OVERDUE FINES: 25¢ per day per item RETURNING LIBRARY MATERIALS: ———-—-—.___. Place in boon return to remove charge from circulation records THE PUERTO RICO WAGE RATE SUBSIDY FOR AGRICULTURAL LABOR By Thomas M. Dickey A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Agricultural Economics 1980 ABSTRACT THE PUERTO RICO WAGE RATE SUBSIDY FOR AGRICULTURAL LABOR By Thomas M. Dickey The decline in production of the agricultural sector of Puerto Rico is due primarily to the sector's inability to adjust to the rapid in- creases in the minimum wage rates for farm laborers that have accompanied Puerto Rico's rapid industrialization. Both the United States and Puerto Rican minimum wage authorities set minima on an industry by industry basis and have vigorously pursued a policy of narrowing the gap between wage rates on the island and the standard federal minimum wage rate applied on the mainland. Puerto Rico's production of food- stuffs has historically been minor due to the island's ability to import the lower cost United States mainland foods. Thus, Puerto Rican agri- culture has been dominated by three export crops which are heavily dependent on low cost supplies of labor. The Wage Rate Subsidy Program was implemented in 1969 for the ex- press purpose of reviving the agricultural sector by retaining the sector's labor force by raising the laborers' wage rates without rais- ing the farm operators' hourly wage costs. This is accomplished by setting higher guaranteed wage rates and subsidizing the increases over the pre-program minimum wage rates. The program's effects in increasing agricultural production and employment are explained in terms of the relative cheapening of farm labor as the costs of other production inputs continued to rise over Thomas M. Dickey time. A two equation model is used to estimate the net effects of the program on agricultural production and employment. The Index of Agricultural Production equation is an adaptive expec- tations model that reflects the sequential nature of farm decision making. The farm decision making process is sequential because of the biological processes involved and because the process of investing or disinvesting in the fixed assets of agricultural production places limits on the extent to which a farm operation may be changed in the short run in response to changing price relationships. Thus, decisions concerning fixed assets, and therefore annual operating levels, must be based on the farmers' expectations of future price relationships. The Employment equation includes the Production Index as an inde- pendent variable (since the adaptive expectations model did not give results consistent with theoretical expectations) as a means of reflect- ing the mutual dependence of the production and employment decisions of the farm operator on his farm's profit function. A single Employment equation is used instead of a labor demand and supply model because the employment of farm laborers is shown to be demand determined as a result of the minimum wage policies in Puerto Rico. In order to exclude any effects of errors in specifying the statis- tical model, the net effects of the program are estimated by taking the difference between the model's predictions of the with and without pro- gram values of production and employment. The with program predictions are estimated by using the observed values of the independent variables (the "fitted" values). The without program predictions are estimated by substituting an assumed behavior of the wage rate variable in place of the observed wage rate data. It was assumed that the wage rate Thomas M. Dickey variable would have continued to increase at a rate equal to its average annual rate of increase during the years prior to the start of the program. The predicted without program values for the production index and employment are then obtained by adding the net effect of the program to the observed with program values in each case. The net effect on the Production Index is converted into nominal dollar values for comparing with program costs. The program's effects on both production and employ- ment were found to be positive, even though both of these indicators continued to decline. Although the program did not fulfill its stated purpose of reviving the agricultural sector, it did slow the sector's decline. TABLE OF CONTENTS Chapter I. INTRODUCTION ...................... Background and Problem Setting ........... Research Objectives ................. Data Sources .................... Analytical Procedures and Organization of the Dissertation ................. II. THE ECONOMY OF PUERTO RICO ............... Trends in the Island Economy ............ Relationship with the United States ......... Governmental DevelOpment Efforts .......... Labor Output Trends ................. Minimum Wage Legislation in Puerto Rico ....... Summary ....................... III. THE AGRICULTURAL ECONOMY ................ Production History ................. Product and Factor Price Determination ....... Product and Factor Price Trends ........... Labor Market Conditions ............... Trends in Agricultural Employment ........ Demand Determined Employment ........... Perceived Labor Shortages ............ Seasonal Responsiveness of the Farm Labor Force . Summary of Conditions in the Agricultural Economy . . IV. THE WAGE RATE SUBSIDY PROGRAM IN PUERTO RICO ...... Background of the Program .............. Operation and Procedures of the Program ....... Wage and Subsidy Rates under the Program ...... Program Costs and Workers Benefitted ........ V. A FRAMEWORK FOR THE ANALYSIS OF THE PROGRAM'S EFFECTS Introduction .................... Effects in the Short Run under Competitive Conditions .................... Assumptions .................. Production Theory ............... The Demand for Labor .............. Theoretical Impacts of the Program in Short Run .................. Effects in the Intermediate Run under Competitive Conditions .............. ii Chapter V. (Continued) Review of the Agricultural Sector, 1961 to 1969 ................. The Farm Decision Making Setting ........ Theoretical Effects of the Program in the Intermediate Run ............... VI. DEVELOPMENT OF THE MODEL: METHODOLOGY AND STATISTICAL ESTIMATION ............... Introduction .................... The Available Data ................. The Index of Agricultural Production (AGPROD) and the Index of Product Prices (AGPRCE) Employment of Agricultural Laborers (AGEMP) and the Na e Rate Variable (WAGE) ....... Feed Prices IFEED) and Fertilizer Prices (FERT) . The Index of Production Equation .......... Methodological Considerations . . ........ The Adaptive Expectations Model ......... Statistical Estimation of the Production Index Equation ................ The Employment Equation .............. Methodological Considerations .......... Statistical Estimation of the Employment Equation ................... Summary ...................... VII. APPLICATION OF THE MODEL: ESTIMATES OF THE PROGRAM'S EFFECTS ON AGRICULTURAL PRODUCTION AND EMPLOYMENT . . Introduction .................... The Results of the Application of the Model v111.' SUMMARY AND CONCLUSIONS ................ Summary ...................... The Index of Production Equation ........ The Employment Equations ............ Estimation of the Program's Effects ........ Conclusions . . .' ................. Areas for Future Research ............. APPENDICES APPENDIX A. INCOME MAINTENANCE PROGRAMS AND THE WAGE RATE SUBSIDY .................. A Review of Income Transfer Programs ........ The Negative Income Tax .............. The Wage Rate Subsidy ............... Lessons from the Puerto Rican Experience ...... Page 55 55 58 60 60 60 60 61 62 62 62 63 67 68 68 72 73 76 76 81 88 88 92 93 100 102 103 103 106 106 107 Page APPENDIX A. (Continued) Comparison of the NIT and MRS Effects ........ 109 Outline of an Alternative Income Assistance Plan . . 112 Summary of the Plan's Advantages ......... 115 Guidelines for Implementation .......... 115 APPENDIX B. TIME SERIES DATA ................. 117 APPENDIX C. SPANISH VERSION OF THE STATEMENT OF LEGISLATIVE INTENT IN LAW NUMBER 142 OF JUNE 29, 1969 . . . . 118 APPENDIX D. COMPUTATION OF FITTED VALUES USING THE PRODUCTION INDEX EQUATION ............ 119 BIBLIOGRAPHY ......................... 121 iv Table 1. Economic Growth of Puerto Rico .............. 2. Average Annual Rates of Growth in the Puerto Rican 10. 11. 12. 13. 14. 15. LIST OF TABLES Economic Indicators ................... . Net Income by Industrial Origin, Puerto Rico, Selected Years ...................... . Employment and Real Gross Product Per Worker in the Agri- cultural and Non-Agricultural Sectors of Puerto Rico, 1961 to 1975 ....................... . Product and Factor Price Growth Rates in Puerto Rican Agriculture, Selected Periods, 1961-1974 ......... . Employment in Puerto Rican Agriculture, 1964 to 1975, in Thousands ....................... . Changes in Employment of Farm Laborers in Calendar Year 1968 and the Origins and Destinations of the Net Movements ........................ Monthly Changes in Total and Farm Laborer Labor Forces in Calender Year 1968 .................. . Puerto Rico Minimum Wage Rates and the Wage Rate Subsidy Program's Guaranteed Wage Rates for General Laborers Not Covered by the Federal Minimum Wage Rates, 1969 to 1975 Direct Treasury Program Costs .............. Selected Program Participation Data - Sugar Cane Portion (Law 141 of 1969) ................ Selected Program Participation Data - Non-Sugar Cane Portion (Laws 142 of 1969 and 20 of 1972) ........ Subsidy Payments per Ton of Cane, by Method of Harvest and Cultivation Status, 1972/1973 Crop .......... Effects of the Wage Rate Subsidy Program on Agricultural Production in Puerto Rico, 1970 to 1974, Values in Real and Nominal Terms and Direct Program Costs ........ Effects of the Wage Rate Subsidy Program on the Employ- ment of Farm Laborers in Puerto Rico, 1970 to 1974, Equation 6.19 Using the Current Production Index ..... V Page 15 27 28 33 34 45 47 48 49 80 82 83 16. 17. 18. 19. A.1. A.2. A.3. A.4. Effects of the Wage Rate Subsidy Program on the Employ- ment of Farm Laborers in Puerto Rico, 1970 to 1974, Equation 6.20 Using the Lagged Production Index .. .,. . . Estimated Effects of the Program, Assumption #1 and Employment Equation 6.19 ................. Estimated Program Costs for the Hypothetical Case in which the Subsidy is Increased in Order to Stabilize the Production Index at the 1970 Level .......... Selected Program Evaluation Indicators .......... Weekly Benefit Levels and Tax Rates for Selected Program Parameters ........................ Time Series Data Used in the Estimation of the Model . . . Other Time Series Data: Farm Laborer Labor Force; Median Weekly Earnings, Laborers, Except Farm; and Agricultural Production Index, 1951 to 1959 .............. Selected Intermediate Computation Results in the Pre- diction of the Production Index for the Without Program Case ........................... vi Page 84 96 98 99 114 117 117 120 LIST OF FIGURES Figure Page 1. Real Gross Product Per Worker in Puerto Rico, 1961 to 1975 ............................ 14 2. Median Weekly Wages for Non—Farm and Farm Laborers, Fiscal Years 1961 to 1974 ................. 17 3. Index of Physical Production of Agriculture in Puerto Rico, 1951 to 1975 ..................... 21 4. Indices of Agricultural Production by Subsectors and Products, 1963, 1966 and 1975 ............... 228.23 5. Agricultural Net Income in Puerto Rico, 1961 to 1974 . . . . 25 6. Labor Force and Employment of Farm Laborers in Puerto Rico, 1961 to 1974 ..................... 28 7. Agricultural Employment in Puerto Rico, 1963 to 1974, by Month .......................... 3O 8. Effect on Employment of the Relationship between the Equilibrium and Minimum Wage Rates ............. 41 9. Diagrammatic Representation of Production Decision ..... 53 10. The Index of Physical Production of Agriculture in Puerto Rico, Observed Values (1961 to 1975) and Fitted Values (1963 to 1975) ....................... 69 11. Employment of Farm Laborers and Foremen, 1961 to 1974, Observed Values and Fitted Values ............. 74 12. The Index of Physical Production in Puerto Rico, 1961 to 1974 and the Predicted "Without Program" Index, 1970 to 1974 .......................... 86 13. Employment of Farm Laborers in Puerto Rico, 1961 to 1974 and the Predicted "Without Program" Employment, 1970 to 1974 .......................... 86 vii CHAPTER I INTRODUCTION Background and Problem Setting The agricultural sector was the dominant sector in the Puerto Rican economy until 1955 when the industrial development programs, begun in the 1940s, finally pushed the manufacturing sector into the forefront. Agricultural production activities were concentrated in producing sugar cane, coffee and tobacco for export to the United States and the world markets. These labor intensive crops came under serious economic pres- sure when the rapidly developing industrial and commercial sectors began to offer higher wage employment and when large numbers of Puerto Rican began migrating to the United States in search of higher wage employment there. Even though the wage rates in Puerto Rico were still lower than in the United States, the island's agricultural producers were unable to offset the rising wages, either through mechanization or other means of increasing productivity, to the extent required to re- main competitive in the United States market. The requirement that they compete in the United States market, for both their exports and their foodstuffs, is due to the free trade provisions of the Common- wealth status. Sugar production began declining sharply in 1953 and although aggregate agricultural production continued to increase tem— porarily, it too began declining rapidly in 1964. The only major sub- sector that continued to increase its production was the animal products subsector, due to its relatively protected local market and its lesser dependence on labor. By the late 19605, official concern about the conditions in the agricultural economy were reaching high levels in the government. In 1967 and 1968, growth rates of the real gross product of Puerto Rico were lower than those of previous years and the efforts to find ways to return to a more rapid rate of growth naturally included a serious look at the agricultural sector. It was believed that the sector possessed unused resources, especially land, that could be used to increase the island's gross product and, since agriculture was relatively labor in- tensive, a revival of the agricultural sector could help alleviate the severe unemployment problem. The role of labor in the decline of the agricultural sector was not perceived to be a problem of the high cost of labor. It was, instead, perceived to be a problem of a shortage of laborers willing to work at the current minimum wage rates, since at that time the farm laborer wage rates were about 40% of the non-fanm laborer wage rates. The combination of the perceived shortage of laborers and the high rates of unemployment on the island naturally led to the idea of a program that would raise the wage rates actually received by the labor- ers, while at the same time avoiding any increase in the wage rates paid by the farm operators through the use of a government subsidy. Thus, in 1969, the Puerto Rican government, through its Department of Agriculture, instituted the "Guaranteed Income for Agricultural Workers"l program. Since the program is, in fact, a wage rate subsidy 1Translated from Ingresos Garantizados a Trabajadores Agricolas. rather than an income guarantee program, it will be referred to in this report as a wage rate subsidy program. As of 1975, when this research was begun, the program had not been evaluated. Research Objectives The objectives of this research are: 1) to describe the economic conditions in Puerto Rico that led to the wage rate subsidy program and that constitute the environment within which the program operates, 2) using existing economic data, to estimate the program's effects on the two principal goals of the program--increasing agricultural pro- duction and employment of farm labor, and 3) to consider the implications of the analysis for the agricultural sector policy of the Puerto Rican government. This research was limited to these three objectives due to the lack of the financial resources that would have been required to obtain data directly from the farm operators and laborers. Thus, questions concern- ing the responses to the program by individual farm operators and laborers could not be answered. Data Sources The Puerto Rico Planning Board prepares, and publishes on a regular basis, the general macroeconomic and selected sectoral measures of economic activity. The Bureau of Labor Statistics of the Puerto Rico Department of Labor conducts a monthly survey similar to the Current Population Survey of the United States Department of Labor and publishes estimates of employment, unemployment and median weekly wages in a variety of regular publications. The Office of Agricultural Statistics of the Puerto Rico Department of Agriculture conducts semi-annual sur- veys of agricultural production and prices received by farmers and also compiles estimates of local consumption of foodstuffs. These data are published annually in the "Facts and Figures on Puerto Rican Agricul— ture.“ In addition to these published sources, this author was able to obtain additional data from the files of the Bureau of Labor Statistics, the Office of Agricultural Statistics and the Wage Rate Subsidy Program Office. Analytical Procedures and Organization of the Dissertation The history of the recent economic growth of the Puerto Rican econ- omy is examined in Chapter II in order to determine how the structural changes have affected the local markets within which the agricultural sector operates. This analysis reveals the significant change in the conditions of the labor market that has accompanied the growth of the industrial sector. The agricultural sector is examined in Chapter III in order to document the changes that have occurred in recent history, to determine the relative influence that the sector has had in its principal product and factor markets, and to isolate the causal relationships that appear to have been the principal determinants in the decline of the agricul- tural sector. This analysis reveals the primary influence of the local market and the general inability of the sector to adjust to the changing labor market. The Wage Rate Subsidy Program is examined in Chapter IV to document the reasons for which it was developed, the rules under which it is implemented and the direct costs which it represented for the Puerto Rican government for the first five years of operation. Chapter V presents a theoretical analysis which is used to estab- lish expectations for the program's effects on production and employment. Static production theory is used in showing that no effects should be observed in the short run. A dynamic analytical framework is developed to be used as the basis for the empirical model. This theoretical con- struct is used to establish the expectation that both production and employment should be smaller without the program than was observed with the program. Equations for the prediction of production and employment are developed and their coefficients are estimated using multiple linear regression techniques in Chapter VI. The production equation requires the use of the Cochrane-Orcutt procedure due to the use of the lagged dependent variable as an explanatory variable. The equations of the model are then used to predict what production and employment would have been if the farm operators' wage costs had continued to increase at an assumed rate of growth in the absence of the program. These predictions are presented in Chapter VII. The sen- sitivity of the assumed rate of growth is also tested. The equations are also used to predict production and employment using the observed data and the difference between the two sets of predictions is used as the estimate of the effect of the program in order to eliminate the model's own error from the results. Chapter VIII summarizes the research and presents the conclusions drawn from the analysis. In recognition of the interest that has been shown in the United States in the use of a general wage rate subsidy program as an income maintenance program, Appendix A provides a brief review of the history of the negative income tax and wage rate subsidy concepts, a few lessons from the Puerto Rican experience, a theoretical comparison of the effects of the two programs, and a proposal for a program which combines the two concepts. CHAPTER II THE ECONOMY OF PUERTO RICO Trends in the Island Economy The economic development of Puerto Rico in recent decades is a well-documented success story. Real per capita personal income has in- creased from $213 in 1940 to $1,106 in 1974--an average annual growth rate of 4.96%. Tables 1 and 2 on the following pages illustrate the results of this development with selected economic indicators. As part of the development process, the economic structure of the island has undergone very significant changes. The manufacturing sector had replaced agriculture by 1960 as the most important sector, on the basis of Net Income. The trade sector has maintained a relatively con- stant share of the economy over the years, maintaining its second place ranking until a few years ago when the growing government sector sur- passed it. As will be shown in greater detail later, the agricultural sector has declined in both relative and absolute terms since the early 19605, leaving it last among the major sectors. Table 3 provides the breakdown by sector of Net Income by Industrial Origin with the cor- responding percentages and rankings. Relationship with the United States This rapid rate of industrialization and the accompanying growth is attributable primarily to the many advantages that Puerto Rico has .mmuwvcmag< quwpmwpmum .enma .Loumccmnow Fm oquocoum mscoecH .cmwumuwmwcmpm mu macaw .m.m "mugzom —mowpmwpmpm .wmmfi . xme.m &m¢.fi xom.o Nmm.H cowumpaaoa xmw.~ flm¢.m &mm.¢ woo.m wEoucH pmcomgwa mpwamu me Puma Rum.¢ xmm.m fiom.m xmw.m puauogm mmogu appamu gm; Puma me.oH ano.ofi xfim.m xma.o_ nuanced mmogw Puma cumfiaomaa ommfiuooma oomfiuommfi ommfluovma .mcoumomucH uwsocouu :muwm opcmaa mg“ cw cuzocw to mmpmm szcc< mmmem>< .N mFQmH .mwuwccmaq< eve : um EcoecH .cmwomuwewcmfia mu macaw .m.m "mucaom Nam.m ofik.~ Nem.m oo~.N mmm.H .maoee cowua_=aoa eoH.H mmm omm hem new a asou=H.Pa=0mcma aueaau Laa Foam mmH.H 44¢.“ omo mom mom a puzooca mmocw ”Began Lea Puma m.¢-.m m.mea.m m.m¢m.H o.MHm .m.m- .Fppz» aeouce pmz Facweoz m.omm.m a.mmm.~ ~.m~¢.~ “.mkm m.mm¢ .przm peacoca ametu Puma H.mom.m m.H~o.¢ 4.0N8.H m.¢m~ ~.om~ ._F32w postage macaw Pacescz ekmfi _onH come ommfl cemfi wave: .ouwm oucmza mo suzocw oesocouu .H anwk .Nmuwccmaa< Pwuwpmwumpm .cNmH .Nocmccmnou Pm oomsocoum macomcH .cowumuwmwcmpa .NLNNNON No NcoNNNNN NN NNNN>N Nu NpcaN .N.N "mugsom NNNN- NNN- NH- HN NN NNNoz No NNNN N N.NN NNNN N N.NN NNN N N.NN NNN N N.NH NN N N.N NN NewscaosoN N N.NN NNN N N.NN NNN N N.N NNN N N.N «N N N.N NN NNUN>LNN . . . . . mpmumm Puma N N NH NNN N N NH NNN N N NH NNN N N N NN N N NH NN NNN NocchN N N.NH NNoN N N.NN NNN N NqNN NNN N N.NH NNN N N.NN NN NNNLN . . . . . NNNNNNNNN NNNNNN N N N NNN N N N NNN N N N NNN N N.N NN N N N NH N coNNNNNONNNNLN . . . . . coNNoacpmcoN N N N NNN N N N NNN N N N NN N N N NN N N N N NNN Nch_z H N.NN NNNN N N.NN NNN H N.NN NNN N N.NN NN N N.NH NN NNNLNNNNNNNNZ N N.N NNN N N.N NNN N N.NN NNN H N.NN NNN N N.HN NN NLNNNNUNLN< NNNoN NNNON NNNON NNNNN NNNON Nana No a N=NN> xcmm mo N maNN> xcmm No N mapw> xcmm No N mspm> xcmm No a wapm> N N N N o N N H N N N N o N N N c N N N .NNNN» umpumrmm .oupm ougmaa .cwmpco NNNcpmaucH Na msoucH umz .m 2an 10 enjoyed under its relationship with the United States. The major features of this relationship include: 1) free movement of goods, ser- vices and capital between the island and the mainland, 2) exemption of Puerto Rico from U.S. Federal taxation, 3) a common currency, and 4) United States citizenship for Puerto Ricans. In addition, the Puerto Rican government has carried out an aggressive industrial promotion policy which has included a number of economic and tax incentives (see following section). Several analyses of Puerto Rican economic growth have suggested that the public works, other anti-depression programs and the Second World War all played significant roles in the pre- Commonwealth stages of Puerto Rico's economic development (Reynolds and Gregory, p. 5, and Friedlander, p. 158). The advantages of this relationship, now called Commonwealth status, have made it possible for Puerto Rico to attract outside investment, giving it an advantage over both foreign countries and the fifty states in the attraction of industrial investment. With no trade barriers, Puerto Rican production can enter the large United States market easily and, of special interest to potential investors, United States equipment and raw materials enter Puerto Rico freely (Riley, p. 41), although Puerto Rico does impose excise taxes on most goods which are imported. The use of dollars as a common currency provides advantages over foreign countries due to the free movement of currency and freedom from problems of depreciation of currency. Exemption from federal taxes is advanta- geous in their competition with the fifty states to attract capital and plants since Puerto Rico can offer a tax-free environment, as they do for limited periods. 11 The free migration of Puerto Ricans to the mainland due to both the cheap air fares and the common citizenship has played an important role in Puerto Rico's growth and development. Although the net migratory flow in recent years has been returning people to the island, the massive migrations of the 19405 and 19505 had numerous effects on economic growth. Studying these effects in the early 19605, Stanley Friedlander found that the emigration: 1) had a significant impact on the rate of growth of per capita income, 2) had a significant effect on the age structure and thus on future growth of the population, 3) contained a considerably higher proportion of labor force members to non-members than the pro- portion of labor force members to non-members on the island, 4) in combination with the increased investment, probably contributed to the reduction of disguised unemployment on the island, 5) contributed to the upgrading of the quality of the labor force by reducing the proportion of unskilled laborers, and 6) contributed indirectly to raising the educational attainment of the population by easing the pressure on the school system (Friedlander, pp. 157-164). Governmental Development Efforts Although Puerto Rico has only had Commonwealth status since 1952, active government promotion of industrialization began in 1942 with the creation of the Puerto Rico Planning Board, the Government Development Bank for Puerto Rico (GDB) and the Puerto Rico Industrial Development Company (PRIDCO). Although originally PRIDCO created and operated manu- facturing enterprises, it now provides a wide range of services to potential investors and newly established plants. PRIDCO has constructed over 962 factory buildings which may be sold or leased to firms, with 12 lease rates depending on geographic location. During a plant's start-up period, Cash Incentive Grants are available to help defray the cost of training supervisory personnel, building rental, transportation of equipment and machinery to the plant site, feasibility studies and others. For plants requiring a work force of 300 or more workers, the Pre-Employment Training Program will pay the entire cost of recruiting and training the workers so that they will be available on the day the plant opens. PRIDCO may also provide secondary financial assistance through loans or minority equity investments to financially sound ven- tures when private financing is not otherwise available. The Government Development Bank may provide loans through mortgages on buildings or chattel mortgages on machinery and eqUipment if a firm is unable to secure financing on reasonable terms. Firms may also receive financial assistance through loans from the U.S. Federal Economic Development Administration for both fixed and working capital. PRIDCO also operates a 32-acre free trade zone near Mayaguez where industrial operations may avoid import duties until the products are sent to the United States or -Puerto Rican markets. Since Puerto Rico is exempt from Federal taxes, except specified ones such as Social Security taxes, the government can use exemption from local taxes as a means of inducing investors to establish plants in Puerto Rico. Through Section 931 of the U.S. Internal Revenue Code, Puerto Rican subsidiaries of U.S. firms can eventually repatriate their earnings without incurring Federal taxes on those profits. Since 1954, Puerto Rico has exempted qualifying manufacturing and some non- manufacturing firms from both Commonwealth and local taxes on corporate earnings, property and licenses. The program permits 100% exemption 13 for periods of 10, 15, 25, or 30 years depending on the location of the plant. Through an arrangement for paying partial taxes during the initial period, firms can receive an extended period of partial exemp- tion. The varying periods of exemption were instituted to induce a more even distribution of plants around the island since a high concen- tration of plants was building up in the main metropolitan areas. In addition, there are tax exemptions for individual investors on dividends and income from machinery leased to tax-exempt firms. Even without the local exemptions, a tax advantage exists since the Puerto Rican corporate taxes are slightly less than the Federal taxes--ranging from 22% to 45% (Pic6, pp. 290-308, and The Witcom Group, pp. 35-48). Labor Output Trends The transformation into a primarily industrial economy, due in part to the migration and the outside investment, has been accompanied by significant increases in labor productivity. Figure 1 and Table 4 Show these increases for the 1961 to 1975 period. Data for prior years con- sistent with these data are unavailable due to changes in statistical reporting. The labor productivity figures are computed by dividing Real Gross Product (in constant 1954 dollars) by the number of employed workers. Over this fourteen year period, total economy labor producti- vity had a 3.46% average annual growth rate, while agricultural and non-agricultural labor productivity had growth rates of 5.56% and 2.26% per year, respectively, but in absolute values, non-agricultural labor productivity remained much higher than agricultural labor productivity. 14 .: o_nmh “ouL30m NN .N on m mm um mm mm :o no Nm J _ a q _ _ - q i coo.— Louuom .mesu_:u_em< 1 ooo.~ 1 ooo.m i 000. >zozoum 4.... 1 mm low 1 mm lom 1 mm cop mum—1mm? $9 $09 mom. unem— NNNN a No-08 >uv ii? 10: 1m: LON— .LNN. xonc. .mNmH cu HmmH .ouNm ougmao :_ mezupaowem< No eowuuzwoca Nmuwmxsa No xmucm .m mesmNN 22 L0u00mnam muuzuon _mE_c< Auv Leuuomnam maoLu .ma_uc_ca any mucuuomaam team: Amy mm mm mm mm we no mu mm mm No. .2 6. .ON ouomnOH ION .GN -OM .0m .0m L .o . o: Nauuom : w o: Lom .om maceu .Gm .ma_ucmca .om .om .ow Jon on I. .65 .8 8 LowooN/ // 6N Lom om menu—:o_em< om GDP ...00— ....................... COP o—F No—_ 0.— ON— .o~_ a cu— muunvoLm . cm— Jom— .mE_c< 6m. Icoxu_:o .o:— .o:_ .o:_ .8. N92 . om. xovc. x02... x09... .mNmH new mmmfi .mmaH .Nuozcogm new Ncouowmaam No cowuuauoga Pocaupzuwcm< mo NmuNucH .e mesa?“ 23 mo_nmuomo> >zoLmum va 1mm 4:: mOOUMHOQ \ “003% mcowsmoo 111411111 mLoN:ME mo_nmuomo> >socmum 141 mc_mucm—m .NeoeeNeeoNV NNNN Nee NNNH .NNNN .Neoeeoea Nee Neoeoooeem Na eoNNoeeoea «mmcmcmm a [III mm oN oN .0m No: .0m 100 Jon 1cm A 1 O O N? M P P Lom— xovc. mu_:cu Nov ecuuomanm Naocu Losuo Avv NN NN NN NN NN .N 111 . . a .— .o— .N JON om .0m _.: .o: .N .om om Nw_eoooNo> Loeeo Low 0 aamocwa .2 ON INN :Mu_:cm e\\mu:coUOu om mu_:cu .ow 11HV11 11 1.11.1 cm 1.11.! .00 NcoLu_u 11.I.I.1. “cucumnsm 111 1. ,oo. Momcmeo H3330mm; >coem~m oo— o—— .o—N w ow— .oN_ Nooooo>< NNN ,oNN .3— .0:— om— -Omp xouc_ xovc_ chzppaowcm< No qu_o:H .e ocamwm 24 permit the examination of underlying trends. Secondly, the Gross Product statistics include fOrestry and fishing and are deflated by an index based on selected products rather than on all products, as is the case in the production index. This analysis was not based on the use of Net Income statistics because their use would lead to a misinterpretation of the actual performance of the sector. This is due to the fact that subsidies to the sector add directly into Net Income and to the extreme- ly large increases in subsidies that the sector has received since 1968 and 1970. In 1967, the agricultural sector received $5.455 million in recorded subsidies. By 1968, they had almost doubled to $10.009 million and have since risen to $50.392 million in 1975 (including the wage rate subsidy). Thus, while physical production has continued to decline since 1963, the figures on Real Net Income level off beginning in 1970. However, if the deflated value of subsidies is subtracted from Real Net Income, the resultant figures of Real Net Income (from Production) con- tinue to decline at more or less the same rate since 1963 (see Figure 5). The sugar cane subsector is very labor intensive and the natural topographic limits to its mechanization have prevented it from keeping up with advancing technologies in competing world regions, except on the best level lands (Ballinger, p. 22). Coffee, on the other hand, while also very labor intensive,has not experienced much mechanization and has had to compete with low-wage producers in South America and Africa. Much of the foodstuffs are produced on small farms since much of the best land belongs to the larger sugar producing farms. 25 .oL:u_:u_cm< mo ucoEucmqoo .m.m .mo_um_umum .meau i_:o_am< mo OUNNNO .mo_p_mn:m "uLmOm mc_ccm_¢ .¢.m .oEOoc. uoz “mooL30m 5N NN N.N N 6N NN mm N.N NN NN i NN N1... 6 t .mu Vom ecuuom .mn _mc:u_:o_cm< .No_e_NN=N tune—moo one; .oo_ oeooc_ uoz _mom .mNN .omp Louoom .NN_ _me:u_:o_cm< .oeouc_ uoz _mom .oo~ Lauuom .mNN .mc:u_:o_cm< .OEOUc_ uoz .mc_eo= .omm .mNN nem—_oo co___mz t—Kma OH Hmma Nouwm OHszm vi. 9200.; Hmz FMLaupzqum’.‘ .m mesmwn. 26 Product and Factor Price Determination Puerto Rican farm operators are generally price-takers in both their product and input markets. On the input side, most of their non- labor inputs, including fuel, fertilizers, feeds and machinery, are imported. The wage rates for agricultural labor, both skilled and un- skilled, are set by either the Administrator of the Wage and Hours Division of the U.S. Department of Labor or the Puerto Rico Minimum Wage Board. In the case of sugar cane workers, the Minimum Wage Board adjusts the minimum wage rates every 28 days based on the average New York raw sugar price (U.S. Department of Labor, p. 17). On the product side, the prices for sugar cane, coffee and tobacco are determined in the respective world markets since the Puerto Rican share of these markets is quite small. With the exception of fresh milk and fruit, much of the local consumption of foodstuffs is imported from the U.S. mainland. Data for the 1961-1974 period from the Consump- tion Section of the Puerto Rico Office of Agricultural Statistics show that imports of starchy vegetables were between 16% and 36% of consump- tion, imports of eggs were between 34% and 55% of consumption, imports of leafy vegetables were between 30% and 67% of consumption and imports of legumes were between 79% and 88% of consumption. In short, any sig- nificant change in the production of Puerto Rican agriculture will have very little, if any, effect on market prices due to the relatively large quantities of products which are exported or imported. Product and Factor Price Trends Table 5 reports the average annual growth rates for the index 0f agricultural product prices and for the available factor prices for the 27 Table 5. Product and Factor Price Growth Rates in Puerto Rican Agriculture, Selected Periods, 1961-1974. 1961-1963 5.00% 3.11% 4.60% 1.34% 1963-1966 -3.07% 2.07% 3.36% 6.36% .1966-1969 2.97% 0.31% 3.28% 8.75% 1969-1973 4.49% 3.23% 3.65% * 1973-1974 75.26% 26.75% 21.57% * * Wage Rates were held constant by the Wage Rate Subsidy Program. periods in which the shifts in production trends are most visible. The actual prices are listed in Appendix B. This table shows that the rates of increase in wage rates paid by the farm operators were considerably greater during the two periods of sharp decline in production (1963- 1966 and 1966-1969) than were the rates of increase in feed and fertil- izer prices. Labor Market Conditions Trends in Agricultural Employment Table 6 documents the significant loss in employment in agriculture during the period of decline in agricultural production. Sixty-nine thousand jobs were lost between 1964 and 1974. Almost 50% of this decline has come in the sugar cane subsector and 30% has come in the coffee and tobacco subsectors. Thus, almost 80% of the decline in employment has been in the export crops subsectors. As would be ex- pected, most of the lost jobs have been those of farm laborers. This category declined by 50,000, or 72% of the 69,000. Figure 6 illustrates 28 Table 6. Employment in Puerto Rican Agriculture, 1964 to 1975, in Thousands. F¢zgal 523:3; iggir Coffee Tobacco Livestock Pineapple 2;??? 1964 122 43 21 11 4 2 39 1965 108 32 21 12 4 * 36 1966 99 31 18 7 4 * 36 1967 88 28 18 4 4 * 32 1968 85 22 19 4 5 * 31 1969 78 21 16 2 4 * 33 1970 68 18 12 * 4 * 32 1971 61 13 12 * 4 * 30 1972 56 11 9 2 4 * 29 1973 50 11 9 2 3 * 23 1974 53 9 11 * 3 * 28 1975 50 * Less than 2, estimates unreliable. Total excludes Forestry and Fisheries. P.R. Department of Labor, revised data. Figure 6. Labor Force and Employment of Farm Laborers in Puerto Rico, 1961 to 1974. Thousand Persons 110 100 90 80 70 Labor Force of Farm Laborers Employment of L 60 Farm Laborers SO 40 30 20 10 FY6162 63 6465 6667 686970 71 727374 Source: Puerto Rico Bureau of Labor Statistics 29 the decline in employment and in the labor force of agricultural labor- ers. These data, however, are the annual averages and do not reflect the high_degree of seasonality of employment, primarily in the export crops. The seasonal increases correspond directly to the harvest periods of both sugar cane and coffee, and are illustrated in Figure 7. Although the average value of output per worker for agriculture increased rapidly between 1961 and 1975 (see Figure 1, page 14, and Table 4, page 15), it seems probable that much of this increase simply reflects the large decline in the employment of farm laborers, parti- cularly for the unskilled export crop harvest laborers. An additional factor would be the expectation that much of the decline in production would have come from the termination of the least efficient farms, thereby also tending to raise the average output per worker. Demand Determined Employment Due to the operation of the minimum wage laws in Puerto Rico, employment of agricultural laborers appears to depend solely on the demand for labor of the farm operators. In studying the 1952-1962 period, Reynolds and Gregory found that increases in wage rates for farm laborers were " . . . probably attributable in good measure to the island's minimum wage law . . ." since even at peak employment seasons agricultural unemployment was sufficiently high to avoid any serious pressure on wages to increase due to labor shortages (Reynolds and Gregory, p. 62). The direct implication is that the minimum wage rates were above the equilibrium wage rates, and the quantity of labor being employed was the quantity which the farm operators could profitably use. 30 «UNumNHmum Loam; mo amocam 00.x Oucoam NouL=Om _:NNN ___ NNNN NNNN NNNN .oNNN _ NNNN _ NNNN_ NNNN _ NNNN _ NNNN_ NNNN _ NNNN_ _ _ _ . . ._ —1t.!?. .. .afi,e..4N.4 N3. or. . .4. on. N . .1 ..1._.1. .14 .. . . . .. V1 0...}. . ”an... ....r. N. m .N ’ J 11 ‘- 411;“: .e... I)“. ..N I I . b v ooNNoN .1. 3 menu cmmsm [loo noNNocm_m ucm >Numoeou .ocau.:u_em< 110w loo— N 1.oNN f.oeN noo>ONQEm unmanagh .eeeoz Na .NNNN oN NNNN .ooNN oNeoea eN NeoENoNNeN NNL=NN=oNLN< .N oeaNNN 31 More recent data tend to confirm this finding. Unemployment rates for farm laborers averaged 19.2% between 1961 and 1969 (when the subsidy program began). The annual average unemployment rate was also rela- tively stable during this period. The lowest monthly unemployment rate was 7.5% in March of 1969, one of the sugar cane harvest months. Even if one considers that the non-farm labor unemployment rate averaged 21.9%, the higher wage rates of this category (see Figure 2, page 17) should more than compensate for the lower probability of finding employ- ment as a non-farm laborer. Perceived Labor Shortages The finding that unemployment rates were high before the enactment of the wage rate subsidy program presents a direct conflict with the complaints of the farm operators that they were unable to obtain ade- quate amounts of labor. The only explanation of this paradox is that the farm operators, faced with higher operating costs for all of their inputs, could no longer afford to hire as much labor as they had pre- viously employed when labor was relatively cheaper. Since the prices paid for their non-labor inputs are determined outside of Puerto Rico, their only feasible alternative was to influence the price of labor by resorting to the political arena. Their appeal via the Department of Agriculture is logical in that the Minimum Wage Board is under the Department of Labor and both of these are concerned with the welfare of the labor groups. The concept of the reservation wage rate does not provide a logi- cal explanation of this paradox. The reservation wage rate is the lowest wage rate which someone will accept for a specific task. For 32 the menial tasks of the farm laborer (cutting sugar cane, picking coffee, etc.), it seems quite plausible that the reservation wage rate would be rising, given the rising affluence and standards of living in Puerto Rico. However, if the reservation wage rate for most farm laborers were above the minimum wage rates, it would be irrational for the laborers to remain unemployed in the agricultural labor force when they could seek employment in the other occupational and industrial categories. The large movements into and out of the farm labor force, which are documented in the next section, suggest that the farm labor- ers do remove themselves from the farm labor force rather than remain unemployed in this category, but that they also return to the farm labor force when they expect to find employment during the harvest periods. Seasonal Responsiveness of the Farm Labor Force An examination of the monthly data on the farm labor force reveals large movements into and out of the farm labor force. Looking at the movements of calendar year 1968 as an example, Table 7 gives the net changes in employment of farm laborers and foremen and shows the origins of the increases in employment and the destinations of those employees that left the ranks of the employed when there were decreases. In nine of the twelve monthly changes, the change in the labor force was greater than the change in the number of unemployed persons. In seven of the twelve monthly changes, the change in the labor force was greater than the change in employment. Further analysis of this table suggests that there is a tendency for laborers to enter into or exit from the labor force in anticipation of an increase or decrease in 33 Table 7. Changes in Employment of Farm Laborers in Calendar Year 1968 and the Origins and Destinations of the Net Movements. Employment Unemplo ment Labor Force I Between the Percent Percent Months of: Change Change Of Change Change lg4§mploy Jan-Feb '688 + 17,229 - 8,805 51.1%. + 8,424 . Feb-Mar + 1,095 - 2,325 212.3% - 1,230 . Mar-Apr - 12,246 + 4,678 38.2% - 7,568 . Apr-May + 888 - 2 .736 308.1% - 1 ,848 . May-Jun - 6,235 + 1,341 21.5% - 4,894 78.9% Jun-Jul - 5,287 + 2,115 40.0% - 3.172 60.0% Jul-Aug - 6,208 - 1,591 25.6% - 7.799 125.6% ug-Sep + 495 + 2,265 457.6% + 2,760 557.6% Sap-Oct + 7,238 + 732 10.1% + 7,970 110.1% Oct-Nov - 1,106 + 394 35.6% - 712 64.4% Nov-Dec - 2,377 - 783 32.9% - 3.160 132.9% Dec-Jan '69 + 1,875 + 259 13.8% + 2,134 113.8% Numbers and Totals of Increases (6) + 28,820 (7) + 11,784 (4) + 21,288 Decreases 6) - 33,459 (5) - 16,240 (8) - 30,383 Net Changg_ - 44,639 - 54:456' - 99,095 Source: Derived from unpublished data from files of the Puerto Rico Bureau of Labor Statistics. employment (Feb-Mar, Apr-May, Aug-Sep, and Dec-Jan'69), and, in those months when a large decrease in employment does occur, for most of the laborers to leave the labor force rather than remain unemployed for several months (Mar-Apr, May-Jun, Jun-Jul, Jul-Aug, Oct-Nov, and Nov- Dec). A comparison of the monthly changes in the total labor force of Puerto Rico and the monthly changes in the farm laborer labor force shows, initially, a different seasonal pattern (Table 8). A further look, however, reveals that of the months in which the total labor force declined, five of the six declines in total labor force were greater than the decline in the farm laborer labor force. Thus, the net flows in those five monthly changes were from the farm laborer and 34 Table 8. Monthly Changes in Total and Fann Laborer Labor Forces in Calendar Year 1968. Between the Chagges in Labor Force Months of: Total Farm Laborers Jan-Feb '68 - 2,333 + 8,424 Feb-Mar - 2,820 - 1,230 Mar-Apr - 15,242 - 7,568 Apr-May + 1,696 - 1,848 May-Jun + 22,954 - 4,894 Jun-Jul - 9,283 - 3,172 Jul-Aug - 11,603 - 7,799 Aug-Sep + 7,089 + 2,760 Sep-Oct + 13,314 + 7,970 Oct-Nov - 3,075 - 712 Nov-Dec + 145 - 3,160 Dec-Jan '69 + 12,916 + 2,134 Totals of: Increases + 58,114 + 21,288 Decreases - 44,356 - 30,383 et Change + 13,758 - 9,095 Source: Unpublished data from files of the Puerto Rico Bureau of Labor Statistics. other occupation labor forces completely out of the total labor force. Although positive proof that the people leaving the farm laborer labor force also left the total labor force cannot be derived from the avail- able statistics, the fact that the net flows resulted as they did is suggestive of the possibility that there is a large movement into and out of the farm laborer labor force in response to the seasonal demands of agriculture. Summary of Conditions in the Agricultural Economy The current state of the Puerto Rican agricultural sector may be traced to its past history of concentration on the labor dependent 35 export crops and the rapid increases in wage rates which have accom- panied the rapid industrialization of the island economy. The concen- tration on export crOps was possible due to the island's access to the mainland United States market for foodstuffs. The vulnerability of the sector to increasing labor costs was compounded by natural topographi- cal limits to mechanization in sugar cane which might have allowed it to remain competitive in both its factor and product markets. No such technological advances were available for the coffee farms. The island's mountainous terrain makes it difficult for Puerto Rico to com- pete with mainland producers in the provision of foodstuffs to all of its population. Although it does seem to be responding in the area of livestock products, especially fresh milk and eggs, this subsector will continUe to be dependent upon imported animal feeds. The Puerto Rican farm operators find themselves to be price-takers in virtually all of the factor and product markets. Their non-labor inputs are mainly imported and their labor input prices are controlled under both the federal and the island's own minimum wage establishments. The rapid industrialization and the burgeoning population are also acquiring the prime agricultural lands which are concentrated along the coastal areas where most of the population lives. On the product side, export prices are determined in the world markets and the food- stuff prices are heavily influenced by the large imports of food from the United States. In short, Puerto Rican agriculture has been unable to compete effectively in any of its markets,with the result that production has declined sharply since 1963. Its dependence on labor makes governmental subsidization of wage rates an obvious alternative in an attempt to 36 revive the agricultural sector, or at least to soften the transition required of the farming population. The following chapter describes the wage rate subsidy program and the manner in which it was imple- mented in Puerto Rico. CHAPTER IV THE WAGE RATE SUBSIDY PROGRAM IN PUERTO RICO Background of the Program In May of 1969, the Senate of Puerto Rico passed Senate Resolution #91, which authorized a study by the Agriculture Committee, jointly , with the Labor and Cooperative Development Committee, to investigate the shortage of cane cutters in the 1968 and 1969 harvests and its pre- sumed connection with the program which places island residents in migrant labor jobs on the mainland. Eight meetings were held between May and October and much of the discussion revolved around the recon- cililation of the Association of Farmers' claim that they could not find people willing to work and the Labor Department statistics on un- employment among farm workers. The idea of bringing in thousands of laborers from a less developed country (to be paid for by the govern- ment) was discussed and later rejected. In June of 1969, the Commonwealth of Puerto Rico passed a law to create the wage rate subsidy program for hired agricultural laborers. In testimony before the Agriculture Comittee of the Puerto Rico House of Representatives, then Secretary of Agriculture, Luis Rivera Brenes, listed four basic objectives of the proposed program. These were: 1) to increase the income of the farm workers, 2) to attract and conserve a labor force, 3) to stimulate the attendance at work and to discourage absenteeism, and 4) to achieve an increase in agricultural production 37 38 (Commonwealth of Puerto Rico, Camara de Representates, p. 6). The statement of legislative intent for the program includes the following paragraphs:2 By this law, it is declared as being public policy of the Government of Puerto Rico, the guarantee for agricul- tural workers of an income greater than the one they are receiving at present, which will persuade them to continue working in the agricultural sector in order to ease the short- age of manpower encountered by this industry, without raising the costs of production. Likewise, it is expected to raise the standards of living of the workers so that they may enjoy a more rewarding experience. Therefore, the purpose of this act is to increase the income of the worker, thereby reducing his absenteeism in agricultural activities, thus increasing the agricultural production through a more intensive labor in planting, cul- tivation and harvesting. However, the economic situation of the agricultural enterprise does not permit the farmers at present to pay wages in an amount equal to or greater than the income de- sired to be guaranteed to these workers. For this reason and with the intention of aiding agri- culture to regain its prosperity and to be able to pay wages equal to or greater than the incomes guaranteed by this act, a supplement is hereby established to the income at present received by certain agricultural workers to be determined in accordance with the provisions of this act, to be defrayed by the Government of the Commonwealth of Puerto Rico, by using the farmer as intermediary so that the income supplement will reach the workers through the mechanism established by this act. The objective of increasing agricultural production hopefully (leads to or is concurrent with lower food costs (if a sufficient pro- portion of local consumption is locally produced), lower food imports, higher agricultural exports, increased income for the island economy and maximum utilization of resources. The objective of an improvement in the standard of living of the rural population will hopefully reduce 2Law 142 of the 29th of June of 1969. Translated from Spanish by the author. See Appendix C for the Spanish version. 39 the rural to urban migration of people, thereby also reducing the social tension in urban centers and the pressure for government-provided ser- vices. In addition, one would expectimproved nutrition, housing facil- ities, choice in consumption and leisure opportunities to result from the higher incomes. Nevertheless, conversations with officials of the Puerto Rican government lead to the belief that the objective of in- creasing agricultural production (by means of improving and increasing the supply of labor at then current labor costs) was at least of equal importance and probably of greater importance than the objective of increasing the social welfare of the hired farm laborers. This belief is supported by the fact that the program was designed and operated by the Agriculture Department rather than the Labor Department and by the fact that when the federal minimum wage coverage was extended in Puerto Rico, the Puerto Rican government passed Law #20 of 1972, to compensate the affected farm operators by absorbing this increase in minimum wages through the wage rate subsidy program. Operation and Procedures of the Program The program operates by setting a new and higher minimum wage rate for farm laborers and then reimbursing the farm operators for this increase in wage costs. The specific rules are that the program will reimburse to the farmer the difference between the new program minimum wage rate (hereafter referred to as the guaranteed wage rate) and the higher of either the regular minimum wage rate as set by the Puerto Rico Minimum Wage Board, or the wage rate at which the farm operator contracted the laborer. 40 As an illustration, assume that the guaranteed wage rate estab- lished by the program for dairy workers is $1.05 per hour. If a dairy farmer employs two workers, one of whom he hired at a wage rate of $0.85 per hour and the other whom he hired at the $0.70 minimum wage rate, he would be reimbursed $0.20 per hour worked by the former and $0.35 per hour worked by the latter. This system of implementation was designed to use the farm operator in the distribution of the benefits, without themselves benefiting directly. If one assumes that the regular minimum wage rage, 'mw' in Figure 8(a), is above the equilibrium wage rate, 'ew', as appears to be the case in Puerto Rico, then the farm operator is forced into passing all of the benefits of the program directly to the workers. Although the amount of labor supplied at the guaranteed wage rate increases, the farmer is already employing the amount of labor, 'Em', that he could af- ford at the regular minimum wage rate. If the regular minimum wage rate is below the equilibrium wage rate and the guaranteed wage rate is above the equilibrium wage rate, the benefits are less than the first case, but they are still passed completely to the worker--since the farmer had to have contracted the workers at the equilibrium wage rate which was above the minimum wage rate. If the guaranteed wage rate is set at a level below the equilibrium wage rate, then the program would receive no legitimate claims for benefits. At first, the program only covered the hourly wage workers, but in 1971, it was extended to piece-rate coffee pickers and in 1972, to piece- rate tobacco steamers. Law #141, which covers workers in the agricul- tural phases of the sugar industry, imposes minimum wage rates but instead of reimbursing the farm Operators on the basis of hours and wage 41 Figure 8. Effect on Employment of the Relationship between the Equilibrium and Minimum Wage Rates. (a) 'mw' above 'ew' (b) 'mw' below 'ew' Wage 9w - - ~5- ew ----- L Ee Employment E = Em Employment E rates of the individual workers, it pays an amount based on the tonnage of cane delivered to the sugar mills and which of the six methods were used to harvest the cane. In the last few years, payments to dairy farmers have been converted to the basis of delivered milk rather than hours on the payroll. To be eligible to receive the reimbursements, each farm operator must submit an annual application. This form requests the following information: 1) Name, address, farm locations and Social Security Number, 2) Employer identification numbers for the State Insurance Fund,Social Security, Unemployment Compensation and the A.S.C.S. Sugar Program, 3) Number of permanent and temporary employees, 4 Farm size, 5) Area, estimated payroll, estimated production and value for each of the farm enterprises, 6) If a union contract determines the regular wage rate, 7) Number of animals, and 8) Sources of borrowed capital. After each quarter, the farm operator applies for the reimbursement on a form which requests, for each worker: 42 1) Name and Social Security number, 2 Work done or farm enterprise, 3 Hours worked, _ 4) Regular wage rate and earnings, 5) Supplement to wage rate and total supplement (subsidy), 6) Social Security tax deducted, and Total received by the worker in the quarter. 7 Separate forms are used for reimbursement for the piece-rate workers and for the subsidies paid on the basis of production (sugar and milk). The applications are checked at the regional program offices and then sent to San Juan where the information is punched on cards. Since the pro- gram pays the Social Security and State Insurance Fund taxes on the reimbursed supplements (subsidy) and interest on the total (the farmers are presumed to have borrowed money to pay the supplements), these amounts are automatically added to the total by the computer. Regular visits are made to the farms to assist the farmers in mak- ing the applications, explain the rules and regulations and to verify the claims for reimbursement. Arrangements are made with the farm operators VNK) have received overpayments due to improper claims for reimbursement. If the amount to be returned is large, the recovery of these funds may be spread over a number of quarters by reducing future reimbursements, thereby reducing the cash flow impacts on the farm operation. A number of problems have resulted in the implementation of the program. First, many of the workers will not give their correct Social Security number to the farm operators. Some of these do not give any number at all so that the tax will not be deducted from their wages, and others will give false numbers in the expectation that other welfare or assistance benefits they might receive will be terminated or reduced. Strict enforcement of the use of Social Security numbers would, in all 43 probability, result in extensive use of false numbers. A second problem which is common is the absence of adequate farm records, especially payroll records. Many of these farmers employ only two or three workers and claim that they know how many hours each one has worked during the quarter. This makes verification of the reimbursement requests extreme- ly difficult, although the claims are normally paid. A third problem, and the one which results in the largest adjustments, results from the automatic inclusion of the Social Security and State Insurance Fund taxes with the reimbursements rather than the payment of these claims upon presentation of records by the employer showing taxes actually paid to the two systems. Due to the lax enforcement by the State Insurance Fund, many of the farmers will declare only the minimum payroll so as to comply with the requirements (cover his workers) and avoid paying the full amounts due. No effort is made to fOrce the farm operators to comply fully with the Social Security and State Insurance Fund require- ments. If an investigation shows that the farmers did not pay the proper amounts of these taxes, the overpayments under the program are returned to the program instead of requiring the farmers to comply with the other institution. A fourth problem, stated by the Acting Program Director, is the payment of supplements for workers who are receiving a wage rate above the program's minimum wage rate (the program's breakeven wage rate). This occurs when the farm operator reports a pre-supplement wage rate which is lower than the program minimum wage rate, even though he is paying a higher wage rate. The farm operator thereby requests a reim- bursement for the worker by intentionally underreporting the wage rate actually paid. Although the program was designed to provide a larger 44 supplement for lower wage rate workers, the incentives (inherent in the design) for farmers to declare the regular minimun1wage rate (of the Minimum Wage Board) as the market rate being paid the workers naturally lead to the farm operators being reimbursed the maximum amount in most cases. This case can also be seen as a simple extension of underdeclar- ing wage rates when the actual wage rate paid is below the program minimum wage rate. Both cases are illegal, but in the latter the farm operator would qualify for some amount of reimbursement anyway. No clear indication is available as to the extent of underdeclaring of pre-supplement wage rates and such situations are often overlooked due to a perceived unfairness and since it is in accord with the original intent of relieving wage costs in the agricultural sector. Wage and Subsidy Rates Under the Program The magnitude of the subsidy in terms of hourly wage rates can be seenin Table 9 for the lowest occupational categories ("other laborers') in each of the agricultural industries except sugar cane. The minimum wage rates in effect in 1969 as set by the P.R. Minimum Wage Board are listed first, followed by the Wage Rate Subsidy Program's guaran- teed wage rates (by period of applicability) and the percentage increase that the guaranteed wage rates represented over the minimum wage rates. The Minimum Wage Board has not revised its minimum wage rates since the inception of the Program. In 1972, the Federal minimum wage rates were raised to a level above the guaranteed wage rates, thereby making the farmers subject to the Federal rates ineligible for the subsidy. The Puerto Rican govern- ment responded by passing Law #20 of June of 1972 which created an 45 N NmNnmuNNa< NoUNNNucmcmo mommcche .NNNooNNm< Neeocewmnach N NoeNNNucmcmw Nommech we ecNuNmo .mumu o: .umgamcmomeNe =.Nueu=mscm ezmmm .NNNN me am omega we NNN .532 No; NN Non NoNNmano NNNoUNNm< NmeNuN>NNu< em NoonNasm mmeoemnmneeh NoN "moezom .chmH\N mUCmm mLmv—FPE $0 cowquuxm m...“ nuts mLmv—LONNN Umppwxm 1:: NoN mm mEeN we» mam: NNNNN mom: ammuceemsm New Nan NeN:NE chmN; we; Nemxgoz umNNNxN cNeucmu "muoz .sz>NNumaNmN .NNNN>NN; ouumnop new mmNNou ecu :N Name: opacimumNa on» men NNNNNN> on» use wasNN msha NmoN New NNN New NNNN NNNN NNN NeNN Nom mNNoNocN N NN.N NN.N NN.N NN.N NNNN NN.N mo. NN.N om.N mam: umoucmeeaw . 111 cu eNNNNN NNN NNN NNN NNN NNN NNN NNN NNNN NNN omeoeoeN N mo.N mo.N mo.N mo.N mo.N mo.N mo. mo.N om.N mom: cmmueegeaa . NNNNNN op NNNN\e NNN Nam NNN New NNN NNN NooN Nom «NemeoeN N oo.N oo.N oo.N oo.N oo.N oo.N oo.N om.N mam: ammueeemsu NNNNNN op NNNNNN Nwm Nme NNN Nmm New New Now Nmm NNNmNoeN N om. om. om. om. om. om. om. mm.N mam: umoNeecaao NNNNNN 6N oNNNNh Noe NNN NNN NNN Nme Nme Noe mmeweucN N NN. ON. ON. ON. ON. ON. ON. NNN: eooeeeeeeN ONNN\m op mmNNNN . . . . . . . . . mam: NN NN NN NN mm mm NNNN ON ON N seeNeNz .N.N Lao: Lao: Lao: Lao: Lao: Lao: NNNNNNN> Lao: «u:E—< NNNNN Nemeamg NNNNN NgNmo N Noam NewmaucN gmcuo Nemzopu -cha Ngupaoa Nwzem ouueaoh mmmwou chzapauNNm< Now mmpmm mom: ammuceceaw N.Emcmoea NuNmnzm mama mum: ecu use Nope: mum: EaeNch ouNm oases; .mNmN op NcoN .Nmuem one: szsNeNz NNNmewN on» Na umgm>ou Noz NewconeN chmemw .m mpnmh 46 equivalent subsidy for the affected farmers. Within months of the 1972 and 1974 changes in the Federal minima, the Program's guaranteed wage rates were raised to meet the Federal minima. Program Costs and Workers Benefited Table 10 reports the amounts expended under the program for the first six years of operation. The Supplements represent the reimburse- ments made to the farm operators for having paid the guaranteed wage rate. The "additional costs“ represent reimbursements made to the farm operators to offset their increased liabilities of Social Security and State Insurance Fund taxes and payments for the interest on the farm operators' funds used to pay the supplements and the taxes on the supplements. For the six years as a whole, the administrative costs represent 4.52% of total costs and the additional costs represent 8.89% of total costs. Table 11 reports the payments made under the sugar cane law (Law 141 of 1969), the tons of cane milled, the supplements paid and the additional costs. Table 12 reports the number of workers, hours worked, total payments, hours per worker and supplement per worker for the non-sugar-cane laws (Laws 142 of 1969 and 20 of 1972). The data on payments and participation under the program are not fully consistent. Nevertheless, they are presented in order to convey the general magni- tudes involved in the program. 47 .mcszauNNm< we acme -Nemawo .m.a .NNNoumcm< Noeoumnmnmch N monNNNucweeo mommech mu NcNuNmo New EON; NNNu coemwNnanca Nmoeaom .NNouNNmno Exam ecu chmesnENme :N NNNmu we» :0 mmcmgxw pmmemucN new Npcm5mpaasm we» no meNN NmNoNaem wean mucmezmeN macaw use Nuwezumm NNNuom no» Nucmexea muzNocN Npmou NaeoNNNuuHk ems Nesos ome No mums ms» an exam ouuenou No NNNNNNN> use wasps gas mass; NNN No mums on» we smxows mmNNoo No NoussNN NouzNust .Nueu os .NNNouvsm< Nmsovmnmnesh N NouNNNNstNw NommsmsN we msNuNwo .NNNNN cmsNNanss: ”mussoms .N1N .Ns .NNNN oe oNNNN =.ooNNNNeNNNN oNeeoaeoue .NNNopwsm< Nmsoumnmsese N Noueuwpsmsmw mommsmsN mu esNuNwo .eszupsuNsm< as ousmssusmsmo Noussomm .emuNusmss<= No mums can ..uNnN ”wussomN .emuNusmss<= No mass NNNNN .mumfi mu oumom< =.~mmH mu om .532 » mama mu mea .E:z meme; mm; Nos moumw>ogs NNPOUNNm< Nmsouenenmsp NoN < NoNNNNsmemNssm NomwsmsN No; mssom mssowsNe .NssNNzuNsms we oNsmENNNNsmo ”mussomN .c.s Nee NNN NNN NNN smxsoz ems NsmEmNsssm .u.s Nmo.N 0mm eNN oNN Lasso: ems mesa: eoe~.NN ome.mN woN.mN omN.N msN.m Nooo.mv NNsNENNs NNNoN NNN.NN NNN.sN emm.N~ smm.e~ mes.om mNooo.V emxsoz mess: ooo.Nm ooc.mm ooo.em ooo.m~ mam.mm umNNNmmsmN Nsmxsoz sstN NNNNN NNNNN NNNNN oNNN NNNN NNoNNN Ammo“ mo om use NNNN No NNN szNv soNNNos mssu Nemsm1soz 1 apes soNNNsNuNNNNs sesmoss smNumNmm .NN NNNNN CHAPTER V A FRAMEWORK FOR THE ANALYSIS OF THE PROGRAM'S EFFECTS Introduction In Chapter III an intuitive analysis of the conditions in the agricultural sector of the Puerto Rican economy showed that the decline in agricultural production was due mainly to the changing price relation- ships over time and a general inability of Puerto Rican agriculture to adopt newer technologies and mechanization in order to overcome these changing price relationships. This analysis implies the need for a dynamic framework of analysis upon which to develop a model which may be used to estimate the magnitudes of the effects of the program on 1 production and employment. The second section of this chapter develops such a framework. The first part of this chapter presents an analysis of the effects to be expected in a short run time frame, utilizing conventional short run microeconomic theory. The instructive aspect of this analysis is the conclusion that the subsidy program should not have any impact on production or employment in the short run. Effects in the Short Run Under Competitive Conditions Assumptions The assumptions of microeconomic theory which are critical to this analysis are: 50 51 1) Product prices are given to the fann operator and are not affected by a change in his level of production. 2) Factor prices are also given to the farm operator and are not affected by a change in the quantities purchased by him. 3) Both product and factor prices remain constant in the short run. 4) The farm operator possesses perfect knowledge of the prices, his production function (i.e., the least-cost combination of inputs for each level of production) and the availability of inputs. Although the Puerto Rican farm operator cannot possess perfect knowledge, it must be assumed that he possesses a reasonable level of knowledge about the forces which may affect his operation. For this analysis to be generalizable to the agricultural sector as a whole, these same assumptions must hold for the sector. That this is generally true has been shown by the finding that: 1) most inputs except labor are imported from the U.S. mainland, 2) the production is either sold for export in quantities insufficient to significantly affect prices or is sold on the island for local consumption in a market dominated by imported foodstuffs, and 3) the wage rates paid by farm operators are determined by the Minimum Wage Board or the Administrator of the Wage and Hours Division of the U.S. Department of Labor. Production Theory Under the assumptions of pure competition, a firm will select its levels of production and of input use so as to maximize its profit function. Profit equals Total Revenue (TR) minus Total Costs (TC): (5.1) Profit = TR - TC. Total revenue equals the Product Price (Py) multiplied by the quantity produced (Y), which is a function (the production function) of the input combinations (X1, . . . , Xn). Total Costs equals the sum over all the 52 inputs of Unit Price (Px ) multiplied by the quantity used (Xi), for i inputs 1, . . . , n. Thus, equation 5.1 may be expressed as: n a g X )_ z P X00 '1 - X (5.2) Profit = ny(X1, . . 1:1 1 1 Profit is maximized when the first order conditions (0 Profit/ 3 aXi = 0, i=1, . . . , n) are met. Thus, (5.3) Profit/ Xi = nylzxi —- Pxi = O, for i=1, . . . , n, and where f is the first derivative of the pro- szi 'duction function with respect to xi. Since nylzxi is the Marginal Value Product of Y for X, (MVPx 'Y)’ then profit is maximized when all 1. inputs are used at levels which equate the MVP»,Y and the factor cost, 1. Px . For fixed input prices, the case of a two variable input profit i function may be shown graphically using Iso-Value-Product and Iso-Cost lines (Figure 9). The slopes of the Iso-Cost Lines are equal to the ratio of the factor prices (Px /Px ) or of the Marginal Physical 2 1 °Y/MPPx 'Y)’ For any given output, the profit maximizer 2. 1. will select the combinations of input quantities which lie on the Products (MPPx Expansion Path since these combinations represent the least cost com- binations. The selection of the optimum level is dependent upon the product price, such that the Marginal Value Product of Y for X1 and X2 in least cost combination is equal to Product Price. 3The second order conditions are met if the law of diminishing returns holds. 53 Figure 9. Diagrammatic Representation of Production Decision (a) Factor-Factor Space (b) Cost and Revenue Curves X2 $ IVP1 IVPZ Maximum Profit per Unit Expansion of Combined Inputs Path 7 ._ / //’/// -¢/ . L.C. . 101 It, I\\\\\\Avp X1 X and X in L.C.C. 1 2 The Demand for Labor A firm's demand for labor, or any input, is derived from its pro- fit function. Since the profit function is maximized when the partial derivatives with respect to each of the inputs is equal to zero (the first order conditions), we have a system of equations: (5'4) M = P f , —- P = O 3 X1 y 1.x1 x1 a Profit ----- = P f , - P = O, a Xn y 1.xn xn which when solved for X1, . . . , Xn, gives the demand schedule for X1 in terms of Px when all the other variables are held constant: i (5.5) x, = gi(Py, x1, . . . , xi_1, xi+1, . . . n Although the precise equation cannot be determined without knowledge of the production function, if the production function behaves according to the law of diminishing returns (i.e., meets the second order conditions), 54 the producer's input demand function may be shown to be always downward sloping (Henderson and Ouandt, p. 70). From this result, we can see that if the price of an input increases with all other prices remaining constant, the quantity demanded will be decreased, and vice-versa. Theoretical Impacts of the Program in the Short Run Under the wage rate subsidy program in Puerto Rico, the farm operator is required to pay the guaranteed wage rate (9) to his laborers and subsequently is reimbursed for the difference between this rate and the legally established minimum wage rate (W)--this difference being the subsidy rate (S). The profit equation may be rewritten as: m (5.6) Profit = ny(X1, . . . , Xm,L) —- P Xi + SL -gL i=1 xi where X1, . . . , Xm are the quantities of non-labor inputs used and L is the quantity of labor used. The first term of the equation is total revenue expressed as product price times the production function, the second term represents the sum of all non-labor input costs, the third term represents the subsidy payment received and the last term is the cost of labor at the rate the farmer paid (the guaranteed rate). How- ever, since the subsidy rate equals the difference between the guaran- teed rate and the minimum wage rate (S = g-W), it follows that: (5.7) SL -gL = 9L -WL -gL = WL and since L is an input quantity and W is its price, the profit equation can be rewritten as: 3 (5.8) Profit = ny(X1, . . . , Xn) —- i=1 PxiXi, where WL is simply included in the general input cost term Px Xi' Since i 55 this equation is identical to equation (5.2), we find that under short run competitive conditions the fanm operator's production and employment decisions are derived from identical profit equations whether or not the subsidy exists and therefore the wage rate subsidy program should not affect either the level of production or the level of employment.5 Effects in the Intermediate Run Under Competitive Conditions Review of the Agricultural Sector, 1961 to 1969 Puerto Rican farm operators are generally price takers in both their input and product markets. Thus, they determined production and input use in accordance with these prices and their particular production function. The unfavorable changes in the relative price relationships and their inability to adopt more productive technologies, especially labor saving technologies, appear to be the primary causes of the sharp decline in agricultural production. Given the high degree of dependence on hired farm laborers, the rapid rises in minimum wage rates appear .to be of greater importance. The Farm Decision Making Setting The biological process of agricultural production requires that decisions concerning desired or planned output levels be made at or by the beginning of a planting-harvesting cycle or period. The existence of fixed assets in the farm firm make this decision making process a 5This would not be true under a general wage rate subsidy program in which the subsidy rate is a proportional function of the difference between the guaranteed wage rate and the equilibrium or minimum wage rate, whichever is higher. 56 serial one--decisions for the current period are dependent or condition- al upon previous decisions, since by definition of asset fixity, investment and disinvestment decisions can only affect future time periods. Thus, the previous decisions, as embodied in the fixed assets of the firm, establish a range within which rational economic decisions "my be made for the current period. The decision at each planting time may be to increase, maintain or decrease output levels. To increase output requires additional invest- ment and to decrease output permits disinvestment of currently fixed assets. The decision will be made based on the current financial structure of the farm and the operator's expectations of what product and input prices will be during the following periods. In particular, the operator will increase his investment in fixed assets if the ex- pected product and input prices will result in total revenues equal to or in excess of total costs over the life of the fixed assets in which he must invest. Conversely, he will begin the process of disinvesting his fixed assets if expected product and input prices will result in total costs in excess of total revenues. If his expected revenue is in excess of his variable costs, he will continue to utilize his fixed assets until such time that he may dispose of them and if his variable costs are expected to exceed his revenues, he will leave the assets idle until he can dispose of them. The process of disinvesting one's fixed assets is commonly more difficult than the process of investing in fixed assets. The salvage value of farm assets, as determined in the market of alternative uses of the asset, will often be below the capitalized value of the asset's ability to produce an income stream in its present use (its Marginal 57 Value Product). For example, a producing stand of sugar cane or coffee has no alternative use and a trailer for hauling sugar cane is not like- ly to be usable except as scrap metal. Thus, the farm operator will continue to use these assets until his marginal variable costs exceed his marginal revenues or until the asset is no longer usable. The Puerto Rican agricultural sector had clearly been undergoing a process of disinvestment between 1966 and 1969 when the subsidy pro- gram began and for three years prior to 1966 had been almost maintaining its level of investment. Although no data are available on the value of farm assets, the data on agricultural production suggest that this has been the case. The changing price relationships, as shown in Table 5, page 25, also lend support to the possibility that for many farmers, total costs may have been larger than total revenues, thereby also supporting the apparent process of disinvestment. Although individual farmers may have been increasing their investment, in the aggregate it has probably declined. The apparent topographical and farm size limitations on the adop- tion of newer technologies also has played an important role in the decline of agricultural production in Puerto Rico. Had it been possi- ble to invest in newer technologies which would have increased the productivity of the variable inputs (particularly labor), the variable costs of a farm could have been lowered to a point which would have justified the new investment and a subsequent increase in production. Given the limitations on the adoption of newer technologies, the Puerto Rican farm operators were generally unable to achieve the signi- ficant increases in input productivity which would be necessary to offset the unfavorable change in relative prices. 68 - Although the farm decision making setting described here deals with the relationships between output and fixed assets, it is still the prices of prodUcts and variable inputs which most directly affect the year-to-year fluctuations in output. The effect of the asset fixity is to introduce a sequential process to farm decision making. which requires the introduction of future expectations into the model which is developed in the following chapter. Theoretical Effects of the Program in the Intermediate Run The change in the time frame of the analysis to the intermediate run does not permit such a simple and determinate analysis as was pos- sible in the short run analysis. This is due to the introduction of changes in the fixed asset structure of the farm or in the combination of variable inputs, either of which represents a change in the techno- logical package being used, and therefore changes in the production function itself. Since the wage rate subsidy holds the farm operator's wage rate constant over time, and since the prices of the other variable inputs have continued to rise over time, the price of labor has become cheaper relative to these other inputs. This change in the relative prices of inputs means that the ISO-Cost lines of Figure 9(a), page 53, change in slope. Thus, a different combination of inputs will be the least cost combination of inputs for each level of output and the expansion path for each year or period will be different. In particular, the relative cheapening of labor should result in a more labor intensive technology through time as the originally fixed assets may be changed, in which case both production and employment should be greater than the levels 59 which would have obtained without the subsidy. It does not seem prob- able that a relative cheapening of labor would be followed by a move to a more capital intensive technology, especially in view of the natural limits to mechanization of sugar cane in Puerto Rico. CHAPTER VI DEVELOPMENT OF THE MODEL: METHODOLOGY AND STATISTICAL ESTIMATION Introduction In order to estimate the effects of the program on production and employment, it will be necessary to develop a model which relates these two measures to the wage rate and then to use the model to predict how these measures would have resulted if the wage rate paid by farmers had taken certain assumed values for the "without subsidy" case. After presenting a brief description of the available data, this chapter will develop and statistically estimate the two equations of the model. The final section of this chapter summarizes the model and the fbllowing chapter applies the model, using a set of assumptions concerning the wage rate, to estimate the effects of the program on production and employment. The Available Data The Index of Agricultural Production (AGPROD) and the Index of Product Prices (AGPRCE) These two variables are measured by the Puerto Rico Department of Agriculture's Office of Agricultural Statistics by means of semi-annual surveys of farms in Puerto Rico. Approximately 1,500 farms are in the sample and the sample is stratified into four categories of small farms and four categories of large farms. In addition, the island is divided 60 61 into five regions. Each of the resulting forty sub-samples has its own expansion factor based on the sampling rate for the estimate of the production of each product in each sub-sample. The prices of the pro- -ducts are weighted averages based on the prices received and the quanti- ties sold by each farm. The annual estimates of output and prices for each product are based on the estimates from the two surveys in each year. The aggregate index of physical production is then calculated using production in the base period (FY 1960-1963). The aggregate price index is computed in a similar fashion using base period value of pro- duction as weights for the aggregated product categories. Emplgyment of Agricultural Laborers (AGEMP)_and the Wage Rate Variable (WAGEji These two variables are estimated by the Puerto Rico Department of Labor's Bureau of Labor Statistics using a survey almost identical to the Current Population Survey of the U.S. Bureau of Labor Statistics. The employment variable is a simple average of the estimates from the twelve monthly surveys for the category of Employed Farm Laborers and Foremen. The wage rate variable is estimated by using an additional question in the surveys for January, April, July and October and gives the median weekly wages for the same category of labor. These median weekly wages were adjusted by this researcher for those April estimates which the Puerto Rico Bureau of Labor Statistics reported as being affected by the Holy Week Observances. This adjustment involved averag- ing the difference between the January and April estimates in the years not affected and adding this average change to the January estimates of the affected years. The four quarterly estimates, if and as adjusted, were then averaged for the annual average. However, Since the wage 62 rate subsidy held constant the wage rate paid by farm Operators, the wage rate variable used in the estimation of the model remains constant at the 1969 value ($18.88) through 1974. Feed Prices (FEED) and Fertilizer Prices(§ERT) The only time series data on prices paid by farmers is the series for "Maximum Prices for Mixed Fertilizers," as published by the Puerto Rico Department of Agriculture in their publication entitled Ventas de Abgggg. The weighted average (maximum) fertilizer prices were used in this research. A second input price variable was constructed by this researcher by using the data on imported feeds which the Puerto Rico 6 Planning Board provides to the Office of Agricultural Statistics. From the data on tonnage and values for imported prepared dfiryvand poultry feeds, a weighted average price series was computed. The Index of Production Equation Methodological Considerations In Chapter V, it was shown that for a given technology, a farm firm under competitive conditions selects that level of output that maximizes its profit equation. In particular, the selection of the optimum output level is dependent upon product price, such that the Marginal Revenue Product when the inputs are in least cost combination is equal to Product Price. Since the least cost combination of inputs is dependent upon 6Office of Agricultural Statistics, Net Income Section, File Number 1.70. 63 input prices, the production decision is dependent upon both product and input prices: (6.1) Y = f(Product Price, Input Prices) where Y is quantity produced. A common approach to estimating this equation, given that the pro- duction function is both unknown and changing over time,is to estimate a supply function where output is a function of the ratios of input prices to product price (Klein, pp. 126-129): ‘ (6.2) Y, = 1' [um/Pym. (PXZ/Py)t’ (Px3/Py),]. This equation was estimated with the data previously described but gave unsatisfactory results in that the signs of some of the price ratio variables were positive rather than negative. Positive signs would mean that the farm operator's demand curves for inputs would be upward sloping, thereby contradicting the logic of economic theory. Another approach is suggested by the serial or sequential character of the decision making process which was described in the preceeding chapter. This approach is to use the concept of adaptive expectations which forms a part of the theory on distributed lags. Although this approach produced much more satisfactory results, there are both con- ceptual and statistical estimation problems associated with its use. The Adaptive Expectations Model The adaptive expectations model (Kmenta, pp. 474-476 and Griliches, pp. 16-49) posits that the value of the dependent variable in the current period is a function of the expected value of the independent variable for the current period: 64 (6.3) Y = a + Bx; + e t t' In the case under analysis, Yt is the level of output which the farmer is to select based on his expectations about the prices which he will encounter in the upcoming period (XE), with the decision being made just before period "t". Since X? is not known, it is necessary to assume a relationship between X? and some known values. The assumption under this model is that x; is a weighted average of X in the previous period (Xt_1) and the value which X was expected to take in the previous period (X{_1): (6.4) x; = (1-)1)xt_l+>1x;_1. Since X§_1 may be expressed as a function of the values of X in period t-2, this relationship may be expanded by using all past values of X. This expanded form gives a geometrically decreasing weight to the values of the previous periods: (6.5) X* = (1 - A)X t + (1 - MAXH + (1 -A)szt_3 + (1 - 11)X3xt_4 t-l + . . . + (1 -X)R"'1xt_n. Thus, equation 6.3 can be rewritten as a function of all previous values of X: (6.6) v = a + (1 - h)13(xt_1 +71xt_2 + 112x . +An‘lx t t-3 + ' ° t-n) + at Since there are too many regressors in this equation to permit statisti- cal estimation, the Koyck transformation may be used. Lagging the equa- tion once, multiplying through by A, and subtracting this from the original equation and adding AYt_1 to both sides gives: (6.7) Y = a + (1 -X)th_1+RYt_1+ n t t’ where nt = et — Aet_1. 65 Although this transformation permits statistical estimation by reducing the number of regressors, it creates a new problem of estimation in that the new error term is serially correlated, thereby causing the estimates of the coefficients to be inconsistent (Kmenta, p. 479). To overcome this problem, the Cochrane-Orcutt procedure may be used (Kmenta, p. 479). This is an iterative procedure which uses ordinary least squares (OLS) to estimate the original equation. The computed residuals are used to estimate the coefficient of serial correlation (p) and this estimate is then used to transform the original equation (6.7) to: (6.8) (Yt - BYt-I) = a(1 - p) + B(Xt - 6Xt_1) + ut. These estimated coefficients are then used in the original equation (6.7) to Obtain a new set of residuals with which a new p is calculated and subsequently is used in the transformed equation (6.8) again to obtain another set of estimated coefficients. This iterative procedure continues until the estimated coefficient of serial correlation (B) converges, i.e., changes by a predetermined small amount. The estimated coefficients which result from the final round coincide with the maxi- mum likelihood estimates (Kmenta, pp. 287-289). Although the use of this procedure results in consistent estimates, the computation of the fitted values becomes complex since all previous error terms enter into the equation. In year t, the equation becomes: (6.9) V =3+’1§x +2111 at nt-l A t t']. +pe+p 9+...‘1'pe t-1 0 1 t-1' Thus, when using the estimated equation with an assumed set of values for the independent variables in order to predict what the levels of production would have been for the "without sibsidy" case, there are 66 no known values with which to compute the residuals which should be used in the subsequent years. The conceptual problem of the adaptive expectations model is con- cerned with the imposition of a distributed lag scheme when it may actually be endogenous to the model. Mundlak provides simple examples of how the geometrically distributed lag scheme implies a geometric decline in output when the price of capital is increased by an amount greater than the increase in product price in the model: Y = f(x,k) (Mundlak, pp. 51-60). Although Mundlak makes his point, he uses a two period analysis and a rise in the price of a firm's fixed assets so that output cannot be affected until the second period. For the pur- poses of this research, however, the behavior of the farm firm is assumed to be more directly related to changes in the prices of vari- able inputs, for which the geometrically distributed lag seems to be a reasonable approximation to what one would expect even if it does not reflect the real response exactly. Additionally, there is no available procedure for statistical estimation of the correct lag scheme, even if this research could avail itself of a more voluminous amount of data with which to use a more accurate method. Thus, the use of a potentially incorrect lag scheme may be seen as a less serious error than the use of a model which ignores completely the sequential character of the farm decision making process. One additional conceptual problem, but one which appears to be less serious, is that the volume of available data do not permit the estima- tion of separate lag structure coefficients for each of the three vari- ables in the equation. Thus, the lag coefficient combines the effects of all three variables and should not be utilized to compute long 67 run elasticities for the coefficients of the independent vari- ables. In summary, the adaptive expectations model, equation 6.3, relates actual production to the farmers' expectations, at planting time, of the year's product and input prices. In order to estimate the coeffi- cients using observed rather than expected prices requires the trans- formation to equation 6.7. However, this results in a serially correlated error term, thereby requiring the use of the Cochrane-Orcutt procedure. The resulting coefficients coincide with the maximum likeli- hood estimates of equation 6.7, but the coefficient of serial correlation (6) must be used in computing the fitted values with equation 6.9. Finally, the equation is estimated using three independent price vari- ables instead of just one since many more data points would be required to estimate three separate lag coefficients. This implicitly assumes that the lag coefficients for the three price variables are the same-- meaning that the farm Operators use the same set of weights for each observed price from the prior years. Thus, the equation to be estimated is: (6.10) Yt = a + lelt-I + 82X2 + B3X3 +AYt_1 + error. Statistical Estimation of the Production Index Equation The results of the statistical estimation procedure are as follows: (6.11) In AGPRODt = 1.53916 + .0414235 In AGPRCEt_1 (.0297) - .247280 In WAGEt_1 (.0341) - .192987 In F550,}1 (.0506) + .357305 In AGPROD (.0785) t-I 68 p = -.401066 (.2541), R2 = .9899, F = 197, D.W. = not applicable, Number of Observations: 13 (1963-1975) Values in parentheses are standard errors. AGPROD = Index of Agricultural Production AGPRCE = Index of Agricultural Prices WAGE = Median Weekly Wages of Farm Laborers FEED = Weighted Average Imported Feed Prices Therefore, the original adaptive expectations model (equation 6.3) becomes: (6.12) 1n Production Indext = 1.5396 + .04142 In expected product pricet - .24728 In expected wage ratet - .19299 In expected feed pricet Figure 10 shows the observed values of the production index and the fitted values resulting from this equation.7 The Employment Equation Methodological Considerations The theoretical derivation of the farm manager's demand for labor function (equation 5.5) gives employment as a function of the product price, the wage rate, the prices of all other inputs and the quantities of all other inputs. This function tells how much labor will be employed at each possible wage rate when all other prices are held constant. For a given production function and a given set of input and product prices, the farm manager selects the level of output which will maximize his profits and thereby automatically selects his input use levels. While 7An observed value for 1975 is possible since the index for 1975 was available even though the other variables did not have 1975 values avail- able and since this equation uses lagged values for the independent variables. 69 NNNNN NN NNNNoNNNNNNN NNNNNNNNN NNNNE ,__—4 .4mm mo:_m> N\uo>somno mo:_m> :vouuNu: oo— Imo— 10: 1am—— low. JNNN 82 a 318 NE .52.. .NNNNN 6N NNNNV NNNNNN NNNNNN use NNNNN 6N NNNNN NosNe> eo>eomeo .ouNm ousmss sN mssuNsuNsms No soNNossoss NNNNNNsN No xmssN as» .oN NLNNNN 70 these selections are simultaneous, they are all dependent upon the profit function (which embodies the production function) rather than being innaw dependent among themselves, as is the common case of supply and demand simultaneously and interdependently determining both price and quantity. This mutual dependence of the production and employment decisions on the profit function would suggest the formulation of an employment equation similar to the production index equation which was developed in the pre- vious section. The following adaptive expectations model for the employment of farm laborers in Puerto Rico was estimated: (6.13) E = a + B X* + B X* + B X* + e , t 1 1t_1 2 2t.1 3 3t.1 t where E is employment, X1 is the product prices index, X2 is the wage rate variable and X3 is the fertilizer price variable. Although this equation resulted in a very good fit (R2 = .9886), the sign of the coefficient for fertilizer prices was positive--implying an upward sloping demand curve for fertilizer--and the coefficients for both the product prices index and the fertilizer price variable were not signi- ficantly different from zero. The more common adaptive expectations model: (6.14) E a + B X* + B X* '+ B X* + e , t 1 1t 2 2t 3 3t t which is predicated on the assumed relationship between the expected prices and the observed prices of: * = - (6.15) xt (1 )1)xt + x;_1, was also estimated on the basis that the bulk of the labor is employed at harvest time when current prices are known with much greater accuracy. 71 This formulation, however, resulted in a lag coefficient greater than one (meaning that prices farther back in time receive greater weights in determining expected price in the current period) and the coefficient on the wage rate variable was extremely close to zero (implying that the wage rate has no effect on employment). An alternative approach might be to simply use current or lagged prices as the explanatory variables, but this would ignore the sequential character of the production decisions and would permit an extreme value of one of the variables to yield an estimate of employment which would be completely out of line with the requirements of the level of produc- tion which is determined separately. A logical choice for a variable which will dampen the effects of an extreme price value is the level of production. The simultaneous nature of the production and employment decisions suggests using the fitted or predicted values for the index of production which resulted from the index of production equation described in the previous section. This is, in effect, a two-stage least squares pro- cedure. Thus, the equation: (6.16) E = f(X , X , X , Y ) t 1t 2t 3t t was estimated, and although it gave a good fit (R2 = .9893), it yielded a positive sign on the wage rate coefficient. This implies that farmers would employ larger amounts of labor as the price of labor increases, ceteris paribus. In that the production and employment decisions are mutually dependent upon the profit function and are not interdependent (as the use of two-stage least squares would imply), the employment equation was 72 estimated using the observed values for the index of production instead of the fitted values. This equation: (6.17) E = f(X1 , X X3 , Yt) t t t resulted in a good fit (R2 = .9599), but the wage rate coefficient was 2t, not significantly different from zero and the Durbin Watson statistic indicated a moderate serial correlation in the disturbances, meaning that the estimators of the coefficients are not efficient and that the variances of the estimated coefficients are biased (Kmenta, p. 278). Another alternative is to use the lagged index of production in place of the current index of production. This equation: 1 X . X . Y ) 1t 2t 3t t-l resulted in a slightly improved fit (R2 = .9732), significant coeffici- (6.18) Et = f(X ents on all variables (t 2.0) and less serial correlation in the dis- turbances (D.W. = 1.8). Although the logic of using current values of the index of production seems to be stronger, the improved statistical results of the lagged formulation may be due to a more direct reflection of the sequential nature of production decisions through the use of lagged values instead of current values. In light of these logical and statistical considerations, the statistical estimates of both of these equations are presented in the next section and both are used in the next chapter to estimate the pro- gram's effects on employment, thereby providing alternative estimates. Statistical Estimation of the Employment Equations The results of the statistical estimation procedure are as follows: 73 (6.19) Employment Equation with Current Production 1n AGEMPt = 17.1965 + 2.30804 1n AGPRODt (.589) + 1.19110 1n AGPRCEt (.605) - .49451 1n WAGEt (.379) - 2.80598 1n FERTt (1.170) 2 R = .9599, F = 53.8, D.W. = 1.5668, n = 14 (6.20) Employment Equation with Lagged Production 1n AGEMPt = 18.4859 ’ + 2.14384 1n AGPRODt_1 (.409) + 1.36951 1n AGPRCEt (.502) - .89010 1n NAOEt (.251) - 2.86476 1n FERTt (.954) R2 = .9732, F = 81.6, D.W. = 1.8402, n = 14 For both equations: AGEMP = Employment of Farm Laborers and Foremen AGPROD = Index of Agricultural Production AGPRCE = Index of Agricultural Prices WAGE = Median Weekly Wages of Farm Laborers FERT = Weighted Average Maximum Fertilizer Prices Figure 11 shows the observed values of employment and the fitted values for both of these equations. Summary The model developed in this research consists of one equation for the prediction of the index of physical production and one equation for the prediction of the level of employment of farm laborers. The produc- tion equation (6.11) makes use of the adaptive expectations model to incorporate the sequential character of farm decision making and includes as independent variables the prices of products, wages and feed, in addition to the lagged dependent variable. This equation was estimated 74 Figure 11. Employment of Farm Laborers and Foremen, 1961 to 1974, Observed Values and Fitted Values. Employment of Farm Laborers (in Thousands) 9o— /\ 80- ....... , 70- Observed Values “Fitted” H Values 60c _. (Eq. 6.19) .unF/N \\ . ”Flttfid' Values (Eq. 6.20) 2...: 7.1:.“ 50— 40- s 30r ZOr 10— J l L l l l l l l l l 1 1 1 FY 61 62 63 64 65 66 67 68 69 7O 71 72 73 74 75 using the Cochrane-Orcutt procedure due to the introduction of serial correlation with the inclusion of the lagged dependent variable. The conceptual problem of the imposition of a specific lag structure was considered, but was disregarded since there was no alternative means of constructing the equation 50 as to reflect the sequential character of the production decision. The supply function formulation was esti- mated and rejected on the basis that some of the signs of the coeffi- cients were contrary to theoretical expectations. Two alternative employment equations are presented. Both include current period prices for products, wages and fertilizers as independent variables, but one includes the current production index (6.19), while the other includes the lagged production index (6.20). The production index is included as an explanatory variable in order to dampen any abrupt changes in the price variables and in recognition of the mutual dependence of the employment and production decisions on the profit function. Both variations of the adaptive expectations model (using lagged employment as a regressor) and a two-stage least squares formu- lation (using the fitted production index values from the production equation) were estimated and rejected since, in each case, some of the estimates of coefficients were contrary to theoretical expectations. CHAPTER VII APPLICATION OF THE MODEL: ESTIMATES OF THE PROGRAM'S EFFECTS ON AGRICULTURAL PRODUCTION AND EMPLOYMENT Introduction The type of evaluation design used in this research is a modifica- tion of the design which Hatry, et al., characterize as "Time Trend Projection of Pre-Program Data versus Actual Post-Program Data" (Hatry. Winnie and Fisk, pp. 45-47). Rather than using only the nine years of pre-program data in the estimation of the equations, it was possible to use all fourteen years of data since the effect of the program on the explanatory variables was only to hold the wage rate variable constant. Although the equations are not simple time-trend relationships, the use of the model requires that assumptions be made concerning how the wage rate variable would have resulted for the "without program" case. The assumption which is made, however, is a time-trend projection and speci- fies that the wage rate paid by farm operators would have continued to increase at 5.96% per year, which was the average annual growth rate for this variable from 1961 to 1969. Equally critical to the correct application of the model is the assumption made concerning the other price variables which enter into the equations. Here, it is assumed that the index of agricultural product prices and the feed and fertil- izer variables would not have been affected by the implementation of the program. This assumption is based on the analysis presented in Chapter III which showed that the prices in the agricultural sector of Puerto 76 77 Rico (with the exception of labor) are primarily determined outside of Puerto Rico as a result of the island's dependence on imported food- stuffs and agricultural inputs and the relatively small share which the agricultural exports represent in the respective world markets. For these reasons, it is argued that these prices are not affected by changes in the level of production (in the case of product prices) or of input use (in the case of non-labor inputs). The magnitude of the problem of farmers' falsely claiming larger subsidy payments, as reported in Chapter IV, pages 43 and 44, is com- pletely unknown. Thus, the variables used in the model cannot be ad- justed to reflect this problem. However, any such increase in subsidy payments would clearly make labor even cheaper relative to the prices of other inputs and would thereby increase the program's effects on production and employment. A comparison of the observed and fitted values for the equations (Figurele and 11) shows that the equations do not capture the rela- tionships of all of the factors which influence production and employ- ment. Although they appear to capture most of the important relationships, either directly or indirectly, some errors remain in the equations. These errors may result from the omission of relevant variables or from errors in the measurement of the included variables (especially the feed and fertilizer variables). Although the produc- tion index equation tracks quite well after the first three years, it does not predict the one year increase in production in 1974. The employment equations show an even greater error, even in the more cri- tical recent years, and should therefbre have less reliance placed on their predictive abilities. The existence of errors in the model 78 requires the use of a simple two-step procedure in applying the model. This procedure is to compute the effect of the program by taking the difference between the fitted values of production or employment for the "with program" and the predicted values for the “without program" data and then subtracting this difference from the observed "with program" values to give the "without program" predicted values. This procedure ensures that the effect of the program will not also include the errors of the model.8 In addition to providing predictions about what production and employment would have been without the program, the effect on produc- tion may be translated into dollar values for comparison with the direct treasury costs of the program. This translation requires a two-step procedure. First, the difference in the production index for each year is multiplied by the value of production in the base period (average of 1960/61 to 1962/63) which was $246.883 million. This intermediate re- sult is the dollar value of the effect on production in real terms. This is then multiplied by the index of agricultural product prices to convert it to nominal or current dollars, in which terms the direct costs of the program are stated. The evaluation of the program's effects 'Hs purposely limited to the five year period of 1970 to 1974. First, the equations were esti- mated with data for the period of 1961 to 1974 and projections beyond 1974 will not be within the range of the data. Secondly, some of the 8The difference between the observed "with program" values and the predicted "without program" values will include the model's errors. In the case of the employment equation in Figure 11, it is clear that the difference will be smaller in the period 1971-1973 than the equa- tion's relationships would suggest. 79 data which were available for this research are of questionable accuracy. Thirdly, and of greatest importance, is the observation that the effects of the program are cumulative in nature and the model is incapable of predicting the significant structural changes which would be expected to occur in a longer period of time as the farm operators become able to alter their technological package through a change in their fixed asset structure. A particular consequence of the use of the adaptive expectations model, which relates current production to all previous prices but not current prices, is that there is a one year delay in the effect of the program on production. However, since the employment equation does not use lagged price variables, the model results in an immediate effect on employment. The unacceptable results of the statistical estimation of the adaptive expectations formulation,which uses lagged price variables in the employment equation,suggest that it does not adequately reflect the realities of the farm operators' employment decisions. In particular, and as was noted in the previous chapter, a large part of the hired labor force is seasonal labor used primarily in harvesting sugar cane and coffee. These tasks occur at a time when current prices are either known or may be predicted with much greater accuracy than was possible at the time when the production decisions were made. Based on improved knowledge about price relationships, the farm operators may generally make limited changes in their harvesting technologies. In the case of sugar cane, for example, there are six alternative methods for harvesting the cane. In addition, the cane field might or might not have been cultivated. The labor content of each of the result- ing twelve methods is reflected in the subsidy program's payment rates 80 which are paid on the tonnage of cane milled, based on the labor content of each method (Table 13). These different methods affect the value of the crop through the raw sugar yield and milling costs, but they do not significantly affect the tonnage of cane, which is the measure which entered into the index of production.9 Table 13. Subsidy Payments per Ton of Cane, by Method of Harvest and Cultivation Status, 1972/I973 Crop. Harvest Method Used Payment/Ton Was Cane Field Cutting Collection Was Field Burned? Method Method Cultivated? Yes No Hand Machine Hand Machine Yes No X X X $ .45 $ .04 X X X .43’ .07' TFT FTXT X77 .57' .097 X X X .47 .11 X X X 1.53 1.17 X X X 2.20 1.52 SOurce: Compendio Estadfstico, Oficina de Ingreso Garantizada a Trabajadores AgriCoTas, Departamento de Agricultura, Julio, 1974, p. VII, (Mimeo). In the case of coffee, an improved price will result from a more selective picking of the coffee beans. Since only the ripe beans are picked under the more selective method, a greater number of passes through the coffee trees is required, although less is picked on each pass. Whether or not the coffee pickers are paid on an hourly basis or a piece-rate basis (since piece-rates are computed according to a per hour productivity conversion), 8 relative cheapening of labor should 9A change to the use of raw sugar tonnage in the computation of the index was being considered by the Office of Agricultural Statistics of the Puerto Rico Department of Agriculture. 81 induce a greater utilization of labor. Although once again the quantity of coffee harvested is unlikely to change, thereby not affecting the pro- duction index variable, a slight rise in the average price of coffee may show up in the product price index when it is computed at a later date. Thus, in spite of the natural expectation that the effects on pro- duction and employment should occur at the same time, the possibility of using alternative methods of cultivation and/or harvesting Show how the impact of the program on employment could occur in the first year without a concurrent effect on production. In an effort to test the sensitivity of the model to the wage rate variable, and thereby to determine how critical is the accuracy of the assumed wage rates in the "without program" case, the model was also used with alternative wage rate assumptions. These assumptions are that wage rates would have increased by either 5% or 7% per year. These results are presented in the lower portions of Table 14, 15 and 16. It may be noted that the annual average rate of increase in wage rates for non-farm laborers was 7.45% between 1961 and 1969 and 6.03% between 1969 and 1974. Since these wage rates reflect the rulings of the Minimum Wage Board1which would have set the minimum wage rates for farm labor- ers if the program had not been implemented, the assumed values appear to be reasonable. The Results of the Application of the Model Table 14 presents the results of the application of the production equation to predict the effects of the program on agricultural produc- tion. Columns 2 and 3 give the estimated index values for the "with 82 Amumsu oNeN 1 >N omn.m~ mmfi.wm ~w~.NN NmN. ONO. NeN. «Hm. Hum. sNNN Nwe.om NNN.NN wmm.~N ooN. Nmo. NNN. NNN. NNm. NNNN Nmm.NN omo.m NoH.m How. mmo. Now. mmm. «mm. NNNH oem.NN mwm.s omm.m emm. oNo. mum. mew. mmm. NNNN oom.e » .0 NV m emm. o Nmm. Nmm. com. onN Neopesonm seem a uNes moves one: sN ope; suzosm1Noo.N .MN sowposomm< ommnmm mmm.NN Nam.NH oNN. Hmo. NNN. «Hm. NNN. eNmN Nwo.om omm.NN «mm.m NNN. mmo. NNN. NNN. NNN. NNNN NNm.NN mmm.e mmm.m on. owe. NNN. mmm. emm. «NNN som.NN ww~.m mom.~ cum. NNo. mmm. mom. Nmm. NNNN oom.e « o o com. o Nmm. Nmm. sow. ONNH NNoNesoQo EseN o uNen Nope; omez sN ope; spzosm Noo.m .NN soNN Esmm< mmmewm Nmmewm meeeN NNN. coo. NmN. NNN. NNN. eNmN Nwe.o~ moo.¢N oNN.NN NNN. moo. «NN. NNN. NNN. NNNH NNN.NN mom.N mNm.o mom. wmo. Now. mmm. «mm. «NNN som.NN mmN.m mem.~ cum. NNo. mmm. mom. «mm. NNNN oom.o » o o a com. o Nmm. Nmm. sow. ONNN NNoNeconm Seem a uNes Nope; «we: sN oNeN suzosm,Nmm.m .NN soNNmsomm< 1:45:28 .6 6:385 FT--- 5 E NNN _ NNN E ANN 1 ENREN $313 NNN. 565.; NNNNI _. NNN. N NNN. Na: 5.5 a 22.25: apesEoz IN News xousN xousN .uoss nglolfiwz sup: xousN .Neo mm: mo soNNoouoss sN soNNoouoss so mm: No NooNusN soNposu soNNoouoss > NNou pooswo omeososN mo o=Ne> uouoNuoNs poom+m1woz 1oss uoueeNNmm, uo>somoo .NNNou Eesmoss poost use mENoN NesNEoz use News sN Nospe> .stN op onN .oon ousoss sN sowuoouoss NesouNsowsm< so Eesmoss Nuwmosm ouem omez esp mo Nuoowmu .sN mpneh 83 NNN.o~ moonw, NNNqu NNNNNN NNN. msonmm uNNN NNN.o~ mNo.N mum.NN Noe.mm oeN. on.NN mNmN moN.e~ mo~.e meo.om eem.mm Now. mom.om «NNN No~.om mem.m Nmm.mm Nmo.Nm NNN. NeN.em NNNN NNNMMN NNN.N mN~.om mum.Nm com. smo.mm ONNN msouesoqm seem Na uNea1NoNeN o e3 sN oues suzos Noo.N .NN soNumssmm< emmnmm waqo mmm.wm mN~.N~ oNN. museum sNNN NNN.N~ on.m Nmo.m~ mos.mm «NN. on.NN NNNN mom.o~ Noe.e mee.om ee~.mm on. Nom.om «NNN mN~.NN Nmm.m NON.em Nme.Nm com. NeN.sm NNNN mom.Nm Now Nme.em mum.Nm eem. ome.mm ONNN msouesomo ELeN No uNea1woNes o e: sN oues space Noo.m .NN soNNsE=NN< mNs.NN NNNNN NmmuNN www.mm NNN. - mee.om sNNN sNN.NN NNN.e No~.N~ moa.m~ NNN. on.N~ NNNN Nmm.mm ~mm.m umm.m~ es~.mm mom. mom.om «NNN oom.om Ne~.m omm.sm Nmo.Nm oew. NsN.sm NNNN Hue.Nm meo.N mmeemm. mum.Nm .emm. www.mm ONNN msouesomo stew Nu uNeQINoNeN o e: sN owes suzoso Nmm.m .NN soNNssomm< 43 QM N: NNN NNN NB 8.: $13 2: NNN deoté mm: NoosNNz NsoENoNQEN NoosNNz suNz xousN Nsoeonosm Neo> NsoENoNoEm so ms: mo 111114ma14H111 soNNosuoss uo>somoo eoNoNeoeN tote .62 Neo 2 EN eoeeeNNNN eoNeeNNNN .xousN soNNoouoss Nsossou osN msNN: NN.e soNNesom .eNNN op ONNN .ooNs opsoos sN msosooee age; No psoeonQEN any so sesmoss NuNNsom owes owe: me» mo Npooymm .mN oNoeN 84 Nom.mN NemeoN NNman mms.mm omN. museum stN Nem.mN mem.N mcm.om mmm.mm Now. on.N~ mNmN mmo.sm omw.m Nme.NN Nom.sm 0mm. mom.om «NNN Noo.om omo.e Nsm.Hm Nmm.mm cow. NNN.sm NNNN Noe.mm NNO.N - mum.mm oom.mm NNN. reme.mm onmN Neouesooo sees No uNea,NoNes one: sN eves suzoem,Noo.N .mN soNNsEsNN< NNMNNN Neonw Nmm.NN Nms.mm sNN. mes.om sNNN mom.N~ moo.e www.mm mmm.mm on. on.NN NNNH smN.m~ NNN.m on.mN Hom.sm oem. mom.om NNNN mmN.Nm smm.m Nmm.mm Nmm.mm New. NsN.em NNNN NNN.em mom.N moo.em oomJMN NNN. ewe.mm oNNN Neouesomo Esee No uNen Noyes owe: sN ope; saxosm Noo.m .NN soNNs93NN< NNN.NN NNm.m meadoN mmsemm NNN. meeuom sNNN Nmm.o~ mmm.e omm.N~ mmm.m~ mom. on.Nm NNNN smo.m~ mmm.m eNo.m~ Hom.sm com. mom.om «NNN Nmo.om omm.m Nos.mm Hum.mm wow. NsN.sm NNNN Hmo.em NNN.N Nom.mm ooe.mm NNN. owe.mm oNNN NsoNesosm,Esew Na uNes Nope; omez sN oNeN sNzos Nem.m .NN soNNmE:NN< NNN E by Na NNN $0 $1: #13 me: NNN 28834 ms: psospNz NsoENoNsEm NoosNNz saw: xousN Nsoexo—sEm Neo> psosonssm so mm: No soNuoouoss uo>somso uoNoNuoss Noommm Noz1, usoENm—QEM uoueemumm uoueewpmm .xousN soNNosuoss uommes ms» msNN: o~.o soNNeoou .eNNN ou oNNN .oon oNsoss sN Nsosooee seem No NsoszoNsEm osN so Eesmoss NuNNoam apes owe: esN No NNoNNNu .eN oNseN 85 program" and "without program" cases.10 The difference between these values (Column 4) is the effect which the program had, as measured by the equation under the three alternatiVe wage rate assumptions. This measured effect is then subtracted from the observed (historical) "with program" values (Column 1) to give the prediction of what the production index values would have been without the program. Column 6 shows the real dollar value of the effects on production, while Column 7 shows the nominal dollar values. The last column shows the direct program costs, also in nominal terms. Figure 12 illustrates the program's effects on production by plotting the observed "with program" indices (from Column 1) and the predicted "without program" indices (from Column 5). Tables 15 and 16 present the results of the application of the two alternate employment equations to predict the effects of the program on employment. Column 1, in each table, gives the observed (historical) "with program" employment values. Column 2 lists the estimated produc- tion indices (from Column 5 of Table 14) that were used in the employ- ment equations to estimate the with and without program employment levels (Columns 3 and 4, respectively), under the three alternative wage rate assumptions. The differences between these values (Column 5) is the effect of the program on employment, as measured by the alternative equations. These values are then subtracted from the observed values (Column 1) to give the predictions of what employment would have been without the program. Figure 13 illustrates the effects by plotting the observed employment (from Column 1. Tables 15 and 16) and the predicted 10Appendix 0 explains the procedure for computing these values with the production equation. 86 Figure 12. The Index of Physical Production in Puerto Rico, 1961 to 1974 and the Predicted "Without Program" Index, 1970 to 1974. Index (FY 60-63 8 100) 1051— 1001—...— 95*- 90t' Observed "Wi th Prograni' 85- \ 80’ Predicted I ‘t, I'Without Progranfl ‘\ ..... 75" L l L P I 1 ii Ll I FY6162 63 64 65 66 67 68 69 70 7172 73 74 Figure 13. Employment of Farm Laborers in Puerto Rico, 1961 to 1974 and . the Predicted "Without Program" Employment, 1970 to 1974. Employment('OOO) 90' 80- 70“ 60- Observed ”With Progranfl '\ 3 "r 7 20 Predicted "n ”Without Progranfl 10 ..... Eq. 6.19 Eq. 6. 20 l 1 l l L I l I l l l l J L_. FY61 62 63 64 65 66 67 68 69 70 71 72 73 7’4 87 employment levels under each of the alternative equations (Column 6, Tables 15 and 16). The following chapter summarizes the analysis of this dissertation, presents the evaluation of the program and the conclusions which may be drawn from the analysis and suggests particular needs for further research. CHAPTER VIII SUMMARY AND CONCLUSIONS Summary Puerto Rico's rapid economic growth in recent decades has been accompanied by a significant change in the structure of the island's economy. As a result of its special relationship with the United States, and its own government's development programs, the manufactur- ing, government and trade sectors now dominate the economy. Although the agricultural sector continued to expand until the early 19605, it then began a sharp decline in production such that by 1974 its physi- cal production was only 82% of its 1960-1963 average. The decline of the agricultural sector can be attributed primarily to increases in wage rates that accompanied the industrialization process and the inabil- ity of the sector to adjust to these higher wage rates. In 1940 Congress exempted Puerto Rico from the universal minimum wage rate of the Fair Labor Standards Act of 1938 and in its place Congress created a more flexible alternative system which established minima on an industry by industry basis. This system was a compromise which would avoid any unfair advantage for the Puerto Rican producers but also would not result in severe economic dislocations on the island had the universal minimum wage rate been imposed on Puerto Rico. In 1941 the insular government established a similar system, and in 1956 expanded its coverage to industries also covered by the federal minimum. 88 89 Vigorous action under both systems resulted in rapid increases in mini- mum wage rates. For the 1952 to 1962 period, Reynolds and Gregory have found that increases in wage rates for farm laborers were " probably attributable in good measure to the island's minimum wage law . . ." (Reynolds and Gregory, p. 62). The available data for the period 1962 to 1969, when the subsidy program began, supports this finding and leads to the conclusion that, in the agricultural sector, minimum wage rates were above equilibrium wage rates, and therefore, the levels of employment in the sector were determined solely by the farm operators' demand for labor. At least since the turn of the century, when the United States acquired Puerto Rico, the agricultural sector has been dominated by three export crops--sugar cane, coffee and tobacco. Its free trade status with the United States has permitted a reliance on the U.S. market for its foodstuffs, thereby removing any pressure on the agricultural econ- omy to produce significant amounts of food for an ever increasing local population. The virtually complete dependence of the agricultural sector on the export crops, and.in turn, the dependence of these crops on large amounts of low wage labor, made it extremely difficult for Puerto Rican agriculture to make the adjustments to the imposition of large increases in the minimum wage rates which have accompanied the rapid growth of the industrial sectors. 0n the (one hand, neither coffee nor tobacco lend themselves to the substitution of machines for labor, and the topography of the sugar cane areas and the organizational structure of the plantations and their labor unions have limited the mechanization of sugar cane. On the other hand, the Puerto Rican agricultural sector is a price-taker in virtually 90 all of its relevant markets. It has a small share of the market for its export crops, it consumes a small part of the U.S. foodstuffs mar- ket, and it also must import most of its non-labor inputs. About the only markets in which prices are determined on the island are for a few special tropical fruits and vegetables and for fresh milk and eggs. A comparison of the rates of increases for product and factor prices since 1961 is consistent with the response in agricultural pro- duction. Between 1961 and 1963, when agricultural production reached its highest level, product prices rose faster than factor prices. Between 1963 and 1966, product prices decreased in absolute value while factor prices continued to increase, and production began a sharp de- cline. Between 1966 and 1969, product prices began increasing again and the rate of decline in production slowed. Table 5, page 27, also shows that the wage rate for agricultural labor experienced a much greater increase between 1963 and 1969 than did the prices of feed or fertili- zer, which were the only other available factor price data. The only subsector to experience an increase in production was the animal products subsector. Although the production of fresh milk and eggs does require the use of imported feeds, it does not require large amounts of hired labor and the product prices are detenmined to a greater extent in the local market since the perishable nature of these products makes them relatively freer from competition from imports. The Wage Rate Subsidy Program for Agricultural Laborers in Puerto Rico was implemented for the express purpose of reviving the agricul- tural sector by retaining the sector's labor force through increasing the laborers' wage rate, but without raising the farm operators' hourly wage costs. The wage rate subsidy program operates by guaranteeing a 91 wage rate to the hired farm laborers. The farm operators are required to pay this wage rate and are reimbursed on a quarterly basis for the difference between the guaranteed rate and the minimum rate as estab- lished by the Minimum Wage Board. The guaranteed wage rates were estab- lished for various categories of farm enterprises and some are paid on a piece-rate basis rather than on an hourly basis. A simple theoretical analysis, assuming competitive conditions, shows that in the short run the wage rate subsidy program, as imple- mented in Puerto Rico, would have no effect on either the farm opera- tor's decisions on the level of production or on his demand for labor schedule. This result is due to the fact that the farm operator is faced with the same factor prices that he would have paid if the program had not been implemented. In the intermediate run, however, the subsidy program maintained the cost of labor at the initial wage rate while the prices of the re- maining factors of production continued to rise. The increases in wages received by laborers were completely absorbed by the program. This cheapening of labor relative to the other factors of production should lead to a more labor intensive technology (the substitution of labor for the other factors of production at any level of output). Each year, as the prices of the other factors continued to rise, the relative use of labor should have continued to increase as long as the program held the labor cost constant. Thus, both production and employment under the subsidy program should be greater than production and employment without the program. In absolute terms, however, both production and employment would continue to decline if the increases in marginal costs which result from the price increases in the other factors of production 92 (in spite of any substitution of labor) are not offset by increases in product prices. I In an effort to determine the effect of the wage rate subsidy pro- gram on production and employment in the agricultural sector, a two equation model was developed to estimate the net effect of the program on production and employment. The model consists of an Index of Pro- duction equation, which is recursive (based on the model of adaptive expectations), and two alternative Employment equations, one using current values of the production index and the other using lagged values of the production index as explanatory variables. The Index of Production Equation The specification of the Index of Production equation is based on two characteristics of the decision making process in agricultural production. First, the decision to commit resources to agricultural production must be based on the farm operator's expectations of future factor and product prices. This is a direct result of the amount of time required in the biological process of agricultural production. Secondly, the process of investment and/or disinvestment in the fixed assets of agricultural production places severe limits on the ability of a farm operator to make significant changes in his operation in the short run in response to changing price relationships. This is parti- cularly true in the case of perennial crops such as sugar cane or coffee in which a large investment is required to plant a stand. Once planted, the stand becomes a fixed asset and will be cultivated and harvested as long as the value of the production less the variable costs 93 of cultivation and harvesting exceeds the stand's salvage value (possi- bly negative) or the opportunity cost of the land, whichever is greater. The dependence on price expectations suggests the use of the "adaptive expectations" model which is estimated statistically by using the lagged values of the dependent variable (the production index) as an explanatory variable. The use of the lagged dependent variable as an explanatory variable also has the effect of placing limits on the year to year changes which may occur in the dependent variable in that 1 its high degree of correlation with the dependent variable gives it a large influence in determining the value of the dependent variable relative to the yearly price variables, thereby dampening the effect of the yearly price variables. Thus, the equation performs well in pre- dicting the values of the index of production, in spite of the concep- tual problems which are discussed in Chapter VI. The serial correlation in the error terms, which is introduced by the lagged dependent variable, required the use of the Cochrane-Orcutt procedure to obtain consistent estimates of the regression coefficients. The results of the statisti- cal estimation procedures may be found in equation 6.11 on page 67. The Employment Equations The mutual dependence of the production and employment decisions of the farm operator on the farm's profit function suggests the use of the adaptive expectations model for the employment equation also, again reflecting the sequential nature of farm decision making. Since neither of the two alternative adaptive expectations models gave results which were fully consistent with theoretical expectations, a variable which would retain the sequential nature of farm decision making was needed 94 to replace the lagged employment variable in the formulation of this equation. The production index variable was selected for this purpose. The simultaneous nature of the production and employment decisions suggests the use of the fitted or predicted production index--in effect, a two-stage least squares procedure. This formulation resulted in a positive sign on the wage rate coefficient--implying a greater demand for labor at higher wage rates--and was rejected for this reason. Since the production and employment decisions are mutually depend- ent upon the profit function and are not interdependent (as the use of the two-stage least squares would imply), the employment equation was estimated using the observed values of the production index. This for- mulation gave a low significance on the wage rate coefficient and exhibited a moderate level of serial correlation in the disturbances. The use of the lagged production index variable yielded better statis- tical criteria, even though the logical considerations would favor the use of the current production index variable. In view of the slight conflict between the logical and the statistical considerations and in view of the small differences in the predicted values which result from the two equations (especially in the last years of the data), both equations are presented and used in the application of the model. The results of the statistical estimation procedures may be found in equations 6.19 and 6.20 on page 73. Estimation of the Program's Effects The type of evaluation design used in this research is a modifica- tion of the "time trend projection of pre-program data versus actual post-program data." Rather than using only the nine years of 95 pre-program data in the estimation of the model's parameters, it was possible to use all fourteen years of available data since the effect of the program on the explanatory variables was only to hold the wage rate variable constant. The other price variables are independent of the program due to their exogeneity to the agricultural sector. The application of the model requires only an assumption concerning how the wage rate variable would have increased if the program had not been implemented. Since the average annual increase in the wage rate vari- able in the nine years preceeding the program was 5.96%, it was assumed that this time trend projection would have continued. In order to teSt the sensitivity of the model to this assumption, the model was also applied with assumed annual wage rate increases of 5% and 7%, each about 1% above and below the rate of increase in the primary assumption. In recognition of the errors which remain in the model's equa- tions, the model is applied by finding the difference in the predicted values of production and employment in both the with and without pro- gram cases, rather than the differences between the observed withv program and the predicted without program values. This procedure en- sures that the effect of the program will not also include the errors of the model. These differences, then, are the net effects of the program on production and employment. To find what the absolute values of production and employment would have been without the pro- gram, one simply subtracts these differences from the observed values. The estimated effects of the program on the production index and employment and the value of these measures which would have obtained in the absence of the program, under the primary assumption regarding 96 the wage rate variable and the first employment equation, are summarized in Table 17. Table 17. Estimated Effects of the Program, Assumption #1 and Employment Equation 6.19. PRODUCTION INDEX EMPLOYMENT bserved Estimated Predicted bserved Estimated Predicted Year Index Effect Index Employment Effect Employment With of without with of without Program Program Program Program Program Program 1970 .864 O .864 38,484 1,063 37,421 1971 .852 .012 .840 34,147 3,241 30,906 1972 .834 .028 .806 30,909 5,352 25,557 1973 .812 .045 .767 27,310 6,196 21,114 1974 .821 .060 .761 30,443 7,968 22,475 Source: Tables 14 and 15. The immediate, though small, impact on employment in the first year may be reasonable in spite of its conflict with the simple short run theory. This difference may be explained by noting that much of the labor used is seasonal harvest labor and if the prices of the other factors of production were rising noticeably during the first year, then a relative cheapening of labor would be noticed and the substitu- tion of additional quantities of labor in the tasks of cultivation and harvesting would result, even without any change in the level of pro- duction in the first year. The estimation of the program's effects are limited to the five year period of 1970 to 1974 since projections beyond 1974 are outside the range of the data with which the model was estimated and since the effects of the program are cumulative in nature due to the program's interaction with the structural changes in the agricultural sector. 97 In particular, the projection of without program employment would reach negative values in a short period of time. An estimate of the cost that would have been required if the pro- gram were to have halted the decline in production by lowering the farm operators' wage costs can also be calculated with the model. For the purposes of this hypothetical case, it will be assumed that the increases in wage rates received by the laborers under the program are unchanged. A trial and error method was used to find the values of the WAGE vari- able (actual wage costs to the farm operators) that would result in a production index value of .864 for the years 1971 to 1974 (due to the one year lag in response, the 1970 index would not be affected). These values may then be used in the employment equation to find what levels of employment would have resulted with these wage costs. Both of these sets of values are listed in Table 18. These wage cost values can be translated into very rough estimates of total program costs by increas- ing the actual direct program costs by the percentage increase in the number of dollars of the laborers' median weekly wages subsidized under this case over the number of dollars of the laborers' median weekly wages that were subsidized under the actual program. The number of dollars of the laborers' median weekly wages that were subsidized under the program is simply the amount they received less the fixed wage costs to the farm operators ($18.88). The increase in the number of dollars of the laborers' median weekly wages that would be subsidized in this hypothetical case is the fixed costs ($18.88) less the wage costs estimated by the use of the production equation if the production index is to be fixed at .864. Table 18 also lists the actual direct program costs, the actual dollars of subsidized median weekly wages, 98 NNo.m~ N~.N~ No.mN Nwe.om NN~.NN om.mN NNNN NN¢.N~ Ne.NN mN.oN NNN.NN mem.mm om.mN «NNN wem.eN NN.NN mm.N eom.NN emN.Nm mm.eN NNNN . _ mme.m N Nm.e~ No.o N oom.o » NNm.wm w m~.NN m m ONNN NNNN.NV NNNN.NN NNNN.N weasees. NNNou uo~NuNNo=m uoNNuNNosm NNNou oN NNNou woe: see sesmoss Nome: sN Nome: No Eesmoss usosonQEN NNxooz seNuoz > uoNeENNNu omeososN N N Neopos Neopos uoNoNuoss uoNeENNNm .Nosos NNNN one Ne xoesN eoNNoaeoea NNN NNNNNNNNN 6N eoeeo sN uoNeososN NN zuNNoom esp sows; sN omeo NeoNNosNooN: osa sow Npmou Eesmoss uouesmumm .mN oNoeN 99 the percentage increase in the subsidized median weekly wages and the resulting total program costs for this higher level of subsidization. A more accurate estimate of what the program costs would be in this case would require detailed knowledge of the relationships among the program's parameters (the wage rates and hourly subsidy rates), the median weekly wage rate variable and the number of hours worked at each of the wage rates in the different subsectors. Although a comprehensive evaluation of the benefits and costs of the program is beyond the scope of this research, several partial com- parisons may be made with the information produced by the model. First, the value of the increased production may be compared with the direct treasury costs of the program. Secondly, the cost per man-year of increased employment can be calculated. These comparisons are displayed in Table 19. Taken individually, such simple comparisons implicitly assume that the benefits (either production or employment) are the only benefits and that the direct treasury costs are the only costs. Table 19. Selected Program Evaluation Indicators. Nominal Value of Direct Program Cost Per Man-Year of Year' Increased Production Treasury Costs Increased Employment ($'000) ($‘OOO) 1970 0 4,300 4,045 1971 3,288 11,364 3,506 1972 7,998 17,364 3,249 1973 14,009 20,481 3,305 1974 32,737 23,336 2,929 Totals 58,032 76,872 -- Eyerage -- -- 3,227 100 A comprehensive evaluation of the program's benefits and costs should consider the following types of effects. A. Monetary Benefits and Costs 1. 2. The increase in agricultural production. The costs of the resources used in producing the additional agricultural output. The net benefit of increased production in the other sectors that can be properly attributed to this program (the indirect effects). The benefits of increased tax revenues resulting from the increased production (both direct and indirect). The benefits of savings in other government programs such as urban public infrastructure, public security and welfare. The benefits of the increased incomes to the farm operators and laborers. B. Non-monetary Benefits 1. 2. 3. The increased employment of farm laborers. The improved welfare of the families of the now employed farmers and laborers, including such things as their health, nutrition and education. The potential reduction in urban social tensions if the indi- viduals would have migrated to the urban areas in the absence of the program. Conclusions The analysis of this research shows that the wage rate subsidy pro- gram did have measurable effects on both agricultural production and employment in Puerto Rico. It accomplished these effects by holding the farm operators' hourly wage costs fixed at the 1969 minimum wage rate levels while the prices of the non-labor inputs continued to rise. Although both of these indicators continued to decline, the decline was less than that which would have occurred without the program if the wage costs had continued to increase. 101 The analysis of the way in which the program achieved its effects has an important implication for Puerto Rico's minimum wage policy. Rather than instituting the wage rate subsidy, the same effects on pro- duction and employment could have resulted from a policy whereby the Minimum Wage Board would not increase the minimum wage for agricultural laborers. Thus, the program could be seen simply as an income transfer program for agricultural laborers. The high probability that such an alternative policy measure would be politically unacceptable in Puerto Rico may be seen as evidence of a revealed preference for higher laborer incomes (in spite of higher unemployment rates) over the revitalization of the agricultural sector as it is today. The objective of maximum labor force utilization (minimum unemploy- ment) at higher laborer incomes and the revival of the agricultural sector in Puerto Rico appear to be incompatible, given the island's limited base of agricultural resources and its ability to import, at low prices, the foodstuffs required by its population. If these natural resources are to remain in production, rather than lie idle while food is imported, significant structural changes will be required in the agricultural sector. These changes will involve a shift out of the traditional, and labor intensive, export crops and into the production of products which can compete on the local market for their labor re- quirements. The most likely candidates will be fresh fruits and veget- ables and animal products (principally milk and eggs) since these products are highly perishable. This perishability adds to the cost of transportation to the island and should therefore allow a competi- tive advantage to local production. 102 Areas for Future Research A number of areas for future research that became apparent during the process of this research are as follows: 1. Extensive farm management research is needed to identify the products that would be best able to compete on the local market for their labor requirements. A better understanding is needed of the apparent tendency for indi- viduals to leave the labor force entirely after the seasonal jobs in agriculture are over and to re-enter the labor force when these .seasonal jobs become available again. Who are these individuals and why don't they seek employment in the other sectors? There is a need for the collection and dissemination of price data on the principal agricultural production inputs. The absence of this type of data would continue to severely limit effective analy- sis of the agricultural economy. More effective means of reducing unemployment are also needed. The agricultural sector is now so small that further farm-to-city migration will be minimal. Additionally, the sector could not re- absorb very many of those that have already left. A particular area for study is the unemployment effects of the minimum wage policies. There appears to be a conflict between the objectives of increasing labor income and of maximum employment of labor. APPENDICES APPENDIX A INCOME MAINTENANCE PROGRAMS AND THE WAGE RATE SUBSIDY A Review of Income Transfer Programs The general objective of the current United States public assistance programs is to provide assistance to those persons who are unable to earn an income sufficient to meet the needs of some reasonable minimum standard of living. Old age, physical or mental disability or illness, either permanent or temporary, and women with young children and no husband or an absent one are the principal categories of acceptable causes for the inability to earn an adequate income. Temporary unem- ployment also became an acceptable cause and the Unemployment Insurance program was implemented. Underlying these programs is the philosophy held by our society that people should obtain their income through gain- ful and productive employment unless,for reasons beyond their control, they were unable to do 50. All of the programs have been implemented with rules and regulations which were intended to limit the abuse of the programs by people who were considered to be able to earn their livelihood or to depend on someone that had a moral obligation to pro- vide for them. The rules and regulations for the programs have become more extensive and more specific as time goes by in order to provide even greater accuracy in deciding who will qualify and who will not. These rules and regulations reflect a wide range of Opinions as to how the poor should be assisted. Many of the rules are paternalis- tic toward the recipient and grow out of the desires of the donors, or 103 104 taxpayers, to impose their moral judgements about how a person should live as a condition on the provision of public assistance. The classi- Cal example is of the person who will give money to a destitute alcoholic only if he will use the money to buy himself or his family a good meal or some other "approved" use. Another expression of this paternalism is the reduction of the grant when the recipient receives income from other sources, either earned or unearned, and usually by an equal amount. Nevertheless, there exists a growing concern about the demonstrated effects of the paternalistic rules on the lives of the recipients. Wel- fare Rights Organizations have formed to protest the "invasion of privacy" which accompanies the information-gathering procedures used to enforce the rules. The reduction of the grant by an amount equal to the amount received by a recipient from other sources has been shown to be a severe disincentive for a recipient to reduce his dependence on public assist- ance. The recipient NNN) does not have an opportunity for a challenging job or one that pays well will naturally ask himself: "Why work when I can get the money without working?" Since most states operate their own programs and set their own benefit levels, there is a wide variation in program availability and benefits between geo-political jurisdictions which may be causing interstate migration, thereby placing an even greater tax burden on the residents of states offering higher benefits. This philosophy that people should obtain their income through employment is based on the belief that the American economic system is capable of providing the opportunity to be gainfully employed to all who seek employment. The basic law of supply and demand was believed to be valid for the labor market. Still following this belief, the War on Poverty of the 19605 placed a major emphasis on programs to expand the 105 opportunities for the poor by providing training and counseling services and by promoting the hiring of the disadvantaged. But the War on Poverty did not provide the results which were necessary to maintain the belief in the system and the belief began to wane,and many people now conclude that our economic system is not, or is no longer, capable of providing an adequate income to a significant segment of our population or even of providing employment to all who seek it. As evidence that our economic system is not capable of providing an adequate income for the employable, President Nixon's Commission on Income Maintenance Programs found that 70% of the non-aged heads of households of poor families worked for part of the year in 1966. Additionally, one-third of the poor live in fami- lies in which the family head works throughout the year. The Commissuwi came to the conclusion that in many cases the possibility for improve- ment is not realistically within the power of the poor (President's Commission, pp. 2-3). The awareness of the undesirable effects of the paternalistic characteristics of public assistance programs, the concern about the inequalities between the states, and the belief that our economic system is not capable of solving the problem of the employable poor have led to attempts to design a broad program, with a minimum of behavior-restricting rules, which could be made available on an equal basis to all of the poor, no matter where they live. Two basic concepts have received attention as a means of improv- ing the present system of public assistance. One is the Negative Income Tax (NIT), and the other is the Wage Rate Subsidy (WRS). The following sections present the basic formulations of the two programs. Subsequent sections present lessons from the Puerto Rican experience with a WRS, 106 a theoretical comparison of the effects of the two programs and a propo- sal of a plan which combines into one formula both of the NIT and the WRS concepts. The Negative Income Tax The NIT guarantees a minimum income to the recipient by providing a basic grant. For persons with some non-grant income, the basic grant (guaranteed income) is reduced by an amount proportionate to his non- grant income. An optional feature is to exclude or "disregard" certain amounts of an individual's non-grant income before applying the propo- tion by which the grant is reduced. The general form of the NIT equation is: Y = WH + (G - a(WH - 0)) for: (G - a(WH - D)) 2 O, and O < a s 1. where: Y = net income G = guaranteed income (the basic grant) W = wage rate a = proportion of income by which 'G' H = hours worked is reduced (the marginal tax rate 0 = disregarded income on earned income) WH may represent both earned and unearned non-grant income. This equation is often stated as: Y G + (1 - a)WH for: Y s G/a, and Y WH otherwise, ignoring the disregarded income feature. The Wage Rate Subsidy, The WRS, in its basic form, does not guarantee a specific level of income. It increases the recipient's hourly wage rate by providing a bonus, for each hour worked, computed on the basis of the recipient's actual wage rate. The general form of the WRS equation is: 107 Y = (W + b(B - W))H for: b(B - W) a O, and O 6 b s 1, Y = WH otherwise, where: Y = net income b = the proportion of the differ- H = hours worked ence between the breakeven wage B = breakeven wage rate rate and the actual wage rate W = actual wage rate to be provided as a subsidy (the marginal tax rate on wage rate increases) This equation is often stated as: Y = (g + (1 - b)W)H for: W 6 g/b, where 'g', the guaranteed wage rate, is equal to (B-b). This form of the equation shows the subsidy as the guaranteed wage rate less a portion of the individual's actual wage rate rather than as an addition to his actual wage rate. Lessons from the Puerto Rican Experience Although the Puerto Rican experience represents an extreme case of the general wage rate subsidy concept, several lessons for a broader application of the subsidy can be found. First, the system of implementation should directly involve both the employer and the employee-recipient, so as to reduce the Opportunity for fraudulent claims. In particular, both must be aware of and af- fected by the wage rates which are declared as the actual wage rate and the subsidy rate (or the rate eventually received by the recipient). Although Puerto Rico used a 100% marginal tax rate and the minimum wage rate as the general bargained (actual) wage rate, an incentive for the employer to under-declare the actual wage rate exists also under a fractional tax rate system. As an example, assume that the employer agrees to pay the recipient $3.75 per hour, including the subsidy. Under the formula: 108 Y = (W + b(B - W))H Where B = $5 and b = .5, the values would be: Y = (2.5 + .5(5 - 2.5))H = (2.5 + 1.25))H = $3.75H. However, the employer can declare an actual wage rate of, say, $2.00 per hour so as to receive a subsidy reimbursement of $1.50 instead of $1.25. The recipient still receives his $3.75 per hour and may not even care or know about what is happening. On the other hand, the employer is also a necessary party to the implementation of the system in order to avoid widespread fraud by recipients if they were the sole source of information regarding wages earned and hours worked. One possible system of implementation that could work to reduce program abuse would be for the employers to pay by check, with the requirement of a check stub reporting hours worked and wage rate paid. The employee-recipient could then present his (endorsed) check and stub to the program office in exchange for a check which would include the subsidy payment. However, adequate disincentives would still be required to avoid the potential collusion between employer and employee whereby the reported wage rate is understated and the reported hours work correspondingly overstated. The employer, in this case, would pay the same amount in wages, but the recipient would be eligible for a larger subsidy payment. This system would also reduce the oppor- tunity for claims on non-existent workers since the employer would not be reimbursed by the program. Secondly, if the program is operated by means of reimbursements to the employers, a decision must be made as to whether or not the subsidy payments will be considered as wages which would be subject to the pay- ment of Social Security and other wage related taxes. 109 Comparison of NIT and WRS Effects The WRS is clarly of no benefit to those poor that have no hourly wage income, while the NIT would provide these people with a guaranteed level of income. The WRS will generally be preferred by people who work an intermediate nUmber of hours, e.g., more than half-time, but receive low hourly wage rates, while people working fewer hours at higher wage rates will generally prefer the NIT. Clearly, there will be a large number of people who would qualify for benefits under either program. The WRS and NIT have quite different incentive characteristics with (respect to both working more hours and seeking increases in wage rates from employers. The NIT "taxes" increases in earned income whether this results from increasing hours or a higher wage rate. For a constant wage rate, the NIT taxes income from increasing hours by the marginal rate of 'a'. Similarly, for a constant number of hours worked, the in- creases in the wage rate are taxed at a marginal rate of 'a'. On the other hand, the WRS taxes only the increases in wage rates at the mar- ginal rate of 'b' and not the increases in hours worked. In both cases, the incentive effects are relative to what the incentive would have been for the person not benefiting from the program--although the effects are quite real and readily visible since an individual will be very aware of his increase in hours or wage rates and the reduction in his benefits caused by the change in hours or wage rates. Unless the mar- ginal tax rates are greater than one, the person will always receive a higher net income from working longer hours or negotiating a higher wage rate. Due to the mathematical relationships among the programs' para- meters, a reduction in the marginal tax rates will enlarge the coverage 110 of the programs--and thereby increase the cost of the programs. For the NIT, the breakeven non-grant income level (at which point the grant becomes zero) is: Yb = G/a. 'Yb' clearly increases as 'a' is reduced in order to improve the incentives to work. Although 'G' can be reduced in order to maintain a lower 'Yb', lowering 'G' also lowers the level of. support provided. For the WRS, the effective guaranteed minimum wage rate is: g = B-b (b s 1). Thus, reducing 'b' in order to increase the incentive to negotiate higher wage rates will either lower the guaran- teed minimum wage rate (lowering costs and support) or require an increase in the breakeven wage rate 'B' (raising costs and coverage). Although recent theoretical comparisons of the NIT and WRS plans have not been able to Show the superiority of one over the other, they have pointed out some of the advantages and disadvantages of each. Among those who have studied the plans are Barth and Greenberg, Garfinkel, Kesselman, Schlenker, and Zeckhauser. Their findings are limited by the static nature of the analyses and all of the articles point out that an empirical analysis is needed to show how the many sets of preferences among the recipients interact with each plan and to answer the dynamic questions of how the labor supply effects will interact with the demand for labor and how the plans will affect the human investment decisions of the recipients. Zeckhauser limited his analysis by using a "repre- sentative" recipient, while the others used multiple preference sets by comparing the plans for specific groups such as: the disabled, the low-wage-rate worker, those unable to find employment and those with backward-bending labor supply curves. Taken together, these analyses show that the WRS is superior to the MIT in terms of work incentive effects for those people who do not have backward-bending supply curves. 111 The problem of obtaining employment or additional hours through which the incentives may become effective will remain under both programs. Due to the varying effects of each plan on the different groups of people, a general conclusion is that both plans should be implemented. However, Rea (p. 26) found that when both a WRS and the Nixon Family Assistance Plan (an NIT plan) are available, the work incentives are improved but at a large budget cost, most of which goes to those with relatively high incomes. Schlenker recommends the addition of an in- come floor to the WRS (presumably if it is the only plan), but points out that a 100% tax rate would result for those persons who would be receiving a grant to raise their income up to the guaranteed level. This high tax rate results from the discontinuity between Schlenker's two equations: Y W + b(B - W)H for Y 2 G and W 4 B, and Y = G otherwise, where 'G' remains a basic income guarantee grant. A similar discon- tinuity problem appears likely to arise for those individuals eligible under both an NIT and WRS and who must choose between them. If the process of changing from one plan to the other is difficult or time consuming, such administrative barriers may provide special disincen- tives to work or take advantage of training opportunities. Consider, for example, a head of household who is offered a full-time but low paying job and is now receiving benefits under the NIT due to a tempo- rary disability or involuntary unemployment. If the changeover period from the NIT to the WRS is too long, he may find that the short run cash flow problems or high tax rate of the NIT prevent him from taking the opportunity. 112 Both the discontinuity between plans and the probable administra- tive barriers to changing plans lead to the idea of combining the WRS and NIT concepts into one single plan. A plan which overcomes these problems is outlined in the next section. Outline of an Alternative Income Assistance Plan The proposed plan would provide support for the employable poor only--those poor unable to work due to medical or other conditions would require support under a plan or plans which are geared to each indivi- dual's needs. The net income for a person under this plan would be: Y = WH + b(B - W)H + (G - a(WH + b(B - W)H) for b(B - W)H 310. O s a s 1, 05 be 1, and (G -a(WH + b(B - W)H))) 0, where W = actual wage rate b = the fraction of the gap between H = hours worked the breakeven wage rate (B) and B = breakeven wage rate the wage rate (W) to be sub- G = guaranteed income grant sidized a = the fraction of earnings (in- cluding the WRS portion) by which the grant is reduced To avoid continuing the income guarantee feature for someone working full-time at the breakeven wage rate, the following inequality must be observed when selecting the plan's parameters: G:‘ ahB where 'h' is the number of hours which constitute full-time work for the accounting period. If the breakeven wage rate is set relatively low and one wishes to continue the guaranteed income portion for people work- ing full-time at a wage rate above the breakeven wage rate, 'G' must be set larger than the quantity 'ahB'. 113 Since the guaranteed income portion of the equation is a function of the wage rate subsidy portion, the marginal tax rates on hours and wage rate increases are no longer single variable terms. The marginal tax rate on wage rate increases becomes: MTRw = a + b - ab for W s B, and (G - a(WH + b(B - W)H)) > 0 while the marginal tax rate on hours worked becomes: MTR = a + b - ab - 1 ’ 3 b3 h W for W s B, and (G - a(WH + b(B - W)H)) > 0 Since the MRTh is a negative function of the wage rate (W) this plan has the advantage of possessing a progressive tax rate on hours. Thus, a low wage rate individual will be taxed at a lower and possibly negative rate while a high wage rate individual will be taxed at a higher rate. The limiting case is the individual earning the breakeven wage rate (W = B), for whom the MTRh = a. A series of examples using weekly in- come are given in Table A.1. If G s ahB, the guaranteed income portion will terminate before or at the point that the recipient works full- time at the breakeven wage rate. When the guaranteed income portion does become zero, the marginal tax rates become: MTRw b for W15 8, and _b(B _ w) (G ' a(WH + b(B - W)H)) = 0 MTRh If G > ahB, the wage rate subsidy portion will terminate before or at the point that the guaranteed income portion becomes zero and the mar- ginal tax rates become: MTRw MTR II 011 for W 3 B, and (G - a(WH)) > O h aw 114 Table A.1. Weekly Benefit Levels and Tax Rates for Selected Program Parameters. Break- B a b G even MTR "TR Income w W=$.5 W551. W=$2. =$B. 3 .4 .4 $48 $120 .64 - .80 - .08 .28 .40 3 .4 .5 48 120 .70 -1.10 - .20 .25 .40 3 .4 .6 48 120 .76 -1.40 - .32 .22 .40 3 .5 .4 60 120 .70 - .50 .10 .40 .50 3 .5 .5 60 120 .75 - .75 .OO .38 .50 3 .5 .6 60 120 .80 -1.00 - .10 .35 .50 3 .6 .4 72 120 .76 - .20 .28 .52 .60 3 .6 .5 72 120 .80 - .40 .20 .50 .60 2.5 .4 .4 40 100 .64 - .56 .04 .34 .40 2.5 .4 .5 40 100 .70 - .80 - .05 .33 .40 2.5 .4 .6 40 100 .76 -1.04 - .14 .31 .40 2.5 .5 .4 50 100 .70 - .30 .20 .45 .50 2.5 .5 .5 50 100 .75 - .50 .13 .44 .50 2.5 .5 .6 50 100 .80 - .70 .05 .43 .50 2.5 .6 .4 60 100 .76 - .04 .36 .56 .60 2.5 .6 .5 60 100 .80 - .20 .30 .55 .60 2.5 .6 .6 60 100 .84 - .36 .24 .54 .60 2 .4 .4 32 80 .64 - .32 .16 .40 O 2 .4 .5 32 80 .70 - .50 .1O .40 O 2 .4 .6 32 80 .76 - .68 .04 .40 O 2 .5 .4 4O 80 .70 - .10 .30 .50 0 2 .5 .5 4O 80 .75 - .25 .25 .50 O 2 .5 .6 4O 80 .80 - .40 .20 .50 O 2 .6 .4 48 80 .76 .12 .44 .60 O - 2 .6 .5 48 80 .80 .OO .40 .60 O 2 .6 .6 48 80 .84 - .12 .36 .60 O 630'“ (D W MTRw = marginal tax rate on wage rate increases MTRh = marginal tax rate on increased hours breakeven wage rate the fraction of earnings by which the grant is reduced the fraction of the difference between B and W to be subsidized the guaranteed income grant actual wage rate 115 Summaryyof Plan's Advantages The NIT and the WRS individually provide different coverages and have different disadvantages. When operated side-by-side, budget costs. are high and the discontinuities of changing back and forth between plans present special disadvantages to the recipients. By combining both concepts into one plan, it becomes possible to: 1) cover a wide range of employable poor, 2) eliminate the discontinuities between the plans, and 3) permit a progressive marginal tax rate on hours according to the recipient's wage rate.- Guidelines for Implementation The program may specify different sets of parameters for different categories such as family size, but adult dependents should be required to qualify independently of the family under either the disability plan or this plan. By operating the program on a weekly or bi-weekly basis, a steadier income flow may be provided for persons who can obtain only occasional employment and it may eliminate the need for separate unem- ployment compensation programs, if processing of claims can be accom- plished in a short period of time. Each individual would be required to make an initial application, at which time he or she would be classified according to the criteria for the selection of benefit parameters. Upon approval of the appli- cation, the recipient would be issued an identification card contain- ing the appropriate personal and benefit parameter information. Each period in which an individual qualifies for benefits, he or she would present his or her identification card and paycheck, if any. The paycheck would be required to contain information on the number of hours 116 worked, the wage rate and Social Security and other taxes withheld. The local office would then issue a check for the net income (post-subsidy) in return for the endorsed paycheck. To operate efficiently and reduce the opportunities for receiving multiple payments, a nation-wide computer system with on-line terminals in each office would be required. Each quarter, the program files can be compared with the Social Security tax records to check on unreported income. All income received under the program would be taxable and all recipients should be required to file a return, but the minimum taxable earnings limit should exempt most recipients. A staff of case-workers would be needed to police compliance with the program rules, but this staff should perhaps be separate from those case-workers that investigate the original applications. Various alternatives exist for altering the benefit parameters according to family size or other criteria. For example, if G = ahB for a family of four, then 'G' could be lowered (2 ahB) for smaller families and raised (> ahB) for larger families. Other alternatives would be to raise 'a' or lower 'b' as family size increases while hold- ing G = ahB constant. If age of the head of household is a criterion, one might wish to provide different breakeven wage rate levels (B) in recognition that persons of lesser experience generally receive lower wage rates. APPENDIX B Table A.2. Time Series Data Used in the Estimation of the Model szgfi' AGPROD AGPRCE FEED FERT AGEMP WAGE 1960 .960 1961 .983 .965 93.51 59.51 89,365 11.88 1962 .990 .995 94.79 60.98 89,177 11.95 1963 1.062 1.064 102.31 63.27 82.598 12.20 1964 1.046 1.009 92.86 62.62 80,929 14.93 1965 1.039 1.008 81.72 63.75 68.504 13.73 1966 1.041 .969 112.98 67.28 ‘ 62,057 14.68 1967 .979 1.047 113.86 68.31 54.225 16.68 1968 .922. 1.094 113.21 68.12 50,462 17.80 1969 .877 1.058 124.47 67.90 ‘44.915 18.88 1970 .864 1.074 120.46 68.64 38,484 24.95 1971 .852 1.110 130.35 68.74 34,147 26.23 1972 .834 1.157 140.35 70.37 30,909 29.03 1973 .812 1.261 143.68 77.11 27.310 32.30 1974 .821 2.210 174.67 97 74 30,443 33.23 1975 .796 See Chapter V1 for a description of the data and sources. Farm Laborer Labor Force; Median Weekly Earnings, Laborers, Except Farm; and Agricultural Production Index, 1951 to 1959. Table A.3. Other Time Series Data: Fann Median Weekly Fiscal Laborer Earnings, Year Labor Laborers, Force Except Farm 1961 106,839 $ 25.25 1962 110,363 25.82 1963 103,270 26.97 1964 96,876 31.88 1965 84,890 30.97 1966 76,711 35.00 1967 68,960 38.00 1968 64,831 42.62 1969 55,525 44.88 1970 48,727 51.03 1971 44,333 53.38 1972 39,251 55.95 1973 34,298 61.45 1974 37,396 60.15 Source: P.R. Bureau of Labor Statistics 117 Agricultural Production Index (AGPROD) Fiscal Year Index 1951 .817 1952 .951 1953 .822 1954 .903 1955 .851 1956 .907 1957 .801 1958 .885 1959 .926 Source: Puerto Rico Depart- ment of Agriculture APPENDIX C SPANISH VERSION OF THE STATEMENT OF LEGISLATIVE INTENT IN LAW NUMBER 142 OF JUNE 29, 1969 EXPOSICION DE MOTIVOS La agricultural ha sido y continfia siendo un reng16n muy importante en la economIa de nuestro pais. Sin embargo. _por diversas consideraciones, esta industria estfi pasando por.momentos dificiles. E1 terreno en cultivo se ha venido reduciendo considerablementei los costos de producci6n se han elevado a pesar de que los jornales agricolas son relativamente bajos, los trabajadores escasean porque buscan trabajo en otras actividades mas remunerativas. A través de esta ley. se declare que es politica pfiblica del Gobierno del Estado Libre Asociado de Puerto Rico e1 garantizar a los trabajadores agricolas un ingreso mayor al que ahora reciben, que los induzca a seguir trabajando en la agriculture a fin de aliviar 1a escasez de mano de obra que sufre esta industria. sin aumentar los costos de producci6n. Asimismo se pretende elevar e1 nivel de Vida de dichos trabajadores para que disfruten de una Vida mas plena. Es prop6sito de esta ley, pues. aumentar e1 ingreso del trabajador para as! reducir su absentismo en las actividades agricolas, aumentando de esa manera 1a producciOn agricola. mediante una mas intensa labor de siembra, cultivo y recolecci6n. La situaci6n econ6mica de la empresa agrfcola. sin embargo, no permite a los agricultores pager a1 presente. salarios iguales o mayores a1 ingreso que se desea garantizar a estos trabajadores. Por tales consideraciones y con el objeto de ayudar a la agriculture a recobrar su prosperidad y ponerse en condiciones de pagar salarios iguales o mayores a los ingresos garantizados por esta ley, se establece un suplemento al ingreso que actualmente reciben oiertos trabajadores angcolas. a determinarse de acuerdo a lo dispuesto en esta ley, e1 cuél seré sufragado por el Gobierno del Estado Libre Asociado de Puerto Rico. valiéndose del agricultor como intermediario para hacer llegar al trabajador dicho suplemento de ingreso a través del mecanismo establecido por esta ley. 118 APPENDIX D I COMPUTATION OF FITTED VALUES USING THE PRODUCTION EQUATION The use of the Cochrane-Orcutt procedure to eliminate the effects of serial correlation of the error terms requires the use of the fellow- ing procedures in computing the fitted values of the equation or in computing predictions with the equation. Let F = a + BllnAGPRCE t + lenWAGEt_1 + B31nFEEDt_1 t-1 + B41nAGPRODt_1, and 9t =‘fite0 +‘Bt'1e1 + . . . +‘Bet_1. A and e = lnYt - lnY t t. The fitted values (lnYt) must be calculated in order, beginning with 1962 (t = 0) since all previous error terms (et_1, et-Z’ . . . ’ e0) must be used in calculating the fitted value for any particular year 't'. A A In:0 = F0 . e0 = InY0 - 1nY0 lnY1 = F1 + 91 e1 = lnY1 - 1nY1 A . . A InYt - Ft + gt et - InYt - 1nYt When using the assumed values of the wage rate variable to estimate the without program values of the production index, the actual values of the error terms for years 1962 to 1970 are still used. However, since there are no observed without program index values (Yt) for 1971 to 1974 119 120 with which to calculate the error terms (et), these error terms are assumed to be zero. The lagged production index variable, as predicted by the equation using the assumed wage rate values, beginning with the 1971 predicted index value, are the ones used as the lagged dependent variable. Table A.4 lists a few selected values as an aid in understand- ing the procedures used. Table A.4. Selected Intermediate Computation Results in the Prediction of the Production Index for the Without Program Case. A Year Yt t lYt-I WAGEt_1 Ft gt Yt 1970 .864 8 .877 118.88 -.16292 +.00821 .8567 1971 .852 9 .864 20.01** -.17568 -.00671 .8333 .21677 +.00269 ' .8073 .25492 .00108 1 .7741 .28518 .00433 J .7522 1972 .834 10 .8333* 21.20** 1973 .812 11 .8073* 22.46** 1974 .821 12 .7741* 23.80** + Note: Values in last five columns are rounded. * Predicted Values ((Yt) from last column, previous year). ** Assumed wage rate values. BIBLIOGRAPHY BIBLIOGRAPHY Ballinger, Roy A., ed. "Economic Behavior in the United States Sugar Market," California Agricultural Experiment Station Bulletin Number 859, 1972. Barth, Michael C. and David H. Greenberg, "Incentive Effects of Some Pure and Mixed Transfer Systems," Journal of Human Resources, 6 (Spring 1971), 149-170. Commonwealth of Puerto Rico, Camara de Representantes, "TranscripciOn de los Testimonios Prestadoes ante la ComisiOn de Agricultura de la Camara de Representantes, en la Audiencia Pfiblica celebrada en el SalOn de Audiencia 3 del Anexo de este Cuerpo Colegislador, el 24 de Abril de 1969, sobre 105 P. de la C. 434 y 445: Garantizar a los Trabajadores en la Fase Agricola,” mimeo, 1969. Commonwealth of Puerto Rico, Departamento de Agricultura, Oficina de Estadisticas Agricolas y Estudios EconOmicos, "Anuario de Estadisticas Agricolas," mimeo, various editions. , "Ventas de Abonos en Puerto Rico, 1950/51-1973/74," mimeo, March 1975. Commonwealth of Puerto Rico, Departamento de Agricultura, Oficina de Ingresos Garantizados a Trabajadores Agricolas, "Compendio Estadistico," mimeo, October 1972. , "Compendio Estadistico," mimeo, July 1974. , "Informe Sobre los Ingresos Suplementarios a los Trabajadores Agricolas Provistos por las Leyes Num. 142 de 1969 y Hum. 20 de 1972," mimeo, August 1973. Commonwealth of Puerto Rico, Junta de Planificaci6n, "Informe Econ6mico al Gobernador," mimeo, various editions. Friedlander, Stanley L., Labor Migration and Economic Growth, Cambridge: The MIT Press, 1965. Garfinkel, Irwin, "A Skeptical Note on 'The Optimality' of Wage Subsidy Programs," American Economic Review, 63 (June 1973), 447-453. Griliches, Zvi, "Distributed Lags: A Survey," Econometrica, 35:1 (January 1967), 16-49. Hatry, Harry P., Richard E. Winnie and Donald M. Fish, Practical Program Evaluation for State and Local Government Officials, Washington, D.C.: The Urban Institute, 1973. 121 122 Henderson, James M. and Richard E. Ouandt, Microeconomic Theory, New York: McGraw-Hill, 1971. Kesselman, Jonathan, "Labor Supply Effects of Income, Income-Work, and Wage Subsidies," Journal of Human Resources, 4 (Summer 1969), 1962. Klein, Lawrence R., An Introduction to Econometrics, Englewood Cliffs, New Jersey: Prentice-Hall, 1962. . Kmenta, Jan, Elements of Ec0nometrics, New York: Macmillan; 1971. Koenig, Nathan, A Comprehensive Agricultural Program for Puerto Rico, Washington, D.C.: U.S. Government Printing Office, 1953. Mundlak, Yair, "On the Microeconomic Theory of Distributed Lags," Review of Economics and Statistics, 48:1 (February 1966), 51-60. Pico, Rafael, The Geography of Puerto Rico, Chicago: Aldine Publishing Company, 1974. President's Commission on Income Maintenance Programs, Poverty Amid Plenty: The American Paradox, Washington, D.C.: U.S. Government Printing Office, 1969. Rea, Samuel Arthur, Jr., "The Supply of Labor and the Incentive Effects) of Income Maintenance Programs," unpublished Ph.D. dissertation, Harvard University, 1970. Reynolds, Lloyd G. and Peter Gregory, Waggs, Productivity, and Indus- trialization in Puerto Rico, Homewood, Illinois: “Richard D. Irwin, Inc., 1965. Schlenker, Robert E., "Optimal Mechanisms for Income Transfer: Note," American Economic Review, 63 (June 1973), 454-457. United States Department of Labor, Employment Standards Administration, Wage and Hour Division, "Report, Findings of Fact, and Recommen- dations of Industry Committee No. 122 for the Agriculture Industry of Puerto Rico," mimeo, 1974. Zeckhauser, Richard J., "Optimal Mechanisms for Income Transfer," American Economic Review, 61 (June 1971), 324-334. mllllllllllllllll lllllllllli ll lllllllllm 3 1293 03011 1217