1 constitutes the standard of food retai1ing in the city. It is here that the prices are lowest, that the greatest variety Of goods are Offered, and that the best quality is available]. There are some exceptions; the lowest price for rice is fOund in the chain-store promotions, some supermarkets offer a better selection of luxury processed foods, and the specialty stores offer consumers in the upper income brackets a better range of cuts of meat and better quality. But in general the Ga1eria Central is the institution to which others are compared. The satellite markets follow close behind, having slightly higher prices and offering a somewhat more restricted selection of foods. One explanation for this is the fact that the public markets approximate the "perfectly competitive market" of the price-theorists. There are large numbers of buyers and sellers, the goods involved are homogeneous (not branded nor otherwise differentiated), perfect infOrmation exists since customers only have to walk a few feet to Obtain other price quotations, and mobility of the customers is very high indeed. There is even a close approximation to frictionless entry and exit on the part of sellers. It is therefore natural to expect that so-called "normal" profits will be returned to sellers and that margins and prices will be low. The stallholders in the public markets also receive a hidden subsidy, in that they do not pay an economic rent for their stalls. PIMUR estimated that EMSIRVA's rental income from stalls only covered about half of the costs of the public market system. 1PIMUR Wholesale-Retai1 Study, Price Study, and Consumer Survey. 156 It should also be noted that the prices in the Ga1eria Central are not, in a sense, the final prices to consumers. The stallholders of the Ga1eria Centra1 provide no place utility, and so the consumer is faced with the expense Of travelling to and from the city center. As we saw in Chapter III, these travel expenses are quite significant. Although the market stalls are small Operations in terms of physical size some Of them do a surprisingly large volume of business. Meat stalls have average monthly retail sales of Col./$26,500, with beef stalls in the Ga1eria Centra1 averaging Col/’$46,SOO monthly when their wholesale business is included. Grains stalls average Col. $16,850 per month, fruits and vegetables COl./$4,765 per month, and street vendors COl./$3,860 per month. These averages conceal significant differences between the Ga1eria Central and the satellite markets. Sales per stall in the Ga1eria Centra1 of fruits and vegetables are about 15% higher than in the satellite markets, while those of grains are about 20% lower, presumab1y reflecting the competition of the concentration of traditional stores in the Ga1eria Central area]. The average stallholder had 3.8 years Of formal education, and only 13% of stallholders had more than 5 years. Thirty percent had been in business two years or less, twenty-five percent had been in business more than twelve years, and the average was eight years. Fewer than ten percent of stallholders are under thirty years Old, their average age being 45 years. They were a rather suspicious and distrustful group, approximately 60% of them believing that their competitors IPIMUR Final Report, p.62 157 sold adulterated goods, cheated their customers in the matter of weights and measures, and were hoarders and speculators. It is somewhat surprising to note that while 80% of them believed that their competitors sought new techniques to improve their businesses only 43% said that they themselves would like to receive technical assistance. Another odd thing, surprising in view of the state of atomistic competition which should theoretically Obtain in the markets, is that only 29% of the Stallholders said they set their prices in accordance with prices prevailing in the market. Twenty percent price on a"cost plus fixed percentage" basis, while 48% use a "cost plus fixed sum" basis. These figures are self-consistent, with a not-unfamiliar picture emerging from them. The market stallholder in Cali is a man whose age and lack of education renders him unwilling to change his customs and/or unable to change them. He sees changes in the environment (Cajas, COOperatives, etc.) but cannot understand them. He prefers to attribute his troubles to sharp practices on the part of his competitors, perhaps because this relieves him from any Obligation to change his own style Of operation. He is inward-looking, preferring to do what he is accustomed to do rather than to keep a check on the state of the market or to find out what his customers want of him. In many ways he is an urban equivalent of the stereotype of the peasant farmer. One conclusion that can be drawn immediately is that any recommendations we might make about introducing changes in Cali's fijpod distribution system will have to be based upon the assumption that the stallholders will Operate in the new system in exactly the same way that they Operate in the present system. While changes 158 might be possible in the public markets, they will have to be changes at the level of EMSIRVA, the agency reSponsible for administering the markets. Hitherto in this chapter we have been able, for various reasons, to dismiss the retailer types discussed from further considera- tion as potential participants in a new food distribution system for Cali, in the sense that none of them had the potential to become the "new wave" in the city's food distribution. While the discussion above would indicate that the public~market stallholders have only a very small cafiaaiiy for adapting to change, we cannot summarily dismiss them from furtherchnsideration. In the first place the public markets account for an important proportion of total retail food sales in Cali. In the second place the meat vendors in these markets are, as we shall see later, intricately intertwined with meat wholesaling activities, and in the third place, it seems that for some time to come a market sta11 system will provide the best method of retailing fresh fruits and vegetables, products which do not lend themselves easily to routinized mass merchandising. Centrally Located Traditional Stores By traditional stores we mean those stores in which customers stand on one side of a counter and are served by a shOpkeeper and/or .vr' . his assistants on the other side. PIMUR found it convenient to dis— mtinguish four types of such stores, tiengag and small grgnergs, which are located in the barrios, and large graneros and wholesaler- retailers, which are located in the near vicinity of the Ga1eria Central. Table IV.4. summarizes some characteristics of these centrally-located traditional stores. Table IV.4. Characteristics of Large Graneros and Wholesaler- Retailers, Cali, February 1969. Large Graneros Wholesaler-Retailers Monthly Sales - Range C01.($) 50 Ubl~r30,(hl 160,0C0- 350,000 Mean Col.($) 102,600 231,600 Selling Area, sq. m. 55 121 No. of Paid Employees 3.0 4.5 Investment - Fixed Col.($) 13,000 34,540 Liquid Col.($) 22,627 49,096 --. W—OR-M-‘n- -—-O--¢‘ ----c~o——. _- Source: PIMUR ldholesalemRetail Study These stores deal in grains, nonnperishables, and processed foods. They do not carry fresh fruits and vegetables, presumably because the competition from nearby stallholdw rs and street vendors is so strong. While they do not sell meat themselves they rent space in their stores to butchers, thus providing izheir meats“ rs with a wider range of offerings. As is to be expected free the intensely competitive nature of selling activities in the Ga1eria Central area, their prices have to be in line with those of the market stallholders, and so their gross margins are low. PIMUR estimated that the margins, inclusive of spoilage, were 9.2% for grain stalls in the market, 9.5% for the large graneros, and 7.2% for the wholesalervretailers. It is not clear whether the prices are set by market stallholders, acting as atomistic competitors, with the traditional stores meeting the price, or vice- versa, with the wholesaler~retailers and large graneros acting as price- leaders. 160 PIMUR counted 152 large graneros and 24 wholesaler-retailers, who accounted respectively for 11.4% and 4.8% of Cali's retail food sales. This means that these two groups of stores together sell rather more food than do the stallholders of the three Ga1eria Central markets, and that slightly over one-quarter of all retail sales of food in Cali are made in the Galeria Centra1 area. These two groups of stores are very similar in their operating styles. Neither gives credit to any great extent, and neither offers delivery service. They can be differentiated on the basis of size, as in Table IV.4., but the primary difference between the two is a functional one. The wholesaler-retailers make a substantial proportion of their sales to other retailers while the large graneros sell entirely to consumers. These other retailers are operators of small tiendas in the barrios, who by reason of their small scale cannot buy by the sack or case, and so need a bulk-breaking service. The specialist wholesalers in the city do not provide such a service, and the wholesaler-retailers have stepped into the space. PIMUR's interviewers reported that most of the wholesaler-retailers were originally retailers who had developed "backward" into wholesale activities rather than wholesalers who had developed "forward". Another report made by the interviewers was that the wholesaler- retailers found some difficulty in estimating the proportion of their sales made at wholesale. This is understandable, since a housewife buying once per week for a large family could well make larger unit purchases than the operator of a small tienda buying daily, so that it would be impossible to distinguish between wholesale and retail sales solely on the basis of size of unit purchase. It indicates, however, 161 that the wholesaler-retailers do not operate separate wholesale and retail departments. This checks with the results of the interviews, which show that only four out of the twenty-four wholesaler-retailers offer volume discounts. But this raises the question of explaining how the wholesaler-retailers attract the custom of the small tienda operators. It is also difficult to explain how the lower margins of the wholesaler-retailers, which presumably are reflected in lower prices, can be maintained in the intensely competitive conditions of the area around the Galeria Central. There are some other interesting points to be noted about the operators of these two groups of stores. As Table 1V.5. shows, the operators of the large graneros are much younger than the operators of the wholesaler-retailers and have been in business for a much shorter time. It is tempting to suppose that there has recently been an influx of young entrepreneurs into the large granero sector. A 2 x 2 cross-split of operators above and below average age against those above and below average time in business lends moderate support to this idea. The four-way split is significant, by the chi-square test, at the 10% level of probability but not quite at the 5% level. Table 1V.5. Characteristics of Operators of Large Graneros and wholesaler-Retailers, Cali, February 1969. Large Graneros Wholesaler-Retailers Average Age, years 33.3 42.3 Average Time in Business, years 4.9 7.6 Average Level of Education, years 6.5 6.3 Source: PIMUR, Wholesale Retail SurVey 162 From the point of view of potential for change this result is very interesting. It indicates that there is a substantial body of people in the food retailing sector who have had experience of operating stores under tightly-competitive conditions,whose level of education is high enough to enable them to benefit from training in management techniques, and who are still young enough to be able and willing to change their Operating methods to a substantial degree. Such people could well play an important role in any changed food distribution system we might recommend. As far as the wholesaler-retailers are concerned, the potential for change is less apparent. However, they take a rather less distrustful view of their competitors than do the market stall- holders or the Operators of small graneros and tiendas, reversing the 60-40 split that the latter show in believing that their competitors cheat their customers, adulterate their goods and hoard and speculate. Again, two-thirds of the wholesaler-retailers said that they would like to receive technical assistance, as compared to only 43% of market stallholders. Non-Centrally Located Traditional Stores The small graneros and the tiendas are the neighborhood stores of the city. PIMUR distinguished the two types on the basis of volume of sales, tiendas having sales up to Col./$20,000 per month and graneros having sales of from Col./$20,00l to Col./$50,000 per month. There are also qualitative differences between the two types. For instance, almost all small graneros offer fresh meat, while tiendas rarely do. Also, the small graneros is a grocery store pure and simple while the 163 tienda, though it may be a very small grocery store in the same sense, may be a specialty store (bakery, delicatessen, etc.) or it may make a substantial proportion of its sales in beverages, to be consumed on or off the premises. In the limit it may be a sidewalk cafe, or a bar. It was estimated that beverage sales in the tiendas accounted for l8% of their sales volume and 35% of their profits]. There is a corresponding difference between the physical plants of the two store types. The tienda tends to be located in the front room of a dwelling, while the small granero is usually located in a building intended (or suitably modified) for use as a store, which may or may not have living quarters attached. Table IV.6. Characteristics of Small Graneros and Tiendas, Cali, February l969. Tiendas Small Graneros Monthl Sales - Range Col.($ 0-20,000 20,001-50,000 - Mean Col.($) 6,900 33,l00 Selling Area, sq. m. 22 35 No. of Paid Employees 1.9 2.7 Investment - Fixed Col.($) 4,450 7,500 - Liquid Col.($) 1,180 7.l36 Source: PIMUR wholesale-Retail Study 1PIMUR Final Report, p. 46 164 Since the tiendas and the small graneros are, for a large proportion of Cali's population, the sole food suppliers they carry all lines of foods - meats, grains, processed foods, and fruits and vegetables. But the selection is very limited compared with that offered by the larger stores, and very often an individual store will not Offer all lines. This is particularly true in the case of tiendas. For example, only 30% of the tiendas carry meat]. PIMUR's Wholesale-Retail Study estimated that at the time of the field work there were 2,700 tiendas and 670 small graneros operating in Cali, accounting respectively for l3.6% and l6.3% of total retail food sales in the city, or, between them, for 29.9% of these salesz. The neighborhood stores thus constitute an important component of Cali's food distribution system. Their margins are high compared with those of the large graneros. Inclusive of spoilage, margins for tiendas were 13.7% and for small graneros 12.2%3. There is little that is surprising about the operators of these stores. Table 1V.7. gives some of their characteristics, showing that, as is probably to be expected, the operators of the larger stores, the small graneros, are rather better educated and have been in business longer than the Operators of the smaller stores. One interesting point, not noted in this table, is that 28% of the small 1PIMUR Final Report, p.46 zThis agrees well with the PIMUR Consumer Survey's estimate of 3l.3%. 3PIMUR Final Report, Table 2.19 (p.62) 165 granero operators and 4l% of the tienda operators have been in business for one year or less]. This checks with PIMUR's experience that about one-quarter of the tienda operators on the interview list went out of business during the three months that elapsed between preparing the list and starting the interviews. There can be little doubt that in many cases operating a tienda is a form of concealed unemployment. While Table 1V.7. Characteristics of Operators of Small Graneros and Tiendas, Cali, February 1969 Tiendas Small Graneros Average Age, years 44.9 44.8 Average Level of Education, years 5.0 5.7 Average Time in Business, years 4.7 6.4 Source: PIMUR Wholesale-Retail Study Data concealed unemployment or underemployment is preferable to no employment at all, this particular form of it has the undersirable aspect of bringing with it the total or partial loss of the money, usually hard-earned savings, invested in the operation. This is not only disastrous for the individuals concerned, it is also a serious matter for the national economy, which is chronically short of investment funds. Admittedly, there.is virtually no machinery set up to attract these funds into the national economy, but we wish to point out the wastage involved in dissipating these saved accumulations. 1PIMUR Wholesale-Retail Study Data 166 Meat Distribution1 As was shown in Chapter II, meat is an important item in the Cali family's food budget. In total, meat accounts for 31.5% of food expenditures in the city, and of this the "red meats", beef and pork, account for 21.5% and 3% respectively. We shall confine our attention in this section to beef distribution. Pork distribution is less complex, yet sufficiently similar to beef distribution that our recommendations will be applicable to both systems. There are two main flows of beef into the city. The larger flow, which amounts to about 80% by weight of the total, goes through Cali's municipal slaughterhouse, while the remainder goes through the slaughterhouses of other cities of Valle. This latter flow, the so- called forranea beef, is handled by dealers licensed by the Cali municipal government and pays municipal taxes. Over half the cattle comprising the main flow come through the cattle market in Medellin, the rest being assembled from cattle grazers in Valle. Note that the activities of the Medellin cattle market provide a price basis for transactions in other areas. This is effectively the only national-scale commodity market in Colombia. This stage of the marketing process is handled by assemblers, of which 30 were active at the time of PIMUR's field work, although their association, ASADEGA, has 40 registered members. 1The source of the data quoted in this section is the PIMUR Meat Study, the findings of which are reported in PIMUR Final Report, pp. l24-l40. 167 Cattle are purchased from the assemblers on the basis of inspection on the hoof. The cattle then pass through the slaughterhouse and the carcasses are delivered to the buyers. Since this is a hot- meat system (refrigeration is not used) the slaughtering takes place at night and delivery is done in the early hours of the morning. The slaughterhouse is operated by EMSIRVA, the agency which is also responsible for the administration of the public markets, and the delivery of carcasses is done only by EMSIRVA's trucks. Deliveries are made only to the satellite markets and to dealers in the Ga1eria Central area. Dealers located elsewhere in the city have to make arrangements with other dealers in the Ga1eria Central area to take delivery of their carcasses and to arrange for shipment to their places of business. The minimum unit of transaction with the assemblers is one animal, so that only butchers who can count on selling at least one carcass per day (one day is the shelf—life of meat in a hot-meat system in a climate such as Cali's) can deal directly with them. This effectively limits their clients to the butchers located in the Ga1eria Central area and to stallholders in the public markets. Although these stallholders usually sell less than one carcass per day, the conditions of close proximity in the public markets make it feasible to form small syndicates for buying purposes. This option is not available to the tienda or the small granero. There were in 1969 172 public—market stalls dealing in beef, and ll9 butchers (the so-called bancos de carne) operating in the Galeria Central area. These bancos de carne, which are stalls like those in the public markets, are located in six famas (establishments devoted exclusively to renting space to butchers) or in the traditional 168 stores (large graneros and wholesaler-retailers), 39 of which rent Space to butchers. Since the average daily throughput of the slaughter- house in l968 was 288 cattle, l54 hogs, and 30 calves, it can be seen that the operators described above usually handle only one or two carcasses per day. The public-market stallholders account fOr 25% of retail beef sales in Cali, and the bancos de carne of the Galeria Central area fOr 8%. They also supply at wholesale, the tiendas and small graneros who account fOr 47% of retail beef sales]. This wholesaling must be considered as most unsatisfactory, since the prices charged at wholesale are the same as at retail, and the buyer gets no choice of cuts. The custom is to sell by the arroba revuelta, a 25-lb. bag of mixed cuts. In effect, the only service provided is that of making beef available in the early- morning hours, when the operators of the tiendas and small graneros are in the Ga1eria Central area making their purchases. The 42 licensed dealers in forranea beef buy at the slaughter- houses of other cities in Valle, and sell to institutions in Cali, to supermarkets, and to the 12 specialized meat stores in Cali. The reasons given for patronizing the forranea beef dealers were lower prices, better quality of beef, and the greater choice of cuts available. The first reason is important to the institutions, and the others to the super- markets and specialized meat stores, who cater to customers in the upper socio-economi c levels . 1959 of these establishments sell beef directly, and 607 rent Space to bancos. IV? The assemblers take a gross margin of about 12% on sales. Slaughterhouse charges take rather over half of this. Other Operating expenses include the transportation of live animals to Cali, losses on weight shrinkage, and losses due to the death or crippling of animals. The gross margin appears to be in line with the services rendered. The margins taken by other Operators in the system are shown in Table IV.8.. Note that the margins for tiendas and small graneros should be added to those of the wholesale-retai1 outlets in considering the prices to consumers. Table IV.8. Average Margins Taken by Meat Dealers, Cali, February 1969. Outlet Type Beef (%) Pork (%) Wholesale-Retail Bancos 15.5 12.2 Tiendas, Small Graneros 8.3 9.5 Public Market Stalls 14.5 9.2 Supermarkets 10.4 8.1 Specialist Meat Stores 14.2 14.0 Source: PIMUR Final Report, Table 3.4, p.133. The margins taken by the wholesale-retai1 bancos and by stall- holders seem high compared with the rather rudimentary services offered. That they are not the consequences of small scale of Operation is evidenced by PIMUR's estimate that the 923593 netted over COl./$13,000 per month and stallholders almost Col./$5,000 per month, which are very good incomes by the standards obtaining in Cali. The margins are much higher than needed to cover the fixed costs of being in business and to support the Operator and his family. 170 All the symptoms of market imperfection are here, high profits, high margins, and poor service. The question is, how are these imperfec- tions kept in existence? We argue that the imperfections can be ascribed to the fact that the minimum transaction with the assemblers is one animal, a unit purchase of about Col./$2,300. This is a sizeable amount of money to tie up in a product with a shelf-life of only one day, unless there is a virtual- ly assured sale fOr it. A new entrant to the beef retailing business has no such assurance; indeed, he will have to throw away a good deal of each carcass each day while he is building up his business, a process which could take weeks or months. Thus the capital requirements for entry are quite substantial. Added to this is the need to acquire a certain amount of knowledge of meat cutting, which can only be done by working for a while as a butcher's assistant. These two requisites for entry, capital and skill, are mutually contradictory, except in rare instances. Someone possessing the amount of capital needed will not work as a butcher's assistant, and a butcher's assistant will seldom be able to command the required capital. Beef supp1iers typically give one or two days credit, which is of little help to someone trying to build up a clientale. In any case, it is by no means certain whether this credit will be given to other than established butchers. ‘ The large unit transaction is also a barrier to the growth of already established enterprises. A butcher selling one carcass per day will only purchase another carcass per day if he is convinced that his potential market is at least twiceas large as his existing market, while for the man selling two carcasses per day the potential increase in sales has to be at least 50%. (A further consideration is that a butcher se11ing three or more carcasses per day will probably have to delegate the meat-cutting to assistants and spend his time in supervision, a substantial change in his method of Operating.) This potential sales increase will be almost impossible for the stallholders and Operators of the bancos to perceive because of their Operating methods. First, since they operate in Open stalls, customers can see from a distance whether they have any meat left to sell, and if they do not the customers will not bother to make known their unsatisfied demand. Secondly, it is their practice, as soon as they have sold what they have in stock, to clean up and close up. They do not keep fixed hours. Thus their sales estimates are self-fulfilling prophecies, and they see no unsatisfied demand. ‘ Therefore, we find a situation in which each seller adapts the logical policy of stocking as many units as he expects to sell, but in which supply is so restricted in comparison with demand that each seller, without collusion, can make excess profits. Sound business policy is socially dysfunctional. ’ The culprit seems to be the practice of dealingin units of one animal at the wholesale level. The associated problems of in- divisibility prevent adequate adjustment of supply to demand. While a shortage of meat stalls in the Ga1eria Central area might be a contributing factor, by acting as a barrier to entry, the fact that in some of the satellite markets, eSpecially Siloe, many meat stalls are unrented, indicates that this particular barrier to entry probably does not play a significant part in restricting supply. 172 Retailers - Evaluation of Performance With the exception of the traditional (personal service) stores and the public markets, we have rejected all the types of retailers in_ -52 _ _ .. -——_.._.._._-...-—-_~____~ ._ -~—‘—..._. ...»— Cali_as candidates for the primary positions in future food distribution systems fOr the city. The prospects fOr the two types excepted from this conclusion will be discussed in Chapter V. In this section we shall make some comparisons between the main retailer types which could not be fitted into our previous type-by-type discussion format. Table IV.9. brings together data that has been used ear1ier, illustrating the relative importance of the principal groups of retailer types in the distribution of foOd in Cali. This table reinfOrces our previous discussion of the overall importance of the traditional (personal service) stores, and the rather specialized part played by the public markets. Table IV.9. Relative Importance of the Main Retailer Types in the Distribution of the Principal Foodstuffs. Ca1i, February 1969 . Percent of Total Sales of Each Product Made by: Public Traditional Self-Service Specialty Product Markets Storesa Storesb Outlets Meat 30.5 54.0 7.0 8.5 Fruits and Vegetables 51.5 24.9 7.0 16.6 Grains 13.7 72.4 13.9 -- Dairy Products 3.0 41.1 10.6 45.3 Poultry and Eggs 4.6 37.5 16.6 41.3 Processed Foods 6.7 65.9 23.5 3.9 TOTAL 20.1 55.2 12.5 12.2 Source: PIMUR Consumer, Retail, and Commodity Studies, 1969 Notes: a. Includes non-centrally and centrally-located stores. b. Includes supermarkets, COOperatives, chain stores and Cajas. 173 An important desideratum of any economic system is that it be efficient, which can be defined roughly as making goods or services available at an acceptably low price while still returning acceptably high profits to the Operators. The difficulty with making this definition Operational is that what is "acceptable" must always be determined by social values on the one hand and the expectations of businessmen on the other. Neither of these is particularly susceptible of measurement, and both change with time in unpredictable ways. However, it is safe to deduce that, of two systems with the same level of profits the one with the lower prices is the more efficient, and if the levels of prices are the same then the one with the higher profits is the more efficient. There are, of course, other factors affecting social ”efficiency. Chief amongst these are the efficiency of use of labor and capital, and what constitutes "efficient" use will depend upon circumstances. If economic prices are set upon these factors they can be subsumed under prices and profits, but very often such economic prices are not socially desirable. Using margins as a measure of prices, Tables IV. 10. and IV. 11. m. IN. ”I n-i-s - 'I4Il ...: would seem to indicate that larger retail estab11shments are more _5,.. ,M‘. , '-n. efficient than smaller ones,' and that out1ets specializing in a narrow 4’“. —r u “A -.‘1-._-_... 1123,9f909d5 are more efficient than retailers Offering a wide line. These conclusions cannot be rigorously justified from these tables, because there are several variables which are not taken into consideration in the tables. Chief amongst these are the differences in the nature (perishability, etc.) of the goods offered by the various retailer types, the management philOSOphies of the COOperatives and the Cajas de Compensacion, and the economic subsidies received by these stores 174 and, to a lesser extent, by the public market stallholders. EMSIRVA does not charge stallholders economic rents, in the sense that the rents do not cover the costs of administration and depreciation of the market buildings]. Table IV.10. Retailers' Gross Margins by Outlet Type, Cali, February 1969. Outlet Type Prlce%5Pread Spoilgge Loss GrosszMarg1n Self-Service Supermarkets 14.5 1.2 13.3 COOperatives 10.9 1.0 9.9 Cajas de Compensacion 8.6 0.8 7.8 Chain Stores 8.0 0.5 7.5 Traditional Stores Tiendas 13.7 0.8 12.9 Small Graneros 12.2 0.9 11.3 Large Graneros 9.5 0.9 8.6 Wholesaler-Retailers 7.2 0.5 6.7 Public Market Stallholders Fruits & Vegetab1es 24.7 12.5 12.2 Grains & Processed 9.2 0.5 8.7 Beef 20.0 5.5 14.5 Street Vendors 26.4 15.0 11.4 Source: PIMUR Retailer Survey, 1969 Table IV.11. shows that Operating tiendas and fruit and vegetable stalls is tantamount to concealed unemployment (the official 1PIMUR Final Report, pp. 66-67. 175 minimum wage in Cali in 1969 was Col¢$450 per month), while beef stallholders are making grossly excessive profits. The low returns to management of fruit and vegetable stalls can be explained as the natural consequence of the great ease of entry forcing profits down to a sort of acceptable minimum. Tab1e IV.11. Financial Comparison of Out1et Types, Cali, February 1969. Profits Before Taxes as: Sa1es Investmenta Outlet Type Col.$ Col.$ % on % of Investment Return Sa1es (per month) to mangt.b Col. Self-Service 385,000 219,700 2.4 4.2 $ 3,739 Wholesaler-Retailer 231,600 83,600 2.9 8.1 4,677 Large Graneros 102,600 35,600 3.6 10.6 2,782 Small Graneros 33,100 14,600 4.1 9.1 978 Tiendas 6,900 6,300 6.7 7.4 306 Public Market Stalls Grains & Processed 16,900 3,500 6.7 32.5 1,050 Fruits & Vegetables 4,800 558 8.9 76.0 410 Beef 45,300 2,600 10.7 189.0 4,773 Source: PIMUR Final Report, Tables 2.20, 2.21, 2.22. Notes: a. Excluding buildings. b. Taken as before-tax profits less Opportunity cost of 2.5% per month on investment. With tiendas, several factors are at work. Tiendas tend to locate as far as possible from competitors, in order to obtain the maximum locational monOpoly, but they are so numerous that any such monopoly is very limited in areal extent. Any attempts they might 176 make to increase their sales volume can be countered by their competitors on the next street corner, so they are cut off from any possible economies of scale. At the same time, their small area of influence makes it impracticable for them to specialise in a narrow line of goods, so they are also cut off from any economies of Specialisation. They are therefore locked into a state of profitless monOpolistic competition. Tab1e IV.12. presents comparative data on prices for some standard food items as sold by various retailer types. Table IV.12. Relative Prices of Foodstuffs by Retailer Type (Base: Ga1eria Central = 100), Cali, February 1969. Super- Tiendas and Small Graneros Food Item Ga1eria Satellite markets Areas Area Area Centra1 Markets Area I II-IV V VI Beef, lst.qua1. 100 100 103 103 102 101 Beef, 2nd.qual. 100 105 111 112 112 114 Beef, 3rd.qua1. 100 99 114 114 113 116 Beef viscera 100 101 108 106 107 110 Rice, lst.qua1. 100 103 104 104 106 110 Calima beans 100 101 103 98 106 108 Eggs, medium 100 102 103 108 110 116 Pan Sugar 100 99 97 100 106 104 Potatoes,guata 100 97 109 110 111 119 Tomatoes,green 100 100 126 132 137 143 P1antain,green 100 115 164 159 138 165 Source: Adapted from PIMUR Final Report Table 2.18, p.59. 177 It can be seen that the highest prices are generally to be found in the tiendas of the barrios of sociO—economic levels V and VI. This lends support to the hypothesis that poor peOple pay more for their food than do the rich. On the other hand, calculations based on the data of the Market Basket Study gave the following results for comparative prices on an average market basket purchased in the indicated outlets: Ga1eria Central 100 Satellite Markets 103 Supermarkets 106 Traditional Stores Central] 103 Non-central 101 These results contradict the figures given in Table IV.12. However, the two sets of figures are not comparable. Table IV.12. is based on similar qualities of foodstuffs, while the Market Basket data referred to the prices of the foods actua11y purchased. It is reasonable to suppose that poor people buy low-quality foodstuffs while rich people buy high-quality foods. The conclusion would seem to be that while poor peOple might not pay significantly more for their food, and might indeed pay less, they receive a poorer quality of food for their money. Another possible confounding factor is that weights and measures might be rigged to bring the apparent prices of certain foods, such as rice and beef, into line with those Obtaining in the city center. 1Since these stores sell only a limited range of foods their index is based on the assumption that the foods not offered (meat, fruits and vegetables) were purchased in the nearby Ga1eria Central. 178 The Wholesaler-Retailer Interface As we will see in the next section of this chapter, wholesalers in Cali carry very limited ranges of goods. The only exceptions are the wholesaler-retailers. The most important wholesalers, those dealing in grains and non-perishables, typically specialize in two or three products, carrying five to ten others as sidelines, while the typical wholesaler of fruits and vegetables handles only one product. Whole- salers also fall into line with the custom, mentioned at the beginning of this chapter, of being passive sellers. They do not advertise their wares, nor merchandise, nor provide support services. They expect their customers to take the initiative in any contacts. In consequence, when the Operator of a tienda or small granero goes to the city center each morning to make his purchases he finds this activity very demanding of time and effort. There are several hundred supp1iers, a11 se11ing homogeneous, unbranded, and undifferentiated goods, Spread over some ten city blocks for all the world like stall- holders in a huge market. This may make for a "perfectly competitive market" in the economists' sense, but introduces inefficiency into what ought to be largely routinized buying. To illustrate, consider the case of a granero who handles five fruits or vegetables. Since he customarily buys these 25 times per month1 and has to buy each one from a different dealer this implies 125 purchases per month. If he gets quotations from four dealers each time before purchasing, a reasonable figure if he wantsto approximate the best price-quality 1PIMUR Final Report, Table 2.33, p.82 179 combination available, he must take part in 500 negotiations in connection with Just a small sub-set of his stock. All this has to be done in time for his tienda to be Opened at 8 a.m., and in competition with 3,370 other tienda and small granero operators and 2,600 market stallholders. Finally, when all his purchases are complete he has to arrange to have the goods collected and transported to his store. Since the tranSportation facilities available are not really adequate for the load put upon them at this time1 the negotiations involved in this stage of the purchasing process could be rather protracted, with the transport Operator holding the better hand. In practice, it is unlikely that the buying process is as complicated as the foregoing description implies. It is improbable that a daily activity should not be routinized to at least some extent. That this is the case is indicated by Table IV.13.,which shows that two-thirds of store Operators and slightly under half of all stallholders Tab1e IV.13. Percent of Retailers Purchasing From Regular Suppliers, Ca1i, February 1969. Self- Traditional Public Service Stores Market Purchasing from regular suppliers 68% 68% 44% Contacting only one supplier before purchasing 11% 38% 30% Source: PIMUR Retailer Study, 1969 1PIMUR Final Report, p.283. 180 habitually buy from the same suppliers. This is not to be interpreted as meaning that each of these Operators uses only one supplier. In fact, only 11% of self-service store Operators, 38% of tienda and small granero Operators, and 30% of stallholders talk to only one supplier before purchasing. The more usual practice appears to be for each retailer to contact favored sub-groups of wholesalers. This practice should effectively maintain a "perfect" market in the sense of preventing price discrimination by the wholesalers, but it could also limit the extent to which retai1ers approximate their Optimal price- quality targets. In keeping with their "passive seller" philoSOphy, wholesalers offer very little in the way of services to their customers. Only 20% provide delivery service on more than 50% of their sales, and 60% provide no delivery service at all]. Credit is rather restricted, as Shown by Table IV.14.,and no re-grading or re-packaging is done, except insofar as the bu1k-breaking service provided by the wholesaler- retailers can be considered in this light. Tab1e IV.14. Extent and Terms of Credit Received by Retailers, Ca1i, February 1969 Self-Service Traditional Public Market % Days % Days % Days Grains 85 14 36 14 ‘ 65 12 Processed Staples 86 12 3O 14 45 14 Fruits and Vegetables 25 10 9 11 32 3 Meat 28 14 l 2 100 3 Poultry and Eggs 58 12 3 10 29 5 Source: PIMUR, Retailer Survey, 1969 1PIMUR, Wholesale-Retai1 Study 181 The general conclusion to be drawn here is that the existing wholesale structure is very ill-conditioned to supply the tiendas and small graneros. Only the wholesaler-retailers are making any attempt to cater for these Operations, and even then in only a rather limited fashion. The passive selling methods of the suppliers compel the operators of the small traditiona1 stores to Spend an excessive amount of time and energy in purchasing, and must tend to induce in them an attitude that buying is more important than se11ing, thus extending the "active buyer - passive seller" philosophy further down the line of distribution. Also, this attitude will unduly inhibit the small Operator from finding or seeking items which could effectively be added to his product mix. On the other hand, the structure of wholesaling is almost ideally suited to the needs of the market stallholders. Each stallholder handles only a narrow line of foods, so that the need to contact severa1 wholesalers in connection with each product purchased is not a significant burden. Transportation is not a problem for him, because even if he does not have his stall in the Ga1eria Central (within walking distance) he can consolidate his shipping requirements with those of other stallholders from the same satellite market. This arrangement is attractive to the transport operators and economical for the stallholders. It can be argued that the nature of wholesaling in Cali is a result of the requirements of stallholders. Denied the opportunities of economies of scale, the stallholders are forced to seek economies of Specialization, and will demand corresponding services from their suppliers. The cheek-by-jowl competition in the markets and the impossibility of differentiating themselves from their competitors in selling leaves buying as the only activity over which they have some control. This in turn forces wholesalers into very specialized behavior. It is not surprising that the wholesale sector in Cali should be stallholder-oriented. In the compact towns of Latin America, until the pOpulations reach levels of about 100,000 the central market is a very efficient mode of food retailing. Everyone in town lives within walking distance of the market and market stalls are the only type of retai1er small enough to exist in numbers sufficient to provide competitive discipline. Such a system is probably at its highest effectiveness when the pOpulation is around 20,000 to 50,000, with tiendas starting to becone viable when the population is about 100,000, and becoming more effective as it increases beyond this level. Until a few years ago cities with populations greater than 100,000 were limited to two or three in Colombia, so that food distribution grew with a strong stallholder orientation. The major cities of the country have grown so rapidly that the system has not had time to change, and in Cali the efforts of the municipal government to extend the public-market system have tended to Obscure the need for change. The situation in Cali in 1969 was that the city needed non-centrally located food stores, but that such stores were incompatible with the stallholder- oriented wholesale sector. Two alternative solutions are available for remedying this situation. Eitherilmepublic-market system must be drastically extended, providing markets within walking distance of everyone inthe~city, or the wholesale sector must be modified to supply non-centrally located food 183 stores in an efficient manner. .1110. mm: 531.959.91951'“. Wholesalers fall naturally into four groups according to the products they handle. The four groups are: a) Meat wholesalers,who have already been described. b) Wholesalers of grains and unbranded non-perishabTes. There are two sub-groups in this category, the Specialist wholesalers and the wholesaler-retailers. c) Wholesalers of fruits and vegetables, who can be further divided into dealers in potatoes, plantain, tomatoes, and minor fruits and vegetables. d) Distributors of branded processed foods. Meat Wholesalers have already been described. In pork there is only one set of wholesalers, the bancos de carne that Specialise in this product. In beef there are two sets, the bancos de carne and the forranea dealers. We have already noted how the size of the unit transaction, one carcass, poses problems of divisibility that create barriers to growth and to entry in the traditional (bancos de carne) wholesaler group. Presumably it is these barriers that perpetuate the high margin - low service mode of operation observed amongst these dealers. As we noted before, they are wholesalers only as a result of being Open for business at the time that retailers are doing their purchasing inthe Ga1eria Centra1 area. The forranea dealers constitute an interesting phenomenon, illustrating how marketing institutions come into being to provide for unsatisfied needs. At present they are supplying only the most desirable customers, 1arge accounts with steady demand. This would be consistent with limitation of their activities by the Cali municipal authorities, who license these dealers, but the PIMUR Meat Study did not state whether such limitation was practised. Wholesale1§_of Grains and Processed Staples _-..._._ There were 46 specialist wholesalers in Cali at the time of PIMUR's work, of which 36 were interviewed. They deal in grains and processed goods for which brand identification is not of marketing importance, staples such as pan sugar, lard, cooking oils, and chocolate]. The typical specialist wholesaler does the bulk of his business in one or a few commodities, and might handle up to a dozen or so others as sidelines. In volume of sales they range from C0146130,000 to COL/$4,000,000 per month, with a mean of COL/$1,200,000 and median approaching Colq6700,000 per month. Amongst those interviewed the four largest (11%) accounted for 32% of total sales, which is only a very moderate degree of concentration. As far as UweCali market is concerned the concentration is even less, since the seven largest wholesalers make over 60% of their sales in other cities of Valle2. The smaller the business the lower is the prOportion of sales made outside Cali. The average age of the Operators of these businesses was 44 years, and ages ranged from 23 to 77. Virtually all had completed elementary school, but only 15% had finished high school. They had been running their businesses for an average of 14 years. The distribution of "length of time in the business” was virtually uniform between 1 year and 22 years, indicating that this is a rather stable sector. 1Although these goods are often branded, in the sense of identifying the manufacturer or packer, this branding is rarely of importance in the marketing of these goods. 2PIMUR Final Report, p.74 Cross margins are low in this field of food distribution. The PIMUR Wholesale-Retail Study quotes margins realized at about 3.4%], and this agrees well with figures quoted by PIMUR's Grain Studyz. Tab1e 1V.15. details the answers given to two interview questions: a) what margin, on the average, do you need to make to stay in business? and b) do you usually make this margin? Table 1V.15. Margins Needed and Obtained by Wholesalers, Ca1i, February 1969 N0. of ReSpondents Gross Margin % 2 3 4 5 7 8 10 15 Needing this margin 4 6 3 8 1 2 10 Needing and obtaining this margin 3 5 1 3 0 O 2 O Source: PIMUR Wholesale-Retail Study, 1969 Not surprisingly, the greater the margin desired the greater the prOportion of those disappointed. Respondents obviously had only a somewhat vague idea of the margins they needed, as evidenced by the tendency of responses to bunch at the round numbers, 5 and 10. Comparison of these responses with sales volume indicated that there was no correlation between the two, although one would have expected to find an inverse relationship. It appears that these wholesalers do not perceive economies of scale. lPIMUR, Final Report, p.88. 2PIMUR Final Report, p.211. 186 Two-thirds of the Specialist wholesalers interviewed said that their selling prices were based Upon competitors' prices, i.e., current prices in the market. Of those who used other ways of deciding prices (principally, adding a fixed sum to the purchase price) ha1f also took competitors' prices into consideration. Three-quarters of those interviewed do no promotion at all. Of the nine who claimed to do some advertising six used only posters (which may have been merely the signs over their doors). Four employ salesmen. Only one-third of these wholesalers provide delivery service, while almost all give credit. Typically they give credit on about 50% of their sales for a period of 15 or 30 days. PIMUR'S interviewers received the impression that most wholesalers intended to tighten their credit policy, and this is corroborated by their statements concerning the uses to which they put borrowed money. Of the twenty- nine who use borrowed money twenty-eight use it to increase their purchases of merchandise and one to give credit. No re-packaging or re-grading services are provided. The selling practices of this group are consonant with our previous observation that the wholesale sector in general is well- prepared to serve stallholders, if not other retai1er types. The grains dealers in the public markets are specialist retailers, who offer a few lines of foods in great depth. They neither need nor want re-grading, re-packaging nor delivery services. Their "active" buying of undifferentiated goods exerts pressure on margins and creates market, as opposed to administered, prices, and their specialist format does not require that their wholesalers stock broad lines of goods but 187 rather great depth in one or a few lines. Purchasing practices of the Specialist wholesalers resemble those of retailers. While 80% consistently buy from the same group of suppliers they contact an average of 3.75 suppliers before buying]. The nature of the people who supply them depends upon the product. Country assemblers are the suppliers for most grains, rice mills for rice, and producers or brokers for products such as lard, oils, and pangsugar. One of the characteristics of the agricultural sector in Colombia is that nobody has reliable information about acreage planted, harvests, or quantities in stock. Figures compiled by G.Trant and his co-workers at Universidad del Valle Show estimates fOr past harvests from different sources varying by factors of two or more. Data for current harvests and stocks must presumably be no better and probably a great deal worse. Another characteristic of Colombian agricultural marketing is that there are no commodity exchanges, except for the Medellin cattle market. In the developed countries market prices for agricultural goods are determined by a process which requires the timely dissemination of crOp information and the recording of transactions, neither of which are available in Colombia. How, then are market-clearing prices determined? The clue to finding the answer to this question was provided by a corn assembler in Cartago, in the north of Valle, who said that he 1PIMUR, Wholesale-Retail Study, 1969 188 set his prices according to the number of telephone calls he received from wholesalers asking for price quotations. He said that the . quantities involved also entered into his price-setting process, but secondarily. Market-clearing prices, then, are set by each supplier individually, each supplier knowing his own stock position and estimating the state of the demand side of the market from the number of requests he gets for price quotations. The wholesalers, by making multiple contacts, and by closing more or less rapidly, provide the feedback that brings the prices of the individual sellers into line. For this system to work effectively and fOr price adjustments to be smooth it is necessary that there be many wholesalers, each making multiple contacts with suppliers. Such a system has some interesting consequences. The lack of crap and stock information makes speculation very risky, particularly for suppliers and wholesalers. The only peOple who could take speculative positions would have to be in a position to do a good deal of travelling in the producing areas to see for themselves the crop situation. While such people may, and probably do, operate in Colombia they are not country assemblers, who only Operate locally, or urban wholesalers. This may explain why PIMUR's Grain Study was unable to discover who carried out the necessary function of storing crOps between harvests. This situation also explains a puzzlingfeature of the Specialist wholesale sector in Cali, the number of operators. Simple economic theory would predict that there would be either very few such wholesalers, say, five or six, or very many, upwards of a hundred. In fact there are forty-Six. The answer to this question appears to be that there are no I! II. 189 economies of scale to be obtained in purchasing. While it is as easy to order ten or a hundred truck-loads as one truck-load, the absence of stock infOrmation makes it too risky to order more than is needed immediately. That is, the Economic Order Quantity is rather small, and each subsequent purchase has to be fully negotiated. Far from there being volume discounts available, it is more likely that asking a supplier fOr a quotation on a large shipment will lead him to believe that demand is increasing and thus to quote a higher price]. Therefore economies of scale in specialist wholesaling taper off rather rapidly, which explains why such economies are not perceived by the wholesalers, as we noted earlier. The Operating format of these wholesalers is such that their primary assets are not financial resources nor estab1ished markets (in the sense of having a customer franchise) but rather a thorough understanding of the markets in which they are buying. This understanding can only be acquired by experience and instruction from practitioners. Outsiders cannot prepare themselves for entering this sector by studying past performance of the markets, since transactions are not made public. The result is that while entry is fairly easy, requiring only a few thousand pesos and a few square meters of storage space accessible to retailers, successful entry is very difficult. The number of specialist wholesalers operating must therefore lie between a lower limit set by the limited economies of purchasing 1This is contrary to U.S. business practices, where buyers placing large orders usually receive quantity discounts. 190 in any one line of food and the risks involved in taking on lines of which the Operator has no experience, and an upper limit set by such economies of scale as do exist (e.g., truckload purchases) and the very narrow margins available in this sector. Economically, this sector appears to be very efficient, though the system of which it is a part (stallholder-based food distribution) is not economically efficient in a city the Size of Ca1i. It can be seen from our analysis of the functions performed by the specialist wholesalers that their primary contribution to food marketing is the establishing of market-clearing prices. Since commodity exchanges cannot instantly come into being the Specialist wholesalers will be needed in the foreseeable future, no matter what type of food distribution system is developed for Cali. However it should be an easy, even natura1, transition for them to move away from being stock- holding wholesalers towards acting as brokers, where they would perform the same price-establishing function but at lower margins, of the order of 1% as opposed to the 3.4% currently obtaining. The enabling factor for this change to take place would be the emergence of large retailers or the emergence of full-line wholesalers catering to food stores rather than to market stallholders. It is likely that the lines of goods handled by this sector will change in the future. They will lose many processed foods, as branding becomes more common and brings with it administered prices. 0n the other hand, they will find more sc0pe for their talents in handling the major fruits and vegetables (potatoes, plantain, tomatoes, etc.) as and when these become subject to grading and can be marketed by Specification rather than byinspection. Already in Cali potatoes are being handled in this manner. 191 Wholesaler-Retailers The retailing aspects of these operators have already been discussed in this chapter (pp.158 to 162). Their wholesaling functions are comparable to those of the wholesale-retai1 beef dealers, the bancos de carne. Like the bancos de carne the wholesaler—retai1ers sell at retail prices and provide no other wholesale services. They are not so much concerned with catering to a target market as exploiting the Opportunity of making sales by being Open for business at the time the retailers are in the Ga1eria Central area doing their purchasing. That is, they are passive marketers, exploiting an existing market rather than creating new markets. As Suppliers to the non-centrally located traditional stores they play a significant but not important part, supplying 19% of the grains and processed staples that flow through these stores]. Although we are not able to verify this, it is likely that they are important suppliers to certain sub-groups of these stores. For example, tienda Operators in the outlying barrios use taxis to transport their purchases to their stores, and might find that being able to pick up all, or almost all, their goods at one location brings significant savings in transporta- tion costs. PIMUR was interested in the wholesaler-retai1ers because they appeared to be the forerunners of a store-based food distribution system. It seems that this expectation must be discarded. They are passive marketers, and their attitudes, average age, and education, which were 1Estimated from PIMUR Wholesale-Retail Study data. 192 discussed earlier, hold out little hOpe that they will constitute a new wave in food distribution. Fruits and Vegetables Wholesalers It is convenient for our purposes to consider the wholesalers of fresh fruits and vegetables as divided into four groups. The first three groups are the wholesalers of the three most important products, potatoes, plantain, and tomatoes, while the fourth group is made up of the dealers in the minor fruits and vegetables, such as onions, carrots, mangos, pineapple, etc.. Wholesalers of the first three groups as a rule have permanent places of business, while those in the fourth group typically rent Space on a daily basis in the bggggas, 1arge halls near the Ga1eria Centra1 which specialise in this activity. Table 1V.16. shows some general characteristics of these wholesalers. The most obvious aspect is that they are small operations, doing about the same volume Tab1e 1V.16. General Characteristics of Wholesalers of Fruits and Vegetables. Ca1i, February 1969. Potato Tomato P1antain Other No. of Operators 39 72 120 220 Average monthly sales, Col.$ 66,700 23,500 32,800 20-30,000 Sales and storage area, sq.m. 32 4 ll 6 Average time in business, years 4.0 5.3 7.6 7.6 NO. of types of products handled 1.1 1.2 1.1 1.6 Average age, years 37 40 43 40 Average years of schooling 3.8 5.8 3.5 3.3 Source: PIMUR, Wholesaler Study, 1969 193 of business as a small granero. Another aspect is their extreme Specialisation by product. As we shall argue below, this is a consequence of the differences of production conditions and rural marketing practices between the various products. Potatoes are grown in two regions of Colombia, in the hills near Bogota and in the extreme south of the country, on the border with Ecuador. Both regions are several hundred miles from Ca1i. In consequence potato wholesalers Operate very much like the Specialist grains wholesalers, buying by telephone from country assemblers in the producing regions. Since there are obvious economies in buying by the truckload this accounts for the small number of dealers in potatoes and their relatively large-size operations. Plantain has some very unusual characteristics for a fruit. It is a staple which is grown as a by-product of coffee cultivation, the plantain's broad leaves being used to shade coffee bushes. It is almost a non-seasonal crop, it can be marketed green or ripe, and it has a long shelf-life. The main growing area is Quindio, about 100 miles north of Cali. This steady source of supply fairly close to Cali reduces the need for high skill in buying and makes for only very moderate economies in truckload buying, which explains the large number and small size of the p1antain wholesalers. Another factor contributing to this situation is the large number of small farmers in the hills around Cali who have a few plantain trees and who send a few stems of fruit into Cali to market. This accounts for only a small proportion of total plantain sales but a large prOportion of trans- actions. 194 Tomatoes consumed in Cali are grown in the flat part of Valle, within a radius of about 60 miles of the city. Producers are small in scale and fairly widely scattered. There are well-deve10ped markets for this product Operating in the Valle cities of Palmira and Tulua. These keep the marketing of tomatoes fairly well-ordered, subject to seasonal variations in production, and the short distance from Ca1i limits economies of scale amongst urban wholesalers in this city. These considerations probably account for the fact that there were 72 urban wholesalers of tomatoes in Cali in 1969, as shown in Table 1V.16., while there were only 33 rural assemblers]. Marketing channels for the minor fruits and vegetables vary enormously. The most distinctive characteristic of most channels is the widely-scattered, sma11-scale production. In most cases the producing unit is a small farmer in the hills with a few fruit trees in a corner of his land, most of his land being used for subsistence crOps. At the other end of the scale are such Operations as orange producers who grow on a commercial scale with fully integrated marketing, a cooperative marketing set-up that handles the output of pineapple growers in Risaralda, and, perhaps the most SOphisticated, a channel which deals in lettuce flown to Cali from Bogota to be sold to upper- income consumers. In general, marketing of the minor fruits and vegetables has to deal with the smallness and inaccessibility of the producing units, seasonality and other variations in the supply, and thinness of the market for individual products, which must make for 1PIMUR Final Report, p.184, Table 3.34. 195 protracted and fully-negotiated transactions at each state of the marketing process. The factor that dominates the structure of wholesaling of fruits and vegetables is the need to buy by inspection from small scat- tered producers. Stallholder-orientation does not shape the wholesale structure, as seems to be the case with grains and processed staples. However, changes are evident in this sector. Potatoes are already being marketed on the basis of specification rather than inspection, and tomatoes for processing are being grown under purchase contracts, which will Spread the acceptance of specifications and grading amongst the growers of this crop. Changes in coffee cultivation will probably lead to plantain being grown as a primary crop instead of as a by-product]. And INCORA, the National Colombian Agrarian Reform Institute, an agency of the Ministry of Agriculture, is encouraging the formation of producer cooperatives, which will Spread the use of grading to other products. It is likely that it will eventually become possible to market some 70% to 80% by value of the fruits and vegetables consumed in Cali on the basis of grades and specifications rather than on inspection. The effects of this will be widespread, but as far as urban food distribution is concerned the principal effect will be to make it practical for food stores to handle a wider range of these products than they can do at present. However, during the foreseeable future there will be many varieties of fruits and vegetables which will have to be marketed in the same way as at 1PIMUR, Final Report p.179. 196 present, through small specia1ist dealers at the wholesale and retail levels. Wholesale Distribution of Branded Foods Although there is in Colombia nothing like the variety of branded processed foods that there is in the U.S.A. these foods form an important part of the sales of supermarkets and a not-insignificant part of the sales of other outlets. Pastas are widely used to thicken soups, and processed oats have been consumed by all classes for many years. Experience in the more deve10ped countries indicates that this sector will grow in importance at the expense of grains and un- branded processed foods. Distribution of these foods is done either by the processor (directly or through wholly-owned subsidiaries) or by agencies. In either case the selling is done by the processor, since the agents are passive sellers to the point of being apathetic. We were told by the marketing manager of a major food processor that there are only two agencies in all Colombia who employ salesmen. Not only are they passive sellers, but Since they are operating in an environment of administered prices they are also, perforce, passive buyers. Surprisingly, the policy of attacking the remaining profit-determining variable, turnover, does not seem to have been deve10ped amongst these operators. The only distributive services provided by the agencies are storage close to the point of sale and, in some cases, credit collection. 197 As can be imagined, this situation has some dysfunctional aspects: a) Small food processors, otherwise viable, could fail to survive through being unable to obtain effective distribution. b) Firms which are too small to justify doing their own distribution are forced into doing it, with unnecessary increases in costs. c) Retailers are bothered by having to deal with large numbers of processors' salesmen, each handling one or a few products. The manager of a Cali supermarket told us that he Spoke to 600 salesmen each month (equivalent to one every 20 minutes of the working day). This seems excessive. Although the situation at present is bad, prOSpects for improvement in this sector are good. Branded foods and other house- hold necessities are a "modern" growth sector of the Colombian economy, and it should only be a short time before some bright young entrepreneurs step in to provide the distribution services so urgently needed. Entrance requirements are not great. Ten thousand pesos (say, U.S. $600) in capital, a guarantor to provide respectability, and the ability to talk persuasively to both processors and retailers should be enough to start such an enterprise. The figure of ten thousand pesos is only a ball-park estimate, but it would appear that the capital requirements involved would be minimal. Summar The fOod distribution system in Cali is based upon the old- fashioned concept of a central market in which the stalls are operated by small-sca1e retailers specialising in individual lines of foodstuffs. 198 Such a system is an efficient distributor of food in towns and small cities, but is not well adapted to the operation of non-specialised food stores. In Calithe public market system supplies only about 20% of the food consumed, the rest being supplied by food stores of various types. The result is that in Cali food distribution is inefficient, in the sense of costing too much, a1though individual Operators are efficient, in that their margins and profits are low. (Meat distributors constitute an important exception in this respect.) To reduce food prices to consumers it is important that the system change from being stallholder-oriented to being store-oriented. Unfortunately, there is no evidence of such a change being presently in process in Cali. f _ CHAPTER V Conclusions and Recommendations In the preceding chapters we have examined the components of the food distribution system of the city of Cali. In this chapter we Shall evaluate the performance of the existing system, discuss the capacity for Spontaneous evolution of its component institutions, and make recommendations concerning a food distribution system which will avoid at least some of the major shortcomings of the existing system. Cali's basic food distribution system, the public market system, is a holdover from the days when Ca1i was a small city. For a small . - ____,..- ..p .. - ____ city a pub1ic-market system is desirable. Only in this way can a small *- —-~——_.__,___ community support enough food retai1ers for their numbers to ensure that market discipline will be enforced. But we show, in the appendix to Chapter IV, that market stalls are less efficient retailers than stores. Therefore a market-based system is less satisfactory than a store-based system for a city as large as Cali. A further advantage of the store- based system is the fact that system coordination (such as, for example, in voluntary chains) is much easier to obtain than in a stall-based system. ‘ Another way of looking at the Situation is to note that market _Stallholders are, because of the physical Size of their operations, _” unable to obtain economies of scale, but that the large number of -L-.._.qn———— u 199 200 Operators that a central market can support enables them to Obtain economies of product-Specialization. Indeed, the cheek-by-jowl conditions in these markets impose competitive pressures that force Operators to take advantage of any such economies. This specialization emphasizes active buying, since small differences in price become critical, and this in turn imposes the same type of behavior upon the suppliers of the stallholders. They also become active buyers and specialize by product. The stallholder and stallholder-oriented sectors of the food distribution system, then, are adapted to each other so that the stall- holders can Operate at close to their potential effectiveness. Unfort- unately, this adaptation by the wholesale sector prevents the store- based food retailers from reaching their potential. Each storekeeper has to assemble his assortment of goods from the elementary offerings of a large number of wholesalers. The costs of thus resolving discrepancies of assortment are duplicated for each store, instead of being incurred principally by one or a few full-line wholesalers and then shared over large numbers of retailers. Socially, these excess costs should be charged to the account of the public-market system. Also from the point of view of the community, the public market system imposes upon the society the costs of resolving the discrepancies of assortment of goods sold through the markets. There are the direct costs to the consumer of getting herself to and from the market, and there are the indirect costs of public funds needed to construct and administer the market buildings (insofar as these costs are not covered by the rents charged to stallholders). We have already noted the active buyer-passive seller behavior 201 shown by almost all food distributors in Cali. This is indicative of product- or technique-orientation, as opposed to marketing- or consumer-orientation. The failure to develop a marketing orientation could well have its roots in the pub1ic-market origin of food retailing in the city. One of the peculiarities of such a system is that it is the market as a whole that attracts customers, not the individual stallholders, and the conditions within the markets inhibit attempts by stallholders to obtain differentiation in their selling methods. This failure to develop a marketing orientation has some undesirable effects upon the operators of food stores: a) it leads them to try to operate like stallholders, concentrating their efforts upon buying. Since they do not specialize by product they have little to gain by such efforts. Therefore there is waste of managerial resources in this respect. b) they are not led to realize that a store must Offer its customers a package, a combination of variety, selection, price, and service. The stallholder, by contrast, need offer only price, the rest of the package being offered by the market as a whole, through the multiplicity of sellers. c) they are induced to regard their market as a given, an exogenous factor. This constitutes a psychological barrier to growth. This attitude will do little harm to a stallholder, whose potential sales are sharply limited by his Operating format in any case, but is a severe handicap to a store Operator. d) their-active-buying behavior has inhibited the development of the broad-line food wholesalers which are needed to support a store- based food retailing sector. 202 In Cali, then, fOod distribution is dominated by the public- market system, in the sense that the operating methods and philosophies of the other components of the food distribution system have been determined, or at least strongly influenced, by the requirements of market stallholders. Yet the public markets do only one-fifth of the city's retail food sales, so that any benefits which might accrue from this confOrmance with the needs of stallholders are being made available to, or being enjoyed by, only a small proportion of the population. Worse, the adaptation to stallholder needs has inhibited the development of efficient store-based food retailing. TherefOre the greater part of the population has not only received little or no benefit from the pub1ic-market system, but has actually been indirectly penalized by it. A particularly unfOrtunate aspect of this is that it is the poorer consumers who have been the more severely penalized. Exaluation of Existing Food Distribution Institutions As we have seen, the existing food distribution system provides consumers with either locational convenience at high prices or with low prices in inaccessible locations. Our assignment is to devise a system which will provide both low prices and locational convenience because, as we have seen in Chapter III, locational convenience is the primary determinant of where the Cali housewife does her Shopping. Also, these low prices and locatiOnal convenience must be offered in conjunction with an adequate selection of foodstuffs. We shall describe such an 'adequate' se1ection later in this chapter. The most desirable way of devising such a system is to design it round any existing institutions that can be suitably modified. This 203 has the advantages of minimising disruption and making the least calls upon human and capital resources. We reviewed the existing food distribution institutions in Chapter IV, where for various reasons we removed from further consideration all but two types of retailers. These two types were the public markets and the non-centrally located traditiona1 stores, the tiendas and small graneros. At first sight the obvious Solution is to extend the public market system. The existing satellite markets exhibit generally good performance in terms of prices, variety, and quality, and the whole- salers are well attuned to the needs of stallholders. However, even a superficial examination of this option shows it to be impractical. 1) The existing five (in l969) satellite markets supply 8% of food consumption in Cali, or an average of 1.6% for each market. On this basis, sixty markets would be required to supply all the city's food. More generously, assuming that each sate11ite's drawing power is increased by 50% by the disappearance of the Ga1eria Central, and that self-service stores (supermarkets, cooperatives, and Cajas) and specialty stores maintain their 25% share of the total market, thirty satellite markets would be needed. Since *t—n ’ 1 Cali can affOrd to build markets only at the rate of one every two 1 I! or three years it would be a very long time before this strategy would Show any results. 2) Current meat distribution practices in Cali are such that the consumer pays excessively high prices for meat, which accounts for about thirty percent of food expenditures. These practices have their origins in the limitations of meat stalls. TherefOre extending the public market system would merely perpetuate the 3) 204 existing high meat prices which so heavily tax the resources of the poor. As was noted above, the package of services offered by a public market is the result of there being a large number Of stalls Operating. It follows that there is a minimum number of stalls a market must have if it is to be an effective food retailing unit. This is particularly true of satellite markets, which do not enjoy a central market‘s advantage of being located at a focal point of transportation. We have no model of public markets which would enable us to predict the number of stalls required for effectiveness, but examination of Table V.l. suggests that Siloe market has too few operators in each product line to be effective. Visual observation of the satellites in operation suggested another criterion of effectiveness, the amount of "bustle". Bustle can best be described as a compound Of crowdedness, busy-ness, and air of purposeful activity, and intuitively can be taken as closely related to effectiveness. Table V.1. Distribution of Vendors in Satellite Markets, by Product Type. Cali, February 1969. Product Type A1ameda C.Colon Floresta Porvenir Siloe Fruits and Vegetables 243 37 65 66 26 Meats and Fish 44 42 38 31 15 Grains and Staples 27 33 29 18 10 Dairy 7 10 - - 4 Poultry and Eggs 4 - 2 6 - Bread and Pasta 2 3 4 - - Others 47 46 23 23 8 Total Stall Operators 374 171 161 144 63 Mobile Vendors 90 219 57 35 31 Total Vendors 464 390 218 179 94 Source: PIMUR Final Report, Table 2.12 4) 205 On this basis Siloe is ineffective, Alameda and Cristobal Colon are highly effective, and Floresta and Porvenir are just about at the level of effectiveness. Translating this into number of vendors, Table V.1. suggests that if a market is to be viable it should have at least 150 vendors in operation. Average monthly sales of market stalls in 1969 were Col./$7,300, so our minimum-sized market would have to have monthly sales of Col./ $1,100,000. This is equivalent to supplying all the food require- ments of a 40-block area in barrios of S.E.L. V or a 67-block area in barrios of S.E.L. VI. A service area of forty blocks means that 55% of the clientele must walk more than three lineal blocks to get to the market, and 20% more than four blocks]. A service area of this magnitude could, at a pinch, be considered as being consistent with locational convenience for the consumer. Unfortunately, this does not help consumers living in barrios of S.E.L. VI. Also a network of markets of minimum size will not be stable. Some markets will benefit more than others from such factors as transport facilities and will grow. This will increase their effectiveness, drawing custom away from the less-favored markets and forcing these below the minimal effectiveness level. The consequence will be that large areas will be left with deficient service. Market stalls are not efficient distributors of foodstuffs whose purchase is susceptible of routinization. The converse of this is that market stalls can Obtain no advantages from hand1ing products 1See Figure V.l. (page 212) 206 which have been graded, standardized, or pre-packaged. In fact, the additional costs of these services will be a handicap to the stall operator, in that they will make him less competitive, less able to obtain any returns from active buying. So extension Of the public market system will certainly not encourage, and will probably inhibit, the development of the food processing industry and the acceptance of grading and standarization in Colombia. These considerations combine to eliminate the extension of the public market system as a serious contender for the role of the basis of an improved food distribution system fOr Ca1i. While the existing public markets can be expected to play a useful part in Cali in the future, their importance will more and more be confined to dealing in fresh fruits and vegetables. What about the existing tiendas and small graneros? Could they, perhaps, modify their Operating methods so as to become the low- price neighborhood stores that are needed? Here again the answer is no. The reason is, quite simply, that they are too small. As can be seen from Table V.2. (page 213 ) a full-line store will have to have sales of upwards of Col./$250,000 per month if it is to operate with low margins. This is an order of magnitude greater than the sales of the average small granero(Col./$33,100) or tienda (Col./ $6,900). Worse,the existing operators cannot grow to the required volume of sales by a program of reducing prices to generate more volume to enable further price reductions, etc. The reason for this is that these stores are perceived as charging high prices, offering low quality, and having no desirable attributes other than locational 207 convenience. Therefore they face demand curves which are price- inelastic around their present Operating points, at least in the short run. Under such conditions price reductions bring reduced money sales volumes, so a program of growth relying on positive feedback between price reductions and sales volume increases cannot work. Large neighborhood stores will have to start large. There is no modification of the existing neighborhood stores that could bring about the needed quantum jump in scale of operation. There are various other reasons why 1arge neighborhood stores will have to start large. Perhaps the most important is that of being visible. PIMUR's Consumer Survey indicated that word-of-mouth diffusion of Shopping information between consumers in Cali was virtually non-existent. So if a consumer is to become aware that she has alternative shopping outlets the alternatives have to be visible, to be obvious. Store Size and appearance are the most effective ways of conveying the information that something new is available. As was mentioned in Chapter IV, in connection with supermarkets, visibility is a potent fOrm of reminder advertising fOr a neighborhood store. It is also an important factor in informative and persuasive advertising. It fellows that if the operator of a neighborhood store is to deveIOp fully his potential market he cannot affOrd to rely on a "better mousetrap" philoSOphy of marketing. ~He will have to be an active merchandiser. Not only must he be an efficient food retailer, he must make sure that everybody in the neighborhood knows that he is an efficient food retailer. This factor, as we shall see later, 208 is of some significance in deciding upon how the new food distrib- ution system should be organised. For the present it is enough to note that this requirement is more consistent with a large store format than with that of a typical tienda or small granero. Again, a small non-specialist store cannot cope with the assortment of goods required to satisfy the demands of all, or nearly all, the consumers in the neighborhood. Meats, produce, and non-perishable stap1es must all be offered, and offered in some degree of depth. They must also be handled with a high degree of skill if they are to be sold at low prices. To devote scarce human resources with these skills to small stores would be wasteful, and in any case the presentation of such an array of goods would require what is, in Cali, a large store. On either count the store would have to start 1arge. An Improved Food Distribution System We have shown that none of the existing food retailers can be deve10ped into the basis of an improved food distribution system for Ca1i. A new form of institution is needed. The basic requirement of this new institution is that it should be viable when located in the residential areas. That this property is essential was Shown in Chapter III. Location in the residential areas means that this store will not be one of a group of stores forming a shopping center (except in a few cases), so it cannot specialise by product. It will have to be a full-1ine food store. While it need not carry all lines of foodstuffs at the depth of Specialist stores it should nevertheless be prepared to supply substantially all the food requirements 209 of its customers. The requirement that the new institution be a full-line distributor has some interesting consequences. The Market Basket Study showed that eighty products accounted for 97.6% of food expenditures. 3 Each product will have to be made available in a number of grades, varieties, and, in the case of most processed foods, brands. So our new retailer will have to stock about eight hundred different items. With this range of inventory efficient stock control will become important in the retailer's profit picture. Operating efficiency will require a high stock turnover rate on these eight hundred items, which implies that there will be frequent, routinized purchases for inventory replenishment. This is inconsistent with the nature Of the existing specia1ist wholesale sector. Therefore a new type of wholesaler will have to be created to support our new retailer. This new wholesaler will be a broad-line wholesaler, able to supply all, or almost all, of the goods required by the new retailers, and able to supply considerable managerial services as well. These new wholesalers will not replace the existing Specialist wholesalers. As we saw in Chapter IV, the principle marketing function of the specialist wholesalers is that of establishing market-clearing prices. This function will still have to be carried out, and the specialist wholesalers are the only people who can do it. It is import- ant that this function not be entrusted to a government agency. Such an agency would concentrate great power into a siegle decision-making unit, so that any mistakes it might make would have far-reaching consequences, and the unreliability of agricultural statistics in Colombia would make mistakes inevitable. 210 At first Sight it appears as though we have merely added another middle-man to an already inefficient system, so that our new system, far from being an improvement, would have higher costs than the existing one. This is not the case. The costs of the new wholesalers will be more than offset by the economies its appearance upon the scene will make available to the other components of the distribution system. Retailers' unit costs will be lowered through the achievement of economies of scale, and the existing wholesalers will be relieved of most of their cost-incurring activities by the new institutions. In effect, the specialist wholesalers will become brokers, and their margins will be reduced correspondingly, becoming of the order of 1% instead of the 35% presently obtaining. In the case of processed foods, distribution costs will probably be directly reduced as the new wholesalers replace the existing, practically non-functional, agencies. Locational Analysis of Proposed Retailers A key factor of the proposed system is that the retailers Should be able to Obtain economies of scale. This implies high sales volumes. It is vital to show that retailers located in barrios of S.E.L. V and VI can realistically expect to generate sales volumes high enough to permit these economies. The reason for locating retailers in the barrios is that, as shown in Chapter III, locational convenience is the determining factor in the consumer's choice of retailer for the majority of the population of Cali. We should note here that by locational convenience we mean accessibility, in terms of both time and money, particularly money. We have noted previously the extra monetary burden carried by those who do 211 not, or cannot buy food locally. The natural criterion of locational convenience is being located within reasonable walking distance of the consumer's home. On this basis the retailer's area of influence, his patential market, would include all homes within walking distance of three or four lineal blocks. For the purpose of determining the viability of the proposed retailers we will take three lineal blocks (about 300 meters) as the distance defining locational convenience. Figure v.1. depicts the geometry of the Situation. It will be seen that the retailer can expect to serve an 18-block area in which no consumer has to walk farther than three lineal blocks to reach the store. The store will not handle all the food consumed in the area. PIMUR estimated that it could reasonably take its realizable share of the market as: 90% of meat sales 90% of grains and non-perishables staples sales 75% of fruits and vegetables sales 54% of dairy products sales which, when weighted by the relative importances of the four product groups, amount to 84% of food expenditures in the service area. The estimates above ref1ect consumer behavior as described in Chapter 111. As was noted there, the local store has great competitive advantages in meat se11ing. Our new retailer, by offering low prices, should be able to get the great bulk of the more-or-less "discretionary" purchases on non-perishabTes. The public markets Should remain strongly competitive in the retai1ing of the minor fruits and vegetables, 212 11.1 .1. V17 I' .11 L11 ll l l T T 1 I 1" 1+! Tr r LII L ' LII “E ii iii Figure V.l. Nobody living within the cross-hatched area has to walk more than three lineal blocks to reach point A, thus defining the service area of a store at A. As can be seen by counting squares, the service area (inside the diamond-shaped figure) comprises 18 city blocks. (Each small square represents one city block.) 213 particularly the more perishable types. And door-to-door sales of milk will not be affected by the advent of our new retailers. Table V.2. presents the results of calculations based on the above estimates and the results obtained from the Market Basket Study. It can be seen that even in the poorest barrios, those of S.E.L. VI, the potential market is large enough to make possible economies of scale. Tab1e V.2. Potential Sales of Retailers Serving l8-Block Areas, by Socio-Economic Level of Location. Ca1i, February 1969 Socio—Economic Level I—III IV V VI Monthly food expenditures 1,730 1,220 789 636 (Col.$ per family) (a) Families per block (b) 20 25 35 30 Food expenditures for 18- 623,000 500,000 500,000 340,000 block unit (001.5 per month) Potential sales of prOposed 523,000 420,000 420,000 285,000 retailers (001.5 per month)(c) (a) from PIMUR Market Basket Study (b) based on house counts. See PIMUR Final Report, Table 2.40 (c) calculated as 84% of total food expenditures. These figures refer only to sales of foodstuffs. Sales of non-foods should add at least ten percent to the potential sales estimated. The results of Table V.2. are average figures which will be locally affected by topographical factors (in some localities the blocks may be laid out in such a way that there are not 18 blocks within the three block walking distance), by industrial areas replacing dwelling houses, by the availability of bus routes, and by proximity to satellite markets. There is an interesting consequence of defining the service area in terms of walking distance. It can be shown that service area and walking distance are related by the equation 214 A = 2n2 where A - service area, in blocks and n defining walking distance, in lineal blocks. Manipulation Of this equation, or inspection of Figure V.l, Shows the great importance of the fringe of the retailer's service area. If the retailer, by increasing the attractiveness of the package he offers, can increase his service area to one defined by a walking distance of four lineal blocks his service area, and hence his sales volume,increases by 78%, from 18 blocks to 32. On the other hand, if he loses the customers living more than two lineal blocks distant his service area falls by 56%, from 18 blocks to 8. To take a less extreme case, suppose that a retailer's service standards fall to the extent that he loses all customers living more than 2% blocks distant. In a barrio of S.E.L. V his sales would drop from Col./$420,000 per month to Col./$292,000, in line with the drOp in service area from 18 blocks to 12;. Consideration of Table v.3. (on page 218) Shows that his profits would be completely wiped out, leaving a loss of Col./$l,000 per month. This points up the need for the proposed retailers to be consumer-oriented. The operators of these stores must be able to sense consumer dissatisfaction before it reaches dangerous levels, they must be able to discover the causes of such dissatisfaction (i.e. they must have a good deal of empathy with their customers), and they must be prepared to remedy them. The analysis also shows how judicious proportioning of equity and loan capital in these enterprises can force the operators to adopt a competitive attitude. If, in our example, the owner/operator had a 40% equity in the business, the rest being borrowed at 30% interest, he would have been able to keep up with his 215 interest payments but would have received no salary, and no return on his investment. Similarly, the owner/operator who can extend his area of influence by as little as half a block of walking distance will reap a rich harvest. The importance of the service area fringe is such as to provide a strong force encouraging retailers to pass any cost benefits accruing from increases in volume along to their customers in the form of reduced prices. A potential danger in this situation could arise at later stages in the development of the system. If rival chains start using advertising as a competitive tool for obtaining larger service areas they could end up in a costly stand-off, with large advertising expenditures being required just to maintain the status quo. Cali is a compact city, its layout characterized by regularly- spaced streets running orthogonally. It Should therefore be suitable for a network of food stores arranged on a grid with stores located six or seven blocks apart. But there will be areas where the regular grid of stores will not be applicable. The most important of these distortions of the grid will be caused by the presence of main roads. These thoroughfares have several effects upon the flow of shoppers. They constitute boundaries between the areas they separate, they facilitate the flow of shOppers along their length, and stores of all types tend to concentrate on them. The potency of these effects is such that the retailer grid Should be laid out using the main roads as the basis. That is to say, in planning the grid the stores along the main roads should be allocated their sites first, and then sites chosen so as to fill in the gaps left over. "Planning" is perhaps 216 too strong a word in this context, since it is not proposed that the municipal government be reSponsible for allocating sites for retailers. Rather, it is meant that prospective retailers should bear these considerations in mind when buying sites for their stores, and whole- salers when accepting affiliations to their chains. These considerations are economic in nature and should prove persuasive to profit-oriented entrepreneurs as long as they are aware of them. There is another topic of some importance in this area. There will be certain Spots, usually determined by the intersections of main roads at which the flows of shOpperS and traffic will be great enough to support shopping centers, concentrations of stores. PIMUR1 has described the assortment of store types which would form suitable nuclei for such shOpping centers. The location of these shopping centers is a matter in which the municipal government should be represented. The reason is that such centers strongly influence traffic flows (as well as being dependent upon them), and can give some shape and coherence to the districts in which they are located. These qualities are absent in most of Cali, particularly the east and south-east in which the bulk of the barrios of S.E.L. V and VI are located. PrOperly situated, shopping centers could be useful tools of urban development, while if poorly situated they could severely hamper desirable development. Therefore the location of shOpping centers must be done in consultation with the municipal government in order to facilitate such things as re-routing of buses and to ensure compatib- ility with urban development planning. IPIMUR Final Report, p.116 217 Economic Analysis of Proposed System Components Tab1e v.3. presents pro-forma income statements for the prOposed retailers for several levels of monthly sales, and Table v.4. presents the corresponding projected investment requirements. The main point about the figures of Table v.3. is that at a sales level of Col./ $300,000 per month the retailer is already enjoying substantial economies of scale. This means that our proposed retailers will be viable even in barrios of S.E.L. VI. Although Table V.2. indicates that retailers in such barrios will sell only Col./$285,000 of food each month, sales of non-foods should push total sales to well above Col./$300,000 monthly. The statements of Table v.3. are based upon profits equivalent to a rate of return on capital of 30%, or 2&% per month, the standard rate on borrowed funds in Colombia in 1969. The salary budgeted for the manager is very handsome by Cali standards, being equivalent to an income of Col./$7,000 per month when fringe benefits are included. This is a good deal more than the manager of such a store would normally receive, but it is not intended to represent the manager's salary so much as the enterprise's return to management. The profit as budgeted is actually an Opportunity cost, a minimal "normal" profit, and any "pure" profits have been subsumed under the manager's salary. This makes sense, because during the early stages of the proposed system these stores will not be operated by paid managers but by the owners, the entrepreneurs. It will only be later, when the system is firmly established and administrative rather than entrepreneurial talent is required, that paid managers will be used. The other estimated salaries and wages are also high by 1969 218 Table V.3. Projected Income Statements of Proposed Retailers, at Varying Levels of Monthly Sales. Cali, 1969 a) Sales Col./$300,000 $350,000 $400,000 $450,000 Purchases 270,125 318,960 362,585 412,285 Gross Margin 29,875 31,040 37,415 37,715 Expenses Manager/Owner 5,000 5,000 5,000 5,000 Meat Manager 2,500 2,500 2,500 2,500 Perishables Supervisor 1,250 1,250 1,250 1,250 Cashier 0 $1,000 1,000 1,000 2,000 2,000 Casual Labor @ $600 3,000 3,600 4,200 4,200 Fringe Benefits 5,100 5,340 5,980 5,980 Total payroll Col/$17,850 $18,690 $20,930 $20,930 Utilities 1,500 1,500 2,000 2,000 Insurance 250 250 250 250 Depreciation 1,300 1,300 1,835 1,835 Transportation 300 200 200 200 Packaging 575 650 725 800 Supplies and Maintenance 300 350 400 450 Promotion Expenses 800 800 1,000 1,000 Miscellaneous 500 500 600 600 Total Expenses C01./$23,275 $24,240 $27,940 $28.065 Profit BefOre Taxes 6,600 6,800 9,475 9,650 Profit After Taxes 6,350 6,500 9,100 9,250 Overall Gross Margin 10.0% 8.9% 9.35% 8.4% Source: PIMUR Final Report, Table 2.41 (p. 100), with modifications. Note: a) All figures in Colombian pesos. Table v.4. Projected Investment Requirements of Proposed Retailers, at Varying Levels of Monthly Sa1es. Cali, 1969 a) Sales Leve1 Col./$300,000 $350,000 $400,000 $450,000 Fixed Equipment 88,600 88,600 122,600 122,600 Working Capital 34,900 40,800 1 46,600 52,300 Lands and Building 130,000 130,000 195,000 195,000 Total Investment Col./$253,500 $259.400 $364,200 $369,900 Source: Adapted from PIMUR Final Report, Table 2.42 (p.100) Note: a) All figures in Colombian pesos. 219 Cali standards. This should help to attract capable people into food distribution and raise its operating standards. The investment estimates are also affected by the same up-grading objective, since rather more equipment has been budgeted than is found in stores outside the self- service group. This has an effect upon the income statement, since the increased investment raises expenses by raising depreciation charges and raises the profits needed to provide the required return on invest- ment. However, the equipment should pay for itself in reduced spoilage, and in facilitating the handling of perishables. The allowances for salaries for meat and perishables managers mean that some savings are available for owner/operators who have specialized knowledge in these areas, but the owner/Operator we envisage, particularly in the early days of the system, is a man with administrative and marketing skills rather than experience in product hand1ing. Later, when the system is firmly established and its Operation well understood, owner/operators could well rise from the ranks of the meat and perish- ables supervisors. It is important that the prOposed retailers sell meat directly rather than rent space to butchers as is currently widely practiced. Existing margins on meats are so high that they can be brought down enough to significantly reduce food prices while still providing a large proportion of the profits of the new retailers. This, together with the desirability of raising the standards of meat-hand1ing in Cali, means that the new retai1ers must keep control of their meat Operations. Some training in meat handling may need to be provided for the new retailers if this control is to be effective. It is, of course, possible that the various departments of the new stores could be 220 operated by separate entrepreneurs, with control and coordination being exercised by the wholesaler through some form of franchise agreement. The key issue here is control and coordination, and any arrangement that can guarantee control and coordination will be satisfactory. The ideal people to hold control would be the wholesalers, but a wholesaler would have difficulty exerting control over, say, a butcher renting Space from a retailer. Although the new retailers will probably be very similar in appearance to the existing supermarkets they will in reality be very different. The supermarkets are located in upper-class barrios and cater to upper-class consumers. They sell staple foods at very competitive prices, but the nature of their clientele permits them to carry large stocks, and presumably to obtain large sales, of luxury items such as liquors, candies, and toilet articles, as well as canned foods, which are luxury items in Cali. These are high-margin articles, and the gross profits provided by their sales allows the supermarkets some Operating slack. Again, the upper—class families who patronize the supermarkets spend far more, per capita, on foods than can the lower-class families to whom our new retailers will cater. Therefore the supermarkets can develop substantial sales volumes even though their clients obtain most of their meat at the Specialist stores and most of their fruits and vegetables in the public markets. The supermarkets can afford, within limits, to handle only the products they wish to handle. In contrast, our new retailers will have very limited sales of those non-staple foods that offer comfortable margins of profits. They will have to run a very tight operation, in the sense of being 221 unable to afford any inefficiencies. At the same time, they will be unable to afford the luxury of offering only easily-handled articles. It is part of their assignment to act as full-line food suppliers, and in any case they cannot ignore anything that will help them to build their sales volumes. Projected income statements for the proposed new wholesalers are given in Table V.5. These Operations are quite different from any existing institutions in Cali's wholesale sector. It is vital that the Operators and their assistants have some training in marketing, for these men will be the keys to the success of the whole proposed system. In addition to their functions as wholesalers they will have to provide management services to retailers in such areas as accounting and stock control, and coordinate promotional activities. They will also have the responsibilities of channel captains, chief amongst which will be that of setting and maintaining policies of low margins and high turnover. All these responsibilities will devolve upon the wholesalers because there are very few peOple in Cali with the appropriate training and ability to carry out the necessary tasks, and these peOple will be most effective when working as channel captains with the wholesalers. In this way their contributions can be made available to the greatest number of firms. Since such services cannot easily be supplied to a randomly fluctuating clientele a chain form of organization will be required. Each wholesaler will be linked in a contractual relationship with ten to twenty retai1ers, supplying about two-thirds1 of their stock. The retailers will benefit from the help they receive in connection with management problems, while the wholesaler's buying will be made easier 1 PIMUR's estimate (PIMUR Final Report, p.102) 222 and risks reduced, by the steady demand from his retailers. Table V.5. Projected Income Statement of Proposed Wholesaler. ' Ca1i, 1969 Sales Col./$2,738.000 Purchase 2,600,0QQ_ Gross Margin COl./$ 138,000 5.0% Expenses Manager/owner 15,000 Meat Supervisor 5,000 Perishables Supervisor 5,000 Operations manager 5,000 Butchers 0 $2,000 4,000 Secretary 1,500 Bookkeeper 2,000 Casual Labor 0 $600 6,000 Fringe Benefits 17,400 Total payroll Col./$ 60,900 Rent 6,000 Utilities 3,000 Transport (incoming) 4,500 Delivery 8,000 Insurance 1,000 Packaging 12,000 Depreciation 1,600 Maintenance 1,000 Bad debts 5,000 Promotion expenses 2,500 Miscellaneous 2,500 Total expenses Col./$ 107,000 Profit before taxes 30,000 Profit after taxes Col./$ 25,000 Source: PIMUR Final Report, Table 2.43 p.103 (adapted) Table V.5. is based upon returning the going rate Of 25% per month on investment as profit after taxes. The figures given are estimated for a wholesaler supplying ten retailers. The returns to management are high, amounting to Col./$21,000 per month (Col./$15,000 salary plus 40% in fring benefits). 223 In conjunction with the 2;% per month return on investment this Should suffice to attract competent entrepreneaurs. The gross margin of 5% in line with existing wholesale margins in Cali, and the services provided (delivery and packaging, which includes grading and throwing-out of spoiled goods) should reduce the expenses of retailers enough to take several percentage points off the margins they require. These services account for one-fifth of total wholesale expenses. PIMUR estimated the investment required for the wholesaler at Col./$l,006,000, made up of Col./$193,000 Of fixed equipment and Col./$813,000 of working capital]. We have already noted PIMUR's estimate that the wholesalers would supply about two-thirds of the needs of their affiliated retailers. The one-third which they would not handle would be made up of non-foods (household supplies), soft drinks and liquors, processed foods, and items like mi1k, poultry, and eggs, for which efficient wholesale distribution is already available in Cali. This figure will certainly apply in the early stages of development of the prOposed system. However, as the wholesalers become established it will probably happen that food processors will find it more economical to deal with them rather than with the ineffective independent agents presently used, and, in many cases, more economical than doing their own distribution direct to retailers. Similarly, the poultry and egg distributors, who in many cases are outlets owned by producers' associations, may find it more economical to deal through the wholesalers and Operate as sales offices instead Of holding stock themselves. 1PIMUR Final Report, Table 2.44 p.104 224 Organisation The basis of the prOposed food distribution system is the chain of stores of ten to fifteen stores associated with a broad-1ine wholesaler. The form of association should be that of the voluntary chain, with the wholesaler as the channel captain. Although fully- integrated chains would be good there are few entrepreneurs in Colombia who command the amount of capital required to start such a chain, and fewer still who have the marketing and other skills required to operate it. Cooperative chains would have problems in that the members would be reluctant to let the wholesaler have the powers that he would need if he were to act as channel captain. The wholesaler will need a good deal of authority over the operations of his affiliated retailers, eSpecially in the early stages of deveIOpment of the system. The prOposed retailers are the very antithesis of the neighborhood stores presently operating in Cali. Instead of avoiding competition they will have to compete vigorously for the custom of people living on the fringes of their service areas, instead of being technique-oriented they will have to be consumer- oriented, and instead of taking their markets as givens they will have to adopt a low price - high turnover philosophy and regard their markets as functions of their own skills as marketers. It will be some time before these ideas are full assimilated, and during this time the new retailers will be frequently tempted to relapseiinto the old ways of doing things. The wholesaler should have enough power to prevent this, particularly with regard to pricing policies. 225 A characteristic of the voluntary chain is that the whole- saler is the channel captain, and can exert a very strong influence over the behavior of his associated retailers. Chains of this type have been in existence in the U.S. for many years, and a considerable amount of experience has been accumulated with them. It is strongly recommended that use be made of this experience, eSpecially in drawing up the contracts between the wholesalers and retailers. We describe below a public body, one of whose functions would be that of aiding in the financing of the new wholesalers and retailers. This power of the purse-strings could be very effective in persuading retailers to accept the idea of having the wholesalers "interfere" in their operation. The public body, in effect, would be in a position to impose a "supervised credit" form Of control over the new system during the time that such control would be most needed. A primary objective of this control would be to make the best use of the managerial capabilities which we propose to concentrate in the wholesale sector. The system we propose would be composed of institutions which would be independan:of each other financially, except for the control of the above-mentioned public body. Each participant would be the owner of his own operation. This would have the disadvantage that each participant would, in order to enter the business, have to be able to invest rather substantial sums of money in his Operation. This would limit very severely the number of potential entrants into our new food distribution system. 226 An alternative organizational and financial possibility is to lease out a proportion of the retail Operations. This alternative is confined to the retail sector. While we can visualise a syndicate type Of wholesale Operation, with many Operating partners, the whole- sale sector is too critical to permit the extension of decision-making and policy-making powers to lessors. The reasons will become clear after we have discussed the application of leasing in the retail sector. Leasing-out of operations in the retail sector could vary widely in degree of application. At one extreme a retailer cou1d rent out a few square meters in a corner of his store to a baker for two or three hours each day. The baker might come into the store at the appropriate times just to sell fresh-baked bread. The bread could even be baked on the premises. At the other extreme the wholesaler might build and equip the retail outlet, as far as physical facilities are concerned, instal an administrator and undertake to supply management services, and lease out the Operations, department by department, to concessionnaires. The principal advantage of this would be that the potential entrants to the retail sector would need much less capital than the new retailers we described earlier. This would greatly widen the ranks of potential entrants. In particular, it would permit some of the existing stallholders, with their specialised prOduct-handling Skills, to enter the new system. Unfortunately, there are two serious dis- advantages to this idea. First we have already noted that there should be a _ ”,,_______e 227 contractual relationship between the wholesalers and retailers of the new system. This is required to permit the provision of management services by the wholesaler. Particularly during the early stages of development of the system, it will not be easy to Specify in the contracts precisely what are the rights and duties of each party to the contract. The addition of a third level of distribution, the lessees, will make it difficult to contractually formalise the relationships between components of the marketing channels. Second, as we have already noted the key to the success of .. ‘9 D“ “a. 'u. .4, .44).: “mm: gr .",«-... the whole system is that the participants must be marketing- -oriented. "H.M“A‘fl -5...” w—‘fiA‘I’ h—n- bul- ‘ _I" "" "' 9 l wtiHBOMrou’h/i . -. 1‘... w This is absolutely essential. But marketing orientation is a new “(...—‘— 4.,” “I!" W concept in food distribution in Cali, which means that exposure to the marketing concept will be through a diffusion process. Diffusion of ... ...... ..L MWMHHJI—rfi'. ...“... a ideas is more rapid amongst the better- educated, and the Consumer m4 “A‘s-M Survey showed that in Cali education is closely correlated with income. mm—m-dw_, _v Ill-WW "WFW' Luis Income, in turn, is closely correlated with wealth, or command over capital. It seems safe to conclude that the lower the requirements of capital to potential entrants to the retail sector of our proposed system the Slower will be the adoption of the marketing concept. In the developmental stages of our proposed food distribution system the risks associated with lack of a marketing concept far outweigh the advantages of being able to select from amongst a larger number of potential retai1ers. We must therefore reject leasing of Operations as a feasible strategy during the developmental stages of the proposed system. On the other hand, once past the developmental stage and into the growth stage-leasing out of operations could well prove to be a very effective 228 strategy. At this stage the operation of the system will be well understood, and the rights and duties of participants well defined through experience. Actual operation in a marketing—oriented environ- ment will probably be the best way of communicating a marketing orientation to potential entrants. The best apprenticeship for potential retailers will be the operation of a concession. On several occasions we have noted that the wholesalers must have some power over the retailers. This can take any of several forms. The most extreme form would be actual dispossession, in cases where the wholesalers owned and leased—out the retail facilities. Less dramatic manifestations of power would be the withdrawal of management services or the refusal to allow use of the chain's trade-mark. As was suggested earlier, the experience of voluntary chains in the U.S.A. should prove a useful guide. The greatest weakness of the system as proposed is the difficulty of getting it started. The proposed wholesalers and retailers are mutually dependent, and in the existing system neither can exist as planned without the other. There will therefore be a difficult period while the first few chains become established. Some subsidization of the first Operators will certainly be needed, but even more important will be a program of Sponsorship by a public body. This body will be charged with a variety of difficult but extremely important tasks. One of the most important will be to recruit retailers for the new system, in order to build up new chains to viable size, in order to keep the needed subsidies as small as possible. Along with this will go the task of finding the loan capital needed by the new entrants. This body will also have to be responsible for administering the programs of training and technical assistance 229 that will be needed. Not the least of its tasks will be that Of being the public relations agent for the ertire system. This will be needed to find new entrants to the system, and also to offset the political pressures that will undoubtedly come from those who feel threatened by the changes in food distribution. Another function of this body will be that of helping to resolve the conflicts that are bound to arise between the members of the chains. This will be more important in the early stages, while the members of the chains are still unsure of their i rights and responsibilities, and while the public body has the consid- erable influence that goes with being the disburser of subsidies and F raiser of funds. It follows from the above that this body must be carefully chosen. It must have a substantia: amount of technical Skill at its disposal, it must have enough prestige to enjoy the confidence of the public, the government (at both national and municipal levels), and the new food distribution institutions it is working with, and it must have extensive contacts with financial institutions in Colombia and, possibly, elsewhere. Organizationally, therefore, the new system will include three closely-linked institutions, the retailers, the wholesalers, and the sponsoring/regulating public body. The public body should fade from the scene as the system becomes established. Any formal contacts with other institutions in the food distribution sector should be straightforward. There will be arm's-length dealings with suppliers and with tranSporters, but these pose no Special problems. 230 Social Consequences When and if the proposed food distribution system comes into Operation it will reduce food prices. PIMUR has analyzed the costs- added structure of food distribution1 and has Shown that the proposed system could distribute feed at prices 6% lower than those under the existing system. This is an undeniable social good. It means that a large proportion of the population of Cali will be better fed. It also implies a large proportional increase in the effective demand for q elaborated goods, manufactures, processed foods of a luxury nature, etc. % Properly handled, this increase in demand could be a powerful stimulant ' to economic growth and to increased employment. Against this must be set the social costs of disruption in the existing food distribution system, particularly the dislocation of the Operators of the tiendas and small graneros. There were 3,300 of these stores Operating in Cali in 1969, giving employment to their owners and perhaps 700 paid employees, thus providing the livelihoods Of 4,000 families. What about these people? Let us examine this figure of 4,000 families rather more closely. The first thing is to note that at least 30% of the Operators Of these businesses can best be described as transient. They will go out of business whether the new system comes into Operation or not, and the only effect of the new system will be to discourage others from coming in to replace them. AS we noted on page 165 , the dissipation of personal savings involved makes this type of concealed unemployment a particularly undesirable one, so that in this reSpect the advent of the new system will probably be a net social benefit. Although the 1PIMUR Final Report, pp.105-lO7 231 underemployment or unemployment of the 1,200 people involved is a matter for serious concern it cannot be attributed to the prOposed system. This leaves 2,800 people whose lives will be affected adversely by the new stores. Against this can be set the employment which these stores will bring about. When it is fully established the new system will sell between Col./$80 million and Col./$lOO million of food each month, and involve from 200 to 250 stores. These stores will employ an average of eleven people each, so that the retail sector of the system will be reSpousible for employing between 2,200 and 2,750 peOple]. In addition, the salaries paid (ranging from Col./$6OO to Col./$2,EOO per month plus 40% fringe benefits) are rather greater than the net profits of the existing neighborhood stores. (See Tables V.l. and IV.ll) So the new system roughly balances out livelihoods lost and livelihoods created. As far as disruption is concerned, this should be minimal. The growth of the system should take at least five years, more probably ten, leaving ample time for normal attrition to reduce the numbers of tiendas and small graneros and allow orderly withdrawal. This also applies to the reduction in numbers of stallholders in the public markets, a1though here some redeployment will be necessary as numbers in individual markets fall below the levels at which the markets are effective. 1The population Of Cali has been assumed constant for the sake of comparability. This does not affect the argument. 232 In fact, when the opportunities for small neighborhood businesses as cafes and high-convenience (e.g. during evening hours) stores are taken into account it would seem that there would be a net social benefit to be derived from the new system entirely apart from the benefits of reduced prices. This is a somewhat surprising conclusion. Usually the introduciion of more efficient techniques causes a good deal of dislocation, However, our argument appears to be sound. The possible weak point is our assumption that underemploy- ment with dissipation of savings is no better than no employment at all, which is a point Open to question. The losses suffered by the centrallvmlocated traditional stores can be offset by the opportunities created by the new whole- salers], and only negligible disruption should be felt by the existing wholesalers as they move upmrhannel to become brokers. This move only relieves them of the task of selling, which they are not good at, and leaves them free to concentrate upon the thing they can do well, buying. A substantial loss of business will be felt by the intra-city transport sector, since it is in transportation between wholesaler and retailer that the new system will realize a great part of its cost savings. However, a) the barrows, an important segment of this sector, will lose most of their business anyhow, if the proposal to construct a wholesale center on the edge of the city Comes to fruition, and --—. -- c.--...— 1These businesses will be hurt by the clearance of the Ga1eria Central whether the new system comes in or not. 233 b) this sector has trouble in keeping up with the demands made upon it by the existing system, so that the major loss will be in excess profits. The personnel transportation sector, the buses, will also lose some trade. This will be fairly serious, since it will bring into prominence in Cali the problem which plagues mass transport- ation everywhere, the problem Of profitably using during the slack hours of the day the large numbers Of vehicles needed to cape with rush-hour traffic. The answer to this, Of course, is to adOpt a policy of staggered hours Of work, but this appears to have dis- advantages which outweigh its benefits tO transportation. There are a further 3700 stallholders in the public markets who will, in the long run, be displaced by the new system. However, this will take a long time to accomplish. It should be noted that the last stall Operators to be affected will be those dealing in fruits and vegetables, amounting to 2400 people. It would appear that natural attrition, through retirement, could be used to phase out the public markets faster that the competition from the new system. The social disadvantages Of the proposed food distribution system would appear, on the whole, to be minor. They are certainly negligible compared with the major advantages the new system Offers in terms Of reduced food prices. The Contribution of Government The various levels of government intervene quite substantially in Cali's food distribution, both directly and indirectly. Direct intervention includes activities such as policing controlled prices and the Operation Of the IDEMA stores by the central government, and the Operation Of the slaughterhouse and administration Of the public markets 234 by the municipal government. Indirect intervention includes economic legislation such as the laws against peculation, licensing Of dealers, and the support and encouragement Of non-profit organizations, the cooperatives and the Cajas. In general these efforts have been either ineffectual or dysfunctional. PIMUR has analyzed the effects of the major activities, and made recommendations for changes as necessary]. We are concerned here with the contribution that government can make towards helping the proposed food distribution system to become successful. Several problem areas can be foreseen in which help from the government will be useful or necessary. One such area will be problems within the new institutions, such as conflicts between retailers and wholesalers. Another will be in public relations between the new institutions and the public, and between them and the existing instit- utions. A third will be in mobilizing the financial and human resources the new system will need, and a fourth will be the question of subsidies to wholesalers in the early stages Of develOpment of the system. A fifth possible problem area lies in conflicts between the concepts involved in the new system and existing laws and regulations. The key contribution from the government will be the establishment and support of the public body discussed earlier, which will be responsible for fomenting the proposed system and for certain aspects of its regulation when it is established. It is particularly 1PIMUR Final Report, Chapter 6. 235 ‘ important that this body enjoy the confidence of those levels of government with which it will have to deal. This implies that the government have a good understanding Of the purpose of the prOposed system and the methods by which it is expected to achieve that purpose. Raising the funds required to create the system will be difficult. In order to combat inflation Columbia has embarked on a policy Of credit restriction such that fully—secured corporate loans cost 25% to 30% per year. Personal loans, secured on real estate, can cost 60%. On the other hand, there are special lines Of credit available for agriculture and for most manufacturing industries, the so-called "development industries". The shortage of funds available for lending will severely hamper the development Of the prOposed food distribution system unless the central government provides a line of credit for this purpose. This will be particularly true for wholesalers in the early stages, because of the novelty of the institutions and the relatively large amounts of capital required for each enterprise. Finding and developing the human resources required will need help from the government. There are three main forms that this help can take. There must be vocational instruction in various aspects of retailing, particularly meat management, which could be handled by SENA, the national apprenticeship system. Technical assistance teams will be needed to advise the wholesalers while they find their feet, and the wholesaler's reSponsibility to provide management services to his retailers will create a need for people with training in cost accounting and Operations research. This training is at university level, so that some scholarships will be needed, probably to universities outside 236 Columbia. University-level instruction in marketing will also be highly desirable for the wholesalers and retailers themselves. Evolution of the Proposed System The major defect of the proposal described is the difficulty of visualizing how the first few chains will come into being. Once they are established and shown to be profitable imitators will enter the field. But how the first chains will start is not immediately Obvious. Historical precedent does not help. Voluntary and COOperative chains started as associations formed by existing institutions with the Objective of meeting the challenge of the integrated chains. They were change-averting rather than change-inducing movements. It appears that present pressures in Cali's wholesale sector are not such as to encourage the natural evolution of wholesalers of the type proposed. In particular, the fact that they will be buying from existing wholesalers will generate resistance to their activities, because the simplistic theory that middlemen only add costs finds wide acceptance in Columbia. An attractive strategy would be for the pioneer wholesalers to start as meat wholesalers only. This is the high-risk stage, because Of the losses through spoilage and the small volume over which fixed costs can be Spread during the period while business is being built up. The next stage would be to provide delivery service to their clients, making a charge for delivery of goods bought from other wholesalers. Retailers' tranSportation expenses are high enough that this service could be made quite profitable through consolidating shipments. A buying service, for which a charge would be made, could be made acceptable to clients by suitably offsetting the delivery 237 charges, and from here it should be only a short step to becoming a wholesaler holding stocks. As their volume increases, the new whole- salers will arrive at a position where they will be able to demand volume discounts from the Specialist wholesalers, thus putting pressure on them to stOp holding stocks and become brokers. It should be noted that the clearance Of the Ga1eria Central, with the subsequent removal (if EMSIRVA's plans come to fruition) of wholesalers to a wholesale center on the outskirts of the city, will ‘rv disrupt the meat wholesale system as it presently exists. This will mtg-is' I I. provide an excellent Opportunity for the new wholesalers to enter meat wholesaling. It will also provide an excellent, possibly unrepeatable, I opportunity to modernize the meat wholesaling function in Cali. A danger with this strategy is that the new wholesalers might, probably will, develop a product orientation rather than a marketing orientation during the time they are entirely dependent upon meat sales. Also, people who would be prepared to enter the business in this manner would probably already be butchers with a strong product orientation. They would be unlikely to be bringing in the new ideas the new system would require. 0n the whole, therefore, this strategy must be rejected, unless some exceptional individuals come forward who want to enter food wholesaling in this manner. A better source Of entrepreneurial talent would appear to lie in the opportunities made available by the moribund state of the agencies which at present handle processed foods. As we noted earlier, it should not be long before some bright young men enter this branch Of commerce, and the better Operators among these should be excellent candidates for the jobs Of channel captains for our proposed 238 voluntary chains. Under this strategy the tactics will be straightforward. The governmental body discussed in the previous section would wait until these new entrepreneurs could be identified, and then would approach them with offers of supervised credit, technical assistance, and such other help as might be useful in starting them Off as wholesalers for our proposed retailers. The evolutionary development of the new whole- salers would not be complicated. Initially, they would act as whole- salers carrying a more-or-less complete line of processed foods. Then they would proceed through stages Of providing delivery service, managerial services, and buyer services (not necessarily in that order), until they end up as full-line, full-service wholesalers. There is, Of course, a final possibility. A sponsor might be found, willing to finance one, or even two, chains of the type described, as a philanthropic action rather than as a profit-seeking action. Although very unlikely, it is possible that such a "deus ex machina" could be found in one of the large charitable foundations. More realistically (but with still only a low probability Of occurrence), one or more of the chain stores might decide to diversify into full-range food distribution. Any which made such a decision would constitute ideal candidates as integrated chains. However, such decisions are unlikely for two reasons. l) There was, as of l969, plenty of room for expansion for these chains. For example, there are six cities in Valle, besides Cali, with populations approximating one hundred thousand, yet only one of the chains had set up a store in one of these cities (Palmira, near to Cali). 2) In 1969, and into the foreseeable future, it would appear that the gross margins available on small manufactured good are far larger than those on foodstuffs, and the handling costs are lower. We have no direct authority for this, except the observation that the laws regulating prices in Colombia are presently concentrated upon the so-called "articles Of primary need", comprising principally foodstuffs. Summary We have shown that the existing food distribution system in Cali is not satisfactory. The growth of the city has outstripped the evolutionary capacity of a fOOd distribution system which was originally designed to serve market stalls. The most important requirement of the Cali consumer is that food retailers be located near their homes. We have shown that public markets cannot fulfil this requirement, and that a food distribution system built around neighborhood stores is needed. Such stores can be large enough to enjoy economies of scale and thus sell food at low prices, but they will have to be operated by people with a marketing orientation and be supplied by wholesalers of a type which does not at present exist in Cali. A food distribution system has been proposed, based upon these new wholesalers and retailers, which could reduce the food bills of three-quarters Of Cali's population by about 6%. An important characteristic of the proposed system is that the new instit- utions will be sufficiently profitable to attract able entrepreneurs. The most serious problem envisaged lies in the novelty (to Cali) Of the new holesalers and retailers. This will slow initial 240 develOpment, and the lack of experience will heighten the probability of making errors. We have outlined strategies for easing the problems of the initial stages. APPENDIX I POPULATION ESTIMATES APPENDIX I Population Estimates The only data available concerning the population of Cali are the censi of 1951 and 1964. These showed populations, as of those dates, Of 241,000 and 639,000, with an implied growth rate of 7.7% per year during the intervening period. If this growth rate is extrapolated the pOpulation of Cali at the time of the field work, February 1969, would be 928,000. PIMUR used slightly different growth rates to arrive at an estimate of 895,000. ‘ Both of these estimates, however, depend upon an exponential or quasi-exponential growth model, in which the observed growth is the resultant of vegetative growth (i.e. the natural net growth rate of the population), plus immigration from other areas. This immigration is also assumed, in the exponential model, to be proportional to the size Of the city. But it is difficult to conceive of a mechanism existing in the rural areas which will adjust the inflow Of migrants to a set prOportion of the current population Of the city. In short, an exponential growth model, while applicable to closed biological system, is not relevant in the case Of open systems‘such hs cities. We suggest the following as a more plauSible model. We assume first that the rural population of Colombia is, to a first approximation, steady. Given the fact that no substantial land areas have been opened recently for exploitation, and the absence of a primegeniture system 242 of land inheritance, this assumption seems to be reasonable. Assuming population constancy, it follows that the movement of people from rural to urban areas must be such as to offset the rural net reproductive growth rate. The mechanism behind this balancing outflow is probably popu1ation pressure on the rural areas. The vast majority of the rural population lives on plots of land which are now (and have been for at least the last twenty years) as small as is consistant with supporting one family, and that employment in commercial agriculture, (sugar, coffee, stock raising) is stable. If this assumption is accepted, we can calculate the number of people leaving the country for the city each year in the following manner. The annual growth rate for Colombia is 3.2%1 , and demographers in Colombia believe that the rural growth rate is higher than the urban growth rate, though this has not yet been conclusively proved. It would seem that taking the urban rate at 3.1% and the rural rate at 3.3% would fit these conditions pretty well. In 1964 the population of Colombia was 17,485,000, and of this 51.5% lived in rural areas. Thus each year the rural population would have an excess of 17,485,000 x .033 x .515 = 297,000 people who would migrate to the cities. Let us now assume as a first guess, that this flow divides itself amongst the various cities in proportion to their populations. It can be argued that the spatial distribution of urban centers with respect to the settled rural areas will be an impOrtant factor. However, 1Pearson, Lester 8. Partners in Develgpment, Praeger, New York, 1969, p.250. Ill" Illl II .III I 244 we feel that this effect will be relatively insignificant since the urban centers in Colombia are interconnected by a11-weather roads and transportation costs are not very high. Cali, then, with 7.5% of the country's urban population would receive 7.5% of the flow, or 21,600 people. Would a yearly flow of this approximate size account for the growth of Cali between 1951 and 1964? Our model of the city's growth is of the form P(n) = aP(n-l) + k where P(n) is the population at the end of year n P(n-l) is the population at the end of year n—l a is the vegetative growth factor (1.031 for Ca1i) and k is a constant influx rate If we solve this difference equation using the populations in 1951 and 1964 as boundary points, we find that k comes out at 19,000. This close agreement, 19,000 versus 21,6000, is probably rather better than is warranted by our first-cut guess about how the flow splits itself between the various cities, but it is very encouraging as a test of our model. Data from the Consumer Survey make another test possible. Respondents were asked whether the household head was native to Cali, and if not how long he had lived in that city. We can assume that the number of immigrants who are heads of households is in a fairly constant proportion, through time, to the total number of immigrants, or that the time distribution of length of residence of immigrant household heads is similar to that of immigrants as a whole, and from the Consumer Survey data we have a sample of this distribution. This was tested against the distribution that would be expected from an immigration 245 rate proportional to the population of the city, and against a rectangular distribution. The differences between the exponential model and the observed rate were significant at the 1% level of probability virtually throughout the entire time period tested, while the rectangular distribution agreed with the observed data to well within the 10% level of probability except for the years 1951 to 1953. Since these were the years when the "violencia" was at its peak, and a vigorous industrial building boom was in progress in Cali, it is to be expected that the immigration rate during this period would exceed the normal rate. We feel, therefore, that we are justified in adopting the constant-influx mode1, perhaps adjusting it to suit temporary conditions. For instance from 1964 until 1969 Cali was in a state of economic stagnation while in 1967 and 1968 there was a construction boom in Bogata in preparation for the Pope's visit to that city in 1968. We will therefore use 17,000 instead of 19,000 for k for the years 1965 through 1969, 21,000 for the years 1970 and 1971 (in con- sideration) of the Pan-American Games which are to be held in Cali in 1971, the preparations for which will involve considerable construction activity) and 19,000 for 1972 forward. On this basis we find that our own estimates of the city's population work out at Year 1969 1979 1989 Population 835,000 1,238,000 1,880,000 The estimates for 1979 and 1989 must be taken as upper limit estimates, since it is to be haped that birth control programs will have started to show results by these dates, at least in the urban sectors of the country. However, there is no way of telling when, or to what degree the programs will become effective (if indeed they ever 246 start), so we have to stay with the figures as tabulated. The above figures refer to the month of July for the various years listed. The field work upon which this study is based was carried out in February 1969, at which time the above model shows the population to have been 813,000. APPENDIX II CONSUMER SAMPLING AND SURVEY METHODOLOGY APPENDIX II Consumer Sampling and Survey Methodology The consumer sampling and surveys consisted of the Consumer Survey and the Market Basket Survey (Market Basket Study). The latter was the basis upon which sample sizes were determined. In the Market Basket Survey the original intention was to record the food purchases of 100 families for one month. At the sug- gestion of Dr. James Shaffer of Michigan State University the objective was changed to that of recording the food purchases of 400 families for one week each. The decision to aim for 100 fami1y-months (or 400 family-weeks) was almost entirely arbitrary, since no secondary data concerning food purchasing was available. Perhaps the main criterion was that of availability of human and financial resources. It was estimated that two-thirds of those asked to participate would do so, and since "those asked" would be the respondents to the Consumer Survey this consideration set the sample size of the Consumer Survey at 600 families. This number appeared to be large enough for the purposes of the Consumer Survey. Simple random sampling offered several advantages which led to its selection as the sampling technique. It was the method which would be least affected by the paucity of demographic data about Cali, it simplified the analysis of the data which was to be obtained, and it guaranteed that there would be adequate representation of the sector in which we were most interested, the families in socio-economic levels V and VI. 248 249 The sampling itself was enormously simplified by the courtesy of Empresas Municipales de Ca1i in letting us use their list of electricity meters. This list was sampled, at a sampling intensity of 0.6% using Monte Carlo methods, giving us a list of addresses in those barrios which enjoyed electric power service. The barrios outside this category were, without exception, of sufficiently recent origin for all housing therein to be registered in the municipal cadastral office, from whose lists population estimates for these barrios were prepared. The houses in these barrios were then sampled at the same sampling intensity of 0.6%, using two~stage random sampling by blocks and houses. The principal weakness of this method of sampling was that it makes no allowance for the case where more than one family occupies one house. There was a housing shortage in Cali in 1968-69, which together with the poverty of the people made the multi-family case one of common occurrence. The Consumer Survey showed that roughly one house in four or five was occupied by more than one family. In many cases all the occupants ate as a unit, but to the extent that they did not the sampling method was not truly simple random. Since the practice of sharing a one-fami1y residence amongst several families is more frequent amongst low-income families the lower socio-economic levels are under- represented in our results. Unfortunately, we did not anticipate how widespread this practice was, and the interviewers were not instructed to obtain the appropriate data, so we cannot make the corrections required. As it happens, the only results affected by this oversight are those constructed for the entire population. Results pertaining to individual families are unaffected, and these are the ones most germane to the present paper. Nevertheless, we should note that the figures presented for total (city-wide) consumption are biased in the upward direction in consequence of the under-representation of the lower—income families. The magnitude of the error involved is unlikely to exceed 10%. In the event, 629 interviews were carried out in the Consumer Survey. The rate of participation in the Market Basket Survey was disappointingly lcw, as only 323 housewives agreed to take part, and of these tnly 240 "finished the course". This high drop-out rate could, we belie e, have been substantially reduced by better preparation of the hevzaeives for the ordeal of having the privacy of their kitchens violated. Interview supervisors should have visited those who offered to participate and explained in detail what was involved, making much of the importance of their cooperation to the success of the study and of the imcnrtanee ef the study to the future of the city. Surveys with a high rate of refusal are always subject to criticism on the grounds that the factors associated with the refusals may also be systematically associated with the variables which are the subjects of the study. That is the sample as finally composed may not be reprisentatiue of the population. To avert as far as possible such criticism we present in Table A.II.1 comparison of several characteristics of the samples involved in the Market Basket and Consumer Surveys. In no case is tne difference between the two samples significant at the 5% level of probability. Table A 11.1. Comparison of Selected Characteristics of Participants in the PIMUR Consumer Survey and Market Basket Survey, Cali, February 1969 Income Level 1 2 3 4 No. of Families in Market Basket 31 56 62 40 Sample Composition (%) Consumer Survey 20.2% 27.8% 28.8% 23.2% Market Basket 19.7% 28.0% 32.6% 19.7% Average Family Size.l Consumer Survey 8.0(2.8) 6.6(2.5) 5.6(2.1) 5.5(2.3) Market Basket 8.5(2.6) 7.2(2.6) 5.6(2.l) 5.6(2.6) Years of Education of Household Head Consumer Survey 4.0(3.5) 5.3(3.9) 7.1(4.2) 10.8(4.7) Market Basket 4.0(2.6) 5.1(2.8) 6.5(3.9) 9.9(4.4) Calinorn Heads of Household (%) Consumer Survey 41% 51% 43% 47%2 Market Basket 35% 53% 45% 62%2 1 Figures in parentheses are standard deviations 2Of all the comparisons in the table this has the highest ”2" value. In this case "2" is 1.9, compared with 1.96 which would be required for significance at the 5% probability level. The evidence therefore is that the Market Basket sample is representative of the Consumer Survey sample as far as these demographic 252 variables are concerned, and we infer that the similarity extends to food purchasing patterns. With the exceptions mentioned above, we are taking the Consumer Survey sample to be representative of the population of Cali, so we consider the results of the Market Basket Survey to be M :M' " representative of the food purchasing habits of this population. _‘wawwv ' . .. - 'cnu..hl'iflt".,‘a_.’a-. y‘....*,..-..._'. I _ ‘ tn 1 “w... ‘7“ 7,- 7? 0;: “9' ..fi; we . “a3”... ...__. .mm—c- . -'|'.‘- "Whit-n. APPENDIX III EXPERIENCE WITH THE ATTITUDE-MEASURING SECTION OF THE QUESTIONNAIRE APPENDIX III Experience with the Attitude-Measuring Section of the Questionnaire This part of the Consumer Survey started with the interviewer saying ”Continuing our talk about places where you can make food #1 purchases, I'm going to read out certain characteristics of these places. F For each characteristic please indicate which place, in your opinion, is be§t_and which is worst." The respondent was then handed a piece of paper on which all r the various types of retailers were listed in a circular (pie-chart) fashion, in order to avoid biases against the first or last items listed, and asked to read the names to refresh her memory. Then the interviewer read out the following list of characteristics: 1) Quality of Foodstuffs 2 Agreeable clientele 3 Gives credit 4 Delivers to the home 5 Hours of business 6 Cleanliness and tidiness of the establishment 7 Weights and measures 8 Price and reductions and discounts 9 Low cost of transport 10) One can serve oneself 11) Service and attention 12) Variety of foodstuffs 13 Convenient location for the housewife 14 Prices of staple foods 15) Prices of non-staple foods These were chosen as being the most important factors in the consumer- retailer interactions. Unfortunately we have not been able to devise any method of determining how exhaustive the factors listed are, and 254 the extent to wh ch they account for observed retai1er-se1ection patterns. However the discussion in chapter III indicates that the important factors are included. Some problems were encountered with the phrasing of the items listed. Item (8), price reductions and discounts, was intended to test the hypothetis thtt ”haggling" is an important element in food buying. The Spanish word ”Ligatear" which is most commonly used to mean "haggle" is considered in Caei to have somewhat pejorative connotations, so the words EEQEJ§EQXH§QS¥H?QIQ§.(“BdUCtions and discounts) were used instead. while these are principally used in connection with bargaining and dickering they rev also refer to any price cuts, whether bargained for or not, and ”any reSpondonts took this meaning. The chain stores, which hold frequent cut»price promotions, were often mentioned as best in this reSpect, although they do no dickering. This dual interpretation invalidates any straight—forward analysis of the results, but the readiness of resiondents to take the wider meaning would indicate that haggling is indeed less important than is popularly be1ieved. Hen1ey1 came to the same conclusion watching transactions in the market-p1ace in La Paz, Bolivia. Item (TU), “one can serve oneself", also invokes two concepts. The first is the simple self service versus personal service dichotomy associated with store format. The second can be illustrated by describ- ing three ways of buying a pound of tomatoes. The buyer has item (10) in the greatest degree when she can select each tomato that makes up .‘H-- 1"Market Processes in La Paz, Bolivia", Research Report No.3, Latin—American Studies Center, Michigan State University, East Lansing, p.46 ff. the pound, has it in lesser degree when she selects from amongst several pre-packaged pounds of tomatoes, and has it in least degree when she is forced to accept the shopkeeper's selection. On this continuum the supermarket will appear about the middle of the scale, with obliging and disobliging personal-service store operators at the extremes. The results of the survey showed that consumers used both concepts when replying to this question. Interviewers reported that this section of the questionnaire was the one that respondents found most exhausting and difficult to answer. In many cases (11% of interviews) the respondent just gave "don't know" answers to the entire section, and it is of interest to note that the prOportion of such replies amongst respondents of S.E.L. VI was significantly higher than that for the rest of the sample (32%, significant at the 1% level of probability by the chi-square test). Survey questions of this type must be considered susceptible to interviewer bias, and therefore the responses were checked for manifest- ations of this. In only one case out of twenty were the responses significantly different from the average of the group: this interviewer was a law student. It would seem that if anything the interviewers leaned over backwards to avoid pressuring the respondents to answer and to avoid "leading" by their mannerisms. On the whole, the high propor- tion of "don't knows" seems a reasonable price to pay for the absence of interviewer bias. There were various other forms of unusable responses to this section, such as that given by people who said that the place they patronized was "best" for all attributes and who had no opinion about "worst", and those who gave only one or two "best" or "worst" ratings and all the rest "don't know". Before the data can be used in a formal quantitative analysis of purchasing patterns they will have to be carefully screened for these and other unusable reSponses. This will involve the exercise of a good deal of subjective judgement. The effect of using these dubious responses in the analysis reported in Chapter III is to obscure the results rather than to produce Spurious results. That is to say, the conclusions drawn err on the conservative side, reducing the apparent importance of those character- istics singled out as the key considerations in selecting retailers. The ranking of importance of characteristics, however, should not be effected. APPENDIX IV COMPARISON OF MARKET STALLS AND STORES APPENDIX IV Comparison of Market Stalls and Stores A market stall is a small establishment, using one or two square meters of storage, selling, and display space, and located in very close proximity to other similar stalls. They handle very narrow lines of goods, with moderate to great depth within the lines. The goods are displayed so that the customer can handle and inspect them, and as a rule only one person tends a stall because space is so severely restricted. A store is much larger than a stall, using five or more square meters of storage, selling, and storage Space, and a counter separates the customers on one side from the goods and attendants on the other. There is moderate to great width of lines carried, with little to moderate depth within lines. Space behind the counter is relatively unrestricted, even small stores being able to use three attendants if pressure of business so demands. Stores of similar types are not usually located immediately adjacent to each other, and even where they are a customer still has to walk ten to fifteen meters to get from counter to counter. Behind these observable differences there are basic Operating differences. The stallholder relies on his knowledge of his products and his skill in buying, because these are the only ways of securing 259 260 competitive advantage that cannot be immediately copied by his competitors. The storekeeper's strategy is based on offering a total package of goods and service, an important component of which is the prospect of one-stOp (or at least few-staps) shopping. When a housewife is buying food, the greater part of her purchases are of a routinized nature. The stallholders depth of line only introduces the need for extra decisions and his narrowness of line means that she has to deal with an irritatingly large number of retailers. The stallholders' relative skills are of little or no significance to her. Shepping of this type is carried out more effectively in stores. When the housewife is looking for shOpping goods the stall is a more effective retailer, but food purchases in this mode will account for only a small proportion of the total, say about l0%. The transaction efficiency of the store is higher than that of the stall. The stall's greater depth of line will make for prolonged negotiation before purchase is consummated, and there is the stage of paying over the money, making change, etc.. with virtually every item purchased. The store's shallowness of line makes for short negotiations and one money—exchanging process can cover many items. In a market, eSpecially in the case of shOpping goods, there is a tendency for buyers to make several negotiations with different stallholders before purchasing. This behavior, which should be distinguished from bargaining or haggling, is an intrinsic part of the process of buying in a market, and is greatly facilitated by the close proximity of stalls to each other. It results in a substantial prOportion of fruitless negotiations for the stallholder, and further reduces his transactional efficiency, the number of sales he can make in a given time. In yet another way does the storekeeper have an advantage over the stallholder, in that he can put several peOple to work behind his counter, thus greatly increasing his rate of sales. The potential productivity of a storekeeper is therefore higher than that of a stallholder, which means that he has the potential to sell at lower prices than obtain in the markets. He can only realize this potential if the flow of customers is high. This explains why the tiendas and small graneros charge high prices, because by locating themselves away from other stores they automatically cut themselves off from the possibility of generating a high density of shOppers. It also explains why the wholesaler retailers and large graneros, located in the Galeria Central area where the shopper density is very high can compete successfully with the stallholders of the Galeria Central, although the stallholders enjoy the benefits of subsidised rents for their stalls. The foregoing discussion has been comparing stalls with traditional food stores. The advantages of the store over the stall are even greater if the supermarket format is considered. BIBLIOGRAPHY BIBLIOGRAPHY Articles and Periodicals Abbott, John C. "Marketing and Area Development Studies". Toward Scientific Marketing Edited by Stephen A. Greyser. Chicago: American Marketing Association, 1964. Bennett, Peter D. "The Role of Government in the Promotion of Efficiency in Retail Marketing of Food Products in Greater Santiago, Chile" Proceedings of the American Marketing Association. Chicago: American Marketing Association, 1965. Carson, David. International Marketing. New York: John Wiley & Sons, T967. Currie, Laughlin. Markets and Marketing in Developing Countries. Edited by Reed Moyer and Stanley HoTTander. Howswood, Ill.: Richard D. Irwin, l968. Drucker, Peter F. "Marketing and Economic Development". Marketing Classics. Edited by Ben M. Enis and Keith K. Cox. Boston: Allyn and Bacon l969. Galbraith, John Kenneth. Economic Development. Cambridge, Mass.: Harvard University Press, 1964. Galbraith, John Kenneth, and Holton, Richard H. Marketing Efficiency in Puerto Rico. Cambridge, Mass.: Harvard University Press, 1955. Heilbroner, Robert L. The Great Ascent. New York: Harper & Row, Publishers, 1963. Henley, 0.5. and Farace, R. Vincent. "Consumer Buying and Communication Patterns in Bolivian Urban Food Retailing". American Marketing Association, Annual Meeting, Washington D.C., 1967. Latin American Studies Center. Food Marketing in the Economic Develggment of Puerto Rico. East Lansing, Mich.: Latin American Studies center, Michigan State University, 1970. Latin American Studies Center. Market Processes in the Recife Area of Northeast Brazil. East Lansing, Mich.: Latin’American Studies Center, Michigan State University, 1969. Latin American Studies Center. Market Processes in La PazgiBolivia. East Lansing, Mich.: Latin American Studies Center, Michigan State University, l969. 263 264 Latin American Studies Center. Market Coordination in the'DeveTOpment Mintz, Pearson, Rostow, Shapiro, Slater, of the Cauca Valley Region - Colombia. East Lansing, Mich.: Latin American Studies Center, Michigan State University, l970. Sydney W. "The Role of the Middleman in the Internal Distribution System of a Caribbean Peasant Economy". Human Organization, Volume 15, No. 2 (Summer 1956). Lester B. Partners in Development. New York: Praeger, 1969. W.W. View from the Seventh Floor. New York: Harper & Row, Publishers,i1964. Stanley J. "Comparative Marketing and Economic DeveTOpment" Science in Marketing. Edited by George Schwartz. New York: JOhn Wiley 8 Sons, 1965. Charles C. "Marketing Processes in Developing Latin American Societies". Journal of Marketing, Volume 23, July 1968. ICHIGAN STATE UNIV. LIBRARIES illiliiiiilliilliillllllilmiINN)“Iillllililli""W 31293101045460