Analysis of changes in Ghana, 1960-70 labor demand and supply in Ghana, 1960-70

ANALYSIS OF CHANGES IN LABOR DEMAND AND SUPPLY IN GHANA, 1960-70 by William F. Steel Series: Studies in Employment and Rural Development No. 45 Division: Employment and Rural Development Department: Development Economics Development Policy Staff International Bank for Reconstruction and Development This paper was prepared for the internal use of the World Bank, and may not be quoted, referred to without permission of the authors or the Bank. The views expressed represent those of the authors and are not necessarily those of the Bank. Washington, D.C. August 1977 CONTENTS INTRODUCTION 1 SOURCES AND COMPOSITION OF DEMAND FOR NEW EMPLOYMENT Sectoral employment shifts 2 Sectoral sources of employment increases 4 Detailed industry sources of employment increases 7 Occupational composition and sources of increase 9 SOURCES AND CONSEQUENCES OF INCREASED LABOR SUPPLY Changes in population and labor force 11 Underutilization of labor 17 CONTRIBUTION TO NEW EMPLOYMENT OF LARGE AND SMALL SCALE SECTORS Employment absorption in large and small enterprises 20 Major small and large scale detailed industries 22 SUMARY AND CONCLUSIONS Synopsis of basis facts presented in each section 27 Analytical and policy conclusions 29 Directions for future research 32 REFERENCES 34 TABLES 1. Changes in sectoral composition of employment, 1960-70 3 2. Sectoral sources of employment growth, 1960-70 5 3. Major detailed industry sources of additional employment from 1960-70 8 4. Employment increase and composition by occupation, 1970 10 5. Growth of population and labor force, 1960-70 12 6. Occupational composition of foreign-origin and Ghanaian workers in 1960 14 7. Distribution of employment and labor force increases 1960-70 by age group 16 8. Underutilization of labor, 1960-71 18 9. Sectoral sources of employment absorption 1960-70, large and small scale 21 10. Major small-scale detailed industries: small-scale share and growth, 1960-70 23 11. Major large-scale detailed industries. small-scale share and growth, 1960-70 25 ANALYSIS OF CHANGES IN LABOR DEMAND ANID SUPPLY IN GHANA, 1960-70 by William F. Steel INTRODUCTION This paper presents and analyzes the major trends in employment, unemployment, and labor force participation in Ghana over 1960-70. The data are primarily from the Population Censuses, as presented in Steel (1977). The principal questions to be examined are: 1. What were the sectoral, industry, and occupational sources of employment increases? 2. What was the relationship between the growth of labor demand and supply over the period, and what were the implications for underutilization of labor? 3. How can differences between employment, unemployment, and labor force participation trends between men and women be explained? 4. To what extent was modernization and industrialization a stimulus for employment growth? 5. What was the composition and relative contribution to new jobs of employ- ment in small-scale as against large-scale enterprises? 6. What policies are appropriate to stimulate productive labor absorption in the future? 7. What research is needed in order to understand trends more fully and to develop appropriate policies? The discussion of these questions is divided into four parts: sources and composition of demand for new mployment (question 1); sources and consequences of increased labor supply (questions 2-3); contribution to new employment of large and small scale sectors (questions 4-5); and summary and conclusions (questions 6-7). Questions 2, 4, 6, and 7 are to some extent analyzed throughout the paper, as appropriate. A synopsis of the basic findings in each sections is presented at the beginning of the summary 'and conclusions, as a concise review of the basis for the analytical conclusions. 2 SOURCES AND CONPOSITION OF DEMAND FOR NEW EMPLOY\ENT Sectoral Employment Shifts The structure of employment in Ghana "modernized" over the 1960s in the sense of shifting from the primary to the secondary and tertiary sectors (Table 1). The share of each primary subsector declined, resulting in an overall decrease from 64% to 58% of employment. The gain in share of manufacturing accounts for most of the corresponding increase; a slight offsetting decrease in construction resulted in a net shift in secondary employment from 13% to 16% of the total. Government admini- stration and education were the other sectors registering substantial gains, with the tertiary sector as a whole increasing from 23% to 26%. Although the shares of male and female employment moved in the same direction for all subsectors except "other services," the share of male employment in the tertiary sector as a whole increased whereas the corresponding female share decreased. This is explained by the relatively large share of female employment in commerce, a declining sector, and the relatively small share of female employment in govern- ment and education, the major increasing sectors other than manufacturing. The share of female workers in manufacturing and in the secondary sector as a whole in- creased more than for males. Females remained much more heavily concentrated in commercial activities than males, while males remained disproportionately repre- sented in construction, forestry and fishing, utilities, government, education, and mining and quarrying. Large-scale recorded employment, which accounted for about 13% of all workers in 1970,1 exhibited a different sectoral structure and pattern of change from employ- ment as a whole. The already small primary share decreased markedly from 26% to 17%, and the secondary share also fell from 30% to 26%, in contrast to the general rise of secondary employment. AlLhough the large-scale manufacturing share did rise, it was overwhelmed by the drastic drop in construction. Tertiary employment came to dominate the modern sector, rising from 44% to over 57% of its workers. This in- crease was led by a doubling of the government share to a size more than twice that of any other subsector, and by a substantial increase in the share of education. Relatively little change occurred in the employment structure of employment outside the modern sector ("small scale"), except for a rise in manufacturing and in the secondary sector as a whole. The lack of change in the tertiary sector iAdjustment for underestimation of the growth of government employment would raise the share to 14% (see Steel 1977, Statistical Table B.6, note w). TABLE 1 3 CHANGES IN SECTORAL COMPOSITION OF EMPLOYMENT, 1960-70: MALES AND FEMALES, SMALL AND LARGE SCALE' (percentage) Total Hales Females Large Smalla Sector 1960 1970 1960 1970 1960 1970 1960 1970b 1960 1970b (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) PRIMARY 63.6 58.0 66.7 60.7 58.6 54.7 26.0 16.5 69.2 65.0 AgricultureC 38.3 36.4 35.3 33.6 42.9 39.7 12.7 8.8 65.5 62.3 Cocoa 20.4 18.3 23.8 21.4 15.0 14.5 forestry and 3.1 2.4 4.7 4.1 0.4 0.3 4.6 2.1 2.8 2.4 fishing Mining and ar d 1.9 1.0 2.9 1.7 0 3 0.2 8.7 5.6 0.9 0.2 quarrying SECONDARY 13.2 16.2 15.1 17.0 10.3 15.3 30.1 26.2 10.7 14.6 Manufacturing 9.2 13.2 8.7 11.7 10.0 15.1 7.3 11.8 9.5 13.5 Utilities 0.6 0.6 0.9 1.2 0.0 0.0 4.3 3.3 0.0 0.2 Construction 3.5 2.3 5.5 4.1 0.3 0.2 18.6 111 1.2 0 9 TERTIARY 23.2 25.7 18.2 22.3 31.1 30.0 43.9 57.3 20.1 20.5 Transport, stor 2.6 2.7 4.3 4.8 0.1 0.2 9.4 7.2 1.6 1.9 age & commun. Commerce 14.7 13.9 6.4 4.1 27.8 25.9 10.4 8.5 15.3 14.8 Education 1.4 2.5 1.8 3.5 0.7 1.4 6.1 10.2 0.7 1.2 Governmente 1.7 3.5 2.7 6.0 0.2 0.6 12.4 24.7 0.1 0.0 Other services 2.7 3.1 3.0 3.9 2.3 2.1 5.6 6.6 2.3 2.5 TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100,0 100.0 100.0 a"Large" scale refers to "modern" sector establishments covered by Labour Statistics (generally 10 or more employees). "Small" scale refers to "intermediate and informal sector establish- ments, calculated as a residual by subtracting Labour Statistics figures from Census data. bFigures differ from those reported in Steel 1977, Statistical Table B.1, because the adjusted total for large scale is used instead of the reported total (see note w to Statistical Table B.6). c"Agriculture" includes livestock farming, agricultural services, hunting, and all crops other than cocoa. d. Mining is placed in the primary sector because it is purely extractive (i.e. negligible proces sing) and is an export industry (as are cocoa and forestry). e"Government" refers to public administration and defence not elsewhere classified (i.e., under education or a particular industry). Details may fail to add to totals because of rounding. Source: Steel 1977, Statistical Tables A.2, A.3, A.4, B.1, B.2, and B.6. I 4 share of small-scale employment indicates that the overall increase in the tertiary share was attributable entirely to the rapid growth of government and education in the modern sector. Sectoral Sources of Employment Increases The agricultural sector was the largest source of increased employment over 1960-70, in spite of its declining share. It provided 36% of the net employment increase in the economy, and 46% of the increase in female employment (Table 2, cols. 4 and 6). It was not, however, nearly as important in absorbing male labor, providing less than 10% of the net increase and growing at just 0.1% per annum (cols. 5 and 8). The figures suggest a need for increased attention to the role of women in agricultural production, based on further research into whether the increase in female agricultural employment represented replacement of male workers, cultivation of new land, or a purely statistical change resulting from increased reporting of female agricultural workers as "employed" rather than as "homemakers." Manufacturing was the second-largest source of new employment and the second- fastest growing sector for both male and female workers, as well as for the total (Table 2). Services, however, was by far the dominant source of new employment for males, providing 75% of the net increases. The expansion of male employment in these two sectors in the face of substantial decreases in the number of male workers in commerce, mining, and construction raises several questions for further investi- gation: (1) did workers leave these latter sectors to seek more attractive employ- ment opportunities in the (public)* service sector and in manufacturing? (2) did government expand its employment in order to absorb laid-off workers? or (3) did surplus labor in declining sectors shift to generally unproductive "informal" service activities? Some initial evidencc toward answering these questions will be presented below in discussing the growth of specific industries and changes in labor force participation. The commerce sector presents an apparent paradox, with a substantial decline in male employment more than offset by a vigorous increase in female employment. The implied take-over by women of male jobs in this sector is most likely a result of the Aliens Compliance Order of December, 1969. This Order was aimed at enforce- ment of laws requiring registration of non-Ghanaians working in certain sectors, and it led to the rapid departure of a large number of non-Ghanaian Africans (see Addo, 1972, for the only available statistical evidence). Traders were particularly af- fected, and the Ghanaian Business Promotions Office was charged with making small loans to Ghanaians wishing to take over trading businesses left behind. Since over TABLE 2 SECTORAL SOURCES OF EMPLOYMENT GROWTH, 1960-70 Sector (ranked by Increase in employment Distribution of increase,(%) Average Annual growth (%) employment growth) Total Males Females Total Males Females Total Males Females (1) (2) (3) (4) (5) (6) (7) (8) (9) Agriculture, forestry, 207,412 14,410 193,002 36.2 9.6 45.6 1.3 0.1 3.3 hunting, and fishing Manufacturing 178,897 64,234 114,683 31.2 42.8 27.1 7.6 4.7 11.6 Services 136,185 111,834 24,351 23.7 74.6 5.7 9.1 9.5 7.7 'Commerce 60,876 -29,512 90,388 10.6 -19.7 21.3 1.6 -2.9 3.3 Transport, storage, and 16,478 15,284 1,194 2.9 10.2 0.3 2.4 2.3 11.1 communication Electricity, gas, water, 6,140 5,787 353 1.1 3.9 0.1 4.3 4.1 20.3 and sanitation Construction -15,089 -15,086 -3 -2.6 -10.1 -0.0 -1.7 -1.8 -0.0 Mining and quarrying -17,235 -16,988 -247 -3.0 -11.3 0.1 -3.6 -3.7 -1.0 Total nonagricultural 366,252 135,553 230,699 63.8 90.4 54.4 3.7 2.4 5.6 Total, all sectors 573,664 149,963 423,701 100.0 100.0 100.0 2.2 1.0 4.3 Note: Details may fail to add to totals because of rounding. Source: Steel 1977, Statistical Tables A.1 and A.3. 01n 6 2 half of male traders in Ghana in 1960 were non-Ghanaians, the decrease in male com- mercial employment probably reflects their departure shortly before the 1970 Census. Ghanaian women apparently dominated the replacement of male as well as female non- Ghanaian traders. The overall 1.6% per annum growth of commercial employment is not un- reasonable in view of real output growth nearly twice that rate3 and of increasing urbanization. The emphasis on expansion of infrastructure and social services in the early 1960s explains the relatively rapid 4.3% per annum growth of utilities. Since transport and communications should -imilarly be tied to the growth of infrastructure, it is surprising that this sector did not grow faster than 2.4% per annum. Female employment grew at exceptionally rapid rates in both these sectors, but from an extremely small base, so that The total impact on female employment was negligible. The decline in construction employment reflects the termination of the post- IndeDendence building boom, after the overthrow of Nkrumah in 1966. Nining is a generally declining sector ir Ghana, given the exhaustioa of the readily-obtainable mineral deposits near the surface. Although nonagricultural employment growth of 3.7% per annum exceeded the growth rates of population and production in the economy as a whole, the slow (1.3%) growth of employment in agriculture held the overall growth of employment to 2.2%. This rate is equal to the growth of the total labor force and greater than the 1.8% growth of the population age 15 and over (total population grew at 2.7%). Whether the implied increase in labor force participation was a supply response to increased demand and wages or represented a supply shift is a complex question that will be discussed below (pp. 11-19 ). The structure of nonagricultural employment changed more radically over 1960-70 for men than for women (Steel 1977, Statistical Table A.2). The drastic decline of men in commerce (from 24% to 10% of the total) was more than matched by the rise of the share in services (22% to 38%). The combined total of services and manufacturing jumped from 46% of male nonagricultural employment to 66%. The decline of the share of women in commerce (from 67% to 57%) was taker up in manufacturing (from 24% to 33%); their combined dominance of female nonagricultural employment, however, staved virtually the same (from 91% to 90%), 2Persons of foreign origin represented 51.5% of male sales workers but only 15.4% of females; foreign-origin males constituted 44.9% of foreign-origin sales workers. (Ghana, Population Census, 1960, Vol. IV, Tables 6 and 19). 3Calculation of the growth of output per capita depands heavily on the choice of end points, as well as of the price deflator. 4The urban population (towns of 5,000 or more) in 1970 represented 29% of the total, as against 23% in 1960. 7 Detailed Industry Sources of Employment Increases Examination of detailed industry sources of employment increases over 1960-70 indicates that the initially large industries rather than the fast-growing ones absorb the most labor. Three of the top five sources of employment increase--food production, retail trade, and cocoa--grew at only 1.8% per annum or less, yet together accounted for 49% of new jobs (Table 3, cols. 2-3). Although half of the top twelve industries were in manufacturing, the only really substantial contributor--local food preparation--could be considered as much a service as a manufacturing activity. It accounted for 11% of the total employment increase, almost as much as the other five top manufacturing industries conbined (13%). All of the manufacturing indus- tries were related to food and beverages or to clothing. In the food and beverage industries, the small-scale sector was growing considerably faster than large- scale employment, while the opposite was true in the textile industries.5 Three of the industries that grew by more than 10,000 workers were related to expansion of the public sector, and accounted for 21% of the total employment increase: public admini- stration, education, and research institutes and community services. 'These activities most closely meet the criteria that a "growth sector" should be relatively large, predominantly "modern," and fast-growing. It is questionable, however, whether tertiary sector services with little scope for raising productivity can serve as a Rostovian "leading sector." In any case, rapid growth of the public sector in Ghana may be a phenomenon of the past, given the fiscal problems and pressures to combat inflation in the 1970s. What other sectors can lead growth of Ghana's employment into the 1980s? Given that all of the remaining industries providing over 10,000 new jobs, except for textiles,6 were over 90% dominated by small-scale employment in 1970, the potential of small-scale activities for absorbing future labor supply increases must be investigated thoroughly. Only three of the ten industries providing an additional 5,000 or more jobs for men over 1960-70 also provided as many new jobs for women: public administration, education, and food production (Table 3, cols. 4-9). Of those, public administration and education were the dominant sources of male employment increase, but relatively minor sources of new female employment. Tertiary sector industries were in general the most important ones for males. Female employment increases, however, came DSteel 1977, Statistical Table B.7. The relative contributions of the large and small-scale sectors are discussed below, pp. 20-26. 6The small-scale share of textiles in 1970 was from 59% to 81%, depending on whether Industrial Statistics of Labour Statistics are used to represent the large- scale sector (both show the industry as 99% small-scale in 1960-62). Source: Steel 1977, Stalistical Table B.7. TABLE 3 MAJOR DETAILED INDUSTRY SOURCES OF ADDITIONAL EMPLOYMENT FROR 1960-70: TOTAL, MALES, AND FEMALES AGE 15 AND OVER k Total, male and female employment Male workers Female workers Industries Jobs Annual % Industries Jobs Annual % Industries Jobs Annual % adding 10,000 ('000) growth adding 5,000 ('000) growth adding 5,000 ('000) growth (1) (2) (3) (4) (5) (6) (7) (8) (9) 1-11 Foodstuffs 164 1.8 9-1 Public administra- 61 14.5 1-11 Foodstuffs 135 13.3 Lion 9-1 Public administra- 67 15.2 9-31 Education 31 10.7 6-2a Retail trade ex- 88 3.2 Lio cept petroleum 6-2a Retail trade ex- 64 1.8 1-11 Foodstuffs 29 0.6 3-18 Local food prpara- 61 n.a. cept petroleum Lion 3-18 Local food prepara- 62 n.a. 3-21 Wines & spirits 13 304.7 1-12 Cocoa 56 3.8 1-12 Cocoa 51 1.0 3-31 Textiles 11 8.8 3-32 Wearing apparel 14 2.9 9-31 Education 43 12.1 9-32,35 Research insti- 9 37.8 3-14 Oils & fats 12 24.9 Lutes, community services 3-32 Wearing apparel 18 2.4 9-34 Trade & other asso- 8 12.2 9-31 Education 12 17.9 ciations 3-21 Wines & spirits 15 309.6 7-la Road Transport 7 1.6 3-22 Malt brewing 12 19.3 3-14 Oils & fats 14 26.9 9-53 Vehicle repair 7 5.9 9-1 Public administra- 6 32.5 tion 0. 3-31 Textiles 13 9.7 3-84, 9-52 Electrical 6 75.8 machinery & repairs . 3-22 Malt brewing 13 19.1 2. 9-32,35 Research insti- 12 44.9 tutes, community services otal, all industries 574 2. 150 1.0 424 4.3 ource: Steel 1977, Statistical Tables A.4 and A.5 oC predominantly from the primary sector, with foodstuffs and cocoa accounting for 53% of the total gain. Although both lists contain several manufacturing industries, there is no single industry that provided over 5,000 new jobs to both females and males. Retail trade was a major source of new employment for women but not for men, while the reverse held for road transport. Employment expansion by industry is heavily sex-biased, and growth of a particular industry cannot be assumed to have neutral employment benefits with regard to sex.7 Occupational Composition and Sources of Increase The pattern of employment increase by occupation resembles that by sector (Table 4, cols. 1-3). The overall ranking corresponds to that for females: agri- cultural and production occupations predominate. For men, however, professional, technical, clerical, and service occupations rank close to agriculture as sources of new jobs. These are the occupations in which the modern sector share is highest and the female share is lowest (cols. 10-11; excluding administrative and managerial workers, a very small category). Ironically, these occupations are relatively im- portant sources of employment for women in the modern sector (col. 9), even though they contribute relatively little to total female employment (col. 5) . Expansion of activities requiring professional, technical, clerical, and service workers thus is important if the modern sector is to grow and if women's employment in the modern sector is to rise, but tends to favor male employment overall. Sales occupations, on the other hand, are dominated by women overall and represent the modern sector occupation with the highest female share (Table 4, cols. 11-12). The low (1.2%) modern sector share, as well as the outflow of male workers, suggests that it may be more of a residual sector for women seeking employ- IID ment than a potential "leading sector." It accounts for over half of female non- agricultural employment. Production occupations appear to be the most balanced in terms of providing a substantial share of employment to both men and women (cols. 2-5) while having a relatively large modern sector component (col. 10). It accounts for over half of male nonagricultural and modern sector employment (cols. 6, 8), although it is less important for females in the modern sector, who constitute only 3% of large-scale production workers (cols. 9, 12). Expansion of activities involving production workers can provide productive income-earning opportunities to both men and women while helping to modernize the economy. A danger exists, however, that overemphasis on modern sector production could have a negative effect on female employment, which is predominantly in small-scale production (see Steel and Campbell 1977). 'See Steel and Campbell 1977 for further discussion of "The Impact of Moderni- zation on Women's Employment in Ghana." TABLE 4 EMPLOYMENT INCREASE AND COMPOSTTION BY OCCUPATION, 1970 upation (in Employment increase Distribution of employmenlt in 1970 (%) Modern Female share- er by employ- 1960-70 ('000) Total Nonagricultural Modern sector sector Total Modern t increase) Total Males Females Males Females Males Females Males Females share (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (i1) (12) riculture 235 40 194 59.8 54.5 --- --- 6.0 10.3 1.4 42.9 15.6 oduction 136 23 112 23.1 15.4 57.4 33.8 52.7 16.8 31.9 35.4 3.3 4les 68 -18 85 2.9 25.7 7.3 56.4 1.1 3.3 1.2 87.8 24.5 ofessional, 60 43 16 5.3 2.0 13.3 4.4 12.9 33.8 49.8 23.5 22.0 technical erical 43 33 10 4.2 0.9 10.6 2.1 12.0 22.4 60.1 15.5 16.7 rvice 35 30 5 4.0 1.5 10.0 3.3 11.5 11.9 50.8 23.3 10.1 ministrative, -2 -2 0 0.6 0.0 1.6 0.1 3.8 1.5 100.0a 5.2 4.0 Dtal 574 150 424 100.0 100.0 100.0 100.0 100.0 100.0 12.7 45.2 9.7 abour Statistics figures are 124.4% of the number counted in the Census for 1970. The excess may represent professional and clerical workers classified as administrative in Labour Statistics Source: Steel 1977, Statistical Table A.7 and Ghana, Labour Statistics 1970, Table 6. CD SOURCES AND CONSEQUENCES OF INCREASES LABOR SUPPLY Changes in Population and Labor Force, 1960-70 Several startling differences between male and female employment and labor force figures require explanation: a. male employment grew at only 1.0% per annum, while female employment grew at 4.3% over 1960-70 (Table 5, line C.1); b. the male labor force grew at only 1.1% per annum, slower than male popu- lation growth, whil- female labor force grew at 4.1% (line B.1), faster than female population growth; c. male labor force participation fell in the age groups 15-19, 20-24, and 25-29, and stayed constant in the 30-34 group, whereas female participation fell only in the 15-19 group and rose sharply in all older age groups (Steel 1977, Statistical Table C.1a). The labor force participation decline in the 15-19 age group for both males and females can be explained by the rapid expansion of education in the 1960s. Even given the bias of higher-level education toward males, however, educational expansion cannot account for the opposite trends in male and female labor force participation in the 20-24 and 25-29 age groups, nor does it explain why the male labor force age 25 and over should grow at only 1.1% per annum while the corresponding female labor force grew at 4.9% (Table 5, line B.3). The evidence presented in this section suggests that more satisfactory explanations may lie in the exodus of immigrants (especially males) after the Alien' Compliance Order of 1969 (discussed above, p.4), plus a possible bias in the benefits of improvements in health. The female population as a whole grew somewhat faster (3.0%) than the male population (2.5%; Table 5, line A.1) over 1960-70. The difference was much. more pronounced for ages 25 and over than for children: 2.4% average annual growth in female population as against 1.4% for males (line A.5b). There is relatively little difference in growth rates in age groups below 25, except for 10-14 (lines E.1-5). The consistently higher growth rates for females in all age groups above 25, and the much higher growth rates for the population in the older age groups above 45 (especially 65 and over: line E.12) suggest that major improvements in health and in disease control may have occurred in such a way as to benefit women more than men.8 Demographic data should reveal whether life expectancy increased more for women than for men. 12 TABLE 5 GROWTH OF POPULATION AND LABOR FORCE, 1960-70: TOTAL, MALES, AND FEMALES (average percentage rate per annum) Total Males Females (1) (2) (3) A. Population 1. Total 2.7 2.5 3.0 2. Foreign-born -3.7 -4.0 -3.3 3. Ghana-born 3.3 3.2 3.4 4. Under 15 3.4 3.3 3.5 5. Age 15 and over 2.2 1.8 2.5 5a. Age 15-24 2.9 3.0 2.8 5b. Age 25 and over 1.9 1.4 2.4 B. Labor force 1. Total, 15 and over 2.2 1.1 4.1 2. 15-24 1.1 0.2 2.2 3. 25 and over 2.7 1.1 4,9 C. Employment 1. Total, 15 and over 2.2 1.0 4.3 2. 15-24 0.6 -0.6 2.2 3. 25 and over 2.8 1.1 5.1 D. Unemployment 1. Total, 15 and over 2.1 3.0 0.5 2. 15-24 3.6 4.6 2.0 3. 25 and over -0.5 0.5 -3,3 E. Population by age group 1. 0-4 2.1 2.1 2.0 2. 5-9 4.2 4.1 4.3 3. 10-14 4.7 4.4 4.1 4. 15-19 L,4 4.5 4.3 5. 20-24 1.5 1.4 1.6 6. 25-29 0.8 0.4 1.1 7. 30-34 1.5 0.9 2.1 8. 35-39 1.6 1.2 2.1 9, 40-44 1.1 0.5 2.1 10. 45-54 2.7 2.0 3,5 11, 55-64 2.8 2.4 3.3 12. 65 and over 4,6 3.9 5.3 Source: Steel 1977, Statistical Table C; Ghana, Population Census: 1960 Vol. II, Table 1, and 1970 Vol. III, Table C.5. 13 The importance of out-migration as an explanatory factor is indicated by the facts that there was relatively little difference between male and female growth of the Ghana-born population (Table 5, line A.3) and that the foreign-born population decreased in size, with a more rapid rate of decrease for males (4.0% per annum) than for females (3.3%; line A.2). The decrease of 142,223 foreign-born males represented 7.5% of the 1960 male population, whereas the decrease of 67,614 females was just 3.7% of the 1960 female population.9 The foreign-born population, however, represents only 62% of all foreign nationals, some of whom may have left the country under the Aliens Compliance Order even though they were born in Ghana. If the net decrease in foreign male nationals age 15 and over of 135,568 is added to the 1970 labor force and population, the annual percentage increases would have been 1.9% and 2.5%, respectively.10 The net decrease of 67,414 in female foreign nationals 15 and over would have raised their labor force and population growth rates to 4.7% and 2.9%, respectively. The actual out-migration -- if it was due primarily to the Aliens Compliance Order and not to a permanent reversal of the long-standing flow of migrants from neighboring countries into Ghana -- would have been of much larger magnitude, but no basis exists for estimating the total in the absence of data on the size of the foreign population in 1969 or its rate of growth over 1960-69. The question still remains as to why the female labor force grew much faster than female population growth whereas the reverse was true for males. Differences in labor force participation between the decreasing foreign-born population and the rising Ghana-born population provides a partial explanation. The labor force par- ticipation rate in 1960 of 95.2% for foreign-born males exceeded the 87.8% for Ghana- born males, while the foreign-born female rate of 42.8% was below the Ghana-born rate of 57.9%.11 Hence, the out-migration of the foreign-born population in response to the Aliens Compliance Order tended to lower labor force participation and growth of the remaining male population and raise it for females. The earlier discu-sion of the impact of the Aliens Compliance Order on male and female employment in commerce provides a further explanation. The foreign- origin population was disproportionately in sales occupations (Table 6). Since Ghanaian males tend to view petty trade as an inappropriate occupation for them, the departure of foreign traders left a vacuum that was filled principally by Calculated from Ghana, Population Census, 1960 Vol. II, Table 1, and 1970 Vol. III, Table CS. 10Calculated from Ghana, Population Census, 1960 Vol. IV, Table 15, and 1970 Vol. IV, Table D7. 11Calculated from Steel 1977, Statistical Table C.2, and Ghana, Population Census, 1960 Vol. IV, Table 15. 14 TABLE 6 OCCUPATIONAL COMPOSITION OF FOREIGN-ORIGIN AND GHANAIAN WORKERS IN 1960 (percentage of employed workers age 15 and over) Occupation (ranked Total, all workers Male workers Female workers by share) Foreign Ghanaian Foreign Ghanaian Foreign Ghanaian (1) (2) (3) (4) (5) (6) Clerical 0.8 1.9 0.9 3.0 0.3 0.3 Administrative, 1.0 0.4 1.2 0.7 0.2 0.0 managerial Professional, 2.0 2.4 2.2 3.3 1.1 1.2 technical Service 5.2 1.6 5.5 1.8 4.0 1.4 Sales 20.6 12.3 11.7 2.6 55.4 25.7 Production and 26.6 17.3 30.7 22.2 10.8 10.6 related Agriculture 43.8 64.1 47.8 66.5 28.2 60.8 Subtotal: clerical, admini- 3.8 4.7 4.3 7.0 1.6 1.6 strative and professional Subtotal: sales and 47.2 29.6 42.3 24.8 66.2 36.2 production Total: all occupations 100.0 100.0 100.0 100.0 100.0 100.0 Note: details may fail to add to totals because of rounding. Source: Steel 1977, Statistical Table A.7; and Ghana, Population Census, 1960 Vol. IV, Table 19. 15 Ghanaian women, not by Ghanaian men. This phenomenon may have occurred in occupations other than trading as well, since the foreign nationals most affected by the Order tended to hold menial jobs with low prestige in Ghanaian society, (as indicated by the relatively high percentage in sales and production shown in Table 6). The Order may thus have led to an outward shift in the supply curve of female workers -- an "encouraged worker effect" induced by the sudden availability of small, informal businesses particularly suitable for female employment in Ghana. A shift in the supply of female labor may have been reinforced in several other ways. Increased urbanization provided more opportunities for small-scale and part- time activities through which women could earn additional income without seeking full- time employment, and often without leaving the home. Stagnation of income per capita may have stimulated greater income-earning efforts by women in order to maintain the level or growth of family income. These changes would have less effect on male labor, given that over 93% of males in the age groups between 25 and 64 were already in the labor force and that part-time or informal employment is more likely to serve as a second rather than a primary income source for men. The increase in female labor supply appears to be concentrated in the child- bearing ages: 32% of both the labor force and the employment increase over 1960-70 came in the 25-34 age group, double the share for males (Table 7). Further research with migration and occupation data would be useful in assessing the extent to which women in this age group entered occupations abandoned by out-migrants and whether out-migrant males were concentrated in these ages. The rise in female labor force participation in these age groups, and increasing education of females aged 15-19, could have an impact on future population growth by encouraging decreased fertility. An indication that such an effect may already have begun is that the rate of growth of children age 0-4 is less than half that for the population in age groups between 5 and 9 (Table 5, lines E.1-3). Further study of demographic data to determine fertility rates in younger as against older female cohorts and comparison of their participation rates at comparable ages is needed in order to analyze the extent to which increased employment opportunities affect the birth rate. In any case, it is clear that a rapid rise in labor supply can be expected for some time to come. Given the outmigration for political reasons just before the 1970 Census, the Ghana-born population growth rate of 3.3% per annum is a more accurate estimate of population trends than the net rate of 2.7%. More important is the dramatic difference in population growzh in the age groups from 5-19 (over 4%) as compared to 20-44 (1.6% or less: Table 5, lines 1-9). These groups will be entering the labor force through 1980, maintaining a probable labor force growth rate above the population growth rate even without increasing participation rates. 16 TABLE 7 DISTRIBUTION OF EMPLOYMENT AND LABOR FORCE INCREASES 1960-70 BY AGE GROUP: MALES AND FMALES (percentage) Age category Additional employment Composition of la- Number Composition bor force increase Males Females Males Females Males Females (1) (2) (3) . (4) (5) (6) 15-24 -20,705 60,222 -13.8 14.2 5.2 15.9 25-34 23,670 134,963 15.8 31.9 15.8 31.7 35-44 33,026 96,379 22.0 22.7 17.8 22.2 45-54 45,237 68,527 30.2 16.2 24.4 15.7 55-64 29,857 32,315 19.9 7.6 16.0 7.3 65 and over 38,882 31,297 25.9 7.4 20.9 7.1 Total 149,963 423,701 100.0 100.0 100.0 100.0 Note: Details may fail to add to totals because of rounding. Source: Steel 1977, Statistical Tables C.2b and C.2c. - ,17 Underutilization of Labor Comparison of unemployment statistics to trends in employment and labor force can throw further light on the nature of changes in labor demand and supply over 1960-70. In spite of the relatively rapid increase in female labor supply and par- ticipation, female unemployment rose at only 0.5% per annum, as against 3.0% for males, and even declined at a substantial 3.3% rate for women age 25 and over (Table 5, lines D.1-3). The female open unemployment rate fell from 5.2% to 3.9% over the decade, exactly offset by the rise in the male rate from 6.5% to 7.6% so as to produce no change in the overall 6.0% rate (Table 8, lines l.a-b).12 A de- crease in unemployment in the face of supply growing faster than the population implies that opportunities for female employment must have been expanding rapidly, inducing more women to enter the market. Evidence on employment expansion, however, indicates that these opportunities were primarily in small-scale activities, especially in commerce, associated with urbanization and with the departure of foreign workers. Given the apparent increase in demand for female workers, the 3.0% average annual growth of male unemployment when the male labor force was growing at only 1.1% appears puzzling. The evidence suggests that the key lies in the behavior of male youths in the 15-24 age bracket. Unemployment in this group rose at 4.6% per annum, as against only 0.5% for males 25 and over (Table 5, lines D.2-3). The male unemployment rate in this group jumped from 15.9% to 22.7% over 1960-70, whereas the female rate fell slightly from 12.6% to 12.3% '(Table 8, lines 3.a-b). Increasing education held the male labor force growth down to practically nil for ages 15-24 (as against a 2.2% annual increase for females: line B.2), so that excessive growth of male labor supply in this age group is not an adequate explanation. In spite of a 0.6% average annual increase in employment demand for youths age 15-24, however, male employment in this category fell at the same rate (line C.2). The rapid growth of unemployment of males age 15-24 must have been attributable neither to excess supply nor to inadequate demand, but to voluntary unemployment while engaging in a job research. Increased education along with expansion of the modern sector may have raised the hopes of many young men of getting a high-paying modern sector job, thereby making it worthwhile to turn down low-paying employment in the informal or intermediate sectors while searching for a job that would pay many times that income. This be- havior does not show up clearly for women because of the bias of education and modern sector employment toward men, and because women may be more responsive to informal/ intermediate sector opportunities. !-2The 1971 Supplementary Enquiry used a one-week reference period, and showed an increase in female unemployment compared to the 1961 Post-Enumeration Survey, which used the one-month period (Table 8, lines 2,a-b). Both, however, show considerably lower unemployment rates than the corresponding Censuses, though the 1971 rate is much closer to the 1970 rate than the 1961 rate is to the 1960 rate (the discrepancy is mainly in the rural agricultural sector). 18 TABLE 8 UNDERUTILIZATION OF LABOR, 1960-71: TOTAL; MALES/FEMALES; URBAN/RURAL; NONAGRICULTURAL/AGRICULTURAL (percentage of labor force) Males Females Total Urban Rural Nonagri- Agricul- cultural tural (1) (2) (3) (4) (5) (6) (7) Unemployment rate Total country a. 1960 (Census) 6.5 5.2 6.0 n.a. n.a. n.a. n.a. b. 1970 (Census) 7.6 3.9 6.0 8.6 4.9 11.5 1.4 c. 1961 (P.E.S.)a 2.7 1.6 2.2 4.5 1.5 4.9 0.3 c. 1971 (Supp. Enq.)b 6.1b 3.6b 4.9b 8.2b 3.6b ll.6b 0.2b . 1970 Census a. Urban 9.5 7.3 8.6 8.6 --- 10.1 0.9 b. Rural 6.8 2.5 4.9 --- 4.9 13.1 1.4 . Ages 15-24 a. 1960 (Census) 15.9 12.6 14.4 n.a. n.a. n.a. n.a. b. 1970 (Census) 22.7 12.3 17.8 22.9 15.3 29.4 3.1 Underemoloyment rates Not working at allc 11.8 6.0 9.2 9.8 9.0 12.7 6.4 Working part-time only 15.3 26.3 20.2 7.8 25.4 8.9 29.5 . Not working full time 27.1 32.2 29.4 17.6 34.4 21.6 35.9 a. Males --- --- --- 15.4 32.2 19.3 33.1 b. Females --- --- --- 20.4 37.0 24.3 29.6 Post Enumeration Survey: Ghana, 1960 Population Census, Vol. IV, Table Gl. A one-week (rather than one-month) reference period was used in the 1971 Supplementary Enquiry: Ghana, 1970 Population Census, Vol. VI, Table H38 (worksheets). Unemployed or did not work on job. ource: Steel 1977, Statistical Tables C.1, C.2, E.1, and worksheets; also see footnotes a and b. 19 Unemployment rates reported in the Census are inadequate measures of labor underutilization, given its very restrictive definition. It defines unemployment to include only those persons who engaged in no employment (including self-employ- ment, part-time employment, and work on one's own farm) during any day of the pre- ceding month and who were actively seeking employment. In practice, few persons other than recent school-leavers with relatives to support them can afford the luxury of full-time job search with no efforts to earn any income. Indeed, 72% of the unemployment recorded in the 1970 Census was in the ages 15-24. Other measures must be used to indicate the actual level of labor underutilization. Measures available from the 1970 Census are those who had a job but did not work in it during the preceding month (this would presumably catch seasonal unemployment); and those who worked only part-time (23 days or less, many of whom might have pre- ferred to work more days). The total number of laborers who did not work was half again as much as the number reported as unemployed, representing a total of 9.2% of the work force (Table 8, line B.1).13 The share of the labor force not working at all was almost twice as high for males as for females and for nonagricultural as for agricultural occupations. Part-time work accounted for a substantial 20.2% of the labor force in 1970, and was disproportionately characteristic of female workers and of agricultural employment (line Lz.2). Altogether, 29.4% of the work force was not working full time, ranging from 15.4% of urban males to 37.0% of rural females (lines B.3.a-b). This is not a definitive measure of unemployment since, on the one hand, many part-time workers may not have desired additional work and, on the other hand, some full-time workers may have been underutilized in the sense either of "disguised unemployment" or of being trained for a more productive job. The underutilization rate of young workers would certainly be even higher than the rates given in Table 8, line B.3 if data were available. The seriousness of this underutilization and the appropriate policies for dealing with it depend, however, on the extent to which it represents voluntary avoidance of undesired or low-wage jobs rather than involuntary unemployment. If the former, the reduction of the wage gap between the modern and intermediate/informal sectors would be most helpful in dis- couraging youths from prolonging their job search. If the latter, then efforts to expand lower-level job opportunities for those without experience would be more im- portant. In either case, reorientation of education toward the types of opportunities most available, rather than toward civil service jobs, would be useful. More research on the attitudes and intentions of unemployed (or underemployed) youth is needed, along with projection of likely patterns of expansion of jobs outside the modern sector. 1- The percentage of the labor force with a job but not working was 5.0% for agricultural occupations but only 1.2% outside agricultural, indicating that the seasonal factor may be prominent in this measure (Steel 1977, Statistical Table E.3a). 20 CONTRIBUTION TO EMPLOYIENT OF LARGE AND SMALL SCALE SECTORS Employment absorption in large and small enterprises The extent to which modernization was an employment stimulus that absorbed new workers and encouraged increased labor force participation may be examined further by examining the absorption of increased labor supply into large-scale, secondary sector activities. Rising income per capita in countries at Ghana's level of develop- ment is associated with a shift in economic structure out of agriculture into secondary activities, and, within the industrial sector, away from consumer goods into invest- ment goods.14 The "modern" (large-scale) industrial sector would, in particular, be expected to show high relative labor absorption under a successful growth "take-off," especially when it is emphasized as heavily as it was in Ghana during the 1960s. Some growth in commerce and infrastructure would be associated with increased urbanization, although the service sector as a whole is not normally the major employer of labor released from agriculture until later in the development process. An excess supply of labor would most likely lead to an increase in small-scale service activities, as well as in unemployment and other forms of underutilization. The initial discussion of economic structure in Ghana (pp. 2-6) analyzed general trends in the structure of employment, and found them to be broadly consistent with the pattern described in the preceding paragraph. The expected, trends were, however, more characteristic of female than male employment, and the structure of large-scale employment shifted rather heavily toward the tertiary sector, with a decrease in the secondary as well as the primary share (Table 1). This section examines more closely the absorption of labor within the large and small scale sectors. The overall structure of new jobs does not indicate a strong shift toward the primary, secondary, or tertiary sectors, given the large shares of agriculture, manu- facturing, and services (36%, 31%, and 24%, respectively: Table 9, bottom, col. 1). A strong bias exists, however, within the large-scale sector: 92% of the net increase in large-scale employment was in services. Large-scale services accounted for 19% of the total net increase; all other large-scale sectors except manufacturing (5%) con- tributed less than 1% (col.2). New small-scale employment was also concentrated, though not quite as heavily: 80% of new small-scale jobs were provided in agriculture and manufacturing, which accounted for 38% and 26%, respectively, of the total net increase (col. 3), Within the small-scale sector, commerce was almost twice as im- portant a source of new jobs as was services. 14Simon Kuznets, "Quantitative Aspects of the Economic Growth of Nations: II. Industrial Distribution of National Product and Labor Force," Economic Develop- ment and Cultural Change 5, 4, Supplement (July 1957). TABLE 9 21 SECTORAL SOURCES OF EMPLOYMENT ABSORPTION 1960-70, LARGE AND SMALL SCALE ctor (ranked by All workers Males Females total increase) Total Large Small Large Small Large Small (1) (2) (3) (4) (5) (6) (7) Increase (decrease) in employment, thousands of workers nufacturing 179 29 150 26 38 3 112 rvices (adjusted) 136 106 30 92 20 16 8 mmerce 61 4 57 1 (-31) 2 88 ansport 16 1 15 1 15 1 * ilities 6 1 6 * 5 * * nstruction (-15) (-12,) (-3) (-12) (-4) (-*) ining (-17) (-4) (-13) (-4) (13) (-*) (-*) tal nonagricul- 366 125 241 105 31 22 209 ral (adjusted) riculture, etc.a 207 (-9) 216 (-13) 27 4 189 tal, all sectors 574 116 457 92 58 26 398 djusted) . .. .. . . . . \..I tal, large and all (adjusted) 574 150 424 Percentage of total net increase nufacturing 31.2 5.0 26.2 17.4 25.5 0.6 26.5 rvices (adjusted) 23.7 18.6 5.2 61.2 13.3 3.8 1.9 mmerce 10.6 0.6 10.0 0.8 -20.5 0.6 20.8 ansport 2.9 0.2 2.6 0.4 9.8 0.2 0.1 ilities 1.1 0.1 1.0 0.3 3.6 0.0 0.1 nstruction (-2.6) (-2.1) (-0.6) (-7.7) (-2.4) (-0.1) 0.1 ining (-3.0) (-0.7) (-2.3) (-2.5) (-8.8) (-0.0) (-0.1) tal, all sectors 63.8 21.8 42.1 69.9 20.5 5.1 49.3 djusted) riculture, etc.a 36.2 (-1.5) 37.7 (-8.7) 18.3 1.0 44.5 tal, all sectors 100.0 20.3 79.7 61.2 38.8 6.2 93.8 djusted) tal, large and all (adjusted) 100.0 100.0 100.0 ess than 500. a"Agriculture, etc." includes cocoa, forestry and fishing urce: Steel 1977, Statistical Table B.2. 22 The contribution of the large-scale secondary sector to new employment was quite small (3% of the total), with the decline in construction jobs offsetting the small increase in manufacturing and the negligible increase in utilities. The large-scale secondary sector had virtually no share in increasing job opportunities for women. It makes a somewhat larger contribution to the increase in male employ- ment, but new jobs for men were overwhelmingly in services (61% of the net increase). The increase in large-scale employment was predominantly a tertiary sector phenomenon, and it must be asked whether this represented a dynamic force leading the economy or simply a means of absorbing surplus labor. The secondary sector accounted for a third of the net increase in small-scale employment, and 27% of the total net increase (Table 9, col. 3). Small-scale secondary employment made the same 27% contribution to increased employment for both males and females, although it was relatively more dominant within the small-scale sector for men than for women (69% as against 28% of the net small-scale increase). Within the small-scale tertiary sector, commerce was almost twice as important a source of new jobs as was services, contrary to the negligible role of commerce in the large-scale sector. Major small and large scale detailed industries Examination of the relative importance of different industries in large-scale and small-scale employment is necessary as a basis for developing specific policies, as well as for more complete understanding of past employment absorption. One ob- jective is to identify industries in which small-scale enterprises are important, productive, and dynamic. Expanded employment in these activities can be promoted, if not directly, then at least by avoiding special incentives to large-scale firms in the same industries. At the same time, activities in which large-scale firms pre- dominate and are major employers should be identified as a focus for future efforts to "modernize" the economy while minimizing the adverse impact on employment. The major sources of employment outside the modern sector are agriculture and retail trade (Table 10). Small-scale employment accounts for an overwhelming pro- portion of the total in both industries (97% or more), and is growing no faster than the 2% overall rate of growth of small-scale employment; the small-scale share of total employment is rising slightly in both. Both have been major sources of ex- panded employment for women more than for men, and both may be suspected of offering easy entry to a source of income which nevertheless is unproductive in terms of the net contribution of additional workers to GNP. Developing policies regarding the role of agriculture in employment absorption is especially difficult. Given its position as the leading employer, the availability of uncultivated land, and the slow TABLE 10 23 MAJOR SMALL-SCALE DETAILED INDUSTRIES: SMALL-SCALE SHARE AND GROWTH 1960-70 de Industry 1970 Small- 1960-70 average emplov- scale annual growth (%) ment share Employ- S-S (%) ment share (1) (2) (3) (4) 1 Agriculture and livestock farming, agricultural ser- 1,672,657 97.7 1.5 0.1 - vices, and hunting 2 Retail trade 404,887 97.0 2.2 0.3 32 Wearing apparel 92,508 98.5 2.2 -0.1 18 Local food preparation 68,020 100.0 34.0b 0.0b 3 Fishing 60,676 95.8 0.8 -0.4 la Road Transport 46,775 93.0 2.1 0.4 31 Education 32,653 41.5 11.4 -0.3 55/59 Other personal services 24,907 97.3 -0.8 -0.1 Construction 23,571 32.0 -1.2 0.6 44 Furniture 23,242 96.0 -1.1 0.0 31 Textiles 21,886 80.7 6.1 -1.9 16 Bakery products 18,925 97.1 -0.1 0.3 14 Oils and fats 18,537b 97.1b 18.3b 0.3b 21 Distilling 14,627 90.8 45.5b 0.5b 53 Vehicle repair 13,432 72.4 7.7 1.1 34 Business, professional 13,322 91.3 10.9 1.8 and trade associations -71 Structural clay products 10,744 97.1b 3.1b -0.1b otal, industrial 367,810 82.5 5.9 0.1 otal, nonagriculturale 946,097 70.3 3.4 -0.2 otal, all industriese 2,683,898 85.7 2.1 -0.1 REsidual of Census minus Labour Statistics, except as noted. Based on Industrial Statistics rather than Labour Statistics. Figures based on Industrial Statistics show a decline at -0.3% per annum. Figures based on Industrial Statistics show a decline at -0.1% per annum. Adjusted to show public administration as 0% small-scale. ource: Steel 1977, Statistical Tables B.6 and B.7. Industries with 10,000 or more employees n 1970. 24 growth of agricultural output in Ghana, it must not be neglected (the fact that local food prices led inflation in the 1960s suggests that it has been). On the other hand, rising agricultural productivity is important to the industrialization process, and this implies limited employment growth (perhaps even employment declinR). Government policy in the 1970s has explicitly discouraged growth of small-scale trading, through price control measures that make trading unprofitable or risky and through direct intervention in distribution in an artempt to eliminate the price mark-ups of "middlemen." Neither agriculture nor trading can be effective sources of new productive employment absorption, but both are likely to remain important as income-earning opportunities for surplus labor. In spite of Ghana's emphasis on large-scale industrialization, manufacturing industries are more prominent in small-scale than in large-scale employment: of the seventeen largest sources of employment in each sector, eight of the small-scale industries are in manufacturing, but only two of the large-scale industries (Tables 10 and 11). Given the evidence that small-scale manufacturing is as productive in using capital as is large-scale manufacturing while absorbing more labor (Steel 1977, Ch. 4), industrialization policies should avoid giving incentives to large- scale production that would compete with major, expanding small-scale industries. Three small-scale industries that have apparently already suffered from such compe- tition are: wearing apparel, textiles, and structural clay products (Table 10). All three grew at or above the overall rate of small-scale employment growth over 1960 to 1970; small-scale firms provide the major share of total employment in all threel5; and yet the small-scale share of total employment has fallen in all three. Failure of the small-scale wearing apparel industry to grow would be particularly disastrous for employment because on the order of half the present workers are apprentices (Steel 1977a: 192), who are learning skills in expectation of becoming productively employed (especially self-employed) in the future. Competition between large and small scale sectors is most apparent in textiles, in that it is one of the major sources of employment on both lists. Vehicle repair also appears on both lists, although small-scale growth has been fast enough to increase the small-scale share of total employment in the industry. Distilling and processing of vegetable and animal oils and fats are other major industries in which small-scale employment growth has been sufficiently rapid to maintain the high small-scale share of total employment, but which are vulnerable to large-scale competition. These industries are ones in which incentives to large-scale production must be avoided if the substantial employment IIndustrial Statistics figures indicate that the small-scale share in textiles may have fallen to 58.5% in 1970, although the proportion based on Labour Statistics remains at 80.7%. TABLE 11 25 MAJOR LARGE-SCALE DETAILED INDUSTRIES: SMALL-SCALE SHARE AND GROWTH 196.0-70 de Industry 1970 Small- 1960-70 average emplov- scale annual growth (%) ment share Employ- S-S ment share (1) (2) (3) (4) 1 Public administration 110,962 0.0 16.8 n.a. Construction 49,993 32.0 -1.9 0.6 31 Education 46,006 41.5 12.6 -0.3 1 Agriculture and livestock farming, agricultural ser- 39,668 97.2 -0.6 0.1 vices, hunting Miningc 25,248 18.5 -1.3 -5.4 33 Medical and other health services16,017 27.4 5.1 -2.1 services 1 Wholesale trade 14,353 93.6d 6.8 n.a. -lb Railway transport 13,886 n.a. 3.1 n.a. -2 Retail trade 13,425 97.0 -3.0 0.3 -41/43 Sawmills and wood products 11,689 42.7 2.3e .1, Insurance, real estate and 10,324 6.7 3.1 -8.5 business services -5 Communications 9,785 n.a. 3.1 n.a. -32,35 Research institutes, libraries, 9,223 22.7 18.9 n.a. etc. -2 Forestry 6,629 40.3 -4.7 3.3 a Electric power 5,652 13.3 4.4 -7.6 -31 Textiles 5,2340 80.70 398.90 1.90 -53 Vehicle repair 5,122 72.4 2.6 1.1 h otal, industrial 367,810 82.5 5.9 0.1h otal, nonagriculturalb 946,097 70.3 3.4 -0.2 otal, all industriesb 2,683,898 85.7 2.1 -0.1 Labour Statistics figures, except as noted. Adjusted on the assumption that public administration is 0% small-scale. Predominantly metal mining. 1960 figure. Industrial Statistics data show a -1.4% annual decrease in employment over 1962-70 and a 39.0% annual increase in the samll-scale share Based on Industrial Statistics figures. Industrial Statistics figures show 1970 employment of 11,255, small-scale share of 58.5%, and 1962 0 annual growth of 855.9% in employment and -5.1% in small-scale share Industrial Statistics figures show a decline at 0.1% per annum Source: Steel 1977, StatisticalTables B.6 and B.7. Industries with 5,000 or more employees in 197 26 in small-scale activities is to expand rapidly as a major source of productive labor absorption. This does not mean that small-scale production in these industries should be subsidized, only that there is sufficient evidence of the economic efficiency of small-scale enterprises in them that they should be allowed to compete on an equal basis with large-scale firms. SUMIMARY AND CONCLUSIONS 27 Synopsis of Basic Facts Presented in Each Section Sectoral emoloyment shifts: The overall employment structure shifted from primary into secondary and tertiary sectors. Modern sector: the secondary sector share declined; only the tertiary sector share rose (to 57% of total), due to doubling of government share; Little change in employment structure occurred outside modern sector (except slight rise in manufacturing share). Sectoral sources of employment increases Overall employment growth was 2.2% per annum (1.3% agriculture; 3.7% nonagriculture). Agriculture was the largest source of overall employment increase, but absorbed only 10% of the male increase (46% of females); Services was the major source of new employment for males (75% of increase); Manufacturing was the second largest absorber for males, females, and overall; Commerce: the decline in male employment was more than offset by female increase. Services and manufacturing rose from 46% to 66% of male employment; Commerce (falling) and manufacturing (rising) continued to provide at least 90% of female employment. Detailed industry sources of employment increases Food production, retail trade, and cocoa accounted for 49% of new employment, but grew at 1.8% per annum or less; Manufacturing industries were-not major absorbers (except for local fooc preparation, which is partly a service activity); the biggest increases were in industries related to food, beverages, and clothing; Public administration, education, research institutes, and community services tgether accounted for 21% of the total employment increase, and were growing relatively rapidly. All non-tertiary industries (except textiles) that provided over 10,000 new jobs were over 90% dominated by small-scale employment. Only three industries were in the top ten sources of new employment for both men and women (public administration, educatioh, and food production); male employment increases came predominantly from tertiary industries, female increases from primary activities Occupational composition and sources of increase Agricultural and production occupations dominated employment increases overall and for females; professional/technical occupations provided the most new male jobs, with clerical and service occupations close behind agricultural work. Professional/technical, service, and clerical occupations had the highest shares of employment in the modern sector. Sales was the only occupation that was more than 50% female overall; it had the highest female share (25%) in the modern sector. Production occupations had a relatively large modern sector component and accounted for substantial shares of both male and female workers (though women represent only 3% of modern sector production workers). 28 Changes in population and labor supuly, 1960-70 Female population growth (3.0% per annum) exceeded male population growth (2.5%) mainly for ages 25 and over, apparently because out-migration after the 1960 Aliens Compliance Order was predominantly male and because health improvements increased female longevity more than male longevity; the growth rate for Ghana- born males was 3.2%, for females 3.4%. Male labor supply grew more slowly (1.1%) than male population age 15 and over, and male participation rates fell (overall and for ages 15-29); female labor supply grew more rapidly (4.1%) than the corresponding population, and female partici- pation rates rose (overall and for all age groups except 15-19). Out-migration of the foreign-born population (presumably in response to the 1969 Aliens Compliance Order) accounted for a decrease of 7.5% of the 1960 male population and 3.7% of the 1960 female population; the reduction in labor force may have exceeded the net observed decrease of 135,568 male foreign nationals and 67,414 females (the excess of male over female out-migration partially accounted for the difference in male and female labor force growth). Both male and female out-migrants who had been in sales occupations were apparently replaced by Ghanaian females rather than males. Continuing rapid labor supply growth is indicated by the Ghana-born population growth rate of 3.3% and the more than 4% growth of population in age groups from 5-19. Underutilization of labor Although aggregate open unemployment remained at 6.0% over 1960-70, the female rate fell from 5.2% to 3.9% whereas the male rate rose from 6.5% to 7.6%; the number of females age 25 and over who were unemployed actually fell at 3.3% per annum. The increase in male unemployment was concentrated in the 15-24 age group, in which unemployment rose from 15.9% to 22.7% (as against a slight decline in the corresponding female rate from 12.6% to 12.3%). In addition to the 6.0% open unemployment in 1970 (72% of which was in the 15-24 age group), 3.2% of the labor force had a job but did not work (predominantly an agricultural phenomenon), and 20.2% worked only part-time (especially females and agricultural workers); a total of 29.4% of the work force was not working full time. Employment absorption in large and small enterprises New employment overall was fairly evenly distributed across agriculture (36%), manu-- facturing (31%), and services (24%); within the modern, large-scale sector, however, 92% of the net increase was in services; outside the modern sector, 80% of new jobs were in agriculture and manufacturing. The large-scale secondary sector provided only 3% of the net employment increase, whereas the small-scale secondary sector provided 27%. Agriculture, manufacturing, and commerce (for females) were much more important sources of increased employment within the small-scale sector than within the large- scale employment increase. S99 Major small and large scale detailed industries Farming and retail trade were the major sources of employment outside the modern sector (especially for women), and were heavily dominated (97%) by small- scale employment; small-scale employment in these industries was growing at only 1.5% and 2.2% per annum, respectively. Manufacturing accounted for 8 of the 17 largest small-scale industries, but only 2 of the to7 17 large-scale industries. Wearing apparel, textiles, and structural clay products were major, growing small- scale industries in which the large-scale sector took an increasing share of the market; vehicle repair, distilling, and processing of oils and fats are especially dynamic small-scale industries which might be threatened by rapid growth of large-scale competition. Analytical and policy conclusions The structure of employment in Ghana did "modernize" during the 1960s in the sense of shifting from the primary to the secondary and tertiary sectors. The shift was led primarily by manufacturing, government administration, and education. Sig- nificantly, however,-the "modern" sector was not responsible for the growing share of secondary employment; tertiary employment in large-scale enterprises rose at the expense of both primary and secondary employment, and reached 57% of total modern sector employment. This bypass of an employment shift into the industrial sector is disturbing not so much because it fails to replicate the "normal" pattern of Western development, but because it offers so little opportunity for future produc- tivity and employment growth. Government services provide relatively little opportun- ity for increasing production per worker, and their expansion is constrained by a limited ability to increase Government revenues, especially in a stagnant economy.16 The fact that all non-tertiary industries that provided over 10,000 new jobs during the period were heavily dominated by small-scale enterprises suggests a limited ability of the modern sector to lead net employment increases in more directly pro- ductive activities. Rapid expansion of large-scale manufacturing and other capital- intensive secondary activities is unlikely to absorb as many new workers as it dis- places in small-scale enterprises. The agricultural sector presents another employment dilemma. It was the principal source of new employment over the decade and was especially important in absorbing the rapid increase in female labor supply. Nevertheless, it absorbed only 10% of the male increase, and agricultural productivity failed to rise fast enough to meet the growth of demand for food (as reflected in the more rapid increase in prices of local foods than of other categories). Traditional lack of government 15The extremely rapid increase in government borrowing (and consequently in money supply and prices) in recent years is one indication of an inability to in- crease tax revenues. 30 assistance and concern for female farmers may have inhibited the growth of agricul- tural productivity as farming became increasingly a female occupation. Greater attention to raising agricultural productivity in the future, however, will tend to diminish its potential for absorbing labor and make it all the more important to provide productive employment opportunities elsewhere. Within agriculture, it will be critical to bring the remaining surplus and irrigable land into production using yield-raising techniques that are labor-using rather than capital-intensive. Only in this way can agriculture continue to provide the necessary foundation for employ- ment absorption while still increasing productivity and meeting food needs. The low growth of male employment in agriculture relative to the rate of natural population increase implies that men were shifting out of agricultural employment. In addition, the absolute numbers in commerce, mining, and construction declined. To some extent, special factors were responsible for these absolute declines: the Alien's Compliance Order, stagnation of mining in general, and the end of major public works building. Attractive income-earning opportunities in the modern manufacturing and public sectors was a more general stimulus to the shift in male employment structure. Employment in the modern sector remained heavily biased toward male workers. The high and increasing open unemployment rate among male youths indicates their cognizance of the benefits of waiting for these jobs (in contrast, the unemployment rate for female youths remained stable, in spite of much faster growth in female than male labor supply). The evidence is not sufficient to indicate whether the government expanded its employment principally for the purpose of absorbing labor laid off or disinter- ested in other sectors, or whether independent expansion of government services simply accelerated the flow of workers seeking the relatively high wage in this sector. The evidence does indicate that men were relatively unwilling to take jobs outside the modern sector: only 59% of the net increase in male nonagricultural employment was in small-scale activities, as against 92% of the female increase. The rise in the male unemployment rate, decreasing male labor force participation, and the slow growth of male labor supply imply that men were increasingly remaining unemployed or dropping out of the labor force when unable to obtain a modern sector job. The willingness of women to take work in the faster-growing informal and inter- mediate sectors helps to explain their falling unemployment rate, increasing labor force participation, and relatively rapid labor supply growth. Female labor supply shifted outward as a result of increased employment opportunities in small-scale activities (especially with increasing urbanization and the departure of non-Ghanaian traders under the Aliens Compliance Order), and due to the high proportion of males among the out-migrant population. The rising costs of raising children associated 31 with increased urbanization and expanded education, along with stagnant real income per capita, provided further stimulus to the trend of rising female labor force participation, as well as to the complementary trend of decreasing fertility. The rapid increases in costs of living in the mid-seventies make it likely that these trends will continue. Research is needed to identify the most productive activities that tend to absorb increases in female labor, as well as the policies most favorable to their expansion, if future increases in labor supply are to be employed productively rather than spill over into unproductive "informal" employment and underutilization of labor. (Much of the past expansion of female employment probably fell into this latter category, althouglh little direct evidence exists.) Designing a strategy for modern sector employment expansion faces difficult problems in addition to that of identifying industries whose productivity and ability to expand are not limited and which will not compete with productive intermediate .sect,r employment. Another question is how to increase female labor productivity and maximize utilization of the potential contribution of women to output growth through greater employment of women in the modern sector. Those occupations in which the female share of modern sector employment is highest are also those that have been the major absorbers of male labor. Thus, expansion of the occupations most favorable to female participation in the modern sector will nevertheless dis- proportionately benefit men. The expansion of production work has been relatively unbiased overall in terms of benefiting male and female employment proportionately, but women constitute only 3% of modern sector production occupations. Emphasis on expansion of modern sector job opportunities almost inevitabl7 implies male-biased employment growth. Given a corresponding bias in education and training, however, failure to provide such expansion may forego productive utilization of the fruits of education. A more suitable measure of labor utilization other than the Western definition of open unemployment is needed. The percentage of the labor force actively seeking work while engaging in no income-earning activitie-' all is not especially high, nor did it increase over the 1960s. The absence ci mp;iloyment compensation means that only those with independent means of support an a. ford this luxury. For the most part, this means recent school graduates without families who are living with relatives while searching out the best employment opportunities. The high rate of unemployment among this group must, in this sense, be considered voluntary and temporary. Open unemployment, however, does not include laid-off and seasonally unemployed workers, and is small relative to the underutilization represented by part-time work (assuming that many part-time workers would prefer full-time employ- ment). Employment policies should focus not so much on the openly unemployed but on those in part-time and low-productivity occupations. P, ,32 Although manufacturing was one of the fastest-growing sectors in Ghana's economy during the 1960s, large-scale industrialization cannot be considered a leading sector for employment: it provided only 3% of new jobs. As previously indicated, the increase in large-scale employment was predominantly in the tertiary sector. The small-scale secondary sector, however, accounted for 27% of the total employment increase, and provides a hopeful basis for productive labor absorption in the future. Major, dynamic small-scale manufacturing and repair industries in- clude wearing apparel, textiles, oils and fats, distilling, vehicle repair, and structural clay products. If these are to continue to grow, however, the government must avoid giving special incentives and benefits to their large-scale competitors, which are likely to absorb less labor without using capital any more productively. The two largest small-scale industries, however, do not offer much hope for rapid labor absorption in the future, though their expansion is important in providing a base for adequate overall expansion. Absorption in small-scale farming is limited by the need to raise agricultural productivity, as well as by the continuing tendency of labor to move out of agriculture to more attractive urban opportunities.17 The growth of small-scale retail trade has been actively discouraged by the government through condemnation of the role of "middlemen," public take-over of certain phases of the distribution system, and other measures that tend to discourage trading. Directions for Future Research In general, research needs to be aimed at identifying productive employment activities that are relatively large and are capable of expanding rapidly. The data presented here have suggested an hypothesis that could bear further testing: that modern sector tertiary employment is unlikely to continue past rapid growth because of b,udgetary constraints and that it is a poor source of rising productivity. Qutside the modern sector, research needs to focus on separating relatively productive intermediate sector activities from informal employment in which additional workers provide little or no net gain in output. Identification of productive manufacturing and other small-scale employment opportunities for women is particularly important, since the trends underlying past increases in female labor supply and participation seem likely to continue. (More precise estimation of the determinants of female labor force participation in Ghana is itself an important research topic.) The wearing apparel industry is especially critical because so many women are employed in it and because the large number of apprentices threatens to overexpand the supply of seamstresses. EAlthough the urban-rural migration flow in Ghana matched the number of rural- urban migrants in 1970, the latter flow was much higher for young labor force entrants (Steel 1977, Statistical Table F.1, cols. 7-8). Urban-rural migration as recorded may largely reflect retirement to one's village after leaving the labor force. 33 Women appear to have been replacing men in commerce and to some extent in agriculture. Reasons for and implications of these shifts are appropriate topics for research, especially in relation to altering agricultural extension policies to pay more attention to female farmers. In commerce and other sectors with dispro- portionate numbers of immigrant workers, investigation is needed into the extent to which out-migrants after the Alien Compliance Order (especially men) have returned to their former jobs. Although more information on the impact of this Order on the size of the foreign-origin population in 1970 would be very useful, such data are probably unobtainable. The 1980 foreign population will have to be compared to that in 1960 rather than 1970 to give accurate estimates of long-term trends. The relationship between demographic and economic change is of special interest and importance. One hypothesis advanced here is that expansion of education, rapid inflation, and slow growth of real output tended both to increase female labor force particpation and to reduce fertility. In particular, the question may be asked: to what extent is increasing female labor force participation a cause, result, or con- comitant of fertility decline? Expansion of employment opportunities for women may be a critical motivator for a birth control program. A further demographic hypothesis to be tested is that health, medical, and other improvements benefited women relatively more than men in terms of increased life expectancy. Measurement of underutilization of labor presents one of the most difficult problems of analyzing the employment problem in Ghana. The Census definition of open unemployment is clearly inadequate. The data on part-time workers are not sufficient to determine what proportion would prefer full-time employment. No basis exists for estimating the share of employment outside the modern ,ector that could strictly be defined as "informal,"or having negligible marginal product of labor. Sample surveys of workers in these groups would be useful in obtaining a more precise picture of underutilization of labor in Ghana. In the long run, more useful Census definitions should be adopted (the Supplementary Enquiry uses a one- week reference period, and contains helpful information on secondary occupations as well as days worked). A final important area for primary research is wages and incomes outside the modern sector. Relative male and female wages between sectors and occupations would help explain differences between the patterns of change in employment structure for men and women. The importance of wage differentials in determining rural-urban migration and duration of unemployment could be better understood. And finally, it would provide information on the income distribution implications of Ghana's employ- ment structure and changes in it. 34 REFERENCES ADDO, Nelson 0. 1972. "Employment and Labour Supply on Ghana's Cocoa Farms in the Pre- and Post-Aliens Compliance Order Era." Economic Bulletin of Ghana, Second Series 2, 4: 33-50. GHANA, Republic of. Industrial Statistics: 1962-64; 1970 (worksheets). Accra: Central Bureau of Statistics. . Labour Statistics: 1960; 1970. Accra: Central Bureau of Statistics. . Population Census of Ghana: 1960 Vols. II, IV, and VI; 1970 Vols. III, IV (worksheets), and VI (worksheets). Accra: Census Office. STEEL, William F. 1977. 'Employment, Unemployment, Labor Force Participation, Migration, and Size of Firm: Trends and Composition in Ghana, 1960-70; Statistical Tables Prepared for the International Bank for Reconstruction and Development." Mimeo. . 1977a. Small-Scale Employment and Production in Developing Countries: Evidence from Ghana. New York: Praeger Publishers. and CAMP'BELL, Claudia. 1977. "The Impact of Modernization on Women's Employment in Ghana, 1960-70." Mimeo.