V¢. ~-\ mamas RDLLIY IoR EASI AIR?) 1" 3 ,5 ’V .IH SPECIAL REFERENCE Io UGANDA j :~~ : ,; Thesis far the Degree Rf PH D. CHIGAN STATE UNIVERSITY, ‘ V a . , . ' . ' v . . . V... H'.:;_ -.""../.n. r.,. ialrtc¢;ov‘.fl 1(- 2. In" . ,,-.,,-,.. r ‘7'":"51."’f'wrr 3&fl:"3’"7{v~" .' "’llr.‘:u;:f~'1;7" : 5'17? - Iv‘ , . f':./;r”,,-: i-m’" -o;_-}~-A“ tr-“Ir... m c hi1 0K - '3‘. 1. 3a .y . 'r‘L; a kiln 1434’]. State University This is to certify that the thesis entitled ‘ INCOMES POLICY FOR EAST AFRICA WITH SPECIAL REFERENCE TO UGANDA presented by Asefa Teferri ,has been accepted towards fulfillment of the requirements for Ph.D. Economics degree in 0-7639 WM \ ABSTRACT INCOMES POLICY FOR EAST AFRICA WITH SPECIAL REFERENCE TO UGANDA BY Asefa Teferri The Unemployment Problem Since the 1950's, rising unemployment has become a major problem in Uganda or East Africa in general. Growth in non-agricultural output, fOStered primarily through a strategy of import-substituting industriali- zation, failed to generate significant employment opportunities. Recorded employment stagnated or, at times, declined. When this dismal record is combined with a rapid growth in population (in the range of 3 to 3.5 percent) and labor force, the resultant unemployment gap assumes signifi- cant proportions. Moreover, the neglect of the agricultural sector as a major area of concern in planning for development and the consequent migration of school-leavers and unskilled youth into urban areas has led to an urban labor force and unemployment growing at a rate much faster than for the economy as a whole. Given prevailing conditions, the prospect is that the unemployment problem will be an even more significant issue in the coming decades with implications having economic, social and political dimensions. The dissertation is, therefore, based on the general premise that reappraisal of economic development objectives, priorities and policies is essential to chart a more stable development path for the ugandan or East African economy. Asefa Teferri An.Ana1ysis of the ugandan Data: Major Findings The Ugandan data over the period mid-1950 to mid-1960, supplemented by Kenyan and/or Tanzanian studies or data necessitated by some gaps in Ugandan data, was used to quantitatively estimate the trends in the magni- tude and rate of growth of demand, supply and unemployment of labor. Population growth (as an index of labor force growth) and changes in recorded employment in some major Ugandan towns are used for a rough estimate of the level and rate of growth of urban unemployment. The find- ings indicate that urban labor force and unemployment have grown much faster than for the economy as a whole. An analysis of the behavior of wages over the same period revealed that average money and real wages of Africans in the economy as a whole rose by 15 and 10 percent respectively. The rates of growth in the public sector are found to be higher than in the private sector. While both unskilled wages and salaries have grown over the period, it was a much faster rate of growth in the former that accounted for a significant proportion of the rise in average wages. Moreover, this growth in wages and salaries is much more than the respective productivities would permit. Given a stagnant agricultural income, the rural-urban differential for comparable labor has widened significantly. These findings are contrary to what one would expect in a situation of rapidly growing unemployment -- namely, declining wages and rural-urban income differentials. Hence, the major pr0position advanced and discussed in the dissertation as a basis for incomes policy appropriate for Uganda or East Africa is that the government itself, as the dominant employer and through the statutory determination of minimum.wages, has been the cause of the rising average wages, rural-urban income differentials and unemployment. Asefa Teferri A Comprehensive Incomes Policy The dissertation recognizes ad-hoc policies already tested (e.g., tripartitite agreements to expand urban employment in return for accept- ance of wage freezes by unions) or which can be introduced (e.g., wage subsidy) to generate urban employment opportunities. But these measures are shown to leave intact the prevailing factor price distortions so that while employers may appear to expand employment in response to these policies, they may also decrease employment by failing to make up for the loss made by normal attrition of the labor ferce. Moreover, the urban employment opportunities generated and the rural-urban income differentials not reduced by the ad-hoc measures (they might in fact rise) induces a continuing process of migrants into urban centers so that the net result may continue to be that of rising unemployment. Consequently, the dissertation attempts to provide a comprehensive incomes policy which will include wage restraint and the elimination of tax exemptions and subsidies traditionally provided to imports of capital goods, both of which will help to restore normal factor price ratios and expand urban employment opportunities; encouragement of labor intensive peasant farms as opposed to mechanized farms attempted in Uganda during the 1960's which, to be successful, requires the provision of higher prices and incomes farmers actually receive. The latter will require elimination of export taxes and Marketing Board surpluses. If this is difficult because of the need for revenue, subsidies for inputs of rural labor and land (which is complementary to labor) are suggested as appropriate alternatives. Asefa Teferri The need for some, if not all, of these policies must have been obvious for some time. A brief discussion of the process of nation build- ing in Uganda during the 1960's is provided to show that over-riding political considerations played a significant role in the perpetuation of an abnormal wage-employment relationship and inappropriate farm organi- zation policies. Unless the current military regime succeeds in establish- ing a stable framework of internal political order, the prospects are for the implementation of incomes policies, as unpopular as they are with the already employed labor force, to continue to be postponed. INmMES POLICY FOR EAST AFRICA WITH SPECIAL REFERENCE TO UGANDA BY Asefa Teferri A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Economics 1973 ACIODWLEDGMENTS I wish to express my sincere gratitude to Dr. J. P. Henderson for the constant and untiring help he has extended to me during my studies and in the preparation of this dissertation. 936;; 65;)?! ii TABLE OF CONTENTS Page LI ST or “BLES O O O O C C O O O O O O O O O O O O O O O O O C C O O 0 v LISTOFPIGURES........................... vii Chapter I IWDmTION O O O O O O O O 0 O 0 O O 0 O O O O O O O O O 0 O 1 II LABOR MARKETS IN EAST AFRICA . . . . . . . . . . . . . . . . . 8 A Historical Perspective . . . . . . . . . . . . . . . . . . 9 The Ugandan Data . . . . . . . . . . . . . . . . . . . . . . 15 The Demand for Labor . . . . . . . . . . . . . . . . . . . . 16 The Supply of Labor . . . . . . . . . . . . . . . . . . . . 30 The Unemployment Gap . . . . . . . . . . . . . . . . . . . . 36 III AN ANALYSIS OF THE BEHAVIOR OF WAGES: THE CASE OF m U 0 . C O O O O O O O O O O C O O O O O O O O O O I O 48 The Normal wage Employment Trade-Off . . . . . . . . . . . . 49 The Ugandan Case . . . . . . . . . . . . . . . . . . . . . . 52 Changes in Average Wages . . . . . . . . . . . . . . . . . . 52 Changes in Unskilled Wages . . . . . . . . . . . . . . . . . 57 The Civil Service Salary Structure . . . . . . . . . . . . . 62 Changes in the Distribution of Income . . . . . . . . . . . 67 The wage-Rent Ratio in Uganda . . . . . . . . . . . . . . . 70 IV EDN-MARKET DETERMINATION OF WAGES AND SALARIES . . . . . . . . 76 The Determination of Salaries . . . . . . . . . . . . . . . 78 The Determination of Unskilled wages . . . . . . . . . . . . 82 V AN APPROACH '10 II‘COMES POLICY IN EAST AFRICA WITH stIAL MEWE m mm 0 O I O O O O O O O O O O O O O O O 94 A General Framework of Incomes Policy for East Africa . . . . . . . . . . . . . . . . . . . . . . . 94 Policies to Generate NoneAgricultural Employment . . . . . . 101 Agricultural Development Policies and the Generation of Employment . . . . . . . . . . . . . . . . 106 Needed Reassessment of Development Policy . . . . . . 106 The Denial of Employment Generation: The Ugandan Case . . . . . . . . . . . . . . . . . . . . . . 114 The Historical Trend towards Mechanization of Farms . . . . 115 The Impact of Export Taxes and Marketing Board Surpluses on Employment . . . . . . . . . . . . . . . . . 118 iii iv Chapter Page VI POLITICS AND RATIONAL DECISION-MAKING: THE EXPERIENCE OF UGANDA DURING THE 1960's . . . . . . . . . . . . . 124 VI I smy Am COMUS ION O O O O O O O O O O O O O O O O C O O O O l 3 4 BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140 LIST OF TABLES Table Page II-l Wage Employment in Uganda, 1954-1966 . . . . . . . . . . . . . l7 II-2 Uganda: Indices of Monetary GDP and Gross Capital Formation (Current values). . . . . . . . . . . . . . . . . 21 11-3 ElaSticity of Demand for Labor (EdL) with Respect to mnetary @P: 1955-1966 e o o o o o o o o o o o o o o o 22 11-4 Agricultural and NoneAgricultural Distribution of Total African Employment: Uganda, 1957-1966 . . . . . . . 24 11-5 Distribution of Non-Agricultural Employment, by Type of Employer, 1957-1967 . . . . . . . . . . . . . . . . 25 II-6(a) Uganda: Total African Employment in Selected Industries . . . . . . . . . . . . . . . . . . . . . . . . 27 II-6(b) Indices (1957 a 100) . . . . . . . . . . . . . . . . . . . . . 28 11-7 wages and Salaries as a Pr0portion of Total Government (Local and Central) Recurrent Expenditure . . . . . . . . . 29 11-8 Prospective African Population of Uganda under Various Rates of Growth (millions) . . . . . . . . . . . . 33 11-9 Percentage Distribution of Age Groups . . . . . . . . . . . . 34 11-10 Uganda: Estimated Labor Force . . . . . . . . . . . . . . . . 37 II-ll Uganda: Estimates of Total Unemployment . . . . . . . . . . . 38 11-12 Uganda: Total African Employment in Five Major Toms ' 1959-1966 0 O O O O O O O O O O O O O O O O O O O O O 4 3 11-13 Uganda: Growth in School Enrollment by Category (Aided Schools Only) . . . . . . . . . . . . . . . . . . . 46 111-1 Average Annual Cash Earnings per Employee: Uganda, 1957-1966 e o o o e o o o o o o o o o o o o o o o o 54 III-2 Average Annual Cash Earnings by Sector . . . . . . . . . . . . SS III-3 Index of Average Real Earnings per Employee by Sector: Uganda, 1957-1964 . . . . . . . . . . . . . . . . 56 vi Table Page III-4 Consumer Prices: General and Food Indices, Uganda (Kampala) I 1963 = 100 o o o o o o o o o o o o o o o o o a o 58 III-5 Percentage Increase in wage Bill for African Employees in Uganda: 1956-1964 . . . . . . . . . . . . . . 59 III-6 Distribution of African Male Employees by Groups within Industry, Private and Public Services . . . . . . . 60 III-7 African Adult Male Employees in Main Towns and Other Areas, by Wage Group, as a Percentage of Total, uganda, 1958 and 1965 . . . . . . . . . . . . . . . . . . . 61 III-8 Comparison of Teachers' and Civil Servants' Salaries, 1966. O O O O O O O O O O I I O O O O O O I O O O O O O I O 63 III-9 Index of Civil Service Salaries . . . . . . . . . . . . . . . 6S III-10 Salary Cost of Primary and Secondary Education . . . . . . . . 65 III-ll Estimate of the Wage-Rent Ratio for Uganda, 1957-1966 . . . . 73 IV-l Government-Determined unskilled wage Rates (shillings ”r mnth) O O O I O O O O I O O O O O O O O O I O O O O O 86 LIST OF F IGURES Figure Page 1 The Relationship between wage-Rent Ratio and Employment . . . . . . . . . . . . . . . . . . . . . . . . 50 2 The Relationship between Marginal Value Product and Price (i.e. Wage) of the variable Input mt (x1) 0 O C O O I O O O O I O O O O O O. O O O O O O O O O 120 vii CHAPTERI INTRODUCTION Unlike the African countries of Zambia, Zaire, and South Africa, dominated as they are by the exploitation of mineral resources, the three East African countries of Kenya, Tanzania and Uganda are primarily dependent on the development of the agricultural sector to support and develop their modern sector. Except perhaps for Kenya, where the predom- inance of the European and Asian community accounted for some degree of industrialization, the non-agricultural sector played a relatively minor role in its contribution to national output, employment and export base. At the time of independence in the early 1960's, the structure of the Ugandan economy was such that 90 percent of the total population were engaged in agriculture, with over 60 percent of its GDP originating in agriculture. Roughly 10 and 30 percent respectively came from secondary services and industries. Over 90 percent of all exports were agricultural products. The contribution of secondary industry in Tanzania was slightly less than in'Uganda. Moreover, the relative position of secondary industry as a preportion of GDP had remained more or less constant throughout the 1950's and into the period of independence in the early 1960's. The decade of the 1950's saw the proliferation of concern about the structure and growth of the economies of the underdeveloped regions of the world. These concerns found expression on the theoretical as well as political level. Early theories of development built models of economic growth on the basis of the historical experiences of the now developed nations -- namely, the transformation of economies such as Uganda and Kenya from predominantly agricultural to predominantly industrial. These models generally envisaged the role of the agricultural sector as the supplier of capital for industrial investment and the supplier of labor for indus- trial employment. The rapid growth of national political parties in the 1950's, culminating in independence in the early 1960's, provided the circumstance in which the industrialization theories were incorporated into national development plans. The strategy of developing import-substituting industries was the dominant characteristic of planning strategy throughout East Africa, with each country resorting to a protection of its domestic market from external competition arising within the region and abroad. The policy of ”industrial fundamentalism” was perceived as a process by which various goals could be achieved, such as rapid increase in per capita income, a more equitable distribution of income, and industrial diversification. While these goals did not preclude growth in agricultural output, this growth was perceived as a means of extracting surplus resources, capital and labor, to achieve the objective of rapid industrialization and trans- formation of the economic structure. The policy of imposed industrialization through imports of capital goods required certain governmental policies, namely low import duties, easy availability of foreign exchange, low interest rates on domestic credits, reliance on long term, low interest, and foreign loans. These policies would have the effect of distorting the relative prices of productive factors fromehat they would have been, given the prevailing resources available in East Africa, such as shortage of capital and abundance of unskilled labor. Moreover, colonial determination of wages on the basis of non-market criteria, primarily in terms of poverty datum line criterion, was intensi- fied during the decade of political independence when planned and deliberate improvement in the relative earnings of workers vis-a-vis investors and farmers became a major goal of economic development. The effect of such a policy was to intensify the distortion of relative factor prices as well as the relative earnings of urban and rural populations. The above strategy of development and related policy measures must have made obvious that the generation of employment was not a primary and major objective of development in East Africa. This was despite the fact that there were unmistakable evidences of emerging excess supply or unem- ployment of labor as early as the mid-1950's and growing at an alarming rate throughout the 1960's. The apparent paradox in decisionemaking and the setting of priorities was not so much because of inherent unwilling- ness to create employment opportunities, but due to the early beliefs of development planners that the process of industrialization could provide a substantial growth of employment Opportunities in the modern urban sector. The basic flaw in this scenario of the development process quickly became all too evident. Despite the positive rate of growth in non-agricultural output since the middle of the 1950's, and extending for more than a decade, the rate of growth in total non-agricultural employment was either stagnant or negative. The failure in Uganda to generate non—agricultural employment at a sufficient rate to absorb the rapidly rising supply of labor provides the central purpose of our dissertation. The inapplicability of the Keynesian model of the cause of unemployment, as well as the two-sector models which envisage the non-agricultural sector absorbing labor at a constant wage rate, suggests that the East African countries, and countries with similar prob- lems of unemployment, urgently require a reorientation of planning strat- egies and policies of social and political problems of mounting unemploy- ment are to be avoided. The key to the generation of adequate employment in the urban sector is the elimination of factor price distortions which prevent development from.occurring that is labor-intensive rather than capital-intensive. In particular, the practice of establishing urban minimum wages, independent of conditions of demand and supply in the labor market, independent of living standards in the traditional sector, and independent of consider- ing their effects on the wage structure as a whole, must be replaced by realistic and pragmatic wage policies consistent with the requirements of overall development and the generation of employment. The government of Uganda, unlike many of the developed nations, is the dominant employer of labor, so that while private employers and labor unions may have had some independent influences on the wage structure as a whole, wage restraint for public employees will have a resultant re- straining effect on the pattern of wages in the private sector. To con- ceal the major role of the government in past determination of the urban wage structure, and changes in this structure in the future, by blaming labor unions and/or private employers would constitute an abdication of responsibility to take effective measures to generate employment and thus avoid potential social and political instability. wage restraint and the avoidance of factor price distortions in the urban sector may succeed in generating employment opportunities. Yet, if past policies of neglecting agricultural development and/or maintaining a strategy of agricultural development based on labor displacing technology (e.g., tractors) continues, despite a rapid increase in population, the net effect of efforts to generate urban employment will be to increase urban unemployment through the stimulation of rural-urban flow of labor. The problem associated with the assumed relationship between urban employ- ment and the gap between urban and rural possibilities to earn income is likely to be relatively more serious for societies like Uganda where a narrow urban base is unlikely to absorb a significant portion of the rise in rural population and labor force. Hence, policies to expand rural in- comes and employment or, at least, minimize interference with the structure of price and output determination in the agricultural sector through a reassessment of the roles of marketing boards and export taxes on agricul- tural commodities, would be crucial for a successful elimination or control of the rapid growth in urban unemployment. 0n the basis of the above outline of the basic structure of the economies of East Africa and the rapid rise in unemployment, the following analysis of "Incomes Policy for East Africa with Particular Reference to Uganda" will be divided as follows. Chapter II will attempt to provide some evidence on the trends of demand and supply of labor in East African labor markets since the mid-1950's, utilizing historical data for the region as a whole, and supplemented by an analysis of available statistical data for Uganda. Chapter III provides the usual theoretical argument to show what the expected trend in average wages should have been, given the growing unemployment gap, or excess supply of labor in Uganda. This normal case is contrasted with the actual trend in the structure of average wages. Wages have risen far in excess of what would be expected if competitive market forces were allowed to operate freely, thus providing support to the hypothesis that institutional forces, primarily legislative determina- tion of minimum wages for the unskilled and the rigid relationship maintained between the level of formal education achieved and salaries, explain the level and structure of average wages. Chapter IV explores in some detail the functional relationship between minimum wages and average wages as well as between level of educa- tion and salaries in addition to reviewing and critically evaluating the historical circumstances and supposed "rationale" for the prevailing wage and salary policies. Chapter V attempts to develop an approach to incomes policy in East Africa with special reference to the case of Uganda. First, a general framework of incomes policy will be developed so as to define the concept, its objectives, scope, and criteria, appropriate for the structure and problems of the ugandan economy. Second, some specific measures to stimu- late urban employment opportunities are considered. Third, policies to generate rural employment as a necessary and integral part of a national and comprehensive incomes policy are analyzed. Since the agricultural sector has been neglected by both development theory and policy, a review of policies and needed modifications in the special case of uganda is pre- ceded by a brief analysis of the failure to consider the impact on small- scale and labor intensive agriculture. Chapters II to V demonstrate the failure to generate adequate employment opportunities in both the rural and urban sectors and the lack of effective policy measures to remedy the problem. In the past decade, a growing number of economists have been writing on the seriousness of the problem and its long term political repercussions. East African poli- ticians were, likewise, conscious of them -- as will be evident.from.the various statements that they make and the objectives of development plans specified in recent years. The consistent failure to provide for cor- rective policy measures during the 1960's must therefore surprise observers from societies where rational economic policy-making is a standard assumption. Chapter VI traces the political considerations in Uganda which have been the over-riding motives of decision-making process throughout the 1960's. While nation-building should be expected to dominate political behavior in a newly independent nation, the ugandan case will show that failure to come up with desired economic policies at the appropriate time will, by intensifying the problems of rural poverty, unemployment, and so on, be counterproductive to the long term goal of creating a politically stable society. The failure of the Obote regime during the 1960's meant it not only failed to come to grips with the basic economic problems, but also lost its right to govern when dissatisfied military leaders used Uganda's economic problems and the related societal instability as an excuse to organize a coup d'etat in 1971. Subsequent experiences in Uganda, similar to most African countries where the military have come to power, would indicate that the economic problems are of a magnitude for a mili- tary regime with uncertain political support to effectively handle. CHAPTER II LABOR MARKETS IN EAST AFRICA The relationship between demand and supply of labor in East Africa has undergone basic changes during the post-WOrld War II period. Avail- able evidence indicates a substantial and rapidly growing excess supply of labor, especially in the larger urban centers where both the unskilled and the educated have been migrating in significant numbers. This chapter provides a quantitative assessment of the changes in demand and supply of labor, with particular reference to Uganda during the mid-1950's to mid-1960's. The main objective is to explore the ‘magnitude and trend of total and urban unemployment. The omission of more recent data is partly due to the difficulty of securing statistical information comparable in scope and methodology to earlier data. However, as some cited references will indicate, the same trend has continued, even more notably, into the 1970's. As any student of developing nations would know, available statis- tical data are not the most reliable or the only sources of information regarding economic trends. Qualitative historical documents also provide useful data. Accordingly, we will begin this part of our analysis by re- viewing historical information on the develOpment of the East African labor market over both the pre- and post-World War II period. This will serve two purposes: first, to provide a historical basis to the analysis, and second, to reinforce the quantitative evidence on the growing imbalance between demand and supply of labor in East Africa since the mid-1950's. A Historical Perspective The prevailing pattern of living in East Africa before the expansion of colonial powers into the region has been described by various writers.- Herskovits,1 for instance, has described how the economic, social and other institutions in pre-twentieth century East Africa were governed by established custom.and tradition. The native population made their liveli- hood with subsistence production either of livestock in the case of pastor- alists, or food crops in the case of settled agriculturalists. Moreover, Herskovits describes how kinship relationships and extended family ties provided a form of social security to any individual who failed to take care of his own subsistence. In short, paid employment as an institution was practically unknown. The colonization of the region around the turn of the twentieth cen- tury introduced a series of foreign influences on traditional lifestyles and economic relationships via missionaries, private enterprises and colon- ial administrators. While the colonial regime abolished the slave raids of the Moslem traders,2 private and government activities introduced new demands for African labor. Therefore, alternative methods had to be found to secure the supply of labor services. This need to maintain a proper relationship between demand and supply of African labor had dominated the history of East Africa in the twentieth century. Initially, private and government employers could not use the pay- ment of money to induce an adequate supply of labor, since the needs of 1M. J. Herskovits, The Human Factor in Changing Africa (New York: Random House, Inc., 1962), pp. 62-72. 2G. St. J. Orde-Browne, The African Laborer (London, 1933). p. 13f. 10 the African were rather limited and could be met from subsistence activ- ities. Hence, direct and indirect measures (taxes to be paid in money, com- pulsory labor) were used to force the African into paid employment.3 Meanwhile, the introduCtion of cash crops into the peasant economic structure had begun an historical trend. In Uganda, for instance, peasant production of cotton for cash sales goes back to the early 1900's, and peasant production of both coffee and cotton had become the dominant fea- ture of the national economy by the 1930's.4 The rapid development of peasant cash crops in East Africa, particularly in Uganda where the ownership of land had remained largely African, demonstrated not only that Africans were quick to adapt to monetary incentives, but also that economic forces pushing the peasant to seek alternative sources of income outside agricul- ture were historically weak, if not absent. Yet, roughly up to about world war II, public and private employers did not actively encourage the develop- ment of market incentives to induce adequate labor supply. A collusion of private interest and Colonial Administrators were, implicitly or explicitly, opposed to competitive determination of wage rates through the free inter- action of the forces of demand and supply of labor. Some governmental regulations governing minimum living and working conditions had existed even before the Wbrld war II period. But whenever 3See P. G. Powesland, Economic Policy and Labour: A Study in Uganda's Economic History, East African Studies No. 10, ed. by walter Elkan (Kampala, 1957). 4See walter Elkan, The Economic Development of uganda (London: Oxford University Press, 1961). 5Jack Wbddis, Africa: The Roots of Revolt (New York: The Citadel Press, 1962), p. 66. 11 these measures were not sufficiently strong to induce the desired amount of labor services, the colonial policy preference had been for an increased application of indirect and/or direct measures such as hut tax, poll tax, and forced labor. An increase of taxes, to be paid in money, in a subsis- tence economy was expected to encourage labor migration in search of paid employment. Moreover, a Ugandan Native Authority Ordinance allowed com- pulsory labor for up to 30 days. Those unwilling to perform this work could commute their labor by paying cash.S Hailey says that compulsory labor ”...has been defended on the ground that it is an adoption of the indigenous usage by which communal labor is used for making village roads and the like."6 This may have been correct to the extent that a certain amount of obligatory communal labor was traditional in African society. But as Noon, among others, has pointed out: "Forced labor in Native society was well— safeguarded by customary usages and that, therefore, it never attained the degree of exploitation characteristic of the early years of European domina— tion of the continent."7 In the 1920's, African hostility to the forced labor system resulted in its being made illegal in Uganda, East Africa and other British colonies in Africa. Moreover, Britain was obliged to ratify the 1930 Geneva Conven- tion on forced labor.8 The Convention was, however, riddled with exceptions, so that, as Woddis pointed out, "...the basis for utilizing various forms of compulsion to secure labor still remained."9 Whatever the negative impact 6Lord Hailey, An African Survey, Revised 1956 (London, 1957), p. 1362. 7John A. Noon, Labour Problems of Africa (University of Pennsylvania, 1944) , p. 39. 8Wbddis, op, cit., p. 65. QIbid. , p. 65. 12 of internal and external legal constraints on the available supply of labor, it was easily compensated for by a more extensive recruitment of workers from outlying areas within uganda or from neighboring territories.10 In economic analysis, the minimum supply price of labor in a coun- try like Uganda tends to correspond to the average income of peasant farmers. At a given wage rate, other conditions remaining the same, a rise in average farm income tends to reduce the amount of labor services for paid employment. For a number of regions in Uganda, Elkan produced data to show "an inverse correspondence between the income from the sale of farm products in an area, and the proportion of its men in employment."11 Yet despite the demonstrated difficulty of securing hired labor, unskilled wages were kept more or less rigid. In 1900, 1913, and 1943, wages in the Buganda region of uganda were reported to have been 6/-, 9/- to 12/- and 12/- shillings, respectively, per month. The above overview may not have resolved the appropriateness or in- appropriateness of labor policies in East Africa for five decades after the initial period of colonization. It was not intended to do so. The objective was to show that private employers, government officials, and even some labor market analysts were unanimous in the view of an excess demand for labor in the area throughout this period, though they were reluc- tant to provide market incentives to induce increased supply. In fact, Colonial Labor Department officials continued to report labor shortages, or, at least, absence of surpluses, up to about the mid-1950's. Thus, Uganda's 10Elkan, pp, cit., p. 19f. 11Walter Elkan, Migrants and Proletarians, Urban Labor in the Economic Development of Uganda (London, 1960), pp. 35-37. 13 Labor Department reported that in 1947, "Literal unemployment does not exist and any able-bodied African can almost certainly obtain some form of manual labor at an average wage if he looks for it."12 Similarly, the Kenyan Labor Department reported that in 1948, "There is little or no unemployment in Kenya in any accepted sense, i.e., invol- untary. There is work for all who seek it."13 In 1951, the same Depart- ment reported that "The continued expansion of industry and agricultural concerns has made unprecedented demands upon the labor force of the colony during 1951. It was particularly in the unskilled category of employees that the demand outstripped the supply and many employers reported short- ages particularly those with considerable development programs."14 The Tanganyikan (now Tanzanian) Department of Labor, in a relatively optim- istic report, stated that in 1948, "There has been no lessening of the demand for unskilled labor, but workers have come forward in greater num- bers than was expected."15 Up to about the mid-1950's, the same attitude that work was avail- able ”for all who seek it” continued to be reported. Such optimistic re- ports may have been made merely to satisfy the concern of the British Parlia- ment which, it may be noted, had passed an Act in 1940 affirming its re- sponsibility to uplift the welfare of the colonial peoples.l6 Moreover, 12Uganda Protectorate, Labor Department Annual Report, 1947, p. 5. 13Colony and Protectorate of Kenya, Labor Department Annual Report, 1948, p. 6. 1492, cit., p. 5. 15Tanganyika Government, Labor Department Annual Report,*1948, p. 63. 16Great Britain, Statement of Policy_on Colonial Development and welfare, Cmd. 6175, 1940. 14 the assessments of Colonial Labor Officials may have failed to observe fully and in time the underlying structural shifts which were occurring in the process of the urban migration of the African population and which were transforming the pre-war labor market condition of excess demand into that of post-war excess supply.17 The possibility of such a lag between official reporting and the actual trend was apparent from the abrupt emergence of concern over increasing employment which had actually been occurring since the mid-1950's. Instead of its previous concern over labor shortages, the 1959 Labor Department Report from Kenya stated, "During 1956, a radical change took place in the labor supply position and by the end of the year surpluses of labor were being reported from almost all urban and rural areas."18 Similarly, the Tanganyikan Reports of 1956 and 1957 expressed con- cern over the emerging excess supply of labor. For instance: "In contrast to previous years a notable development in the territorial labor situa- tion has been the increasing unemployment problem experienced in the larger urban areas and in particular in Dar-es-Salaam municipality."19 As with Kenya and Tanganyika, the Ugandan Labor Department Reports of the period indicate a growing problem of unemployment, particularly in the towns of Kampala, Jinga, and Masaka. A 1958 report claimed: "There was a surplus of unskilled labor, semi-skilled workers, domestic, clerical, and other non-manual workers in the main towns."20 17See Derek Byerlee, Research on Migration in Africa: Past, Present and Future, African Rural Employment Study No. 2 (East Lansing, Michigan: Department of Agricultural Economics, Michigan State University, 1972). 18Colony and Protectorate of Kenya, Labor Department Annual Report, 1956, p. 3. 19Tanganyika Government, Labor Department Annual Report, 1957, p. 3. 20Uganda Protectorate, Labor Department Annual Report, 1958, p. 11. 15 The problem of excess supply of labor, particularly in the main towns, had continued to cause mounting concern among scholars and policy-makers. Rapid population growth, educational policies, minimum wage legislation, export taxes on agricultural commodities and a whole range of urban indus- trial policies have been recognized as having some casual relationship with the growing problem of excess supply of labor in East Africa during the 1950's and 1960's. Yet, for reasons to be elaborated later, recogni- tion of these interrelationships was not reflected in the actual process of decision-making in East Africa. As a participant was reported to have remarked in a recent conference held in Kenya: ...although the problems of unemployment, the overly academic school curriculum, etc., had been recognized for many years, the concern for the interrelationship of the issues of education, employment and rural devel- opment had only recently commanded serious attention. The Ugandan Data Turning now to a quantitative assessment of the changes in demand and supply of labor in Uganda, data is presented for the period 1954-1966. Given the close similarities in the economic structures and labor market conditions, it is believed that the conclusions derived from an analysis of Ugandan data could broadly apply to Kenya and Tanzania as well. How- ever, comparative data from the latter will be presented whenever they are convenient and/or illuminating. The objective, it should be pointed out, is not to provide a precise estimate of the magnitude of excess supply of labor, for this would be 21James R. Sheffield (ed.), Education, Employment and Rural Develop- ment, Report of the Kericho (Kenya) Conference, 25th September to let October, 1966 (Nairobi: East African Publishing House, 1967), p. xii. 16 impossible given the current state of statistical knowledge in East Africa. Rather, it is to quantify the broad order magnitude of the problem inasmuch as existing data will allow, and hence corroborate the qualitative assess- ment of the unemployment situation which is causing great concern among economists familiar with the area. Moreover, an empirical perception of the problem, even if imprecise, would be very useful for a meaningful or con- vincing evaluation of prevailing incomes or wage policies. Statistical techniques for estimating changes in demand and supply of labor in developed countries have undergone significant growth and re- finement. Much empirical research has been done to identify and analyze the independent effects of factors functionally related to the demand and supply of labor. In East Africa, comparable research has hardly begun. A.more practical and useful approach to quantitatively estimate the magni- tude and changes in demand and supply in the latter would be to use total recorded employment to serve as an index of demand, and the trend in the size and characteristics of population as a basis for estimating the supply of labor. Such an approach is, no doubt, crude and subject to a wide margin of error. When used in conjunction with historical information presented above, however, it would be adequate to draw conclusions useful as a framework for evaluating prevailing wage or incomes policies and recommending necessary changes in these policies. The Demand for Labor Table 11-1 shows the trend in total recorded employment of Uganda over the period 1954-1966. It reveals that over the entire period the level of total employment has remained more or less stagnant. Total employment 17 Table II-l wage Employment in Uganda: 1954-1966 Number of Index of wage Employees Employment Year (thousands) (1957 = 1954 236 98 1955 237 99 1956 237 98 1957 241 100 1958 243 101 1959 240 99 1960 245 101 1961 236 98 1962 .237 96 1963 222 92 1964 225 93 1965 242 101 1966 246 102 Source: Uganda Government, Statistical Abstract (Entebbe: Government Printer, various issues). 18 increased gradually during 1954-1958, reaching in 1958 a level about 3 per- cent above that of 1954. There was a slight decline in 1959, which re- turned in 1960 to a slightly higher level than in 1958. Over the next three years, however, a sharp decline occurred, reaching its lowest level in 1963, almost 6 percent below the 1954 level. Since 1964, total employment indicated an apparent recovery, but the 1966 level was still only some 4 percent above that of 1954. In short, total employment over the entire period analyzed had stagnated, showing only minor increases or decreases. Over the same period, total employment in Kenya increased by 11 percent while that of Tanzania declined by almost 18 percent. If, as was suggested above, the total employment figures are used as indicators of the trend in total demand for labor, it would be neces- sary to assume that e§_3g£e and e§_p2§5_magnitudes of employment are iden- tical. In other words, it must be assumed that all vacancies created during a given period areand can be filled in the same period. In reality, the planned level of employment may be in excess of the actual or realized level. If this holds true, recorded employment figures would underesti- mate the level of demand for labor. In the context of East Africa, however, we can safely ignore the possibility of such a discrepancy between planned and actual levels of employment, since, as was indicated in the review of historical information, and as will be evident below from an analysis of the supply side of the labor market, excess supply of labor has been and continues to be a basic characteristic of these countries. Another problem of using total employment figures as indicators of demand for labor relates to possible inaccuracies in the reported statis- tical data. In uganda, as well as in Kenya and Tanzania, the source of 19 reported employment data has been an annual Enumeration of Employees,22 which may have failed to include a large number of seasonally-employed persons, particularly in processing industries and other agriculturally related activities. To the extent that such errors exist, the employment figures would underestimate the real amount of demand for labor. For instance,23 the Enumeration data for Uganda reported a total employment figure of 35,700 persons in processing and manufacturing activ- ities, while a Survey of Industrial Production for the same year reported 39,510 persons, even though the latter excludes enterprises employing less than ten people. Furthermore, in cotton and coffee processing industries the Enumeration reported only 7,300 persons employed while the Survey re- ported 17,200 persons. Such a wide margin of discrepancy appears to reflect statistical errors due to the problem of seasonal employment. The appar- ent increase in total employment as officially reported since 1965 indi- cates an upward revision of Enumeration estimates rather than real im- provements in the demand for labor in Uganda.24 The poor quality of statistical information is one major reason why it is extremely difficult to study and analyze the performance of the East African economy. It is unlikely, however, that the errors and omissions in reported employment figures are of a magnitude and nature to significantly 22Uganda Government, Ministry of Planning and Economic Development, Statistics Division, Enumeration of Employees (Entebbe, annually). 23E. J. Stoutjesdijk, Uganda's Manufacturing Sector: A Contribution to the Analysis of Industrialization in East Africa, East African Studies No. 29 (Nairobi: East African Publishing House, 1967), p. 25. 24E. J. Stoutjesdijk, ibid., p. 25f, claims that as a result of the findings from Industrial Surveys carried out in 1963 and 1964, employment data derived from Enumeration of Employees have been scaled upwards. 20 distort the broad conclusion revealed by the data given in Table 11-2 -- that the total demand for labor in Uganda (similarly in Kenya and Tanzania) had remained morecm less stagnant over the entire period 1954-1966. What is perhaps most disturbing in the apparent stagnation of total employment is that it occurred despite a relative increase in reported monetary Gross Domestic Product and Gross Capital Formation (Table 11-2). These data do not show the rate of growth in output in real terms, but when compared with the data in Table 11-1, they do provide some indication that the elasticity of demand for labor with respect to output has been sig— nificantly less than unity. Table 11-3 shows the coefficients of such elasticity calculated for Uganda, Kenya and Tanzania over the period 1955-1966. The basic output- employment relationship indicated by these coefficients would not be funda- mentally different if real, instead of monetary, changes in output were used. Growth in real national output would still be higher than growth in employment. In comparison to stagnation in employment shown for Uganda, real output (in terms of 1960 prices) was estimated to have grown by about 1.4 percent in 1954-1962 and 9 percent in 1962-1965.25 Thus, whether real or monetary growth in output is the frame of reference, it is clear that one cannot lay much hope on a planned growth in national output per se to generate adequate employment opportunities in Uganda and East Africa as a whole. Development models which were popular during the 1950's and 1960's assumed that rapid growth in non-agricultural output would successfully 25R. H. Green, Tabular Appendix, Economic Development Research Paper, No. 103 (Kampala: Mekerere Institute of Social Research, not dated). 21 Table II-2 Uganda: Indices of Monetary GDP and Gross Capital Formation (current values) Index of Index of Capital Monetary GDP Formation Year (1957 = 100) (1957 = 100) 1954 85 94 1955 94 113 1956 95 107 1957 100 100 1958 97 96 1959 99 84 1960 102 92 1961 102 89 1962 99 87 1963 119 98 1964 137 119 1965 138 159 1966 159 151 Source: Statistical Abstract, various issues. 22 Table II—3 Elasticity of Demand for Labor (EdL) with Respect to Monetary GDP: 1955-1966 Year EdL: Uganda EdL: Kenya EdL: Tanzania 1955 0.04 4.0 -1.85 1956 0.00 -0.44 3.56 1957 0.21 4.70 -l.78 1958 -0.40 -2.75 0.77 1959 -0.89 0.18 0.10 1960 0.08 0.05 0.29 1961 -6.10 -4.80 -0.51 1962 0.71 -0.52 -0.63 1963 -0.75 -l.54 -l.48 1964 0.15 1.01 0.27 1965 1.21 -0.04 -l.80 1966 0.09 2.20 0.08 Note: The negative elasticities are due to a fall in employment. Source: F. I. Ojow, Employment! Output and Labor Force Trends in East Africa, Economic Development Research Project No. 138 (Kampala: Mekerere Institute of Social Research, May, 1968). p. 10. 23 absorb surplus labor from the agricultural sector. Even if total employ- ment had remained stagnant, as indicated above, structural redistribution could have brought about growth in non-agricultural employment. The data indicated in Table II-4, however, do not show such redistri- bution in total employment to have taken place. The absolute and relative magnitudes of non-agricultural employment are shown to have remained more or less stagnant. In fact, if the data is to be believed, total employ- ment in Uganda grew slightly faster than non-agricultural employment over the period 1957-1966. Thus, emphasis put on industrialization in the development strategies of East Africa had not succeeded in generating a demand for labor. At this stage, it would be useful to consider the trend in non- agricultural employment disaggregated between public and private sectors as well as between some selected industries in Uganda. The employment data used covers only African employees. Omission of non-Africans should not distort the results of our analysis as they constitute an insignifi- cant proportion of total employment.26 Table II-S indicates that the public sector accounts for about 50 percent of total non-agricultural employment in Uganda -- a characteris- tic shared by East Africa as a whole and indeed most developing nations. The data also show that public non-agricultural employment over the period 1957-1966 had declined, absolutely and relatively, to that of the private sector. While public employment declined by 9 percent over that 26In 1957, for instance, Asian and European employees numbered 4.5 and 9.3 thousands respectively out of a total employment of 241 thou- sand given in Table II-l. The corresponding figures for 1966 were 3.5, 10.6 and 2.46 thousands respectively. 24 Table II-4 Agricultural and NOn-Agricultural Distribution of TOtal African Employment: Uganda, 1957-1966 Index of Non-Agricultural Agricultural* Non-Agricultura1** Employment Year Employment Employment (1957 a 100) 1957 49,127 177,789 100 1958 50,637 177,762 100 1959 49,637 174,623 98 1960 52,966 175,923 99 1961 48,532 172,467 97 1962 51,418 165,355 93 1963 49,941 158,409 89 1964 51,306 161,040 90 1965 50,347 177,839 100 1966 52,206 179,667 101 *Employment in agriculture, forestry and fishing excluding employment in peasant agriculture. **Exc1uding domestic employment. Note: Since Asian and European employees are left out of Table II-4, the totals will fall short of the total employment figures given in Table II-l. Source: Statistical Abstract, various issues. 25 Table II-5 Distribution of African NoneAgricultural Employment by Type of Employer, 1957-1967 Index of Non-Agricultural Employment NoneAgricultural (1957 = 100) Year Private Public Private Public 1957 89,560 88,239 100 100 1958 86,370 90,504 96 103 1959 83,432 91,191 92 103 1960 90,547 88,376 101 100 1961 84,253 82,217 93 93 1962 82,055 83,300 91 94 1963 80,053 78,356 89 89 1964 77,159 83,881 86 95 1965 93,914 83,925 104 95 1966 98,972 80,395 110 91 Source: Statistical Abstract, various issues. 26 period, private non-agricultural employment grew by about 10 percent. If the 1965-1966 data were excluded, the relative trends in public and private non-agricultural employment would be seen to be similar. Both would indi- cate a fall in employment. The data in Tables II-6(a) and II-6(b) provide a clearer picture of the trend in private non-agricultural employment, showing that while employment in manufacturing, commerce and mining industries increased over the period 1957-1966, in construction and transportation it declined. Here again, the apparent increase in the former group of industries would not show if the 1965-1966 data were excluded. The apparent sharp increases in employment during 1965-1966 may have been due to changes in the statistical method used in reporting employment rather than any real changes in magnitude.27 Hence we can broadly conclude that changes in non-agricultural employment -- public and private -- have not succeeded in generating a demand for labor that would indicate the rural to urban employment redistribution trends sug- gested by develoPment models during the 1950's and 1960's. The same con- clusion was indicated by Frank28 with respect to most African countries. Moreover, as the proportion of wages and salaries in total recurrent expenditure are significant in the budgets of East African countries (see Table II-7), the government sector should not be expected to show a better employment picture in the future -- at least, given the prevailing policies. Even if government policies do succeed in scaling down the 27See footnote 24 above. 28C. R. Frank, Jr., "Urban Unemployment and Economic Growth in Africa," Oxford Economic Papers, xx (July, 1968) , p. 256. 27 Table II-6(a) Uganda: Total African Employment in Selected Industries Manufacture Transport Mining Misc. of Food and and Year Manuf. Products Constr. Commerce Quarrying Commerce 1957 17,512 7,513 35,184 10,403 5,536 9,022 1958 18,202 7,031 63,642 10,378 3,949 10,246 1959 17,240 7,071 31,778 9,677 5,136 9,918 1960 16,955 7,620 29,204 10,105 5,322 10,696 1961 17,896 7,768 28,554 9,631 5,769 11,142 1962 17,092 7,603 28,635 8,985 5,095 10,483 1963 17,177 7,424 25,819 9,606 4,853 9,663 1964 16,485 9,370 24,091 8,995 4,830 9,681 1965 18,612 8,159 31,573 9,698 6,084 9,209 1966 20,865 8,672 28,253 9,322 6,156 10,509 Source: Statistical Abstract, 1965, 1970. 28 Table II-6(b) Indices (1957 = 100) Manufacture Transport Mining Misc. of Food and and Year Manuf. Products Constr. Commerce Quarrying Commerce 1957 100 100 100 100 100 100 1958 103 93 104 100 71 113 1959 98 95 90 93 93 110 1960 97 101 83 97 96 119 1961 101 104 81 92 105 123 1962 97 101 81 87 93 117 1963 98 99 74 92 89 108 1964 94 125 68 87 87 108 1965 106 109 90 93 111 102 1966 119 116 80 89 111 117 Source: Same as in Table II-6(a). 29 Table II-7 Uganda, 1960-1966 wages and Salaries as a Proportion of Total Government (Local and Central) Recurrent Expenditure: Total Nin-Recur- Total R:current wage: and ent Expenditure Expenditure Salaries 3 % 2* Year (thousand shs.) (thousand shs.) (thousand shs.) % 1960/61 115,824 622,694 232,504 37 1961/62 107,015 654,965 264,051 40 1962/63 93,644 770,307 280,533 36 1963/64 125,768 806,516 307,317 38 1964/65 211,447 896,224 366,921 41 1965/66 234,433 1,013,513 402,248 40 *The percentages would have been slightly higher if "overseas addition" to expatriate staff paid by the U.K. Government since 1962/63 were excluded. Source: Statistical Abstract, 1965, 1970. 3O proportion of wage and salary expenditure, the chances of generating addi- tional employment are not optrmistic, since the public sector is already very labor intensive. But, as will be more evident below, such changes in public policy would affect private employment opportunities positively and significantly. The Supply of Labor In most developing nations, an analysis of the supply side of the labor market is made very difficult by a lack of systematic demographic and labor force studies. In some countries, such as Ethiopia, even a critical minimum of such data are not available. The East African states of Uganda, Kenya, and Tanzania have been relatively fortunate in this area. The colonial regime established a tradition of gathering some population data, particularly in the post-world War II period. Censuses of population were available even for the period before the war, but these were nothing more than intelligent guesses based on the number of taxpayers, or counting people ordered to assemble in a given central area. In theory, compulsory registration of deaths and births was required of most Ugandans, as early as the 1930's. In practice, however, the system was highly ineffective, as an overwhelming proportion of the populace either did not know of the requirement's existence or did not bother to register.29 The establishment of the East African Statistical Department in 1947 represented the first real attempt at systematic collection and compilation 29This background information and the next few paragraphs draw heavily from T. E. Smith and J. G. C. Blacker, Population Characteristics of the Commonwealth Countries of Tropical Africa (University of London: The Athlone Press, 1963). 31 of population data in Uganda, Kenya and Tanzania. In 1948 a complete enum- eration of Africans and noneAfricans in these countries was attempted. It supplemented by re-enumeration of a sample of selected areas -- to cross— check and to obtain more detailed characteristics of the population. A second census was undertaken for Tanganyika (now Tanzania) in 1957, for Uganda in 1959, and for Kenya in 1962.30 Since independence, no significant improvements have been made on the census data left by the colonial regime. Obviously, most of these data are by now outdated. Nevertheless, they provided a basis on which East African governments and independent researchers could estimate the trend in population during the 1960's and later. Since the two colonial censuses followed practically the same procedure, it facilitated the task of esti- mating intercensal changes in size and characteristics of a population, as well as future projections. The rate of population growth in East Africa, between the two colon- ial census years, was assumed to have remained constant in the compilation of official population data throughout the 1960's. In Uganda, for instance, a constant rate of growth of 2.5 percent per annum was assumed, equal to the rate of growth of the African population from 4,917,555 in 1948 to 6,449,558 in 1959.31 In fact, the Second Five Year Plan of Uganda assumed this rate to hold through 1981.32 Most students of East Africa, however, 30T. E. Smith and J. G. C. Blacker, op, cit., pp. 57-60. 31F. J. Gormely, Some Implications of Population Growth in Uggnda, Economic Development Research Project No. 97 (Kampala: Mekerere Institute of Social Research, April 4, 1966), p. 2. 32Uganda Government, work for Progressi¥Second Five Year Development Plan, 1966-1971 (Entebbe: Government Printer, 1966). 32 estimate the actual rate of growth per annum to lie somewhere between 2.5 and 3.5 percent. A provisional census of population in Uganda under- taken in 1969 estimated the total population to have been about 9.5 mil- lion, which supports a rate of growth nearer to 3.5 percent between 1959 and 1969.33 Thus, while the colonial p0pu1ation census and the various census- based projections of the trend throughout the 1960's gives some idea of population growth in Uganda as well as East Africa as a whole, the actual growth rate remains unknown. The best one can do, e.g., for Uganda, is accept the officially reported population growth trend which applies a constant 2.5 percent rate of growth to the 1959 population estimate, as the minimum estimate, and that calculated on the basis of a 3.5 percent constant rate of growth as being the maximum estimate, while the actual population growth may lie somewhere in between. P. J. Gormely34 adopted this approach in his recent attempt to calculate the size of African popula- tion in uganda over the period 1959—2000. Table II-8 presents his estimates on the assumption of alternative constant rates of growth between 2.5 and 3.5 percent. It shows that even under the minimum constant rate of growth assumption, the population of Uganda in 1959 would double in just 28 years. The total labor force growth rate would depend primarily on the popu- lation growth rate as well as on the labor force participation rate. The latter, in turn, depends on such factors as the age-sex composition of the population, education, income, etc., and how these factors change over 33Uganda Government, Statistical Abstract, 1970 (Entebbe: Government Printer, 1970), p. 14. 34F. J. Gormely, 92, cit., Table II, p. 3. 33 Table II-8 Prospective African Population of Uganda under various Rates of Growth (millions) Year 2.5% 2.6% 2.7% 2.8% 2.9% 3.0% 3.5% 1959 6.425 6.425 6.425 6.425 6.425 6.425 6.425 1965 7.451 7.495 7.539 7.583 7.628 7.672 7.865 1970 8.429 8.522 8.612 8.705 8.800 8.894 9.380 1975 9.537 9.689 9.840 9.993 10.153 10.311 11.031 1980 10.789 11.016 11.242 11.473 11.714 11.954 13.232 1985 12.207 12.525 12.844 13.171 13.514 13.858 15.715 1990 13.810 14.240 14.674 15.118 15.628 16.066 18.664 1995 15.624 16.191 16.765 17.359 17.988 18.626 22.167 2000 17.676 18.409 19.153 19.928 20.754 21.593 26.080 Years Needed to Double Population 28 27 26 25.1 24.2 23.4 20.1 Source: P. J. Gormely, op, cit., Table II, p. 3. '..£ ‘ 34 time. Very little is known about the relative importance of each factor on the labor force participation rates in East Africa. As in most devel- oping nations, a significant proportion of the East African population is believed to be young. Table II-9 shows that well over 50 percent of the population of the three East African countreis are estimated to lie in the 0-21 year age group. Moreover, roughly 50 percent of the total population 35 are estimated to female. Both characteristics imply a relatively low labor force participation rate. In the context of the predominantly rural and Table II-9 Percentage Distribution of Age Groups Age Group Uganda (*3) Kenya (%) Tanzania (%) 0-21 41.5 46.5 42.5 15-21 25.2 26.9 27.0 0-21 66.7 73.4 69.5 Source: F. S. Arkhurst (ed.), Africa in the Seventies and Eighties, Appendix, Table 2 (New York: Praeger Publishers, 1970). peasant economy of East Africa, however, children and women play a rela- tively important role in economic activities. Women, for instance, partici- pate actively in the production of food crops, while children may engage in activities such as picking cotton. With future increases in urbaniza- tion, education and/or family income, one would expect the behavior of the labor force to gradually approach that prevalent in the developed nations. Beyond these general remarks, however, the quantitative estimate of 35See T. E. Smith and J. G. C. Blacker, op, cit., p. 29. 35 the trend in the actual labor force in East Africa remains a difficult task to undertake. Except for Tanzania,36 the East African nations have yet to undertake even a limited labor force survey. Nevertheless, on the basis of information derived from the Tanzanian Labor Force Survey and the 1962 Kenyan Census of Population (indicated above), some attempts could be made to provide tentative quantitative estimates of the trend in the total labor force. In what follows, an estimate of Ugandan labor ferce is made based on a recent study by F. I. Ojow,37 covering all the three East African countries. The Ugandan case will suffice to draw broad conclusions applic- able to the other countries as well because of the close similarity in the estimated proportion and trend of the total labor force. The Tanzanian Labor Force Survey of 1965, referred to above, found that 37 percent of the total population (after deducting those keeping house, in school, unable to work, in prison, etc.) defined the size of the total labor fbrce in that year. Ojow assumed the same prOportion for Oganda (and Kenya) to estimate the size of the labor force before, during, and after 1965. He made three alternative estimates assuming (1) that the labor force in any one year was a constant 37 percent of the total population, (2) that while the labor force was 37 percent of the total popu- lation in 1965, it was growing at the same rate as population before and after 1965, and (3) that the average of the two alternative estimates was taken. The results of the third alternative were found to lie between those of the first two. For this reason alone a presentation of Uganda's 36R. S. Ray, Labor Force Survey in Tanzania, Ministry of Economic Affairs and Development Planning, The United Republic of Tanzania (Jan- uary, 1966). 37F. I. Ojow, op, cit., Tables 1, 2 and 3, pp. 4, 5, 6. 36 estimated labor force is made in Table II-10, Column 8, for the period 1954- 1966, calculated under the third alternative assumption. Column C provides the adjusted labor force estimate on the assumption that 50 percent of the Ugandan population is female, with only 5 percent actively seeking employ- ment. The latter percentage approximates the proportion of females in paid employment. In reality, the female participation rate may be somewhat higher. Furthermore, the population data used are government estimates which, as was indicated above, are calculated on the basis of what was at best a mdnimmm.2.5 percent constant rate of growth. It is therefore possible that the calculated figures for the unadjusted and adjusted labor force underestimate the actual trend. Nevertheless, a broad picture is clearly apparent -- that, over 1954-1966, the rate of growth of the labor force has been significant and perhaps approaching 100 percent or about 8 percent per annum. The Unemployment Gap, When the total recorded employment for a given year (our indicator of the total economy's demand for labor) is subtracted from the estimate of adjusted labor force (our indicator of the total supply of labor), an estimate of the magnitude of the total excess supply of labor or the total unemployment gap is derived for the economy as a whole. Table II-ll provides the trend in the estimated gap for Uganda over the period 1954— 1966. As the adjusted labor force, for reasons indicated above, may have been under-estimated, it is likely that the indicated unemployment gap is an under-estimation of the actual magnitude. Even then, the gap is shown to have increased from about half a million to well over a million over the 37 Table II-lO Uganda: Estimated Labor Force Estimated Adjusted POpulation Labor Fbrce Labor Force (millions) (millions) (mdllions) Year A B C 1954 5.7 1.54 .809 1955 5.9 1.66 .872 1956 6.0 1.76 .924 1957 6.2 1.83 .960 1958 6.4 2.00 1.050 1959 6.5 2.11 1.108 1960 6.7 2.23 1.171 1961 6.8 2.33 1.223 1962 7.0 2.45 1.286 1963 7.2 2.57 1.349 1964 7.4 2.69 1.412 1965 7.6 2.81 1.470 1966 7.7 2.92 1.533 Source: F. I. Ojow, 92. gi_t_., Table l, p. 4. 38 Table II-ll Uganda: Estimates of Total Unemployment Adjusted Labor Force Employment Unemployment Gap Year (thousands) (thousands) (thousands) 1954 809 236 573 1955 372 237 635 1956 924 237 687 1957 960 240 720 1958 1,050 243 807 1959 1,108 240 868 1960 1,171 245 926 1961 1,223 236 987 1962 1,286 231 1,055 1963 1,349 222 1,127 1964 1,412 225 1,187 1965 1,470 242 1,228 1966 1,533 246 1,287 Source: From Tables II-l and II-10 above. 39 period 1954-1966. Moreover, the gap has been consistently increasing every year of the entire period. The same general conclusion applies to both Kenya and Tanzania. The trend in the aggregate unemployment gap fer the economy as a whole, though useful in defining the overall magnitude of the problem, offers limited insight in a situation of disaggregated labor markets with distinct characteristics. In the context of East Africa, an analysis of the unemployment gap in terms of urban and rural labor markets offers a more useful approach to follow. In the rural economy of East Africa, the supply of labor is predom- inantly unskilled and more or less homogeneous. Moreover, an overwhelming majority of the total labor force is rural. Hence, the conclusions of our analysis with respect to the total economy's demand and supply of labor could apply to the rural sector without significant danger of distortion. We need only recognize that rural demand for labor is subject to seasonal fluctuations and that, given rural supply of labor, the unemployment gap is likewise subject to fluctuations. It is generally recognized that excess supply in the rural sector assumes the form of hidden unemployment or underemployment, so that seasonal fluctuation of demand for labor would be reflected in seasonal fluctuation of the degree of underemployment. This is related to the fact that the rural sector is characterized largely by self-employment on a family or communal farm. In contrast, the urban labor market is characterized by the problem of open-unemployment and, excepting self-employment in petty trade and/or small business ventures (where underemployment may be a problem), wage 40 employment is the predominant form of making a living. Moreover, the supply of labor is far from homogeneous. There are the white collar vs. the blue- collar workers, the unskilled vs. the skilled workers and/or the educated vs. the uneducated. While applying such categories to the East African situation is not without limitation, the existence of non-homogeneous labor supply in the "modern” or urban labor market is clearly evident. Moreover, the structure of the urban economy is not homogeneous. we have the public vs. the private sector, the large—scale vs. the small- scale enterprise, trade vs. manufacturing, export-oriented vs. internal market-oriented producers, etc. Hence, the nature of demand for labor may be different between different units of the urban economy. These classifications of the urban labor market may not be exhaus- tive. Nevertheless, they do indicate that an aggregative analysis of the total economy's labor market and unemployment gap would probably lead to distortions if its conclusions are uncritically applied to the urban set- ting, and that an aggregative analysis of the total urban labor market could also result in misleading conclusions. Part of the problem of devel- opment economics in the past has been the tendency to treat the total and/or urban economy as if it were homogeneous and uncritically apply aggregative models developed for advanced economies. Quantitative assessment of open-unemployment in the urban labor market in East Africa is extremely difficult both for conceptual and statistical reasons. Conceptually, there is the problem of defining the urban labor ferce actively seeking employment. What may be regarded as self-employed could easily become part of the labor force if the demand for workers in- creases significantly. Some housewives, students, etc., may be only 41 temporarily out of the labor market because of unavailability of jobs. Such behavior responses should not be ignored in any meaningful analysis of the urban labor market in developing nations. While recorded total urban employment may serve to indicate the extent of demand for labor (though not necessarily for a specific category of labor, e.g., school-leavers), the difficulty of measuring and defining urban open-unemployment presents a critical problem for analysis. Infor- mation from labor exchanges is highly unreliable, as most of the unem- ployed may be reluctant to identify themselves out of fear of deportation to rural areas.38 Moreover, given rural to urban labor migration and the resultant problem of maintaining a representative sample, survey techniques are likely to give unreliable results. Furthermore, the lack of sophisti- cation of questioners and respondents would present additional limitations. Recognition of the difficulties, however, should not prevent us from having a broad picture of the magnitude and nature of urban excess labor supply. For instance, Frank39 has suggested the utilization of urban popu- lation growth as an indicator of the growth in urban labor force. A re- cent Ugandan study4o assumed an urban population growth rate of between 5.8 and 6.9 percent when projecting the urban population trend up to the year 2000. These rates may not be precisely accurate but they serve to indicate that urban population and labor force has been growing much faster 38 . . M. P. Todaro, "Income Expectations, Rural-Urban Migration and Employment in Africa," International Labor Review, v01. 104, No. 5 (1971), pp. 4-5. 39C. R. Frank, Jr., 220 Cite, p. 2520 4 0Uganda Government, Kampala-Mengo Regional Planning Study No. 1, Population and Land Requirement (Kampala, February 1965), Appendix A, Table 1. 42 than for the nation as a whole. Gormely estimated that the urban population in uganda had about doubled during 1959-1970, assuming a constant 5.8 percent annual rate of growth.41 When this is compared with the official recorded African em- ployment in five major towns as presented in Table II-12, it is apparent that the magnitude of urban unemployment has been significant and growing rapidly over time. Moreover, it indicates the danger of distortions in applying the aggregate economy's unemployment gap to describe the urban labor market's unemployment problem. That is, due to the differential rates of population and labor force growth in urban and rural areas, it is possible for urban unemployment to rise over time even if the total econ- omy's unemployment gap remains constant. Various studies undertaken in the past few years support the thesis that in East Africa and other parts of Africa, "by far the most important contributing factor [of rising urban unemployment] has been the massive increase in the number of migrants arriving from surrounding rural areas."42 Cities like Kampala (uganda), Nairobi (Kenya) and Dar es Salaam (Tanzania) are reported to be swarming with job seekers of all types and descriptions. It would be a mistake to lump together the total urban labor force in an analysis of urban excess supply. While comparison of urban recorded employment and total urban labor force as indicated in Table II-12 may be sufficient to describe the demand supply conditions, or excess supply of unskilled urban labor, the skilled and educated category of urban labor would require additional analysis. We indicated above that an imbalance 41F. J. Gormely, gp, cit., Table III, p. 4. 42M. P. Todaro, gp, cit., p. 387. 43 Table II-12 Uganda: Total African Employment in Five Major Towns, 1959-1966 Urban Population (mil- lions) Constant 2.5t Year Employment* Year Annual Rate of Growth** 1959 52,728 1959 248 1960 50,925 1965 348 1961 48,885 1970 461 1962 49,710 1975 611 1963 49,425 1980 811 1964 52,328 1985 1,074 1965 55,828 1990 1,492 1966 56,327 1995 1,888 2000 2,503 Sources: *Statistical Abstract, various issues **P. J. Gormely, gp, cit., Table III, p. 4. 44 exists regarding high level manpower in East Africa, not so much from lack of demand for skilled workers but because of failure to educate the young with skills demanded in the job market. As a recent United Nations study puts it: The manpower imbalances show surpluses of young school- leavers whose fbrmal education has ill-equipped them for employment in either the traditional or the modern sector and has oriented them toward white-collar jobs; an acute shortage of middle level and high level man- power, especially those of technical, managerial and supervisory training: and professions giving less pay and prestige than administrative jobs, especially in the early years of independence and Africanization. Furthermore, since development and investment in East Africa during the 1950's and 1960's were concentrated in urban centers, the young school- leavers were a significant proportion of the rural to urban migrants. Rempel,44 in a 1970 study of Kenya, reports that the highest propensity to migrate occurs in the 15-19 age group and that most of these are primary and secondary school dropouts. Moreover, Sabot,45 in a Tanzanian study of the relationship between migration and education, not only con- firmed Rempel's findings but added that the propensity of the educated to migrate has increased over time, with secondary school-leavers forming a growing and relatively higher proportion of the rural to urban migrants. The number of school-leavers, and how much of the urban population they constitute, would be difficult to determine quantitatively. Nevertheless, 43United Nations, Economic Bulletin for Africa, Vol. X, No. 1 (June, 1970), p. 64. 44H. Rempel, "Labor Migration into Urban Centers and Urban Unemploy- ment in Kenya” (Unpublished Ph.D. dissertation, Madison, University of Madison, 1970). 45R. H. Sabot, "Education, Income Distribution and Rates of Urban Migra- tion in Tanzania" (Unpublished paper, Economic Research Bureau, University of Dar es Salaam, 1972). 45 the growth rate of school-leavers could be inferred from the data contained in Table II-13, which indicates the growth of primary and secondary school enrollment in Uganda over the period 1955-1969. The total primary and secondary enrollment increased by 383,102 over the period 1955-1969. The increase alone is substantially greater than the total enrollment in 1955 or about equal to that in 1960. If the majority of these had settled in urban areas after leaving school, as graduates and/or school-leavers, it would provide the strongest evidence to support the general conclusion that the young and formally educated constitute the largest single group of urban unemployed in the past few years. It would follow that the lowest unemployment rates in urban East Africa may actually be found among illiterates. Related to this phenomenon is the fact that a relatively high urban family income structure would make it relatively easier to support the young and unemployed. Hence, the school-leavers can afford a longer "wait- ing period" before accepting a less prestigious and low-paying job. If this hypothesis can be accepted, it would imply that urban unemployment in East Africa would tend to grow with educational levels and with increasing real incomes of urban families. Unless education begins to be supported substantially by payment of fees, thus shifting the current social cost of education into private cost, the number of school-leavers is bound to increase in the future despite the indicated trend of unemployment. The redistribution of current educational expenditure in favor of non-formal education would not only meet the demands for skilled and semi- skilled workers, as the United Nations study quoted above suggests, but also help restrain the growth rate of school-leavers by tending to increase 46 Table II-13 Uganda: Growth in School Enrollment by Category (Aided Schools Only) Total Junior Senior Year Primary Secondary Secondary Total 1955 258,887 7,359 5,756 271,992 1956 280,269 12,468 3,256 1957 313,391 13,821 4,623 1958 333,785 17,352 4,999 1959 354,473 19,997 5,555 1960 363,013 21,341 6,457 390,811 1961 387,295 25,840 6,699 1962 404,759 30,236 7,826 1963 466,562 39,136 10,195 1964 482,470 43,397 12,652 1965 569,219 9,237 17,323 1966 633,546 21,548 655,094 1969 709,708 35,924 Growth in Total Enrollment: 1955-1969 383,102 Source: Uganda Government, Statistical Abstract, 1962, 1965, 1970. 47 the private cost of formal education. Further elaboration of the problem of urban unemployment in East Africa cannot be made until more data become available. The analysis of the labor market in East Africa (historically and empirically, as much as the available data would allow), with particular reference to ugandan data, suggests the following broad conclusions. (1) The excess supply of labor for the economy as a whole has been substantial and is growing in the face of stagnation of total employment and rapid growth in total population and labor force. (2) The rate of growth in urban population and the labor force has been much faster than for the economy as a whole, reflecting the effects of both natural growth rate and rural to urban migration. Hence, urban unemployment or excess labor supply has been growing at a relatively faster rate, given the stagnation in urban employment as well as total employment. (3) The characteristics of the urban-unemployed are so complex that a mere comparison of the aggregate urban employment and aggregate urban supply of labor would lead to distorted conclusions. While the excess supply of unskilled urban labor force can be described by a comparison of the number of unskilled laborers with unskilled employment, the excess supply of educated urban labor reflects an over-supply in quantity as well as a failure to relate the content and orientation of educational pro- grams to the development of skills in short supply. Consequently, an excess supply of school-leavers exists side by side with an excess demand for a skilled labor force. CHAPTER III AN.ANALYSIS OF THE BEHAVIOR.OF WAGES: THE CASE OF UGANDA 0n the basis of the rapidly growing excess labor supply described in the previous section, and in the absence of monopoly or monopsony elements, economic theory would have predicted a downward trend in average wages. While average wages may have indicated a secular decline, wages in some distinct categories of the labor market could have increased. For example, the critical shortage of skilled labor in Uganda and East Africa could lead to a rise in skilled vs. unskilled wages. Moreover, the excess supply of school-leavers with formal education would.have fbrced down administrative or clerical salaries so that, as above, the differential between skilled wages and clerical salaries would have secularly shifted in favor of the former. Given a downward pressure that would have been expected to operate on both salaries for school-leavers and wages for the unskilled, the direction of the change in the differential between the two could not, a_priori, be easily predicted. It would narrow if salaries de- clined faster than unskilled wages, or widen if unskilled wages declined faster than salaries. Thus, the overall impact of these adjustments in the level and struc- ture of wages and salaries would have been to reduce, or at least restrain, the average earning level in the "modern" sector, keeping the differential between urban and rural incomes at a level justifiable by economic forces such as differences in productivity levels, differences in cost of living, 48 49 and/or the cost of transferring labor between sectors. The impact of these expected changes in wages and salaries on employ- ment would be dependent on the extent of the decrease in average wages relative to the cost of capital -- the other factor of production. Theo- retically, the relative cost of labor could remain constant, decline, or increase despite the fall in average wages, depending on the magnitude and direction of change in the cost of capital (i.e., rent). The shortage of capital is one of the basic characteristics of a developing nation like uganda. In the absence of distortions, one could, therefore, assume that the secular fall in average wages would lead to a corresponding fall in the relative cost of labor or the wage-rent ratio. Even if some fall in the cost of capital (rent) was allowed, it was un- likely to have been extensive enough to alter the expected decline in the wage-rent ratio. The Normal wage Employment Trade—Off The fall in the relative cost of labor should have been expected to lead to an increase in the employment of labor, contrary to what was observed in Uganda. The theoretical relationship between wage-rent ratio and employment of labor can be shown by using a simple graphic illustration as in Figure l, where the horizontal and vertical axis indicate units of labor and capital. Let the initial wage-rent ratio be w'o/ro where wb is the initial level of average wages and r0 is the initial level of capital cost. Let AA be the corresponding factor-price line and 9095 represent the isoquant curve showing the various alternative combinations of capital and labor which 50 ’7‘ FIGURE 1 CAPITAL A B go I I I ' I I Qo . I I I I I I . I I . 1 x O 7 L L A B o l LABOUR Fig. The Relationship Between Wage-Rent Ratio and Employment 51 could be used to produce the same output Q0. The producer will minimize cost by producing Qo at the point indicated by m, where the factor-price line would be tangent to the isoquant curve. In other words, a cost- minimizing producer will employ 0Lo and 0K.o units of labor and capital, assuming, of course, some flexibility to substitute labor for capital and still produce efficiently. Now let the relative cost of labor fall from the initial ratio w'o/ro in such a way that the original factor price line AA shifts to BB. Let the desired output remain the same as under initial ratio wb/ro. The cost-minimizing equilibrium will now be at point n. Hence, the same output Qo will now be produced by using 0L1 and 0K1 units of labor and capital. As shown in our figure, the movement from.m to n is a result of pure sub- stitution effect. The diagram could easily be altered to show both the income and substitution effects. Nevertheless, it is shown that under normal assumptions a fall in the relative cost of labor (wb/ro) would be expected to lead to an increase in employment of labor. Employment could also increase even if it is assumed that each firm or industry required fixed capital-labor ratios. A fall in the relative cost of labor could mean that potential investors would tend to favor labor-intensive industries in their investment plans, thus creating poten- tial employment opportunities. The extent of the expected increase in employment of labor as a results of a fall in the relative cost of labor would, of course, depend on the elasticity of demand for labor. Harris and Todaro46 have recently 46J. R. Harris and M. P. Todaro, "Wages, Industrial Employment and Labor Productivity: The Kenyan Experience,” East Africa Economic Review, Vol. I, No. l (Nairobi, June, 1969). P. 36. 52 estimated the coefficient of elasticity with respect to private industry and commerce in Kenya to be about 0.76. In other words, a 10 percent reduction in wages would be expected to give rise to a 7.5 percent in- crease in employment. It could be assumed that the result applies more or less to Uganda and Tanzania whose labor market conditions are very similar to Kenya's. The Ugandan Case In the above analysis, the expected trends in wages if competitive market forces are allowed to operate have been described. In practice, however, the trend in wages may be influenced by non-market forces. In what follows, the predictions of the theoretical analysis to a condition of excess supply of labor in Uganda will be tested with the actual exper- ience of Uganda. It is crucial to know how well theory can explain the actual behavior of wages because any significant divergence between the actual and the theoretical outcomes is the best clue to the existence and importance of non-market or institutional forces in the determination M of wages. Changes in Average wages An analysis of the actual trend in wages, and the forces which shape it is a difficult and as yet an unsettled area in economics. The lack of basic data in a developing society such as Uganda, covering a suffi- ciently long period of time and broad sectors of economic activities, compounds the difficulty. However, the limited information available will be utilized to provide broad, but useful, conclusions. With respect to 53 Uganda, the studies by J. B. Knight47'48 are very helpful, particularly in providing a systematic presentation of fragmentary data from published government statistics and information not readily available. The data in Table 111-1 show the trend in average money earnings per employee for all sectors, all employees, and Africans alone, while Table III-2 presents the data disaggregated as between private and public sectors. It will be seen that average earnings of all sectors over the period 1957-1966 rose by about 146 percent or about 15 percent per annum. Moreover, this growth rate was much faster over 1962-1966 (17% per annum) than during 1957-1962 (10% per annum). These rates are substantially lower when all employees are considered. For instance, average earnings of all employees in all sectors was about 10 percent over the period 1957-1966. Average earnings of Africans in the public sector (Table III-2) over 1957-1966 increased by 159 percent or about 16 percent per annum in compar- ison to 107 percent or 11 percent per annum in the private sector. The data support the suspicion often held with respect to East African countries -- that increases in African wages, particularly in the public sector, accounted for a significant proportion of the rapid rise in average money wages. Table III-3 shows the trend in the index of average real earnings in Uganda. Over the period 1957-1964, average real wages of Africans in 47J. B. Knight, "The Determination of Wages and Salaries in Uganda," Bulletin of the Oxford Institute of Economics and Statistics, Vbl. XXIX, 48J. 8. Knight, "Earnings, Employment, Education and Income Distribu- tion in Uganda," Bulletin of the Oxford Institute of Economics and Statis- tics, Vol. xxx, No. 4 (Nevember, 1968). pp. 267-297. 54 Table III-1 Average Annual Cash Earnings per Employee: Uganda, 1957-1966 Average Earnings Index All Sectors All Sectors All Sectors All Sectors and and and and Year All Races Africans All Races Africans (1957 = 100) 1957 91 55 100 100 1958 98 59 108 106 1959 104-7 63.3 115 114 1960 110.2 68.4 121 123 1961 124.9 78.5 137 141 1962 132.7 88.2 146 159 1963 142.0 98.7 156 178 1964 155.5 116.8 171 210 1965 170.3 129.2 187 230 1966 181.0 138.0 199 246 Source: For 1957-1964, J. B. Knight, Determination, p. 237. For 1965-1966, Ugandan Government, Statistical Abstract, 1969. Table III-2 55 Average Annual Cash Earnings by Sector Average Earnings Index 1957 = 100 All Races Africans All Races Africans Year Public Private Public Private Public Private Public Private 1957 94.2 88.2 60.8 51.9 100 100 100 100 1958 103.0 94.5 66.8 53.2 109 106 110 103 1959 106.8 103.2 69.0 58.9 113 116 114 114 1960 116.4 106.0 77.0 62.2 124 119 127 120 1961 131.6 120.0 87.5 71.6 140 135 144 138 1962 143.3 125.0 101.0 78.4 152 141 168 151 1963 148.4 137.8 112.3 89.3 158 155 185 172 1964 168.3 146.3 141.8 97.8 179 165 233 188 1965 178.2 165.2 -- -- -- -- -- -- 1966 194.0 173.0 158.0 124.0 -- -- 259 207 Source: Same as Table III-l. 56 Table III-3 Index of Average Real Earnings per Employee by Sector: Uganda, 1957-1964 All Races Africans All Public Private All Public Private Year Sectors Sector Sector Sectors Sector Sector (1957 = 100) 1957 100 100 100 100 100 100 1958 104 105 102 102 106 99 1959 111 109 112 109 109 109 1960 128 130 126 130 133 126 1961 115 117 113 118 121 116 1962 154 160 148 167 177 159 1963 161 163 160 183 191 178 1964 164 172 158 202 224 181 Source: Same as Table III-1. 57 all sectors increased by 102 percent compared to 110 percent for average money wages over the same period (Table III-l). When all employees are considered, average real and money wages rose at rates of 64 and 71 per- cent respectively. Hence, the rise in consumer prices, shown in Table III-4, and the cost of living, accounts for only a small portion of the rapid increase in money wages. Given the stagnation in total and African employment, the rapid rise in average money wages leads to significant increases in the total wage bill fer Uganda. Table III-5 shows that the increase in the total African wage bill for the economy as a whole was about 11 percent over the period 1956-1964. The increase has been particularly rapid since 1960, as indi— cated by a 13 percent annual rate of growth over the period 1960-1964. Table III-5 also shows that the increase in the public sector wage bill was greater than for the economy as a whole, while the reverse was true for the increase in the private sector wage bill. This indicates that in the public sector, wages have risen faster and stagnation in employment has been less pronounced than in the private sector. Changes in Unskilled wages Much of the rise in average African wages was explained by significant increases in unskilled wages. This is clear from the data contained in Table III—6. For the economy as a whole, it shows that a relatively larger percentage of the total African male employees moved up from the 0-99 shillings earning bracket than from the higher income brackets. The improvement in the relative earning position of the unskilled was even more pronounced in the main urban towns of Uganda. As shown in Table III-7, comparing 1965 with 1958, the percentage of African male employees earning 58 Table III-4 Consumer Prices: General and Food Indices Uganda (Kampala), 1963 = 100 African European Year General* Food General Food 1960 97.9 96.7 90.3 93.9 1961 113.5 118.7 92.2 94.6 1962 97.9 96.7 96.1 97.4 1963 100.0 100.0 100.0 100.0 1964 109.4 112.1 102.9 101.7 1965 127.1 134.4 106.5 105.5 1966 121.9 129.7 110.9 108.4 1967 128.1 138.5 113.0 108.8 1968 121.9 124.2 119.0 116.2 *General excludes rent. Source: ILO, Bulletin of Labor Statistics, 1970, p. 67 59 Table III-5 Percent Increase in wage Bill for African Employees in Uganda: 1956-1964 1956-64 Annual Rate 1960-64 Annual Rate Public Sector 140.3 11.6 77.6 15.4 Private Sector 107.0 9.5 49.6 10.1 Economy 123.6 10.6 61.9 12.8 Source: A. Baryaruha, Factors Affecting_Industrial Employment: A Ugandan Experience, 1954-1964 (Nairobi: Oxford University Press, 1969), p. 14. less than 100 shillings dropped more in the main towns than in other areas, while the percentage in the higher income brackets rose more in the former than in the latter. Thus, there was a tendency for a distinct narrowing in the urban wage structure, while the urban and non-urban earnings dif- ferential appeared to have widened. In the absence of classified data on employment by category of skill, it is difficult to be precise on the relative decline of employment as a result of the increase in average wages. Baryaruha49 estimates that the number of unskilled employees in the economy as a whole dropped by about 15 percent over the period 1959-1964, while in construction and local gov- 50 ernment it dropped by 22 and 41 percent respectively. J. 8. Knight has indicated that the relative fall in the number of unskilled employees has 49A. Baryaruha, Factors Affecting Industrial Employment: A Ugandan Experience, 1954-1964 (Nairobi: Oxford University Press, 1967). 50J. 8. Knight, Earnings, pp. 268-270 and pp. 278-282. 60 .mmsmmfi maceum> .muomnumnd Hmoeumfiumum .ucoacuu>ow pronoun “condom .muoxocun ca ooumowosa one mooa>Hom OHHnsmc Home HmNV HvNV Acme Have Ammo HmHV ANNV AmNV mN vH oH om me I we vN em Nv Hayes .NNV .ch .Hmc HoHc And love .mm. lens AeHv Ame .oNo Hove .Hmc Aves Ammo ucaeauo>oo HuooH Acme “was Acme ANNC AHHV onV .HNC lave Ammo .omv inc Am. and lec Ammo aquacuosoo Houucmo Ammo Ammo HHmo ANNV Hch Ammo Ammo ANmV AeHc AmH. HNV Ame HmHV HHov ANev HN mH HN eH NH me we on «H mH mN mm oe on we mmoH>umm .omHz $8 38 82 1va 3d :8 :3 NS .an And Go 8. 8o NS 83 an em me we HN HH mN mH NH NN mH ov em as am mooH>umm .eoz can .om loos Ach loNo HaHv ANH. “Hes loo. Ammo lame ANNV Ans lac HoHv ANec Ammo Hm me me on NH on we no mm Hm m m N 6m 6m .seoo can .mcmua 14H. AHHV Ame AHNV Ammo Ammo revs me. “no Ammo Amer lev em Hm oN mH oH em we Hm mN mH 5 MN «N om He mouoaeoo HmHv AvHV AvHv Ame Ame ANme Ave. Anne .vHo Hch AHNV ANHC AvHV Home keno mN oH «H m o No mm mm 6N HN NH mN «N me me :oHuonuumcoo Ammo Ana AHNV Leno mev ANmV Am. Ava .oNv AmHo Ave “Hay HHV Ammo Ahmo Hm mN HN NH a am mm mm 0N NH NH eH mN No we .mucH .Hscmz .omHz NH m oH N N me H» mm MH m vH HN mm om om nauseous coca mo .macmz AHHV Hue Loo Ame Ana Ammo keno knee AmHv Ava Ammo Ammo Lave Acne .Neo mm «H mH m H on Nm Hm we a NH mm 4N Ne om ocHNuumso can mchHz AmN. AHHV AmHo .mo Ame Ammo 1mm. Amer AvHo Ame HoNc ANNV ANHV lees .Hmc mcHuesm can m HH mH N N mH mN NN mN HH an we me He em ecHsmHm .Nnumouom m 0H s e N mH NH NH N q me Na Na am am mcHnaHo :oauoo ANmV ANNV ANNV HmHv Awe Ammo Ammo Home AmHV AHNV Ave AHHV AoHv “one AHBV H H H u H mm Nm me n N mm me am No em musuHaoHuma moocucmonom moooucoouom moooucooumm momH vomH NomH ommH ommH momH vomH momH oomH omaH mmmH vomH momH oomH ommH NuumsecH uo>o new mmaHHHHsm ooN umeHHHHnm mmHuooH macHHHHsm mmuo race: “on umcHHHHem «mmfouo emu: am00a>uum cannon one ous>wwm .huumsosH segues museum an mooaoameu was: confined no coeusnauumwo OIHHH OHQUB 61 Table III-7 African Adult Male Employees in Main Towns and Other Areas, by Wage-Group, as a Percentage of Total, Uganda, 1958 and 1965 (shillings per month) Private Sector Public Sector 0-99 100-174 175- 0—99 100-174 175- 1958 Main Towns 77.6 11.5 10.9 64.1 18.1 18.8 Other Areas 80.8 9.0 10.1 72.7 16.3 11.0 1965 Main Towns 17.0 53.9 29.1 5.8 40.7 53.5 Other Areas 36.7 40.3 23.0 24.1 44.9 31.1 Difference (1958-1965) Main Towns -60.6 42.4 18.2 -57.3 22.6 34.3 Other Areas -44.0 31.3 12.9 -48.6 28.6 20.1 Source: J. B. Knight, Earnings, Table 6, p. 246. 62 been substantial in both the private and public (central and local) sectors. In short, the substantial rise in average earnings was achieved at the expense of employment, particularly for the unskilled. The Civil Service Salary Structure Table III-8 provides some indication of salary levels and structure in uganda in 1966. It shows that the minimum and maximum salary scale for civil servants with eight years of education51 (e.g., Clerical Assistant) were 114 and 200 pounds per annum, or 1.3 and 2.2 times the minimum wage for the unskilled workers. Thus, primary school graduates desiring em- ployment in the private sector would tend to ask a starting wage substan- tially higher than the minimum wage for the unskilled lacking formal primary education. The demand for clerical assistants requiring formal primary education is limited in the private sector, while the irrelevance of formal primary education for factory (or manual) labor does not justify wages much higher than the legal minimum. The minimum and maximum salary scales for Junior Secondary graduates (12 years of education) in the civil service (e.g., Clerk) were 208 and 598 pounds per annum or 2.3 and 6.6 times the minimum wage for the unskilled. It seems logical to assume that private employers would prefer to employ primary school graduates for factory jobs (even if some education is de- sirable for the job), primarily because they will be cheaper than junior secondary graduates at a wage rate comparable to that of the civil service. The minimum and maximum salary scales in the civil service for senior 51As salaries are incremental, the length of service and age are also factors determining salary differential. 63 Table III-8 Comparisons of Teachers and Civil Servants' Salaries, 1966 Formal Education Completed Salary (P.A.) Classification Level Years Minimum Maximum Increments Teachers Secondary Honors Teacher Degree 17 798 1752 22 Secondary Teacher Degree 17 738 1752 21 Grade 5 HSC + T2 16 612 1080 16 Grade 3, 4 CSC + T2 14 300 828 24 Grade 2 JSC + T4 12 189 354 13 Grade 1 P6 8 132 276 17 Civil Servants A Honors Degree 17 798 1791 19 G3-1 Degree or Equivalent 17 687 1671 23 Executive Officer HSC 14 473 1671 32 Clerk CSC 12 208 598 22 Clerical Assistant JSC 8 114 200 11 Source: J. 8. Knight, Salary Structure in Uganda, Economic Development Research Project No. 106 (Kampala: September 9, 1966), Table 9. See also footnote 48. Key: HSC = High School Certificate, CSC = Commercial School Certificate, JSC = Junior School Certificate. 64 secondary school graduates (14 years of education) are shown to be about four and eight times the minimum and maximum for primary school graduates and about two and three times the minimum and maximum for junior secondary graduates. The salary differential between junior and senior secondary graduates is too large for a mere two years differential in education. An even wider differential exists between college and lower level graduates. The salary scale for teachers follows very closely the salary scale in the civil service. Given the same level of formal education, a some- what higher salary scale for teachers is explained primarily by the extra training required to acquire a teaching certificate. In addition to fixing the minimum and maximum for each level of sal- ary scale, the per annum increments and the number of years it would take to reach the maximum pay are clearly stipulated. Thus, the rise in total cost of salaries is also a function of changes in staff numbers and staff regrading. Table III-9 shows that over the period 1956—1964, the total civil service salary cost rose by 114 percent. Growth in staff numbers accounted for a major portion of the rise, while increments in salary scales were also important. The influence of staff regrading has not been significant. The rise in the total salary cost of education has been faster rela- tive to the civil service salary cost. Table III-10 shows that over the period 1955-1965 the total salary cost of secondary and primary education rose by 348 and 227 percent respectively. An increase in the number of teachers has been important in both increases, but more so for secondary education. The impact of staff regrading was relatively less important in secondary education, while the rise in salary scales was significant in 65 Table III-9 Index of Civil Service Salaries, 1956-1964 Index of Salaries Attributable to: 1956 1960 1964 Staff numbers 100 129.8 184.0 Staff regrading 100 99.1 102.2 Salary scales 100 105.0 113.6 Total salary cost 100 135.1 213.6 Table III-10 Salary Cost of Primary and Secondary Education, 1955-1965 Index of Salaries Primary Education Secondapy Education Attributable to: 1955 1960 1965 1955 1965 Teacher numbers 100.0 136.1 179.0 100.0 303.9 Teacher regrading 100.0 108.7 115.4 100.0 105.0 Salary scales 100.0 116.0 157.9 100.0 139.3 Total salary cost 100.0 171.6 326.9 100.0 447.9 Source: J. B. Knight, Salary Structure in Uganda, Tables 6, 9, 10. 66 both, particularly for primary school teachers. It is obvious that total salary cost could increase substantially by upgrading existing or new staff to higher scales in the salary structure. Moreover, the rise in the level and total cost of salaries has not cur- tailed growth in the number of civil servants and teachers. Africanization of the civil service, and extension of government involvement in directly pro— ductive activities through Development Corporations, accounted for the rise in the number of civil servants. The rapid growth in the number of teach- ers was explained by the rapid increase in school enrollment described earlier. The fact that the government, unlike private employers, is not guided by profit-maximizing objectives (excluding perhaps its behavior in direct- ly productive activities), could partly explain the rise in public employ- ment, despite the rapid increase in level and total cost of salaries. But such policies can only be maintained over the short run. Sooner or later, the capacity to generate additional employment opportunities in the public sector would be constrained by limited capacity to raise additional reve— nue, either through taxation or borrowing (internal and/or external). In Uganda this limit appears to have been reached already. As evidence, the Minister of Finance was recently reported to have said: "Government has decided to continue the embargo on recruitment within the civil service, and, if need be, reduce employment opportunities within the government."52 Contrary to the predictions of a competitive market model, the above 52See African Development (African Buyer and Trader Ltd., Wheatsheaf House, Carmelite Street, London EC4Y OAX, June 1972), p. U. 12. 67 analysis demonstrates that the actual levels of wages and salaries in Uganda have been rising rapidly despite the demonstrated upward trend in the aggre- gated unemployment gap and urban open-unemployment. The rise in unskilled wages has substantially reduced already-employed unskilled labor, and added to the excess supply resulting from a rapidly growing population and labor force. The growth in salaries has been accompanied by some growth in the level of employment of both teachers and civil servants. But an increase in salary cost must indirectly and negatively affect total employment by limiting the government's ability to expand productive investments in agri- cultural and industrial activities. Changes in the Distribution of Income The above non-market changes have affected the relative distribution of incomes within the "modern" or urban sector as well as between rural and urban areas of Uganda. An estimation of these changes is difficult because of statistical inadequacies and the conceptual problem of pricing peasant outputs. Moreover, non-uniform distribution of free and/or sub- sidized public services among different areas and/or groups limits the validity of estimating changes in relative position on the basis of earned income. Despite these limitations, however, some estimation of the broad changes in the distribution of income within and between areas is possible. J. B. Knight53 has attempted to estimate the changes in the distribu- tion of income in Uganda over 1954-1964 by considering private sector in- comes and households by category of recipients. He shows that peasant in- comes per household, on the average, remained more or less constant at 53J. B. Knight, Earnings, pp. 290-293. 68 76 and 77 pounds per annum, while the share of peasant income in total private income dropped from about 70 to 62 percent. As these estimates are aver- ages, the relative positions of some peasants may have been much worse. In contrast, the average income per household of wage and salary earn- ers over 1954-1964 is shown to have risen from 100 to 189 pounds per annum. Disaggregated between taxpayers (mainly salariat) and non-taxpayers (mainly unskilled) this shows that income per household increased from 936 to 1436 pounds for the former and from 71 to 135 pounds for the latter. The share of wage and salary earners as a group rose from about 18 to 23 percent and, if shown separately, the share of unskilled workers would have been even more marked. Finally, Knight estimates the share of profits in total private income to have grown from 13 to 15 percent over the per- iod 1954-1964. He indicates that the growth was mainly for companies rather than self-employed business. In comparison, Stoutjesdijk54 estimated that for the total manufac- turing sector in Uganda, annual rate of growth of the sum of profits, interest, and depreciation has been -7.7 and -2.6 percent respectively in 1957-1962 and 1963-1965. When he considered five individual firms owned by the Uganda Development Corporation, the results regarding profits were mixed.55 For Uganda Cement (1953-1966), and Nyanza Textiles (1959- 1966) net profits before taxation as a percentage of total capital were positive for each year. For Nyanza Textiles the percentage had increased from 20 in 1959 to 32 in 1965. For Uganda Cement, the percentages fluctuated 54E. J. Stoutjesdijk, 9p, cit., p. 27. 55Ibid., Tables 12, 15, 18, 21, 24. 69 widely from near zero to high levels so that no clear trend was discernable. In the remaining three firms, metal products, chemicals and fertilizers, and distilleries, the profit levels were, on the average, negative. In sum, then, Knight's suggestion that the share of profits in total private income had shown only slight improvement appears to be reasonable. If we had considered the distribution of disposable income, the rela— tive position of unskilled wage employees would be better than indicated above. While export taxes and income taxes have significant impacts on peasant in- comes and the incomes of salariats and profit receivers, respectively, un- skilled workers are affected by income tax relatively less. This would tend to narrow the real earnings within the urban sectors while the rural-urban differential would tend to widen. Furthermore, while indirect taxes which fall more heavily on high income earners in the urban areas would tend to reduce the urban real income differential, their impact on the urban-rural differential is not at all clear for the higher urban area cost of living would tend to be cancelled by the concentration of public services it enjoys. To conclude, it is shown that the differential between income of salary earners and unskilled wages have tended to narrow from below. The evidence suggests that the rise in average wages have been at the expense of primarily peasant incomes and, to a lesser extent, profits. These changes were contrary to what one would expect on the basis of normal economic analysis. Before this section is concluded, some broad estimates of the wage-rent ratio in Uganda will be presented on the basis of the actual trends in the cost of labor described above and some estimate of the trend in the cost of capital to Ugandan economy. I) ('1' 'n E! fl) ’[1 r:- "r u, 70 The Wage-Rent Ratio in Uganda It was indicated above that given the prevailing factor endowments in East Africa, the secular fall in average wages predicted on the basis of a growing excess labor supply should result in a fall of the wage-rent ratio and a rise of labor employment. Implicit in the analysis was the assumption that changes in aver— age wages or the cost of capital would make labor or capital more or less costly. In reality, this need not be the case if there are corresponding changes in the productivity of labor or capital to cancel out the changes in factor cost. Ideally, then, the estimate of wage-rent ratio should be based on indices of wages and cost of capital deflated for changes in productivity. In the absence of usable productivity data, however, the estimate of wage-rent ratio in Uganda will be based on the simplified assumption that labor and capital are of constant quality. The index of average money wages in Uganda over the period 1957-1966 has already been provided in Table III-l above. What remains is to esti- mate the cost of capital over the same period. This would be difficult, since it involves knowledge of a number of elements -- the rate of inter- est on borrowed funds (domestic and foreign); depreciation allowances provided on investments and/or expansion of existing facilities; import duties on imported capital goods; the lifetime of the investment in capital goods and the ease with which foreign exchange is made available, including information on the official rate of foreign exchange, so that the cost of capital could be expressed entirely in local currency, comparable to the cost of labor. 71 Unavailable or inadequate statistical data on these elements of the cost of capital make it difficult to measure accurately the trend in the cost of capital. The importance of some of these elements can, how- ever, be safely ignored. For instance, since the foreign component of Uganda's gross capital formation is estimated to approach 90 percent,56 it is possible to assume that imported capital goods serve as the approxi- mate indicator of the trend in investment. This will allow us to define the cost of capital as r = f(i,n)[PC] where f(i,n) is the fraction of the original investment which must be earned annually over n years, the life- time of the investment, to cover both interest and depreciation, and PC is an index of the supply price of imported capital goods. Even this simplified formula to calculate the cost of capital would require knowledge of the domestic and foreign components of funds used to purchase the imported capital goods and the corresponding domestic and foreign rates of interest. It would also require detailed knowledge of the extent of depreciation write-offs, profit-tax privileges, import duties and other domestic costs so that PC, the supply price of imported capital goods, could be properly adjusted. While accurate statistical data concerning these requirements may not be available for Uganda, and for Kenya and Tanzania as well, it is pos- sible to indicate broadly what the trend has been. Since the 1950's and particularly since independence in the early 1960's, the East African governments had embarked on a policy of rapid industrialization. To encour- age capital investment, interest rates were generally kept low, foreign 56J. 8. Knight, Earnings, p. 272. a 72 aid policies were such as to encourage imports of capital goods, fbreign exchange was made easily available to importers of capital goods, and import duties on capital-goods were non-existent or relatively minor.57 The artificial cheapening of capital -- the relatively scarce factor in East Africa -- was intensified by the competition among the three coun- tries to attract capital. The effect of all these policies was to deflate the effective price of imported capital goods, PC, and the cost of capital as defined above in the simple formula. The Ugandan government provides annual data on the price of imported capital goods -- specifically, machinery and transport equipment. The fact that it is given in local currency makes it readily comparable with average wage data. The index of the price of imported capital goods has been used to indicate the trend in the actual cost of capital goods, bearing in mind that the true cost would have been smaller if adjustment for the effects of capital-cheapening domestic policies were considered. The estimate of the wage-rent ratio or the relative cost of labor based on the straight use of unadjusted index of import price of capital goods would underestimate the actual magnitude of the relative cost of labor and should be interpreted with this in mind. Table III-ll brings together the Ugandan index of average money wages, for Africans and all races, and the price index of imported capital goods, machinery and equipment, for the period 1957-1966. These have been used to calculate an indicator of the trend in wage-rent ratio over the same 57 . . . . . See D. Pearson, Industrial Development in East Africa, Studies in African Economics No. 2 (Nairobi: Oxford University Press, 1969): PP. 167- 192. 73 Table III-ll Estimate of the wage-Rent Ratio for Uganda: 1957-1966 (l) (2) Index of Average Money Wages P.A. Price Index wage-Rent Ratio (a) (b) of Imported _1_(_a_)_ L(_b_)_ Year All Races African Capital Goods* 2 2 1957 100 100 100 1.00 1.00 1958 108 106 94 1.15 1.13 1959 115 114 98 1.17 1.16 1960 121 123 101 1.20 1.22 1961 137 141 99 1.38 1.42 1962 146 159 106 1.38 1.50 1963 156 178 98 1.59 1.82 1964 171 210 105 1.63 2.00 1965 187 230 115 1.63 2.00 1966 199 246 99 2.00 2.49 Source: Ugandan Government, Statistical Abstract, and Table III-1 above. * Includes only machinery and transport equipment. : r! ...‘h l r (b ca 1'? (EV ff (3 74 period. It can be seen that the ratio or the relative cost of labor over the entire period 1957-1966. These have been used to calculate an indicator of the trend in wage—rent ratio over the same period. It can be seen that the ratio or the relative cost of labor over the entire period doubled if the money wage for all races is used, more than doubled if the wage scale for African employees alone is used. In both cases, how- ever, the estimates would be higher if the index of the price imported capital goods were deflated to take account of policies to reduce the cost of capital investments. Nevertheless, the estimate presented is suf- ficient to indicate that the relative cost of labor has been rising de- spite the abundance of labor and scarcity of capital which characterizes the Ugandan economy. In summary, the Ugandan case illustrates a situation where average and unskilled wages have been rising despite a secular and significant rise in sectoral and total unemployment, and contrary to the normal wage-employment trade-off presented above. Thus, there was a great dis- parity in income between urban and rural areas. Moreover, the differen- tial in income between manual and clerical employees and between both and white collar workers is much greater than is the case in developed nations, or what would have been the case on the basis of productivity differen- tials. This situation comes from maintaining a rigid relationship between the level of formal education achieved and earnings. In a country such as Uganda where the bulk of the cost of public services is wages and salaries, rapid growth in those areas will signifi- cantly limit the growth of employment in the public sector. Since market 75 forces of demand and supply of labor fail to explain the actual trends in wages and employment, the next chapter will examine the role of non- market forces in the determination of wages and employment. CHAPTER IV NON-MARKET DETERMINATION OF WAGES AND SALARIES An ingrained distrust of the market mechanism as a means of resolving national economic issues has been one of the basic characteristics of British colonialism in East Africa. National economic decisions were perceived as legitimate civil service functions. For example, Colonial Labor Commissioners dominated the field of labor policy in general and the determination of wages in particular. There was no pre-Wbrld war II legislative attempt to fix minimum- wages for unskilled labor in East Africa. Yet it would be wrong to assume that during this period wages were determined by the free interaction of demand and supply of labor. A collusion of non-African private employers and generally pro-employer Colonial Administrators effectively prevented wage flexibility by controlling the flow of labor supply. As indicated in Chapter II, compulsory and/or indirect measures (e.g., taxes) were put into effect whenever employers faced a shortage of labor. Such non- market determination of unskilled wages was rationalized by claiming that African peasants were unresponsive to monetary or wage incentives. In reality, the motive was exploitation, since the rapid expansion of peasant cash crops in the pre-Wbrld war II period had already disproved the "unresponsiveness" hypothesis. In regard to high level or administrative-supervisory labor, where Africans and non-Africans were in potential or actual competition for the same or related jobs, competition was prevented or at least retarded by 76 77 the colonial policy of separating employees into racially distinct labor markets and salary categories. Related to this practice was the fact that salary scales tended to be higher for jobs normally held by non- African employees (e.g., administrative, supervisory, clerical, etc.). For Africans, moving into these job classifications meant movement into a higher status and salary category. Since these jobs required some formal education, Africans aspired toward acquiring formal education in mission schools or government schools, rather than technical skills. This practice of equating job status and salary to level of formal education has generally been carried over by the African elite who assumed political power in East Africa after independence in the early 1960's. In the post-werld war II period, the salary scales of African civil servants and high-level employees improved substantially, not because of a sudden shift in market conditions but, as shown below, as a consequence of colonial abandonment of the old policy of maintaining separate basic salary scales for the different races. The structure and principle behind the determination of these salary scales has been maintained to this day despite changes that have been occurring in the market conditions of East Africa. Moreover, the post-world war II period saw the gradual introduction of legislative determination of minimum wages for unskilled labor -- a pro- cess which was intensified after the political independence of the East African regimes. The pre-Wbrld war II pro-employer bias in the determina- tion of wages gradually gave way to a post-war bias in favor of urban work- ers. The primary forces behind this shift in wage policy were political and 78 social rather than economic. Where unskilled workers are the dominant segment of the total labor force, minimum wage determination divorced from the economic criteria of the level of productivity achieved by workers implies a substantial distortion of factor-pricing with substantial and negative consequences for the generation of employment opportunities. The post-werld War II institutional or non-market determination of wages and salaries in East Africa will be illustrated below with particular reference to Uganda. The experience of Uganda in the determination of wages and salaries broadly reflects the practices that were prevalent in Kenya and Tanzania as well. The Determination of Salaries Up to 1954, colonial policy set African salary scales at three- fifths of that for expatriates doing comparable work. The differential was explained officially on two grounds: (1) that the expatriate workers had to be paid salaries comparable to what they could earn in their home countries: and (2) that it was unwise to create "any class of Africans... divorced in income and interests from their fellows"58 in the countryside. With the growing African elite and Africanization of the civil ser- vice, the racial differential began to cause resentment. An explanation like (2) above, with implicit racial overtones, in addition to the pattern of social life and housing which separated the races in urban areas, seri- ously weakened the acceptance by Africans of the economic argument for the 8Report of the Commission on the Civil Service of Kenya, TanganyikaL Uganda and Zanzibar, 1947-1948 (Nairobi: Government Printer, 1948), p. 92. (Otherwise known as the Holmes Commission) 79 salary differential indicated by (1) above. These resentments, along with growing political awareness and organization on the part or urban Africans in Uganda and the other East African countries, led to the establishment of common basic salary scales for all employees regardless of race. Thus, in 1954, the Lidbury Commission59 suggested the establishment of basic salary scales applicable to all civil service staff -- with an "overseas addition" to induce expatriate personnel to come and work in East Africa. The immediate impact of this policy was to tie the structure of African salaries to those of expatriates -- the latter reflecting demand and supply conditions in overseas labor markets -- and hence narrowing the gap between African and expatriate earnings. In Uganda, for instance, the impact of the Lidbury Commission Report was said to "more than halve the gap between local and expatriate earn- ings"60 over the period 1954-1960, although the gap later widened again, reaching in 1965 the same level as in 1954. The widening of the gap was not due to downward adjustment of African salary scales, which were in fact rising, but to increments in the "overseas addition" to expatriate employees paid by the home government, Britain. The African elite, since independence, maintained the colonial basic salary scales in addition to assuming non-salary benefits such as paid vacations, free or subsidized housing, and so on, without much considera- tion to the needs of economic development -- supposedly the basic objective 59Report of the Commission on the Civil Services of the East African Territories and the East African High Commission, 1953-1954 (Nairobi: Government Printer, 1954). 60J. 8. Knight, Earnings, p. 253. 80 of East African regimes since independence. Another important aspect of the salary structure, inherited by Uganda from the colonial period and maintained since independence, was the policy of rigidly tieing salary scales to the level of formal education achieved, with an arbitrarily wide differential maintained between higher and lower levels of education. This was illustrated in Chapter III. The point to emphasize here is, in the words of Henderson: The income differentials between groups with contrast- ing educational attainment are greater than any cost- benefit analysis of the value of education would sustain, and the earnings of manual-clerical laborers, as a per- centage of the wages of white collar workers, are ex- tremely low. In other words, the size and structure of salary differentials had very little to do with merit or the actual productivity of civil service employees. The market for high-level manpower in Uganda was not only monopsonistic, but the monopsonist, being the government, need not at- tempt to maximize profits or set salaries on the basis of productivity. The tendency for some private sector salaries in Oganda to exceed the salary levels of comparable educated manpower in the civil service may create an impression that the private sector acts as a leader in salary determination. Such differentials exist in some specific cases, particularly in the large expatriate firms. Yet, given the fact that the Ugandan government "pays the salaries of about 82 percent of high level manpower,"62 its dominant position cannot be disputed, as a general rule. This position was supported by a Ugandan government commission which 61J. P. Henderson, "wage Policy in Africa," in African Conference on Economics (Mimeograph, Temple University, 1968), p. 3. 62J. 8. Knight, Earnings, p. 259. 3,. uvn' he ‘w'he 81 reported that there was "evidence to show that it is the level of salaries in the government service that determine the general salary pattern in Uganda."63 Consequently, basic salary scales of high—level manpower in the private sector tended to follow closely those in the government sector, even though some changes may occur following promotions and/or improve- ments in the productivity of the worker or the firm. Moreover, as Henderson argues, the paternalistic relationship between government and the private sector, before and after independence, had created a climate where the private sector was quick to emulate the policies of the govern- ment in return for state aid and protection.64 The East African governments had a policy of increasing the employ- ment of high-level manpower through the generation of new jobs and African- ization of existing jobs. Since, as in Uganda, expatriates accounted for over 50 percent of the total high-level manpower, Africanization alone would have required the substantial expansion of high-level education and teachers. But, as was indicated in Chapter III, the level and struc- ture of the salary of teachers follows closely those of civil servants and moreover, the salary cost of teachers constituted a major portion of the total educational cost. Hence, "the manpower needs cannot be met because to increase enrollment to the necessary levels would bankrupt most govern- ments,"65 such as Uganda. 63Report of the Uganda Civil Service Commission (Entebbe: Government Printer, 1963), p. 28. (Otherwise known as the Ani Commission) 64J. P. Henderson, 2p, cit., p. 7ff. 65Ibid., p. 5. 82 In short, the arbitrary determination of salaries on colonial criter- ia, which were primarily non-economic, continued to frustrate the high- level manpower needs and the long term goals of economic growth in Uganda. The Determination of Unskilled wages Like the determination of salaries, the post—world War II determin— ation of unskilled wages in Uganda was inherited from the colonial era. The government, as the dominant employer, and with statutory determina- tion of minimum wages, sets the general pattern of unskilled wages. The colonial basis of wage policy in the post-World war 11 period can be traced back to the Colonial Development and Welfare Act (CDWA) passed by the British parliament in 1940 so as "to promote the prosperity and happiness of the peoples of the colonial empire.“66 The colonies were instructed to take measures to raise the standard of living of Afri- cans "for humanitarian, political, ...and economic reasons."67 Growing labor disturbances in the main towns of East Africa during the forties were interpreted by the colonial regimes as being caused by inadequate housing and low wages and these provided the "humanitarian and political" reasons for legislative determination of minimum wages. The economic justification was quickly formulated by the Colonial Labor Commissioners. For example: Long-term policy is directed towards the maintenance of stabilization between agricultural and industrial areas. This means the development of independent wage ‘ workers without subsidy from the native land unit.... The first step towards stability is the general adaption 66Great Britain, 9p, cit., p. 8. 67Great Britain, "Colonial Office Circular Dispatch of 5th June, 1941," quoted in A. H. Amsden, International Firms and Labor in Kenya: 1945-1970 (Frank Cass & Co., 1971), p. 12. 83 of a 'living wage' standard for an African divorced from his native land unit. Individual territories set up committees69 while the British Govern- ment set up regional commissions70 to study and make recommendations for the criteria, level, and scope of minimum wages. Their reports supported the basic humanitarian, political and economic justifications for a legal minimum wage already laid down by the various labor officials in each territory of East Africa. The first minimum wage order in Uganda was made in 1949 when the minimum wage of 33 shillings per month was set in the main towns of Kampala and Jinga.71 In Kenya, it was enacted three years earlier.72 Initially, the minimum wage was based on the "need" of a single worker and limited to a few large cities. As could be expected, the private employers were unhappy about the emerging trend of post-war wage policy. Colonial officials were somewhat divided in their opinions. Sir Philip Mitchell, a Colonial Governor, warned that the post-war increase in Afri- can wages "is bound to involve, in time at least, a serious consequential 73 problem of unemployment." In contrast, the colonial Labor Ofifcials were not satisfied with a 68Colony and Protectorate of Kenya, Labor Department Annual Report, 69Colony and Protectorate of Kenya, Report of the Committee on African Wages (Nairobi: Government Printer, 1954). (Otherwise known as the Carpenter Report) 70Great Britain, The East African Royal Commission Report, 1953-1954, Cmd. 9175 (London, H.M.S.Q., 1955). J. 8. Knight, Earnings, p. 244. A. H. Amsden, 2p, cit., p. 12. 73P. E. Mitchell (Sir), African Afterthoughts (London, 1954), p. 237; quoted in A. H. Amsden, 2p, cit. 84 "bachelor" minimum wage. For instance, the Kenyan Labor Department re- ported that "...the general level of wages must be described as low."'74 In a report of the latter view, the Carpenter Report stated: "we are con- scious of a feeling not so much of 'complacent satisfaction' as of grow- ing discontent and frustration,"75 since 50 and 25 percent of all urban workers in Kenya were receiving "...insufficient wages to provide for their basic essential needs."76 Implicit in these justifications for in- creasing the level and scope of minimum wages was the need for stabiliza- tion of labor. As indicated in Chapter II, the various labor departments in East Africa had by this time begun to report signs of an emerging excess supply of labor -- a condition which should have led to stabilization or lower turnover of labor. Yet they were still arguing for higher wages to stabilize labor. Following the Carpenter and East African Royal Commission Report, a "family" minimum wage -- sufficient for a husband, his wife and two children -- was adopted in uganda, Kenya and Tanzania. The law did not cover agricultural workers. The effect was to widen an already existing disparity between rural and urban unskilled workers. The basic criterion for a minimum wage was the concept of the Poverty Datum Line -- i.e., an income to enable the basic and elementary needs of the family to be met. As Henderson pointed out, the minimum requirements 74Kenya, Labor Department Annual Report, 1952, p. 8. 75The Carpenter Report, p. 55. 76Ibid., p. 32. 85 suggested for East Africa tended to be significantly above the actual ex- penditure of families according to government surveys.77 The actual mini- mum.wage adopted tended to lie in between the two.78 Initially, the governments were content to use the "family wage" as a goal to be grad- ually achieved. For instance, the second minimum wage order in Uganda was not enacted until 1959. Yet, as Table IV—l indicates, the minimum wages paid by local governments and the central government in Kampala were 38 and 68 shillings per month, respectively. The latter was about 130 percent above the 1949 level or a 13 percent annual rate of growth. A comparable pattern of changes occurred in the other East African countries.79 Increases in the Ugandan minimum wage since 1959 were rapid, partic- ularly after independence in 1962. Over the period 1958-1965, a short period indeed, the legal minimum in Kampala had grown by 355 percent or about 44 percent per annum. Over the same period, the minimum wage paid by the central government to its own employees had grown by 122 percent, or about 15 percent per annum. While the absolute level of minimum wages in local governments was much lower than in Kampala and other main towns, the percentage growth was roughly parallel. The significant improvement in the relative earnings position of African male employees in the main towns and in the lowest income brackets (see Table III-7 above) could be partly and primarily explained by the rapid rise in minimum wages. The consequence in terms of a significant fall in unskilled employment has already been indicated. Furthermore, if we compare the data in Tables II-l, III-1 and IV-l, 78J. B. Knight, Earnings, p. 245. 7 . 9For Kenya, for instance, see A. H. Amsden, 9p, c1t., p. 18. 86 Table IV-l Government-Determined Unskilled Wage Rates (shillings per month) Minimum wage in Legal Minimum Central Government Minimum wage wage in Year Kampala Elsewhere in Local Govt. Kampala 1957 68.0 -- 38 33.0 1958 68.0 -- 38 33.0 1959 75.0 -- 38 75.0 1960 81.0 -- 38 75.0 1961 81.0 -- 47 75.0 1962 104.0 104.0 47 75.0 1963 121.0 104.0 60 120.0 1964 151.0 135.0 78 120.0 1965 151.0 135.0 78 150.0 1965 222 -- 205 455 (1957 = 100) Source: J. B. Knight, "The Determination of Wages and Salaries in Uganda," Bulletin of the Oxford Institute of Economics and 87 a rough correlation between the trend in minimum wages, total employment and average African wages would be evident. Over the period 1958-1961 the latter increased by about 35 percent or by slightly more than 8 per- cent per annum while total employment dropped by roughly 3 percent over that period. The drop in non-agricultural employment was about the same (Table II-4). The impact of the sharp rise in minimum wages in 1962, from 81 shillings per month in 1961 to 104 shillings per month in 1962, was immediate and significant. Average African wages (Table III-l) rose by about 18 percent over 1961-62 while total employment (Table II-l) dropped sharply in 1962 to a level below that which prevailed in 1954. Non-agricultural employment (Table II-4), which had been falling since 1959, reached a level about 7 percent below that which prevailed in 1957. The same general relationship between the changes in minimum wages, aver- age wages, and employment was evident since 1962. Thus, over the entire period 1958-1965, legal minimum wages in Kampala, the minimum wage paid by the central government in Kampala, and that paid by local governments rose by 355, 122, and 105 percent respectively, while average African wages rose by 124 percent and total, as well as non-agricultural, employ- ment more or less stagnated. From the above data, there can be no doubt that the rapid rise in minimum wages contributed to the rapid rise in the level of average wages observed in Uganda in the past. The argument is particularly strong as the Ugandan Government is the dominant employer of unskilled labor as well as high-level government. It may be hypothesized that the premature unionization of labor might, in fact, have been the major cause of the rise in unskilled and average 88 wages. Trade union membership in Uganda since 1954 has grown very rapidly -- from 680 members in 1954 to 65,000 members in 1966.80 On the surface, this would tend to suggest an increase in the power of trade unions vis-a-vis employers and hence in the ability of workers to secure higher wages through collective bargaining. Indeed, negotiated wages in a large number of industries in Uganda were at levels well above minimum wages. Hewever, given the subordination of labor unions to political parties in Uganda and East Africa as a whole, and the limited right to strike, the mere increase in union membership would have an insignificant impact on union power. Moreover, unions in East Africa were and still are finan- cially weak and, therefore, could not sustain an extended strike even if allowed. For these reasons, it is reasonable to argue, as Knight does,81 that the excess of negotiated wages over minimum wages was more probably due to the decisions of the profit-maximizing firms following increases in productivity or preceding productivity increases under the belief that higher wages might lead to lower turnover and higher efficiency. Moreover, in some of the industries, the fact that unskilled wages have been a small proportion of total cost provides a strong explanation of the tendency to observe negotiated wages above the level of minimum wages.82 In general, Knight suggests that trade unions have had, if any, a minor influence on the level of average wages in Uganda. Henderson argued in a similar manner with respect to Nigeria, adding that the trade unions' "...weak economic position is paralleled in Uganda."83 8OJ. 3. Knight, Earnings, p. 248. 81Ibid., p. 249. 82 . J. B. Knight, Earnings, pp. 249-250. 83 J. P. Henderson, 9p, cit., p. 9. 89 Henderson84 develops more fully the paternalistic relationship between colonial regimes and private enterprise which was referred to in Chapter II above. Moreover, he argues persuasively that the same relationship had continued since independence, although the reasons for cooperation might have changed. Implicit in this relationship was government aid and pro- tection extended to the private sector, and private employers' quick adjustment of wages to the level dictated by government policy or minimum wages. Hence, trade unions were not required to force employers to emu- late the wage policies of the government for its own employees. Therefore, the effect of unions on average wages, if any, could have been through influence on government wage policies, as a powerful voting bloc and/or as a member of Minimum wage Advisory Boards. The voting in- fluence of unions was relevant to the early 1960's and before the emer- gence of one-party politics. Such was the case in Uganda where the mini- mum wage was raised substantially just before the 1962 general election (see Table IV-l). In the context of a one-party system, however, which became the practice in Uganda and throughout East Africa soon after inde- pendence, the voting influence of unions was effectively neutralized. There was no choice but to vote the party line. As Advisory Board members, however, unions were given the opportunity to influence minimum wages to their advantage. As the employer repre- sentatives tended to be large expatriate firms -- who were in most cases already paying wages above the legal minimum -- they tended to lend their support to increasing minimum wages. This was the observation Knight 84Ibid., pp. 7-14. 90 made with respect to uganda.85 As the impact of minimum wages or government wage policies on aver- age wages and on employment, particularly those of unskilled labor, is clear, one must evaluate the supposed benefits of "need" based minimum wages against the various social costs in order to determine the net bene- fit or loss to society. In what follows, therefore, a brief evaluation of popular arguments in support of minimum wages will be provided to show the basic flow in these arguments and the net loss to society of per- petuating the policy of minimum wages under the traditional or colonial criteria. The most popular argument in support of raising minimum wages was that it would increase consumption of goods and services produced in the domestic economy and thus permit domestic firms to produce to full capacity. Moreover, a small increase in domestic consumption was at times suggested to cause a much more significant increase in national product through the operation of the "multiplier" a la Keynes. This would be true, however, if we assume a closed economy or a low marginal propensity to import on the part of those who received wage increases. In reality, an economy such as Uganda's is open to international trade, and imports of consumption goods generally sell at prices competitive with domestic prices even under a protected internal market. Moreover, imported goods may be of superior quality and durability. Hence, unless imports are significantly and quan- titatively restricted, unlikely in a system of normal international trade, the marginal propensity to import of wage earners is likely to be high in 85J. B. Knight, Earnings, pp. 250-251. 91 East Africa. If so, increased wages would lead to increased output, and increased employment abroad and not so much in East Africa. As we have shown earlier, the rise in average wages in East Africa was primarily at the expense of peasant income. Even if the proposition that increased wages would lead to higher consumption of domestic goods and services were true, one would have to show that the marginal propen- sity to consume of urban wage earners is higher than that of peasant con- sumers to argue that there has been net addition to domestic consumption. There is no convincing evidence to prove a higher propensity to consume of urban workers. On the other hand, peasant receivers of income should be expected to have a consumption pattern which could more easily be satisfied by domes- tic goods and services. Furthermore, part of the increase in peasant incomes would go for purchases of investment goods (e.g., fertilizer, tools, etc.) available domestically. If so, purchases of domestic goods and services would more effectively be increased through policies to increase prices paid to peasant producers of cash craps, rather than policies to increase urban wages. In a country like Uganda, character- ized by a predominantly peasant economy, the argument on the side of wage restraint appears much stronger. Employment generation, balance of pay- ments, expansion of domestic markets, exports, all appear to require wage restraint. On the above grounds, it would be unreasonable to argue that the benefit to society from increasing minimum wages would be higher than the social cost in terms of rising unemployment. Another popular argument in support of raising minimum wages was that it would lead to increased efficiency and stabilization of labor. We have 92 already indicated that in a situation of mass unemployment, higher wages would not be necessary to keep labor turnover at a minimum. Moreover, there is no valid basis to argue that increasing wages would automatically lead to improved health, efficiency or productivity of workers. There may be an observed correlation between wage changes and productivity, as Knight has shown for Uganda,86 but this does not prove which is the cause and which is the effect. Furthermore, it is extremely difficult to reduce to quantitative measurement the ”minimum need" of a given worker or family. First, how much protein, calories, etc., an individual would need depends on variables such as age and the intensity of work done. Second, there is no guarantee that increased wages would lead to the purchase of more nutritious food.87 Even if the proposition that increased wages would lead to improved diet and health of the worker were valid, it does not necessarily follow that the efficiency of the worker would improve, since efficiency of a worker would also depend on the efficiency of other factors of production -- in particular, managerial efficiency. In summary, then, the various social benefits supposed to derive from raising minimum wages in Uganda or East Africa do not appear to stand up under close examination -- either theoretically or empirically. Whatever the social benefits, if any, they are unlikely to compensate for the social loss in terms of increasing unemployment and/or overall needs of economic 86J. 8. Knight, Earnings, p. 250. 87N. N. Franklin, "The Concept and Measurement of Minimum Living Stand- ards," International Labor Review (April, 1967), pp. 271-298. 93 development. It is, therefore, essential to consider the framework, content and scope of incomes policy as well as specific measures to deal with both urban and rural unemployment appropriate for an economy such as Uganda. This is the subject matter of the next chapter. CHAPTER V AN APPROACH TO INCOMES POLICY IN EAST AFRICA WITH SPECIAL REFERENCE TO UGANDA A General Framework of Incomes Policy for East Africa The first task towards the formulation of incomes policy for East Africa is to define some of the concepts or terms that are involved and to state clearly the assumptions or conditions on which the discussion of policy will be based. Incomes policy in the context of East Africa would deal primarily with the wage policy of the government, since, as was indicated earlier, the government is the largest employer of all categories of paid labor and its policy determines the pattern of wages and salaries in the econ- omy as a whole. Moreover, the emphasis on government wage and salary pol- icy is justified because governments in East Africa, both before and after independence, have intervened in the determination of non-government labor earnings through arbitration tribunals. These tribunals have, in most cases, been guided by the pattern of wages and salaries in the civil service. Incomes policy, as used in developed nations, was also concerned with non-wage incomes (profits, interest, rent) and prices. A concern with price movements was essential because the control of inflation has been the cen- tral objective. A concern with other incomes was necessary because of their close and casual relationship to the pattern of wages and salaries, which derives from the bargaining process between labor unions and private 94 95 employers, and because of the need to have the incomes policy appear equit- able and politically acceptable. In the context of East Africa, very little consideration will be given to non-wage incomes and prices in the modern sector. Because of the non- market determination of wages, very little causal relationship exists with non-wage incomes in the private sector. Any concern with profits and inter- est should relate only to the government's fiscal and monetary policies which have artificially cheapened the price of capital and thus encouraged the use of capital intensive methods of production. Furthermore, infla- tionary pressure has not been a major problem in East Africa, so that in- comes policy would not have an anti-inflationary Objective or the control of modern sector prices, as is the case in developed nations. However, incomes policy in East Africa must deal with the pattern of agricultural prices and incomes, since the majority of people in these coun- tries earn their livelihood from agriculture. The relationship of these earnings to urban wages is the major factor influencing the rate of rural- urban migration and unemployment. In the past, taxes on agricultural ex- ports and Marketing Boards pricing policies, coupled with general neglect of agricultural development, have helped to stimulate rural-urban migration and thus intensify the level and growth rate of urban unemployment. Thus, while wage, fiscal and monetary policies may succeed in creating urban employ- ment, in the absence of price incentives and other policies to develop employment opportunities in agriculture, the net result may be in expansion of urban unemployment. In short, incomes policy in East Africa would be different from that 96 in developed nations in its objectives, contents and scope. While the objective of incomes policy in East Africa is clearly the generation of employment in urban and rural areas, it must have clearly stated criteria to apply in pursuing this objective. The criteria must be derived from the broader policies of development, which in turn depend on the conditions prevailing in East Africa, the national priority given to the various objectives (maximization of output and/or employment, equit- able distribution of income, etc.) and the strategy adopted in the pursuit of these objectives. The magnitude of the unemployment problem in East Africa is such that employment creation can no longer be ignored or given less priority than maximizing output. Given the general poverty of the East African economies, equitable distribution of income can best be achieved through expanding urban and rural employment opportunities instead of de- pending on the actual redistribution of non-wage incomes. While employment and growth were recently accepted as important objec- tives of policy, there has been a tendency to view them as competitive so that one objective could not be achieved without some sacrifice in the other. It is here assumed that this conflict is more apparent than real, so that employment and output can both be expanded. The traditional belief in an unavoidable conflict is related to the belief that capital intensive techniques are essential to a rapid rate of growth and the conventional use of output per unit of labor input as the measure of efficiency. The assumption that both output and employment can be increased rests on the fact that a rapid growth rate could also be achieved through emphasis on labor intensive techniques of production, particularly in small-scale farming -- 97 the most prevalent in East Africa.88 In the absence of adequate statistical data to evaluate the economic efficiency of peasant small-scale farms, the historical success of cotton and coffee production provide irrefutable evidence. The experiences of many countries, including Taiwan, Korea and Uganda, appear to show that small-scale labor intensive farms could be as efficient as large-scale capital intensive farms and have the advantage of requiring much less cap- ital investment.89 The choice of labor intensive technology should not be viewed as discounting the importance of investment, but meaning that mini- mizing capital investment and maximizing labor input in agriculture and in- dustry should be the guiding principle of efficiency, development and in- comes policy in a capital scarce society such as Uganda. Having clarified the concept of incomes policy and the assumptions as well as conditions on which it would be based, it becomes necessary to specify the variables -- the level and structure of wages and the distribu- tion of income -- it seeks to alter. This has been done in Chapter III above. On the basis of these variables and the concept of incomes policy thought to be appropriate for East African conditions, a set of propositions to define the general framework of incomes policy will be suggested on the basis of which specific policy recommendations would be advanced. These would include: (a) The basic level of wages for unskilled and inexperienced workers 88See Carl Eicher 3E_21, Employment Generation in African Agriculture, Institute of International Agriculture, No. 9 (Michigan State University, East Lansing, Michigan, 1970), pp. 18-34. 89See Owens Edgar and Robert Shaw, Development Reconsidered (D. C. Heath and Co., London, 1972). This point will be elaborated below when employ- ment generation in agriculture is considered. 98 in East Africa should not be an obstacle to expanding employment opportun- ities. (b) To achieve (a), the unskilled wage rate should be fixed at a level that will provide a standard of living comparable to the unskilled labor in the subsistence and/or non-wage sector of the economy and not on some arbitrary criteria or "poverty datum line" as recommended by doctors, nu- tritionists, and so on. (c) The wage and salary differential between different categories of workers should not be larger than is required to provide incentives for workers to acquire skill, to accept responsibility, and to make reasonable effort in their work. (d) Preposition (c) implies that the wage and salary structure should favor employees with technical skill (e.g., doctors, agronomists, etc.) over administrative and/or political bureaucrats with formal education, but no demonstrable technical skill: that existing differentials favoring non-manual workers (typists, clerks, etc.) over manual workers (mechanics, electricians, etc.) on the basis of formal education attained should change in favor of the latter. (e) If a particular firm or industry's capacity to pay wages appears to exceed the "minimum wage" criteria proposed in (b), the correct policy would be not to force it to pay wages and salaries much higher than the capacity of the economy or the industrial sector as a whole, but to tax the excess earnings. Otherwise, it would give rise to different minimum rates for the same work in different industries with a resultant danger of distor- tions in inter-industry wage structure and inefficient resource allocation. Moreover, it would be extremely difficult, particularly within the context 99 of insufficient data and manpower in East Africa, to measure capacity to pay in each employing unit and allocate equitable weights to the various factors of production involved. The above propositions should not imply that wages and salaries be rigidly fixed within and between firms or industries. They could and should be expected to rise over time, but roughly in line with the produc- tivity or standard of living of the economy as a whole. In parts of the economy, however, wages may not rise at the same rate as productivity. This would not be undesirable if we consider that that increased propor- tion of the national income may have to be devoted to expanding investment, employment and growth. Finally, any plan for incomes policy must have some clearly defined procedure or machinery of implementation. In the context of East Africa, where the government is the dominant employer, it would be expected of the government to pay its own workers wages and salaries on the basis of the criteria it wants followed in the economy as a whole. Moreover, the government must require labor arbitrators who it appoints to strictly fol- low the incomes policy guidelines with respect to the level and structure of wages. This would not be a major problem as long as the government itself behaved as indicated. The officially sanctioned incomes policy should be widelprublicized and publicly known if private employers are expected to implement it on their own. It would be unrealistic, however, to expect the government to publish detailed wage and salary structures applicable to all firms and all categories of employees, or to have representatives at each employ- ing unit to watch over the observance of incomes policy criteria. Instead, 100 it would be more practical and realistic to establish the rate for a few key occupations likely to be found throughout the wage sector -- for example, the rate for mechanics, clerks and unskilled factory workers, perhaps indi- cating in broad terms how experience is rewarded. The last suggestion should not imply that a decentralized and sec- tional collective bargaining would be preferable in East Africa, if only broad guidelines were suggested by the government, but that the control and guidance must not be cumbersome and rigid from over-centralization. The actual machinery set up to administer the incomes policy plan should include representatives of employers and unions. Such representation should be extended at both the formulation and implementation stages of policy to achieve the maximum of cooperation. These suggestions are not peculiar to the East African economy. The conditions in East Africa, as specified above, are different from those in developed nations. In particular, wage employees are only a small proportion of the total labor force. Hence, the scope of incomes policy in East Africa should be so comprehensive as to include prices, incomes and employment in the rural or non-wage sector. Consequently, the machin- ery of incomes policy should include representatives of farmers groups to make sure their interests are taken into account at both the planning and implementation stages. This would represent a significant historical de- parture from the usual process of decision-making and meet the peculiar needs and conditions of East Africa. 101. Policies to Generate Non-Agricultural Emplpyment The most important step towards increasing urban employment in Uganda, as in the rest of East Africa and other developing nations, would be an elimination of factor-price distortions -- i.e., the existence of too high wage rates and too low interest and foreign exchange rates despite the abundance of labor and the scarcity of capital. The rapid increase in unemployment, partly because of a decrease in those already employed, is a consequence to be expected when factor prices are kept away from their equilibrium levels for a long period of time. Direct government action to, at a minimum, close the gap between offi- cial and market-clearing interest and foreign exchange rates will eliminate some of the windfalls which are received by preferred firms or industries in the non-agricultural sector and force them to economize more on scarce capital resources in their production plans. In the short run, there may not be that much room to move towards more labor intensive technologies and sectors, either because such technologies are not available or because existing industries or processes may already be labor intensive. But these difficulties themselves result from a general lack of research on labor-saving technologies, at least partly caused by prevailing factor- price distortions. On the other hand, the rapid rise in wages and salaries observed in the past needs to be restrained. Governmental action to directly lower real wages for unskilled urban workers is virtually impossible given the political strength of organized labor. The most feasible approach would be to keep money wages from rising as fast as they did in the 1960's, and attempting to lower, or hold down the rise, in the cost of wage-goods 102 (e.g., food and clothing) so that money wages relative to other factor prices do not increase. The Ugandan Second Five Year Plan (1966-1970) introduced an incomes policy which called for a zero rise in incomes over £600 per annum, 1.0 per- cent per annum for the group 590-210, and 3.5 percent per annum wage in-‘ creases for those earning less than 590.90 This would constitute a sig- nificant improvement compared to the 10 to 15 percent per annum increase observed from the mid-1950's to mid-1960's. Such low increases in wages could be sustained if the corresponding rise in the cost of living is kept down sufficiently. The latter condition was not achieved in Uganda. By 1966, the government announced a significant shift to the "left" in national ideology and economic policies. The sudden rapid increase in government participation in economic affairs financed through borrowing from the Central Bank (in addition to rapid rise in taxes) led to a substantial in- crease in the cost of living index, especially for food and clothing.91 Michael Tribe92 maintains that this sharp rise in the cost of living led to labor union pressure for a reduction in taxes and an increase in the official minimum wage from 150/- to 250/- shillings per month. In 1970, the government appointed a Commission of Inquiry which, contrary to the wage restraint proposed in the Five Year Plan, recommended l85/-, 125/- and 104/- shillings per month minimum wages for workers in main towns, trading centers, 90Ugandan Government, WOrk for Progress (Entebbe: Government Printer, 1966) I p. 149. 91Ugandan Government, A Five Year Capital Develgpment Plan; 1955-1960 (Entebbe: Government Printer, 1955). 92Michael Tribe, "Uganda 1971: An Economic Background," Mwazo, Vol. III, No. 1 (Kampala: Mekerere University, June 1971), p. 52. 103 and all other categories respectively. In regardtx>the civil service, the government attempted to redefine the structure of salaries and rank so as to eliminate unwarranted differ- entials between different agencies. This was bitterly opposed and proved impossible to implement.93 The civil servants were sceptical about the claim that there was a need to freeze and standardize salaries in the face of extravagant unbudgeted expenditures such as the CAD Conference Center, corruption and high living among ministers. The Tanzanian experi- ience with the problem should have indicated to the Ugandan Government that civil servants would not readily accept cutbacks in privileges if the political elite is not also made to do the same. An alternative approach to encourage labor usage in industrial enter- prises concerns the payment of a wage subsidy or a tax deduction of a wage write-off for investments made in more labor intensive methods. This should be expected to stimulate employment opportunities as the effective labor price to producers is reduced by the amount of the subsidy or tax write-off (a reasonably high and positive elasticity of labor is assumed). The weakness of such a policy relates to the fact that the actual earnings or wages paid are not reduced. The differential between urban wages and rural incomes are maintained. The resultant rural to urban migration of labor attracted by higher earnings and better employment opportunities could lead to a rise in the net urban unemployment. Thus, while a wage sub- sidy or tax write-off could expand the short run employment of labor, the indicated limitation suggests that these policies are not effective substitutes __‘ 93Ibid.. pp. 48-49. 104 for policies that deal with the underlying causes of factor-price distortion. The Ugandan government could attempt to create employment through an agreement or understanding between itself, private employers, and unions, whereby the government and private employers agree to raise the amount of labor they will hire in turn for a promise by unions to accept wage freezes for a specified period. On the surface, the policy is appealing and simple. In practice, the contribution of tripartite agreements to the reduction of unemployment has not been significant. The experience of Kenya, where the policy has been tested, indicates that while employment has not been affected sig- nificantly, the level of urban unemployment has probably increased sub- stantially as a result. D. P. Ghai, for instance, makes the following observation with respect to the 1964 tripartite agreement in Kenya: Owing to the financial stringency, the government found itself unable to carry out its pledge. In all, 34,000 jobs were found...But it is doubtful whether these represented a net increase in employment; for while many of the private firms adhered to the letter of the agreement by taking on employees, they vio- lated its intent by not hiring new employees to make good the loss caused by normal attrition of the labor force. Furthermore, even if the agreement was suc- cessful in creating some new jobs, it did not make any particular impact on the level of urban unemploy- ment, for prospects of wage employment stimulated an additional flow of migrants from rural areas.94 Like subsidy of tax write-off policies, tripartite agreements fail to deal with factor price distortions and rural-urban income differentials. It is only a short run (political?) expedient trying to deal with the 94D. P. Ghai, "Employment Performance, Prospects and Policies in Kenya, prepared for the Overseas Study Committee Conference (Cambridge University, 1970), quoted in M. P. Todaro, pp, cit., p. 399. 105 problem of urban unemployment by circumventing economic forces. In light of these facts, it is hard to see why the Kenyan Government had another tri- partite agreement in 1970. It could only understood as another political gimmick to stem off urban discontent. If meaningful solutions are de- sired, the Ugandan Government must avoid tripartite agreements and their negative consequences for correcting unemployment. Following the experience of Tanzania, Uganda might take measures to physically restrict the rural to urban migration of labor taking place in response to efforts (wage subsidies, tripartite agreements) to expand urban employment opportunities. The urban unemployed could be repatriated to farm settlements. The Tanzanian experience has shown that it is diffi- cult to control labor migration to the level of urban job openings95 as long as there is an incentive to migrate. Effective control of migrants would require large expenditures to maintain adequate enforcement manpower. Moreover, potential migrants, particularly young school-leavers, may revolt against sustained restriction on their freedom to migrate in search of better opportunities. More policy alternatives to generate urban employment opportunities can be advanced. The few that are evaluated above are, however, suffi- cient to indicate that ad hoc measures to expand urban employment are not substitutes for correcting the underlying factor-price distortions and rural-urban earning differentials as effective solutions to control unem- ployment. This is particularly true in a country like Uganda where the narrow urban base, even under normal circumstances, could not be expected 95M. P. Todaro: pp, cit., p. 401. 106 to absorb more than a very small percentage of the total labor force. The last point raises a very important fact. That is, while urban- oriented incomes policies are necessary in Uganda, policies to directly stimulate agricultural growth and employment opportunities remain an over- riding consideration. In the final analysis, a predominantly agricultural economy must create most of the opportunities for employment and decent living in that sector. In this case, the concept of incomes policy must be broad enough to include a system of rural incentives to absorb the rising labor force. What follows is a treatment of this issue. Agricultural Development Policies and the Generation of Employment Needed Reassessment of Development Poligy The analysis in Chapter II has already shown that Uganda must make a fundamental reassessment of past development policies and/or strategies in such a way that the agricultural sector can contribute more effectively to national product and employment. While no argument was made against industrialization as a long term approach to diversify the structure of the Ugandan economy and reduce its dependence on exports of primary commod- ities, it was argued that basic short or medium term problems -- primarily rising unemployment -- could not be effectively handled by a policy of hasty or premature "industrial fundamentalism." Various urban-oriented pol- icies were shown to, in fact, accentuate the rate of growth of urban unem- ployment unless national development policy puts proper emphasis on the development of the agricultural sector so as to maximize the income earning and employment opportunities of the overwhelming majority of the population 107 who happen to reside and make their living in the countryside. In recent years, there has been a general agreement to assign agricul- ture the major task of job creation in East Africa. Too often, however, this view is put forward by planners and politicians without proper and neces- sary modifications of theoretical assumptions underlying national planning and the governmental decision-making process. While the objective of agricultural development may have actually been implemented in East Africa, failure to reassess the theoretical basis of planning or developmental strategy and policy has very often led to stagnation of agricultural out- put and employment: or, where output may have been successfully expanded, it was achieved without corresponding expansion in rural employment oppor- tunities. In the past, when the tendency was to look at the agricultural sector in terms of extracting surplus resources for industrial investment, there has been a bias towards medhanized or capital-intensive agriculture. Im- plicit in the choice of this planning strategy was the belief that small- scale peasant farms are inherently unsuitable for the purpose of maximiz- ing agricultural productivity, and thus providing the marketable surplus for rapid industrial investment and growth. A more detailed review of specific anti-employment measures undertaken to develop agriculture in the post-werld War II period and needed modifications in these measures in light of the growing problem of unemployment will be provided below, with a particular reference to Uganda. At this point, it is important and desir- able to evaluate the theoretical and empirical validity of the proposition that with respect to expanding agricultural output, small-scale farms are 108 relatively inefficient compared to large-scale, mechanized, or capital inten- sive ones. Such an evaluation is crucial, since if the proposition were true it would suggest a basic conflict in the objectives of maximizing agricultural output and employment opportunities. Hence, the hepe of assigning to agriculture the major task of job creation would be undermined. In the United States and some other major industrial nations, where capital and organizational talent has been abundant relative to labor, rapid growth in agricultural productivity has been achieved through intensive mechanization of farms. The assumption that the pattern of devel- opment in these countries would be repeated in the now less developed na- tions explains the bias to capital intensive agriculture in the latter, despite the prevailing factor scarcity of capital relative to labor. This assumption has been contradicted by empirical evidence derived from the developmental experience of such ”labor-surplus" economies as Japan, Taiwan, and South Korea. Japan, Taiwan and South Korea were reported to have had 87, 79 and 79 workers per 100 acres in 1965, compared to India, the Philippines, Ceylon, Cblombia, and Brazil, which had 36, 29, 49, 20 and 17 workers respectively per 100 acres during the same period.96 Yet Japan, Taiwan and South Korea were among the most productive nations in Asia and the world. The reasons for achieving high productivity despite the relative intensity of labor are to be found in the improved varieties, fertilizers, and improved implements used to augment human labor rather than displacing it (e.g., implements to break seasonal bottlenecks).97 96FAO, Production Yearbook, 1969 (various tables). 97B. F. Johnston, "Agriculture and Economic Development: The Relevance of Japanese Experience," Food Research Institute Studies 6, No. 3 (Stanford University, 1966), p. 288. 109 In the past, experts, planners, and politicians have been inclined to the view that small peasant farmers in East Africa do not provide a basis for productive investment partly because their farms are too small. Implicit in this belief was the contention that a small farmer would not be able to take advantage of economies of scale and that the cost of investing in each of the numerous small holdings would be higher than the expected benefits. consequently, the huge subsistence sector such as that of Uganda has been neglected, while agricultural investment tended to be concentrated in the development of large-scale commercial farms where big machinery (e.g., tractors) was thought to raise output or productivity per worker. The alternative possibility of raising the productivity of small peasant farms through labor intensive techniques of production, following the experience of Japan, Taiwan, and South Korea, did not receive adequate attention from planners and policy makers in East Africa. In addition to the mistake of assuming that a higher density of workers would necessarily imply lower productivity in agriculture, extensive re- search carried out in the past decade tended to show that output per worker could be generally higher on small farms than on larger ones.98 Moreover, investment costs per dollar of market sales have been shown to be lower in Mexico for small farms than for large ones.99 The above citations tend to confirm that labor intensive small-scale farms could in fact be more efficient units of agricultural development in 98For example, see Lester Schmid, Relation of Size of Farm to Produc- tivity (Land Tenure Center, University of Wisconsin, 1969). 99Folke Dovring, "Land Reform in Mexico," Spring Review of Land Reform (AID, Washington, D.C., 1970), pp. 44-48 and 50-54. 110 Uganda or East Africa with respect to both output and employment maximiza- tion objectives. Due to the relative inadequacy of statistical data, it is much more difficult to conduct similar studies for East Africa. The impact of crop intensification on the employment capacity of small-scale agriculture has, therefore, not been fully documented. The few studies that are available, however, tend to support that a more intensive farm- ing of existing small holdings results in the most employment and supports the largest number of persons (including dependents) per unit of land at a comparably low cost. For example, a recent study by Clayton100 shows that intensive farming of existing small holdings in Nyeri District of Kenya (at about the mid-1960's) created about 10,000 new employment oppor- tunities, providing about 1,072 persons with employment and supporting 2,906 persons per 1,000 hectares, at an estimated cost of 1,000 shillings per hectare. This compares with 116 persons employed per 1,000 hectares at an estimated cost of 760 shillings per hectare on large-scale African farms. The study also indicates that the Mwea irrigation scheme -- a land extensi- fication scheme -- provided employment for 2,807 persons and supported 4,529 persons per 1,000 hectares, but at a cost of 7,160 shillings per hectare. Thus, the most effective policy -- with regard to rural employment gener- ation at minimum cost per unit of employment -- appears to be an intensive farming of existing small-scale farms. This policy has the added advantage, particularly in Uganda, of requiring a minimum of institutional and struc- tural adjustments, since most of the people are already in small-scale farm- ing. These conclusions are supported by a more general and qualitative study 100E. S. Clayton, "Agricultural Reform, Agricultural Planning and Employment in Kenya," International Labor Review, November 1970. 111 by Eicher and others.101 While it is reasonable to assume that employment in small-scale agricul- ture would keep pace with growth in output, there is evidence to show that this has not been the case with employment in plantation agriculture. A report of an ILO mission to Kenya102 -- where plantation agriculture has historically been more prevalent than in the other East African countries -- indicates that over the period 1963-1969 there was a marked reduction in the number of permanent workers employed by coffee and sisal estates. Employment in coffee estates was shown to have declined from an estimated 100,000 persons in 1964 to 80,000 persons in 1969. In sisal estates, employment dropped from 25,000 to 12,000 over the same period. In compar- ison, employment on tea estates rose from 34,000 to 37,000 persons. In general, the observed stagnation of employment in plantations or estates, while output has expanded significantly, was because technology has been in use which is both labor saving and output increasing. For some estates such as sisal, the rate of decline in labor employment has been shown to exceed the rate of decline in output primarily because many such estates have ceased replanting while continuing to harvest existing stock. While small-scale farming, with appropriate land and technology pol- icies, offers the best hope of meeting the goals of expanded output and employment in East Africa, the key to its success will rest on the nature 101Carl Eicher £5.31, pp, cit., especially pp. 18-34. 102P. S. Narasimhan, Kenya: Conditions of WOrk and Life in Plantations, Report of an ILO Mission (mimeographed, May 1969). 112 of agricultural price and taxation policies. Eicher and others103 have shown a historical tendency in many African countries, including East Africa, to reduce the price of export crops to the grower below world market levels and to raise the price of food crops above world market prices. Taxes on export crops have been an administratively easy way of raising revenue, while the conventional strategy for agriculture has been to aim at self- sufficiency in food production. Neither price movement is conducive to employment. On the one hand, expensive food encourages wage demands and hence substitution of capital for labor. Taxes on export crops discourage their production with a resultant negative impact on employment. In the past, pricing policy has tended to assume, perhaps implicitly, that peasant farmers are unresponsive to price and other market incentives. Export taxes and Marketing Board surpluses have been used to transfer re- sources from agriculture for the financing of industrial development. This seems to suggest that peasant farmers are relatively more inclined to consume savings instead of utilizing them for investment, or that the gov- ernment is better suited to make investment decisions than peasant farmers themselves. Perhaps the best available evidence to contradict such a view regard- ing peasant behavior is the historical success of peasant farmers in East Africa in effectively adapting to the production of cash crops such as coffee and cotton. Furthermore, the few empirical studies available tend to support the fact that peasant farmers in East Africa respond positively 103Carl Eicher pp_pl, pp, cit., pp. 27-30. 113 to market incentives.104 Much more research needs to be done in this area. Nevertheless, in the face of uncertainty, the best policy is to avoid price distortions and let market signals reflect the true scarcities so that individual decision—makers can choose whatever strategy of agricultural development best suits their interests. The role of the government would then be limited to policies designed to offset some specific problems of development -- for example, the provision of infrastructure and credit. Another aspect of pricing policy related to agriculture is the proper pricing of agricultural inputs. Domestic production of inputs such as fertilizers, implements, and so on (which are complementary to the input of labor in agriculture), under a policy-protected domestic market, would result in unnecessarily high input prices. This would discourage produc- tion and employment. The right policy would appear to be that which will reduce the effective price paid by peasant farmers for inputs complementary to labor either through the encouragement of imports or subsidy. In fact, the bias towards mechanized farms in the past has given rise to artificially low prices paid by the farmer for imported machinery. This was achieved either through low import duties or subsidized tractor-hire services or both. The availability of foreign tied-loans has aided large scale imports of machinery which would have been impossible given the meager foreign exchange balances at the disposal of East African governments. In this regard, the reassessment of developmental policy and strategy would also require a reassessment of conventional foreign loan and/or aid terms. 104See, for example, J. x. Maitha. ”Productivity Response to Price: A Case Study of Kenya Coffee," East African Economic Review, vol. II, No. 2 (December 1969), pp. 31-37. 114 The above analysis on reassessment of agricultural development policy to achieve the goals of expanded output and employment has been rather general and depends heavily on non-Ugandan or East African experiences, where the better availability of data and/or better awareness of the issue has permitted more extensive research. While the general analysis of the East African situation that is presented is a significant step in the understanding of the overall problem, the formulation of specific measures or policies applicable to expanding agricultural output and employment in a country like Uganda in the coming decades would require a more exten- sive collection of data so that the consequences of alternative policy options could be tentatively projected into the future. In this connection, Uganda or the East African countries would bene- fit greatly from recent studies and projections made with respect to Nigerian agriculture, where much more data and agricultural surveys appear to be available. These Nigerian studies105 tend to support the general conclusions indicated above: that policies and programs which provide incentives to agricultural producers while supporting them with infrastruc- ture, research and modernizing services, offer the best alternative strat- egy for expanding agricultural output and employment opportunities. The Denial of Epployment Generation: The Ugandan Case This section will review Ugandan development policies in the past inasmuch as they represented a denial of employment generation in 105See, for instance, Consortium for the Study of Nigerian Rural Devel- opment, Preliminary Strategies and Recommendations for Nigerian Rural Development (Michigan State University, East Lansing, Michigan, July 22, 1968). 115 agriculture. First, the historical tendency in Uganda towards a policy of mechanized group farms will be delineated. Second, the implications of export taxes on agricultural commodities and Marketing Board surpluses for the generation of employment in agriculture will be considered in some detail. The Historical Trend towards Mechanization of Farms The 1947-1956 WOrthington Plan and its revision in 1948,106 the first real attempt at development planning in Uganda, correctly viewed the development of peasant agriculture as crucial for growth in national out- put. Yet the fact that the agricultural sector was allocated about 10 per- cent of the total planned expenditure107 indicates the lack of faith in expanding national output (the objective of the plan) through investment in peasant agriculture. While the Plan proposed essential schemes such as the control of tsetse fly and livestock disease, forestry expansion and increased water supply to bring more land into cultivation, it also sug- gested pilot schemes to test the viability of state or collective farms -- thus introducing an approach conducive for mechanization and which, as will be evident below, was to become the major factor influencing planning strategy in Uganda throughout the 1960's. The worthington Plan was followed by a Five Year Development Plan108 (1955-1960) which concentrated on capital expenditure. While the Plan envisaged the raising of a large part of the financing from internal and 106E. B. worthington, A Development Plan for Uganda and the 1948 Revision of the Plan (Entebbe: Government Press, 1949). 107Ibid., p. 11. 108Uganda Government, A Five Year Capital Development Plan, 1955-60 (Entebbe: Government Press, 1955). 116 external loans (about 43 percent of the total required resources), the agri- cultural sector was expected to provide about 44 percent of the resources -- about equally from export duties and Development Assistance funds, diverted from the Cotton Price Assistance Fund. Yet, of the total planned expendi- ture, only about 5 percent was allocated to the agricultural sector.109 As before, pilot programs of mechanized farms continued. The process of development planning in Uganda took a major step for- ‘ward with the introduction of the 1961/62-1965/66 Plan.110 In contrast to earlier plans, this was more comprehensive in that it embraced the public, parastatal corporations and the private sector of the economy. The plan was based on the recommendations made by the world Bank Mis- sion Report in 1961.111 The report had recognized the primary role of agri- culture in the development of Uganda. In fact, it did not see great prospects for industrialization. Regarding policy, the Report suggested aiding progressive farmers through subsidies, credit schemes, extension services, training and research. While the Mission said: "We foresee no great advantages from mechanization of production on peasant farms for some years,"112 it nonetheless recommended mechanization on an experimental basis just as the worthington Plan had proposed earlier. In contrast to the Mission Report which had suggested diversification 109Ibid., The First Revision, Sessional Paper No. 13 (1956-57): p. 58. 110Uganda Government, The First Five Year Development Plan, 1961/62- l965l66 (Entebbe: Government Press, 1963). 111IBRD, The Economic Development of Uganda (Baltimore: Johns Hopkins Press, 1962). llzIbid. A I 117 of exports through expanding outputs of cotton and livestock, the government of Uganda aimed at achieving diversification through an expanded manufac- turing sector. While planned allocation for agriculture was greater than the Mission pr0posa1, the Five Year Plan made a major shift to large scale agriculture, whereby a considerable increase in tractor-hire services and group farming were proposed. In fact, at the end of the Plan period, actual expenditure on mechanization and group farming were greater than planned. The results of mechanization and group farming were considered a major failure in terms of both profitability and the generation of employ- ment.113 On both grounds, one would have expected less reliance on mech- anization and group farms in the next or Second Five Year Plan (SSYP).114 Yet the SSYP proposed establishing 100 more group farms and expanded tractor-hire services. Despite the indicated poor results, the government argued that mechanization "...be made available to as many farmers as "115 . . 116 . . possible. The deCision, as David Vail indicates, was made primar- ily on the political ground of creating a united "socialist state". But it is hard to see how an economically non-viable proposition could have led to a more desirable society -- by whatever criteria desirability is judged. 0n the contrary, the failure to generate rural employment would lead to 113Hall Mattrick, Aid In Uganda--Agriculture (London: The Overseas Development Institute Ltd., 1967). 114Uganda Government, pp, cit., p. 48 115Ibid., p. 59. 116 David Vail, Group Farming and Statecraft in Uganda, Development Research Project No. 71 (Kampala, Mekerere Institute of Social Research, 118 expanding urban unemployment. This would involve a waste of human resources as well as potential political instability. The Impact of Expgrt Taxes and Marketing Board Sugpluses on Employment The development of a simple production function model will facilitate an analysis of the effects of export taxes on agricultural crops and/or Marketing Board surpluses on the generation of rural employment opportun- ities. Let Y represent output of cotton or coffee exported by Uganda and x1 represent input of labor in the production of the export crop, cotton or coffee. A particular farm could be represented by a simple production function given as, (1) Y = f[x1|x ,...,xn] 2,X3 where Xi (i=2,...,n) are inputs of factors other than labor and assumed constant. The marginal physical product per unit of labor input X (MPP)x , I l 1 may be given as, dY (2) (MPP) = —- "1 ‘th1 where (MPP)x is the marginal physical product of labor, the variable input. 1 Assuming the existence of competitive markets, the marginal value pro- duct of labor (MVP)x may be given as, 1 dY (3) (MVP) =p -— "1 Y “1 where Py is the world market price of the export crop, cotton or coffee. 119 For a farm having a production function represented by equation (1) , the profit maximizing point can be determined by equating the marginal value product of labor, the variable input, to the price of labor (le) or wage rate (W). That is: dY (4) P dx - P — W where Px = W is assumed given. The relationship between (MVP)x and 1 1 Px s w is depicted in Figure 2, where the profit maximizing farm is shown 1 to employ x° amount of labor. 1 The Ugandan government, like some other ex-colonial countries in Africa, withdraws a portion of rural or peasant incomes through the impos- ition of export taxes and/or the accumulation of Marketing Board surpluses. The administrative cost of these Boards adds to the withdrawal of resources from the rural sector. Consequently, the prices actually received by farmers would be less than the world market prices for the export crops. Let the rate of withdrawal be represented by ty so that the price received by the farmer is given as, (5) (1 - t )P = BP Y Y Y where O < ty < 1 and B < 1. 1 If the farmer had been subsidized per unit of labor input, the actual price or wage paid by the farmer could be given as, (6) (1 - s )P = (1 - s )W = yW x1 x1 x1 where sx is the rate of subsidy per unit of labor input and O < sx < 1 so 1 1 that y < l. 120 Figure 2 3i 0 x x xllxz...xn Fig. The relationship between marginal value product and price (i.e., wage) of the variable input labor (X1). 121 The profit maximizing condition given by equation (4) could be easily altered to accommodate the introduction of export taxes, the accumulation of Marketing Board surpluses and subsidy per unit of labor input so that, d! (7) (MVP) - 8P --— = yw x1,ty y dx1 Equation (7) shows that the profit maximizing farmer will expand the em- ployment of labor up to the point where the marginal value product after taxes and/or Marketing Board surpluses equals the purchase price of a unit of labor less subsidy received. since 0 < B < l, the marginal value product after export taxes and/or Marketing Board surpluses, (MVP)x1't , would be less than the marginal value product in the absence of taxes and/gr surpluses. In Figure 2, this is represented by a downward shift of the marginal value product curve so that, in the absence of labor subsidy, the profit maximizing farmer will reduce the amount of labor employed from.0xi to OXi. The figure also shows that if, after subsidy, the actual cost of labor per unit employed is reduced from W to yw, the original level of employment, oxi, is maintained. This is a case where the amount of the sbusidy was just sufficient to compen- sate for the total distortion in the allocation of resources caused by the presence of taxes and/Or Marketing Board surpluses. In reality, the amount of labor employed would depend on the relative weight of the taxes I-§ 1, so and/or Board surpluses and subsidy -— that is, on whether or not B o 1. The existing (and past) policies in Uganda contain no subsidy for the that the new employment equilibrium could be % ox 122 input of labor. The prevalence of export duties117 and/or Marketing Board surpluses would, therefore, have the effect of reducing the amount of labor already employed in the production of export crops or keeping a higher level of potential employment from being realized. This would contribute towards the rural to urban migration of labor. The rate of migration would be particularly high in a country such as Uganda where the gap between the potential off-farm earnings (represented by urban minimum‘wages) and on-farm earnings from family farms or paid agricultural work is significant. In light of the suggested reassessments of development policy and the distortions in resource allocations created by mechanization or export duties and/or Marketing Board surpluses, the following specific recommenda- tions could be advanced as appropriate for Uganda and East Africa. (a) A set of agricultural price and production policies should be estab- lished, to provide incentives to farmers to produce and employ more than they did in the past. This will require a gradual elimination of export duties and Marketing Board surpluses. (b) If the need for revenue makes the elimination of export duties 117The Uganda Government, Statistical Abstract (Entebbe: Government Printer, various issues) does not provide the revenue from export duties and the total value of exports for the same period. For instance, while the export duty related to the 1961-62 and 1962-63 period, the total value of exports was given for 1961, 1962 and 1963. The same was true for other periods. In some years the value of export duty was shown separately for cotton, coffee, hides and skins, and copper. In others, it appears as an aggregate figure. A rough calculation of export duties as a percentage of the total value of exports indicate: 7 and 4 percent for cotton and coffee during 1961-62; 20 and 8 percent during 1962-63, or 14 and 11 per- cent respectively during 1966-67. Granting some error in these percent- ages, it is clear that when other income taxes, Marketing Board surpluses, and so on, are added to the export duties, the total withdrawal of income from the rural sector would be significantly higher than the percentages indicated above. 123 and/or Marketing Board surpluses difficult or impossible, the distortions in the actual or potential employment of labor in the rural sector could be at least partially minimized by the introduction of labor subsidy and/or urban wage restraint. Both should be expected to increase rural employ- ment and the incentive for rural to urban migration. (c) Eliminate factor price-distortions in the agricultural sector by dropping such policies as subsidized tractor-hire services. (d) Government intervention in the rural sector should concentrate more on the provision of basic services such as feeder roads, water supply, and/or credits so that peasant farming can become more profitable. (e) Inasmuch as land is a factor complementary to labor, the distribu- tion of unutilized government land to those who need it, free or at subsidized prices, is likely to lead to more employment of rural labor. Factors other than land and labor (e.g., fertilizer) could also be subsidized. The po- tential to expand output and employment would, however, depend on the extent to which the subsidized inputs contribute to value added in agri- cultural production. Since the share of labor and land in value added are significant in Ugandan agriculture, a subsidy on these inputs would have the most potential for the generation of rural employment. CHAPTER.VI POLITICS AND RATIONAL DECISION-MAKING: THE EXPERIENCE OF UGANDA DURING THE 1960's In what follows we will discuss a question that has come to occupy an important, though seldom explicitly expressed, position in the whole developmental issue relating to East Africa in general and Uganda in par- ticular. Simply stated, the question is whether the idealized development process, reflected in academic models and emulated by published national economic development plans in Uganda and the rest of East Africa, correctly reflects the process of economic decision-making. It is common knowledge that the East African countries, in common with most developing nations, have set out in five year plans fundamental objectives of development -- to raise incomes, to raise employment, to raise the level of education and training, to achieve a more egalitarian society, and so on. A quick look at the Ugandan development plans during the 1960's will support this. On the surface, at least, the political disposition of Uganda and other developing nations indicates basic commit- ment and urgency towards fulfillment of the objectives set out in national plans. For example, it is not uncommon for these countries to have separate and prestigious ministries of planning and development, staffed with local and foreign experts, and very often placed directly under the office of the prime minister or head of state. But does the actual process of decision—making on matters related to 124 125 specific issues (as contrasted to overall plans), employment, investment, balance of payment, agricultural price and rural development, reflect the development process as idealized in development plans? It has been demon- strated in previous chapters that this is not necessarily so. we have shown that the actual incomes policy (defined as wage and agricultural incomes and price policy) in Uganda during the 1960's has not been consistent with the stated objectives of development -- increasing the incomes and employ- ment opportunities of the majority of the people as well as a.more egali- tarian distribution of these incomes and opportunities. Moreover, the educational structure continued to put emphasis on formal rather than technical education despite the fact that Africanization of technical jobs has been the stated objective. The government, explicitly and implicitly, created a climate not conducive to the effective participation of foreign private investment; yet such investments are crucial in a society where capital is an extreme- ly scarce factor. Rural poverty and lack of jobs, in addition to better urban wages and social services, was causing mass migration of people into urban areas. Given the narrow urban base in Uganda, as in the rest of East Africa, the consequence was mass unemployment. Yet the government continued to give relatively low priority to the rural sector in the allo- cation of public expenditure. Where expenditure was made on the rural sec- tor, it was on capital intensive projects such as group or collective farms and contrary to the factor endowment of the rural sector. Instead of pro- viding market incentives to stimulate peasant producers, farmers were penalized through significant export taxes and Marketing Board surpluses. In the previous chapters, we analyzed the problem of unemployment 126 in Uganda which was developing throughout the 1960's. We suggested in- comes policies to alleviate the problem. It is naive to assume that the government was not aware of the magnitude of the problem and the appropri- ate and effective remedies suggested earlier. various studies cited in this paper indicate that economic advisors and independent researchers were voicing the seriousness of the unemployment and other economic prob- lems in East African journals, or foreign publications easily available to policymakers in East Africa. Thus, the inconsistency between stated 'objectives and actual behavior of policymakers must be explained by forces outside normal economic analysis. It is the contention of this writer that over-riding political con- siderations throughout the 1960's were not conducive to rational economic policy. In what follows we will review internal political evolution in Uganda since independence in 1962 and demonstrate the motives that pre— cluded rational decision-making. Uganda approached independence with a vigorously competitive party system. There were the Uganda Peoples Congress (UPC), the Democratic Party (DP), and the Buganda Kabaka Yekka (YK). None had or hoped to have a clear majority to assume executive power after independence. Member- ship in each party was strongly motivated by tribal and religious consid- erations. In particular, the YK party was formed around loyalty to the traditional king of the Buganda tribe, the most numerous and dominant group in Uganda. The Bugandans, in addition to being intensely tribal, were the rich- est and the most educated group. Hence, they were hostile to the possibil- ity of the other tribes uniting into a mass party, for fear of losing their 127 traditional autonomy and their favored position in the administrative and economic structure of Uganda. This explains why the YK party, which won 21 parliamentary seats, was willing just before independence to join the UPC, which won 37 seats, in coalition against the DP party (24 seats). Mr. Obote, the leader of the UPC party, became the prime minister, and the King of Buganda, the effective leader of the YK party, became the pres- ident. Once he became the chief executive of independent Uganda, Obote immed- iately set out to alter the precarious political relationship between the UPC and the YK. His objective was to secure a clear parliamentary majority. His tactics were to encourage YK party parliamentarians -- his allies in a coalition government -- to defect to the UPC and to aggres- sively confront the UPC party and its platform throughout the regions other than Buganda. To appeal to the various tribal and religious groups in these regions, each with its set of issues and demands, the UPC tended to adopt an essen- tially ad hoc and non-ideological platform. In the allocation of recur- rent and development funds, patronage was extensively practiced by the UPC through its dominant position in the government. In 1964, with many members of Obote's opposition crossing the floor, he was able to bring the precarious UPC-YK alliance to an end. The next two years, 1964 to 1966, was a period of uncertainty, during which the DP and YK parties attacked not only Obote but also the top offi- cer of the army at that time, Colonel Amin. With the army apparently on his side, Obote was able (in 1966) to imprison a few ministers, suspend the constitution, and appoint himself President of Uganda. The Buganda 128 traditional assembly (the Lukiko) countered by demanding that the Obote government remove itself from the autonomous Kingdom.of Buganda. When the national army, led by Deputy Commander Amin, stormed the Bugandan King's (the Kabaka) palace in May of 1966 and forced him into exile, Obote's dictatorship was complete. Uganda, though divided into hostile camps on religious, tribal, ethnic, regional and other issues, was declared a unitary one-party state. The period from 1962 to 1966, characterized by extreme political instability, could hardly have provided the socio-political infrastruc- ture necessary for rational economic decision-making with respect to Uganda's underlying economic problems of employment, rural development, investment, and so on. The decisions to expand minimum wages for urban unskilled workers, to expand group farming utilizing a cooperative institutional structure and stressing mechanization, to maintain the paternalistic and colonial structure of marketing agricultural crops through a State Marketing Board or monopoly, etc., were motivated by political and economic considera- tions. These decisions must be seen as part of the partisan competition between Obote's UPC and the opposition groups for rural support, and as a means of eroding citizens' traditional loyalties. Group fanming and cooperatives provided the medium for the UPC to expand its rural party cells. UPC candidates could promise capital for mechanization and the private pockets of rural leaders if only they supported the Obote regime through the activities of group farms and cooperatives. Not only were these policies uneconomic,as we have shown earlier, but they also ran counter to the sentiments and historical patterns of farming family plots 129 in Uganda. This may explain why individual family plots were maintained on the group farms.118 Uganda demonstrates the conditions which prevailed in most of the African countries for some time following independence. The economic objectives of the state, which required rational solutions to prevailing economic problems, were in conflict with the political imperative of building a nation state under the firm control of whoever happened to assume power at independence. In a one-party state, where diverse groups exist, to attain and reinforce power through solidarity becomes the over- riding goal, in contrast to rational economic decisions arrived at after serious debate of the issues. Though the attainment of ultimate social goals and economic development were stated in official plans and documents, short-run and non-economic considerations were paramount. Under these conditions, National Plans tended to become instruments for legitimizing the ruling parties' statebuilding activities. In the case of Uganda, urban employees and rural farmers were vul- nerable to exploitation for short-run political ends. Despite evidence showing the unsuitability of group farms and urban minimum wages against prevailing conditions of excess supply of labor and the long run develop- mental goals of Uganda, they were rationalized and exalted by political leaders. Since 1966, once political consolidation was seemingly achieved, the Ugandan government began to sharply increase its involvement in economic affairs, culminating in 1969 with an official declaration of a 118Aaron Segal, "The Politics of Land in East Africa," Economic Develop- ment and Cultural Change, V01. XVI, No. 2 (January 1968), pp. 283-84. 130 "Move to the Left," part of which provided "The Common Man's Charter"119 -- an ideological framework of Uganda's new political culture. This was an imitation of Tanzania's declaration at Arusha a few years previously. The charter called for the eradication of tribalism, economic equality, de- cline in social privilege, welfare for the peasantry, and national unity. These objectives were, in themselves, perhaps desirable. But as we will show below, the means used to attain these goals -- partly motivated by political considerations as in earlier years -- were largely counterpro- ductive in coming to grips with the problems of employment and growth in general. Proceeding step by step, the government passed legislation and issued proclamations to implement the "Move to the Left." In a May 1970 Labor Day Speech, Obote announced the government's intention to acquire 60 per- cent of the ownership of over 80 major industries, including banks, insur- ance companies, manufacturing firms, transport companies, and so on. Com- pensation was to be paid on future profits over a period not exceeding 15 years.120 Furthermore, exports and imports were to be a state monopoly. In short, Obote's "Move to the Left" implied the belief that consolidation of Uganda's independence (meaning his political control) required economic independence from foreign economic and financial dominance. Foreign included Ugandan Asians who had been in uganda all their lives. In addition to Bugandans who, as one would expect, interpreted the "Move to the Left" as an attempt to weaken their dominant economic position, l 1 9A. Milton Obote, "The Common Man's Charter," one of four documents in Uganda's Move to the Left (Entebbe: Government Printer, 1969). 120Selwyn Douglas Ryan, "Balance Sheet of the Revolution," Mwazo, Vol. III, No. 1 (June, 1967), pp. 37-64. 131 Obote's policies alienated European and Asian industrialists and traders. Thus, not only did political instability and internal opposition intensify, but a significant reduction in the inflow of foreign private capital re- sulted, as well as a sharp increase in the outflow of domestic foreign capital. In addition to increased imports of machinery for group farms, the net capital outflow created serious strain in the balance of interna- tional payments. Moreover, the "Move to the Left," involving a tremendous increase of government participation in economic affairs, effected a rapid rise in budgetary deficits. Taxes and internal borrowing -- primarily from the Central Bank -- rose sharply, at the expense of funds available for private investors. Though there was no significant inflationary pressure apparent for the economy as a whole, the Kampala cost of living index rose by about 6-1/2, 9-1/2, and 18-1/2 percent, respectively, for high, middle and low incomes earners over the period 1968-1970. In particular, food and clothing prices were reported to have increased dramatically.121 Thus, price movements in Uganda towards the end of the 1960's had a regres- sive effect on distribution of real income. As a result, pressure for higher minimum wages, the major factor in explaining rising unemployment, was intensified.122 The internal political and economic conditions came to closely resemble what it was in Ghana before the overthrow of President Nkrumah by the army. Obote must have been aware of this similarity for he became increas- ingly suspicious of the intentions of his own army. Consequently, Obote 121Michael Tribe, 22, cit., pp. 19-21. 1221bid., p. 20. 132 appointed separate commanders fer the army and the air force, in addition to creating a special security force under the direct control of his office.123 The dynamics of problempridden Uganda at the dawn of independence would have required mature leadership, interested less in the perpetua- tion of its own power and more in gradual introduction of policies to contain the fear and uncertainty displayed by the diverse tribal and econ- omic interest groups. Decisive and constructive solutions to the problem of the Asian minority, and the confidence of the private entrepreneurs in general, were required. Instead, these internal conflicts were inten- sified. Obote's tactics achieved neither the conditions for a stable nation nor the framework for rational economic policies. An unpopular and insecure political leader, Obote could not legislate policies (e.g., wage restraint) demanded by the prevailing problems, but unpopular among politically powerful and affected groups. Sooner or later the mounting economic problems and political dissatisfaction would provide an excuse for opportunistic military groups to assume political power. It has hap- pened in other African nations and it happened in Uganda in January 1971. As one would expect, the new military government announced that increased taxes, the lack of confidence of foreign investors following the declaration of the "Move to the Left," urban and rural unemployment, and political imprisonment, etc., were among the major factors motivating the overthrow of Obote. 123J. H. Mittelman, "The Anatomy of a Coup: Uganda, 1971," Journal of Indian Council for Africa, V01. XI, No. 3 (Oct.-Dec., 1971). pp. 184-202. 133 Subsequent events in Uganda demonstrated what has been true in most other countries of Africa: that a military government which has to depend on the same coercive political tactics followed by the regime it overthrew is often incapable of bringing about the fundamental reforms in the socio- political structures that are essential for successful economic development. very often, as happened in Uganda, the internal situation worsens, making the successful elimination of problems like unemployment even more diffi- cult. we will not go into the analysis of the policies of the military regime in Uganda. Our objective was to demonstrate that President Obote was overthrown because of his failure to bring about needed policies for the interrelated objectives of political and economic development. CHAPTER VII SUMMARY AND CONCLUSION The Ugandan case illustrates one of the most significant features of the development process as experienced in most developing economies in recent years -- namely, a substantial rise in the unemployment growth rate. Stagnation of total and sectoral employment opportunities have been a disappointing feature of the post-independence period, largely because the growth in non-agricultural output (the major objective of development plans during the period) has not produced any comparable expan- sion of employment. On the other hand, the rate of population growth has been rapid -- close to 3 percent annually. Rapid growth and a relatively young population have led to rapid growth of the labor force seeking employment. Given the stagnant level of wage employment, the consequence has been the indicated substantial growth in the level and rate of growth of unemployment. On the supply side, a reduction in labor force growth would be a long term process. The active labor force up to the 1980's has already been born, therefore, already determined. A population control policy, if adopted, will have a restraining influence on the rate of growth and size of the labor force after a long interval. This dissertation has not dealt with measures to control p0pulation growth, since the objective has been the generation of employment opportunities and a reduction of the unemployment gap in the short or intermediate run period of a decade or two. 0n the demand side, the stagnation in employment in Oganda or East 134 135 Africa may be explained by private-social discrepancies or monOpolistic practices which have biased the factor-mix against labor. This has taken a variety of forms: labor prices or wages have risen primarily because of government intervention through statutory determination of minimum wages; capital goods prices have been reduced below the level they would have attained through over-valuation of exchange rate or subsidies of various kinds; interest rates have been kept low; choice of techniques have been biased against labor-intensive ones by aid-tying arrangements or the mistaken belief in the superiority of production techniques of developed nations; and so on. The result is that, especially but not exclusively in the modern sector of the economy, one finds a higher capital/labor ratio than would be the case if factor prices were market-determined and deci- sions were based on them. In sum, the resultant stagnation of employment, in the face of growing supply of labor, has led to a significant level of unemployment. Unemployment of educated manpower has also become a serious problem in East Africa. Analytically, this category of unemployment indicates a case where investment has been misdirected, in the sense of leading to excess supply of certain types of training or skill (e.g., formal primary education). The colonial practice of maintaining a rigid relationship between the level of formal education and salary scales, inherited by the post-colonial regimes, induces and perpetuates the desire for formal educa- tion in the face of shortages in such categories of work as machinists, electricians, plumbers, and so on. The key to the generation of employment opportunities in the modern sector of uganda and, by extension, East Africa, is the elimination of 136 factor-price distortions which have discouraged the use of labor-intensive production techniques. Various ad hoc measures (including wage subsidies and tripartite agreements) to generate employment in the urban sector have been tried in East Africa. Since the urban-rural income differential continues to be maintained (assuming stagnation in rural incomes) or even widened, the ad hoc approaches have been shown to lead to a significant migration of the rural labor force in search of employment opportunities and higher earnings in the urban sector. The consequence of this process is likely to be a much higher level and rate of growth of urban unemploy- ment than would have been the case without the ad hoc measures. More- over, if the educated youth are induced to accept these urban created jobs at the given wages -- as they may likely do when confronted with a long period of unemployment -- the uneducated and older migrants would find it that much more difficult to secure urban wage employment. In short, the ad hoc measures are not effective substitutes for the elimination of factor price distortions. The fact that the urban labor force in an economy such as Uganda is a function of the prevailing wage rate, the rural-urban real income dif- ferential, and the assessment of individuals of the probability of getting a job, makes the concept and scope of incomes policy much more comprehen- sive than it would be in developed countries. In other words, incomes pol- icy in the context of Uganda must deal with prices and incomes received by farmers in the rural sector in addition to wages received by urban workers. In the past, development policy has been aimed at the gradual transfer of peasant farmers to employment in the urban or non-agricultural sector, on the assumption that mechanized farms, requiring much less labor input, 137 are much more productive. A significant proportion or rural incomes were withdrawn through export taxes and/or Marketing Board surpluses to be used fer growth in industrial output and employment. The result, however, has been disappointing. urban employment failed to expand adequately to absorb the rapidly growing rural to urban migrants and urban labor force. The narrow urban base of an economy such as Uganda, even under normal con- ditions, could not be expected to generate employment opportunities for more than a small proportion of the total labor force. In the final anal- ysis, the rural sector in uganda must provide the bulk of employment oppor- tunities. Consequently, and ideally, the removal of agricultural export taxes and Marketing Board surpluses were shown to be the appropriate policies. The removal of export taxes may not be possible given the revenue needs of a developing nation like Uganda. Then subsidies for agricultural labor and other farm inputs may be necessary to compensate for the re- sources that were withdrawn. This will maintain a higher marginal value product and hence, a higher level of employment than would be the case without the compensatory payments. The effectiveness of such policies depends on the relative contributions of the various inputs to agricultural value added. Labor and land are the most important farm.inputs in Uganda in terms of value added, so that subsidies for the purchase of these inputs should be expected to generate substantial employment opportunities. Under- lying this analysis is the premise, supported by the historical achievement record of such countries as Taiwan and Korea, that small-scale family farms in Uganda are responsive to market incentives. The trend towards mecha- nized or anti-employment farming techniques are not a necessary or inevitable III I ll.- [1 i I III" (III I Ill-ll 138 aspect of the Ugandan agricultural sector. In fact, the effectiveness of subsidizing labor inputs to the generation of employment depends on the degree to which subsidized mechanization of farms is eliminated. The above policies that are essential for the generation of employ- ment opportunities in a country like Uganda have gained wide support from economists familiar with conditions in the area. Policy-makers, too, should be assumed to have been aware of the need for such policies. Var- ious statements they have made and the inclusion of employment generation as one of the primary objectives of development plans in recent years indi- cate their awareness. Yet high urban wages and anti-employment policies in the agricultural sector continue to prevail, though at various inten- sity in the different countries of East Africa or elsewhere. The post- independence period in Uganda has been used to illustrate that political considerations primarily explain the persistent divergence between de- sired and actual policies for the generation of employment. In this sense, the effective adoption of the recommended policies to generate employ- ment in both rural and urban areas depends on the extent to which popular political institutions are developed and internal contradictions between the interests of different segments of the population are resolved. Uganda failed in this endeavor. Whether the new military regime will succeed remains to be seen. Its record since 1971 does not, however, offer hope- ful expectations. The Tanzanian socialist regime and the popularity of President Nyerere hasoften been cited as a model for other African countries facing similar economic problems. It is not necessary to go into a discussion of ideological 139 issues pertaining to the applicability of socialism to conditions in devel- oping nations such as Tanzania. The issue is complex and the intellec- tual debate is far from settled. There is no doubt that Tanzania had a highly stable government, but it is not clear whether this stability was due to the political philosophy of socialism or the political skill and charisma of President Nyerere. In any event, Tanzania was able to impose effective restraints on salary and non-salary benefits for civil servants and politicians. Accordingly, the acceptance of urban wage restraint has been enhanced. Moreover, Tanzania has made rural self-help a major ob- jective of employment generation and the elimination of mass unemployment. Yet some doubts persist. Unskilled wages have not been effectively curbed. Despite a variety of rural programs such as "villagization" schemes which started as early as 1963, the colonial structure of export taxes and the consequent negative impacts on agricultural prices and in- comes continue to be maintained. Rural aid programs partly compensate for the resources withdrawn and minimize the negative impact of rural employment. Yet, as Helleiner has shown, rural peasants in Tanzania have not fared as well as the wage earner even if the impacts of export taxes are not considered.124 Consequently, rural to urban migration, particu- larly of the young school-leavers, and urban unemployment continue to exist at a significant level.125 One could, however, argue that it would take some time before the various rural programs show the intended results -- higher incomes and employment. In short, the success of the Tanzanian model of development in the creation of adequate employment opportunities remains to be demonstrated. 124G. K. Helleiner, Agricultural Export Pricing Strategy in Tanzania, Paper 66 (Dar es Salaam: Economic Research Bureau, University College, December, 1966), pp. 2-9. 125See p. 46 above. BIBLICXERAPHY African DevelOpment. London EC4Y OAK: African Buyer and Trader Ltd., Wheatsheaf House, Carmelite Street, June, 1966. Amsden, A. H. International Firms and Labor in Kenya: 1945-1970. London: Frank Cass and Co., Ltd., 1971. Beer, Werner and Michel E. A. Herve. "Employment and Industrialization in Developing Countries," Thefiggarterly Journal of Economics, Lxxx (Febru- ary, 1966) ' pp. 88-107. Baryaruha, A. Factors Affectinngndustrial Employment: A Ugandan Exper- ience, 1954-1964. Nairobi: Oxford University Press, 1967. Berg, Elliot J. "Major Issues of wage Policy in Africa," Industrial Rela- tions and Economic Development, A. M. Ross (ed.). London: Macmillan, 1966. Byerlee, Derek. Research on Migration in Africa: Past, Present and Future. African Rural Employment Studies No. 2. East Lansing: Department of Agricultural Economics, Michigan State University, 1972. Callaway, Archibald. "Creaing Employment for Nigeria's School-Leavers," Qparterly Review of the Institute of Administration. University of Ife, October, 1968. . "Education Expansion and the Rise of Youth Unemployment," in The City of Ibadan, P. C. Lloyd, A. L. Mabogunje and B. Awe (eds.). Cambridge: Cambridge University Press, 1967. Clayton, E. 8. "Agricultural Reform, Agricultural Planning and Employ- ment in Kenya," International Labor Review, CV(4), 1972, pp. 309—334. Colony and Protectorate of Kenya. Labor Department Annual Report, 1947. Nairobi: Government Printer. . Labor Department Annual Report, 1956. Nairobi: Govern- ment Printer. . Labor Department Annual Report, 1946. Nairobi: Govern- ment Printer. . Report of the Committee on African wages (otherwise known as the Carpenter Report). Nairobi: Government Printer. 140 141 Consortium for the Study of Nigerian Rural DevelOpment. Preliminary Strategies and Recommendation for Nigerian Rural Development. East Lansing: Michigan State University, July, 1968. Dovering, Folke. "Land Reform in Mexico," Spring Review of Land Reform, Washington, DuC.: Agency for International Development, 1970. Edgar, Owens and Robert Shaw. Development Reconsidered. London: D.C. Heath and Co., 1972. Elkan, walter. The Economic Development of Uganda. London: Oxford University Press, 1961. . Migrants and Proletarians, Urban Labor in the Economic Development of Uganda. London: 1960. . "Urban Unemployment in East Africa," International Affairs, XLVI(3), Iondon, 1970. . "Out-of-School Education and Training for Primary School- Leavers in Rural Kenya: A Proposal," International Labor Review, FAO. Production Yearbook, 1969. Fisk, E.K. "Labor Absorption Capacity in Subsistence Agriculture,"Economic Record, XLVII(ll9), 1971, pp. 366-378. Frank, C.R., Jr. "Urban Unemployment and Economic Growth in Africa," Oxford Economic Papers, xx, (July 1968), pp. 250-274. Franklin, N. N. "The Concept and Measurement of 'Minimum Living Standards'," International Labor Review, XCV(4), 1967, pp. 271-298. Ghai, D.P. "Employment Performance, Prospects and Policies in Kenya," The Overseas Study Committee Conference. Cambridge University, 1970. (mimeographed) . "Incomes Policy in Kenya: Need, Criteria and Machinery," East African Economic Review, IV(new series)(l), 1968, pp. l9-34. Gormely, P. J. Some Implications of Population Growth in uganda. Economic Development Research Project No. 97. Kampala: Mekerere Institute of Social Research, April, 1966. Government of Kenya. "Report on the Tripartite Agreement on Measures for the Immediate Relief of Unemployment," Inter-African Labor Institute Bulletin, XII(1), 1965, pp. 94-98. Great Britain. Statement of Policy on Colonial Development and Welfare. Cmd. 6175. London: H.M.S.Q., 1940. 142 . The East African Royal Commission ReporgJ 1953-1954. Cmd. 9175. London: H.M.S.Q., 1955. Green, R.H. Tabular Appendix. Economic Development Research Paper No. 103. Kampala: Mekerere Institute of Social Research, (not dated) . Urban Growth in Sub-Saharan Africa. Kampala: Mekerere Institute of Social Research, 1970. Gwyer, G.D. "Employment Opportunities in Kenya Agriculture," East African Journal, IX(3), 1972, pp. 23-27. Hailey, Lord. An African Survey. Revised 1956. London: 1957. Hall, Malcolm. "Mechanization in East African Agriculture," in Agricultural Planning in East Africa, G. K. Heilbroner (ed.). Nairobi; East African Publishing House, 1968. Hence, William. Population, Migration, and Urbanization in Africa. New York: Columbia University Press, 1970. Harbison, F. H. "The Generation of Employment in Newly Developing Countries," in Education,_Employment and Rural DevelOpment, Sheffield (ed.). Nairobi: East African Publishing House, 1967. Harris, J. R. and M. P. Todaro. "wages, Industrial Employment and Labor Productivity: The Kenyan Experience," East African Economic . "Migration, Unemployment and Development: A Two-Sector Analysis," The American Economic Review, LX(1), 1970, pp. 126-142. . "Urban Unemployment in East Africa: An Economic Analysis of Policy Alternatives," The East African Economic Review, IV(2), 1968’ Pp. 17-36. Helleiner, G. K. Agricultural Export Pricing Strategy in Tanzania. Research Paper 66. Dar es Salaam: Economic Research Bureau, University College, 1966. Henderson, J. P. "wage Policy in Africa," African Conference on Economics. Temple University, 1968. (mimeographed) Herskovits. The Human Factor in Changing Africa. New Yerk: Random House, 1962. IBRD. The Economic Development of Uganda. Baltimore: Johns Hopkins Press, 1962. 143 ILO. Report to the Government of the United Republic of Tanzania on wages, Incomes, and PricesAPolicy. Government Paper No. 3-1967. Dar es Salaam: Government Printer, 1967. . Epployment Policy in Africa! Part I: Problems and Policies. Third African Regional Conference (Accra). Geneva: International Labor Office, 1969. . Minimum nge Fixing and Related Problems with Special Refer- ence to Developing Countries. Geneva: International Labor Office, 1967. Johnson, G. E'.. and W. E. Whielaw. "Rural Income Transfers in Kenya: An Estimated Remittance Function." Institute of Development Studies, University of Nairobi, 1972. (unpublished) Joy, J. L. (ed.). gypposium on Mechanical Cultivation in Uganda. Kampala: Uganda Argus Ltd., 1960. Knight, J. B. "The Determination of wages and Salaries in Uganda," Bulletin of the Oxford Institute of Economics and Statistics, XXIX(3), 1967, pp. 233-264. . "Earnings, Employment, Education and Income Distribution in Uganda," Bulletin of the Oxford Institute of Economics and Statistics, XXX(4): PP. 267-297. . Salagy Structure in Uganda. Economic Development Research Paper No. 106. Kampala: Mekerere Institute of Social Research, 1966. ‘ Morris, J. D. Ireri and J. Heyer. Rural Development in Kenya. Nairobi: East African Publishing House, 1971. Lefeber, Louis, "Planning in a Surplus Labor Economy," The American Econ- omic Review, LVIII(3), 1968, pp. 345-373. MacArthur, J. G. "The Economic Study of African Small Farms," Journal of Agricultural Economics, XIX, 1968, pp. 193-205. Maitha, J. K. "Productivity Response to Price: A Case Study of Kenya Coffee," East African Economic Review, 11(2), 1969, pp. 31-37. Mattrick, Hall. Aid in qganda-eggriculture. London: The Overseas Devel- opment Institute, Ltd., 1967. Mitchell, P. E. (Sir). African Afterthoughts. London: 1954. Mittelman, J. H. "The Anatomy of a Coup: Uganda, 1971," Journal of Indian Council for Africa, XI(3), 1971, pp. 184-202. 144 Narasimhan, P. S. Kenya: Conditions of werk and Life in Plantations, Report of an ILO Mission. 1969. (mimeographed) Noon, John A. Labor Problems of Africa. University of Pennsylvania, 1944. Obote, A. Milton. The Common Man's Charter. Entebbe: Government Printer, 1968. Ojow, F. I. Employment, Ougput and Labor Force Trends in East Africa. Econ- omic Development Research Project No. 138. Kampala: Mekerere Institute of Social Research, 1968. (mimeographed) Orde-Browne, G. St. J. The African Laborer. London, 1933. Palekar, S. A. Problems of wage Policyifor Economic Development: With Special Reference to Uganda. Asia Publishing House, 1962. Pearson, D. S. Industrial Development in East Africa. Studies in African Economics No. 2. Nairobi: Oxford Printing Press, 1969. Peterson, William. "Urban Policies in Africa.and Asia," Population Review, Powesland, P. G. Economic Policy and Labor: A Study in Uganda's Economic History. East African Studies No. 10. Kampala, 1957. Rado, E. R. "Manpower Planning in East Africa," The East African Economic Ray, R. S. Labor Force Survey in Tanzania. Ministry of Economic Affairs and Development, The United Republic of Tanzania, 1966. . "The Structure of Employment in Kenya," in Education, Employ— ment and Rural Development, James R. Sheffield (ed.). Nairobi: East African Publishing House, 1967. Rempel, Henry. "Labor Migration into Urban Centers and Urban Unemployment in Kenya." Unpublished Ph.D. dissertation. Madison: University of Madison, 1970. Report of the Commission on the Civil Service of Kenya, Tanganyika, Uganda and Zanzibar, 1947-1948. Nairobi: Government Printer. (Otherwise known as the Holmes Commission Report) Report of the Commission on the Civil Service and the East African High Commission, 1953-1954. Nairobi: Government Printer. Report of the Uganda Civil Service Commission. Entebbe: Government Printer, 1963. 145 Reynolds, Lloyd G. "wages and Employment in a Labor-Surplus Economy," The American Economic Review, LV, 1965, pp. 19-39. Ryan, Selwyn Douglas. "Balance Sheet of the Revolution," Mwazo, 11(1), 1967 ' pp. 37-64. Sabot, R. H. "Education, Income Distribution and Rates of urban Migration in Tanzania." Unpublished paper, Economic Research Bureau, University of Dar es Salaam, 1972. Schmid, Lester. Relation of Size of Farm to Productivity. Land Tenure Center, University of Wisconsin, 1969. Sheffield, James R. (ed.). Education, Employment and Rural Development. Report of the Kerico (Kenya) Conference, 25th September to let October. Nairobi: East African Publishing House, 1967. Scot, R. The Development of Trade Unions in Uganda. Nairobi: East Afri- can Publishing House, 1966. Segal, Aaron. "The Politics of Land in East Africa," Economic Development and Cultural Change, XVI(2), 1968, pp. 273-84. Smith, A. D. Wage Policy Issues in Economic Development. Edinburgh: RaR Clark, Ltd., 1969. Smith, T. E. and J. C. Blacker. Population Characteristics of the Com- monwealth Countries of Tropical Africa. University of London: The Athlone Press, 1963. Stiglitz, J. E. "Rural-urban Migration, Surplus Labor and the Relation- ship between Urban and Rural wages," East African Economic Review, 1(1), 1969' pp. 1-28. Stoutjesdijk, E. J. Uganda's ManufacturingiSector: A Contribution to the Analysis of Industrialization in East Africa. East African Studies No. 29. Nairobi: East African Publishing House, 1967. Tanganyika Government. Labor Department Annual Report, 1948. Dar es Salaam: Government Printer. . Labor Department Annual Report, 1957. Dar es Salaam: Gov- ernment Printer. Thomas, R. L. "High Level Manpower in the Economic DevelOpment of Uganda," in Manpower and Education, F. Harbison and C. Myers (eds.). New York: McGraw-Hill Co., 1964, pp. 297-324. Todaro, M. P. "Income Expectations, Rural-Urban Migration and Employment in Africa," International Labor Review, CIV(5), 1971, pp. 387-413. 146 Tribe, Michael. "Uganda 1971: An Economic Background," Mwazo, 111(1), 1971' Pp. 15-62. uganda Government. Enumeration of Employees. Various issues. . work for Progress, Second Five Year Development Plan, 1966-1971. Entebbe: Government Printer, 1966. . Kempala—Meng Regional Planning Studies No. 1, Population and Land Requirement, 1965. . A Five Year Copital Development Plan, 1955-1960. Entebbe: Government Printer, 1955. . The First Five Year Plan, 1961-62/1965-66. Entebbe: Gov- ernment Printer, 1963. Uganda Protectorate. Labor Department Annual Report, 1958. Entebbe: Gov- ernment Printer. . Labor Department Annual Report, 1947. Entebbe: Government Printer. United Nations. Economic Bulletin for Africa, X(l), 1970. New York. United Republic of Tanzania. wagee,,Incomes, Rural Development, Investment and Price Policy. Government Paper No. 4-1967. Dar es Salaam: Gov- ernment Printer, 1967. Vail, David. Group Farming and Statecraft in Uganda. Development Research Project No. 71. Kampala: Mekerere Institute of Social Research, 1968. woddis, Jack. Africa: The Roots of Revolt. New York: The Citadel Press, 1962. worthington, E. B. A Development Plan for Uganda and the 1948 Revision of the Plan. Entebbe: Government Printer, 1949. HICHIGRN STRTE UN V. BRRRI 9311191 9191111911919 911