Productive transformation, employment and education in Tanzania

37 downloads 1396 Views 1MB Size Report
Jun 24, 2013 - agriculture to industrial and service sectors, from low to higher .... Numbers of students in technical and vocational education and training.
T F A

R D

This paper is a draft submission to the

L2C - Learning to Compete: Industrial Development and Policy in Africa 24-25 June 2013

Helsinki, Finland

This is a draft version of a conference paper submitted for presentation at UNU-WIDER’s conference, held in Helsinki on 24-25 June 2013. This is not a formal publication of UNU-WIDER and may reflect work-in-progress. THIS DRAFT IS NOT TO BE CITED, QUOTED OR ATTRIBUTED WITHOUT PERMISSION FROM AUTHOR(S).

Productive transformation, employment and education in Tanzania Theo Sparreboom* Research and Statistics Department, International Labour Office, Geneva

Irmgard Nübler Research Department, International Labour Office, Geneva

June 2013

Paper prepared for L2C – Learning to Compete: Industrial Development and Policy in Africa 2013 UNU-WIDER Development Conference 24-25 June in Helsinki, Finland.

Abstract This paper empirically examines the role of education in facilitating structural change, increases in productivity and the creation of decent work. Major change in the economic and employment structure in Tanzania from 2001 to 2006 was only to a limited extent translated into higher productivity and the creation of decent work, while the pace of increase in educational attainment was barely sufficient to keep up with this change. Analysis of qualifications of workers indicates high levels of skill mismatch, and both mismatch analysis and rate of return analysis point in particular at demand for workers with secondary education. We conclude that a slow pace of increase in educational attainment may be a major factor limiting the diffusion of decent work.

Key words: employment, education, industrialization JEL Classification: I25, J24, O14 *Corresponding author, email: [email protected]; responsibility for opinions expressed in this paper rests solely with the authors, and does not constitute an endorsement by the ILO.

1. Introduction

Structural change contributes to economic development by enhancing productivity and creating more and better jobs. As economies grow and develop, activities shift from agriculture to industrial and service sectors, from low to higher complexity technologies and from low to higher value added activities. A wide and growing body of literature provides evidence for the important role of structural change in the economy in enhancing productivity and employment.1 Evidence shows that productive transformation is at the heart of economic development, and that some countries followed patterns and paths of structural and technological change that led to high and sustained growth in productivity and jobs while other countries could not achieve such processes. McMillan and Rodrik identified productivity-reducing patterns of structural change in both Africa and Latin America since the 1990s, while they found productivity enhancing patterns of structural change in in Asia. Productivity-reducing effects in Latin America and Africa were attributed to large labour productivity gaps between the formal (modern) and the informal economy and a shift of labour from modern to informal economies (McMillan and Rodrik, 2011). Structural change is also considered as instrumental in the creation of decent work (ILO 2013a; Nübler 2013a). Higher value added economic activities and more complex technologies with higher productivity levels allow for better conditions of work, better jobs, and higher wages. But some (parts of) sectors contribute more to decent work than others, and they do so in different ways. Experience from many developing countries shows that growth in real GDP was only to a limited extent reflected in the creation of decent, wage-paying jobs (ILO, 2013a). High rates of economic growth for more than a decade hardly reduced the share of vulnerable employment in sub-Saharan Africa, which stood at 77.5 per cent in 2011 (81.8 per cent in 2000). This underlines the need for a better understanding of the patterns of structural transformation and their contribution to increased productivity and creation of productive and good quality jobs. Education and skills are intrinsically linked to these processes. Structural change into higher value added sectors, upgrading of technologies in existing sectors and the development of non-traditional sectors applying more complex production technologies and research and development activities increase demand for higher levels of education and skills. In addition to enhancing efficiency and productivity, education and skills also play an important role in facilitating innovation and technological progress. Educational attainment levels and the particular mix of skills, knowledge and competences in the labour force determine the dynamics and pace of structural transformation (Nübler 2013). Education and skills training themselves, however, do not create (decent) jobs, and an increase in education attainment levels may also result in unemployment, overqualification and the underutilization of skills. It is not uncommon to find high rates of unemployment among the better-educated in developing countries (Fares et al., 2005; ILO, 2013b). A study by the Asian Development Bank concluded that, if historical trends 1

Hausmann et al. (2011); Imbs and Wacziarg (2003); Klinger and Lederman (2004); Kucera and Roncolato (2012); Kucera and Tejani (2013); McMillan and Rodrik (2011); Nübler (2013); Ocampo et al. (2009).

2

in structural change and educational attainment are taken as benchmarks, some countries were raising educational attainment levels too fast, and increases in educational attainment of the employed were not necessarily driven by productivity imperatives (ADB, 2007). This paper is exploring the link between structural transformation in the economy and in employment on the one hand, and the change in educational attainment on the other. It aims at better understanding how different patterns of structural transformation in the economy and changes in educational intensity and skills profiles of jobs are related to productivity and the quality of jobs created in the economy. The paper analysis empirical data based on two nationally representative labour force surveys in Tanzania conducted in 2001 and 2006, respectively. The analysis identifies those (parts of) sectors in Tanzania that reflect most of the increase in educational intensity and therefore contributed to value added growth, and those sectors in which education intensity decreased. In addition, it analyses the extent to which this has been achieved by the reallocation of resources between sectors, that is, the higher growth of education intensive sectors relative to less education intensive sectors (‘between sector effect’) or by increasing the share of education intensive jobs within the sector (‘within sector effect’). Additional insights regarding the role of education in structural change can be gained from an analysis of changes in the occupational distribution. Structural change is reflected in the occupational distribution of the employed, which is linked to levels of education and skills. We therefore analyse changes in the occupational distribution in Tanzania, in order to identify patterns of undereducation and overeducation. In a next step, the paper analyses the different levels of education that are in demand in different sectors. It calculates (sectoral) private rates of return to education to identify the relative demand for different levels of education. We demonstrate that the major major structural change in the economic and employment structure in Tanzania from 2001 to 2006 was only to a limited extent translated into higher productivity and the creation of decent work. Much of the sectoral change in employment involved shifts from low-quality employment in agriculture to almost equally low-quality employment in other sectors. At the same time, the pace of increase in educational attainment in Tanzania was barely sufficient to keep up with the structural change in the economy and the increasing demand for higher levels of educational attainment. The analysis of qualifications of workers indicates high levels of skill mismatch, which is principally due to underqualification. Mismatch is decreasing only slowly, and is therefore likely to limit the scope for technological advancement, productivity increases and productive transformation. Finally, rate of return analysis shows increasing returns to secondary education over time, while returns to primary education are decreasing. Increasing returns to secondary education appear consistent with the analysis of mismatch, which points in particular at demand for workers with secondary education. We conclude that low levels of educational attainment constrain the scope for productivity increases and the creation of decent work as it limits structural change into higher value added sectors, and technological upgrading in existing sectors (which 3

together define productive transformation). A slow pace of increase in educational attainment may therefore be a major factor limiting the diffusion of decent work. This paper is organized as follows. Section 2 summarizes educational strategies and provision in Tanzania. Section 3 describes economic growth, structural change and labour market outcomes. Section 4 provides an empirical analysis of structural change, education intensity and the creation of decent work, complemented by an analysis of measures of skills mismatch and rates of return. Section 5 concludes.

4

2. Education policies, outcomes and issues 2.1

Policies and outcomes

Following a period in which economic crises took center stage, the Tanzanian government recommitted itself to the development of the education sector in the 1990s. A cornerstone of education policies was the 1995 Education and Training Policy (ETP - MOEC, 1995). The ETP aimed to realize universal primary education, to eradicate illiteracy and to increase enrolment in tertiary education and vocational training in order to create high quality human capital resources. Priorities included spending on basic education and more equitable access to quality secondary education; postsecondary and higher education should be market-oriented and demand-driven (United Republic of Tanzania, 2001). Strategies for growth and reduction of poverty adopted since 2001 also reflect strong commitment to development of the education sector and include education objectives and targets, such as raising primary school enrollment as well as the transition rate from primary to secondary school, and reducing drop out. In a reversal of past policies, primary school fees were abolished starting in 2001/2002 (United Republic of Tanzania, 2000 and 2005). The 2010 strategy mentions universal access to pre-primary and primary education, but also to lower secondary education (MOFEA, 2010). The emphasis of education policies on primary education is visible in the distribution of students by level of education: in 2009, 8.1 per cent of all students were enrolled in preprimary education, 75.9 per cent in primary education, 13.6 per cent in secondary education, 1.1 in technical and vocational training and 1.0 per cent in tertiary education; teacher training accounted for 0.3 per cent.2 The number of enrolled students in primary education increased from 4.4 million in 2000 to 8.4 million in 2009, and during the same period the net enrolment rate increased from 58.8 per cent to 95.9 per cent (MOEVT, 2012). Enrolment in secondary education also increased strongly and reached 1.5 million in 2009. Numbers of students in technical and vocational education and training (TVET; 123 thousand students) and in higher education (115 thousand students) are low in comparison with enrolment in primary and secondary education (annex table A1). Although school enrolment has increased at all levels, enrolment in upper secondary and tertiary education remains low in comparison with other African low-income countries. According to MOEVT (2010), the gross enrolment rate was 0.22 per cent in 2001/2002 and 3.0 per cent in 2008/2009, which is still well below the average in subSaharan Africa.3

2 The distribution excludes adult education/non-formal education. 3 World Bank (2012a) shows a gross enrollment rate of 6.3 per cent in tertiary education in sub-Saharan

Africa in 2009, and 2.1 per cent in Tanzania in 2010.

5

2.2

Issues in education

According to UNESCO’s education sector analysis, government spending on education in Tanzania has risen above the average for countries with a similar level of development. The high priority given to education has helped putting the country on track to achieve universal primary education, to increase enrolment at all levels and in particular in higher education. The analysis also concludes that higher education and TVET sectors are well positioned to diversify supply and respond to labour market needs (UNESCO, 2012a). An assessment in 2008 led to the conclusion that there were many areas of the ETP that had not been implemented properly, due to financial constraints, weak implementation structures and non-linkage among the education sub sectors (MOEVT, 2008). Furthermore, a number of policy areas that needed immediate attention were identified, such as curriculum improvement, the involvement of the informal sector in education and training and the provision of technical education and entrepreneurial skills. The newly formulated objectives in the 2008 revision of the ETP included the improvement of the quality of education and training at all levels. In the second Primary Education Development Program (PEDP) it was acknowledged that, due to the expansion in enrolment, insufficient attention had been paid to quality improvement, capacity building and institutional arrangements (MOEVT, 2006). PEDP II lists a number of specific challenges regarding the quality of education, including inadequate in-service teacher training and lack of qualified teachers in primary education. In the years following the launch of the first PEDP, the pupil-teacher ratio worsened from 1:46 in 2001 to 1:58 in 2004 and there were individual schools that reported a ratio of 1:200 (Wedgewood, 2007). Issues in secondary education concerned the way the curriculum is offered, focusing too much on reproducing facts and too little on creative, critical thought, while human and material resources did not keep up with the rapid expansion in enrolment. The National Higher Education Policy (MOEVT, 1999) identified a number of problems in higher education, including low student enrolment; unregulated, uncontrolled proliferation of tertiary training institutions; imbalance in science relative to liberal arts, while technological development was needed in order to meet targets of the development agenda; gender imbalance; inadequate financing and material facilitation; and a tendency to distort the real worth of academic programs that are offered. Gender disparities characterize much of the education system, with the exception of primary education, and in particular higher education (UNESCO, 2012a). Urban-rural disparities are also substantial, especially at higher levels of education. To conclude, the rapid expansion of the educational system in Tanzania during the 2000s increased the share of the population with education levels above primary education, which is reflected in the educational profile of the labour force. Nevertheless, the proportion of the labour force with levels of educational attainment above primary education remained low at 5.9 per cent in 2006. Hence, secondary and tertiary education graduates are still very scarce in the labour force which is in accordance with an L-shaped educational structure characterized by a median category of primary

6

education (Nübler, 2013). In addition, increases in enrolment resulted in strong pressures on the quality of education.

7

3. Growth and structural change Tanzania experienced significant growth and structural change between 2001 and 2006. Average annual growth of value added was 7.2 per cent between 2001 and 2006, which was led by industrial and service sectors. Growth of industrial value added averaged 10 per cent annually, compared with 7.8 per cent in services and 4.5 per cent in agriculture. During the same period, employment was growing significantly in industry and in services (15.4 per cent and 10.9 per cent annually, respectively) and only slightly in agriculture (1.6 per cent). Employment growth thus exceeded growth in value added in services and particularly in industry. This is reflected in a declining level of labour productivity in these two (broad) sectors (figure 1). [Figure 1 around here] As a result of the reallocation of workers, the sectoral distribution of employment changed considerably between 2001 and 2006. The share of the labour force in agriculture dropped by 7.5 percentage points, reflecting very strong structural change. Services absorbed most of the relative decline of employment in agriculture as its share of the labour force increased by 5.8 percentage points, but industry also gained 2 percentage points. The share of the labour force employed in manufacturing nearly doubled in the five year’s period. The share of the unemployed (the unemployment rate) decreased from 5.2 to 4.9 percent (see table 1 in sub-section 4.2.2, first three columns). Evidence also shows that the quality of employment, that is, decent work, increased only slightly. To assess the quality of jobs we use the classification by status in employment, and distinguish between ‘vulnerable’ and ‘non-vulnerable’ employment. The analytical strength of this distinction derives from the fact that it overlaps to an important extent with the notion of dualism. Dualism in developing economies refers to the coexistence of a formal segment, which uses reproducible capital and employs regular, full-time wage employees, while the non-formal segment relies much more on unskilled labour together with natural resources and simple tools or implements. Contrary to the formal segment, workers in the non-formal segment are self-employed or engaged in casual/irregular wage work. Vulnerable employment consists of the sum of the status groups of own-account workers and contributing family workers. These workers are less likely to have formal work arrangements, and are therefore more likely to lack elements associated with decent work such as adequate social security and recourse to effective social dialogue mechanisms. Vulnerable employment is often characterized by inadequate earnings, difficult conditions of work that undermine workers’ fundamental rights, or other characteristics pointing at decent work deficits, including low incomes (Sparreboom and Albee, 2011). For example, average monthly income from paid employment in Tanzania was much higher than the income from self-employment in both 2001 and 2006 (ILO, 2010), which also suggests that productivity of paid workers is higher than those in selfemployment. Nonetheless, it is important to note the limitations of the use of nonvulnerable employment as a proxy for decent work. Some workers in wage employment, and in particular those in casual/irregular wage work, are likely to face similar decent

8

work deficits as own-account workers. Conversely, some own-account workers, for example those in professional occupations, may not be vulnerable at all. The vulnerable employment rate in Tanzania is very high, at 90.4 per cent in 2001, and decreased by not more than 2.6 percentage points from 2001 to 2006. These findings are confirmed by the marginal drop of the working poverty rate between 2000 and 2007, with almost one third of workers still counted among the working poor (ILO, 2010).4 The share of workers in informal employment is estimated at 93 per cent, and virtually all workers in rural areas are in informal employment. Finally, the widespread low quality employment in Tanzania is reflected in the high employment-to-population ratio, which stood at 85.4 per cent in 2006, which is more than 20 percentage points above the average for sub-Saharan Africa (ILO, 2013a). To sum up, the data show structural changes in the economy and in employment, increasing value added in agriculture, industry and services with highest growth in industry followed by services. Employment increased in all three sectors, with low increase in agriculture and highest increase in industry, followed by the service sector. Most important are the sectoral changes in labour productivity. Productivity slightly increased in agriculture, but strongly dropped in industry and services. In other words, labour productivity declined since the growth in employment in these two sectors exceeded growth in value added. The productivity gains at the national level, which averaged 3.3 per cent during the period under review, are therefore almost entirely due to ‘Baumol’s structural bonus’, or the shift of labour from less productive to more productive sectors.5 Higher levels of education are considered as a key determinant of productivity increase by enhancing efficiency in use of existing technologies and by facilitating structural transformation from low to higher value added activities. Therefore, important questions concern the effects of the increase in educational attainment levels of the labour force, in particular in relation to structural change in the economy, rising value added in all three sectors, but declining productivity in industry and service sectors. Furthermore, it is important to understand why rising educational attainment levels are reflected in high levels of educational attainment of the unemployed. In Tanzania (mainland), the unemployment rate for those with secondary education and above has been consistently higher than the rate for those with lower levels of education (ILO, 2010, table 1). We will consider these questions in the following section.

Based on the nationally-defined poverty line; according to the internationally-defined US$ 1.25 poverty line, the working poverty rate was 64 per cent in 2007 (ILO, 2011). 5 See Baumol (1985). 4

9

4. Empirical analysis of education and patterns of structural change We develop a framework for the analysis of the effect of structural transformation on the creation of more productive and good jobs, and on the employment of workers with higher education levels. The analysis allows to determine whether the increase in educational attainment levels of the labour force accommodates these changes in the economy. The framework builds on a sectoral approach and distinguishes between the ‘between sector effect’ and the ‘within sector effect’. This approach allows to address two main questions: First, which sectors contribute most to changes in educated employment and to decent employment? Second, to which extent do within sector effects and between sector effects account for changes in educational intensity of employment?

4.1

Analytical Framework

The framework is based on the premise that the level of educational attainment of workers is an important determinant of productivity and the quality of jobs. Jobs are created in different sectors which apply different technologies. More complex technologies are associated with higher human capital levels and education. A job is defined as intensive in education when it requires at least lower secondary education. Similarly, education intensity of a sector or of the economy is defined by the proportion of the employed who have completed at least lower secondary education. The change in education intensity of employed workers in the economy or in a sector therefore is assumed to indicate a change in the nature of jobs. An increase in education intensity of workers in a sector, that is, an increasing share of educated workers, suggests the creation of good jobs, while a decrease suggests that more jobs were created with low levels of technology and productivity. Furthermore, within this framework it is argued that the change in education intensity can be explained by two distinct effects. One effect, the so-called reallocation effect or ‘between sector effect’ is due to structural change and the shift of educated workers between sectors. Sectors differ in the educational profile and educational intensity of their workers, with some sectors having higher education intensity than others, that is, they are associated with a higher portion of jobs with secondary and tertiary education levels. Consequently, education intensity of the economy increases when education intensive sectors grow more than those with lower education intensity. Positive reallocation effects are achieved when sectors grow with shares of educated workers above average education intensity in total employment. The growth of sectors and jobs with high education intensity enhances demand for educated workers and therefore provides employment opportunities for educated graduates. In contrast, the growth of low quality jobs (e.g. in informal sector activities) creates employment opportunities for low educated workers. The second effect, the ‘within sector effect’, measures the change in educational intensity within a particular sector. A positive within sector effect results when the growth of the share of educated employment exceeds growth of total employment in this sector. In a similar way, negative within effects result when the share of educated workers declines relative to total employment in the sector. 10

The following decomposition methodology is applied to identify the importance of the two effects in changes in education intensity in the economy: (1)

e 2006 – e 2001 =

(2)

∑i(α i2006 e i2006) – ∑i (α i2001 e i2001) =

(3)

e i2001 ∑i (α i2006 - α i2001) + α i2006 ∑i (e i2006 - e i2001)

In equation (1), e 2001 and e 2006 are the education intensity of labour force for 2001 and 2006, which each equal the sum of the sectoral education intensities e i weighted by the share of each sector i in the labour force, denoted by α i in equation (2). Rearranging of terms yields equation (3), in which the first part is the change in education intensity due to changes of sectoral employment shares (the between sector effect), and the second part is the change in education intensity due to changing intensities within sectors. In other words, the between sector change shows how much the aggregate education intensity would change if the education intensities of the individual sectors would have remained at their 2001 levels. Subtracting the between sector changes from the overall change in education intensity in each sector between 2001 and 2006 yields the remaining ‘within sector’ change of education intensity.

4.2

Empirical Findings

4.2.1 Educational profile of sectors The distribution of educational attainment levels by economic sectors shows that the agricultural sector employs the least educated workers (figure 2). Most workers have achieved primary education, while the share of workers with secondary and tertiary education is extremely low. In contrast, the service sector has the highest educational attainment levels of employed workers, but it also demonstrates a polarized pattern. Services employ the highest shares of workers with lower and higher secondary and with tertiary education, while at the same time more than 17 per cent have no or incomplete primary education. This reflects the difference in job quality and productivity between the modern service sector (e.g. tourism, international trade) which employs corporate managers, professionals and clerks, while the low productivity segment of services often consist of informal activities which require limited education levels. [Figure 2 around here] Manufacturing, and construction and mining, two sub-sectors in industry, show similar educational attainment profiles. They employ high shares of primary education while the shares of workers with secondary and tertiary education remain lower when compared to the service sectors. While the share of tertiary education is extremely low in all sectors, there is also a striking difference between manufacturing and services with regard to the share of tertiary education. Manufacturing in Tanzania employs hardly any tertiary educated workers, indicating relative low levels of complexity in management and technological tasks performed in this sector. It also mirrors the 11

dominance of micro and small enterprises which apply crafts production mode and low levels of division of labour. This sector provides mainly jobs for primary and lower secondary educated workers in elementary occupations and craft and related trades workers (see annex B for the classification of occupations by educational levels). Few firms in the modern segment of manufacturing apply more capital and technology intensive production modes that require occupations such as machine operators, or even technicians and clerks.

4.2.2 Education intensity by sector These different educational attainment levels in economic sectors are also reflected in the education intensity of sectors (measured by the share of workers with lower secondary, higher secondary and tertiary education). Columns 4 and 5 in table 1 show the education intensity in different sectors in 2001 and 2006. Education intensity is highest for both years in services, and lowest in agriculture. In 2001, only 1.6 percent of workers employed in agriculture had obtained at least lower secondary education, while this is the case for 11.1 and 18 percent in industry and services, respectively. Hence, this suggests that most jobs in agriculture demonstrate extremely low productivity and only few jobs, in particular in the large farms, provide high productivity. Furthermore, the higher education intensity of jobs in industry and services suggests higher productivity of these jobs. The change in education intensity between 2001 and 2006 provides an indication for the change in the nature of jobs created in the economy and in particular sectors. The education intensity at the national level increased by 1.3 percentage points, from 4.6 to 5.9 per cent between 2001 and 2006. This suggests that, on average, more jobs were created that employed workers with at least lower secondary education. In other words, the educational intensity of the new jobs was above the average educational intensity of employment in the economy in 2001. Where and how were these jobs created? This question is explored by analyzing the changes in education intensity by sector. The analysis shows that education intensity remained unchanged in agriculture, increased in industry by 0.2 percent points and decreased in services by 0.1 percent points. The two subsectors of industry show relatively large changes. Education intensity in manufacturing decreased by 2.5 percentage points, while the largest increase (4 percentage points) occurred in the mining and construction sector. The decrease in the manufacturing sector suggests that job creation in this sector involved relatively low-skilled jobs, with educational intensity below the average educational intensity in this sector. In contrast, the strong increase in educational intensity in the mining and construction sector implies that many good jobs were created with an educational intensity above the average intensity in this sector. The education intensity of the unemployed is relatively high, around the same level as in industry. Furthermore, between 2001 and 2006 the educational intensity of this group increased strongly by 3.6 percent points, almost as much as the increase in the mining and construction sector. This increase adds to the graduate unemployment phenomenon, and we will return to this phenomenon further below when we discuss employment quality.

12

Table 1. Structural change and education intensity Sectoral distribution of the labour force (%)

Education intensity (%)

Between sector effect

Within sector effect

Contribution by sector

Contribution by sector (%)

(7)

(8)

(9)

(10)

(1)

(2)

(3)

(4)

(5)

(6)

2001

2006

Change

2001

2006

Change

77.4

70.0

-7.5

1.6

1.6

0.0

-0.1

0.0

-0.1

-10.8

Industry

2.9

4.9

2.0

11.1

11.3

0.2

0.2

0.0

0.2

18.7

Manufacturing

1.6

3.0

1.4

12.9

10.4

-2.5

0.2

-0.1

0.1

8.6

Mining and construction

1.3

1.9

0.6

8.8

12.8

4.0

0.1

0.1

0.1

10.1

14.5

20.3

5.8

18.0

17.9

-0.1

1.0

0.0

1.0

80.8

5.2

4.9

-0.4

9.1

12.7

3.6

0.0

0.2

0.1

11.2

100.0

100.0

0.0

4.6

5.9

1.3

1.1

0.1

1.3

100.0

88.3

11.7

Agriculture

Services Unemployed Aggregate

Percentage of intensification due to between and within sector effects

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: Industry includes manufacturing, mining and construction.

Columns (7) and (8) in table 1 indicate the extent to which changes in education intensity are due to structural change in employment, or are due to changes within (broad) economic sectors. This analysis looks into the dynamics of productive transformation in Tanzania and the relative importance of structural change versus changes within sectors for explaining changes in the educational intensity of jobs and economic activities in the economy. This analysis helps to explain the pattern of economic change that resulted in high growth in annual value added and the much more limited productivity increases. Only in the cases of agriculture and the unemployed was the between sector change less than the within sector change. The negative between sector effect in the case of agriculture results from the decrease of the size of the agricultural sector in accordance with the first term in equation (3) above, while the within sector effect is negligible. Both industry and services increased their share in employment, and therefore contributed to the between sector change in education intensity. The change in education intensity in manufacturing is entirely due to an expansion of this sector, as the within sector effect is negative. In contrast, both the within and the between sector effects are positive in mining, suggesting increasing productivity of workers (in an expanding sector). Mining and construction is unique in that it is the only sector in which the within sector effect was positive (apart from the unemployed). In other words, in all other sectors, job creation within sectors did not involve jobs with relatively high (above average) levels of education. Overall, more than 88 per cent of the overall change in education intensity is accounted for by between sector change. Column (9) in table 1 indicates how much each sector contributes to the national change in education intensity of 1.3 percentage points (which is the sum of the between and within sector effects for each sector). In column (10) the contributions by sector are expressed as a percentage of this national change. As shown 13

in the final two columns, 1.0 percentage point out of 1.3 or more than 80 per cent of the increase in education intensity of 1.3 percentage points is due to the increase in education intensity in the services sector, despite the fact that only around one of five workers are employed in this sector. Furthermore, this increase is entirely due to the expansion of services, as the within sector change amounted to zero in the case of services. The results in table 1 are sharply different from the findings by the ADB for selected Asian countries (ADB, 2007). In all four countries (India, Indonesia, Philippines and Thailand) it was found that most of the change in education intensity over time, up to 90 per cent in the case of Indonesia, was due to within sector increases in education intensity (table 2). Obviously, in making such comparisons the higher levels of GDP per capita and extent of structural change that has already occurred in many Asian countries should be taken into account. Table 2. Structural change and education intensification in selected countries

Period

Between sector (%)

Within sector (%)

Education intensity (end of period, %)

Share of agriculture in the labour force (end of period, %)

GDP per capita (constant 2000 US$; end of period)

Tanzania

2001-2006

88.3

11.7

5.9

70.0

392

India

1993-2004

37.3

62.7

21.5

51.3

525

Indonesia

1994-2004

9.6

90.4

45.9

40.5

876

Philippines

1991-2004

29.0

71.0

50.9

33.1

1,153

Thailand

1995-2005

17.0

83.0

36.5

42.0

2,360

Source: Table 1; ADB (2007); World Bank (2012a).

The comparison between low-income Tanzania and the experiences in selected Asian countries suggests that between sector effects are more important at the early stages of economic development. At later stages, productive transformation within sectors becomes more important, and such transformation requires additional education over and above the education intensity which is needed in accordance with structural change in the sense of sectoral reallocation of workers. However, the results in tables 1 and 2 suggest that the increase in education intensity in Tanzania was barely sufficient to keep up with the sectoral reallocation, and by 2006 had not yet reached a stage in which increases in educational attainment within sectors are significant. Consequently, only a small part of the increase in educational attainment from 2001 to 2006 was available to accommodate technological change within sectors.

4.2.3 Employment quality and education The framework used in the previous sub-section is based on the premise that the level of educational attainment of the labour force is an important determinant of productivity and the quality of jobs. This sub-section examines this premise based on an assessment of the quality of jobs. It will be demonstrated that such an assessment adds to an understanding of the decrease of productivity in industry (despite the increase in educational intensity in this broad sector, see table 1), the polarized pattern of

14

educational attainment in services (section 3) and the graduate unemployment phenomenon in Tanzania. As discussed in section 2, the quality of jobs is assessed using the classification by status in employment. Vulnerable employment is considered as low quality while nonvulnerable employment is associated with better jobs. Given the high vulnerable employment rate in Tanzania, the results in table 1 are primarily driven by changes in the sectoral distribution of vulnerable employment. Tables 3a and 3b show similar information as table 1, but separately for non-vulnerable and vulnerable employment, and excluding the unemployed. Perhaps the most striking fact is the large difference in education intensity: more than 27 per cent of wage and salary workers (and employers) have at least a secondary educational qualification, while this is true for only between 2 and 3 per cent of own-account workers and contributing family workers. In other words, to the extent that non-vulnerable employment is accepted as a proxy for qualityemployment, there is indeed a strong link between education intensity and job quality. A comparison between tables 3a and 3b shows that structural change in non-vulnerable employment is much smaller than in vulnerable employment. The share of agriculture in non-vulnerable employment decreased by 2.7 percentage points, compared to 6.4 percentage points in vulnerable employment. Furthermore, while non-vulnerable employment expanded in industry, services expanded more strongly for own-account workers and contributing family workers. In fact, the share of services in paid employment decreased, and most of the structural change in total employment is due to a shift from vulnerable employment in agriculture to vulnerable employment in services. Even though such a shift signifies structural change (sectoral reallocation) in employment, the extent to which it reflects productive transformation is questionable. Unsurprisingly, table 3b resembles the overall labour market situation in table 1 more closely. The percentage change in education intensity due to between sector change is lower in table 3b than in table 1, but still accounts for most of the change. This contrasts with table 3a, as the education intensity in non-vulnerable employment decreased by 0.2 percentage points at the national level due to a decrease in within sector intensity in all sectors except mining and construction. In other words, the education intensity in the more productive part of the economy decreased, and this decrease is likely to be an important explanatory factor behind the drop in productivity levels in both industry and services. A second factor consists of the effects of changes in the distribution of employment in each sector. In both industry and services, vulnerable employment expanded as a proportion of all employment in these sectors, which would reduce sectoral productivity levels even if the education intensity of employment in nonvulnerable employment would have remained unchanged. The decrease in education intensity in paid employment at the national level also indicates limitations on the demand for graduates at higher levels of education, which helps explain the increase in graduate unemployment. Many graduates are likely to prefer a paid employment over own-account work, and are prepared to queue for jobs in the formal segment of the economy (see e.g. Ghose et al., 2008). Manufacturing experienced a decline in education intensity in non-vulnerable employment, and in 2006 had a lower education intensity than mining and construction, which may be related to the low and slightly decreasing share of medium and high-tech 15

manufacturing activities in total manufacturing value added (see figure 11 in UNIDO, 2012).6 At the aggregate level, the services sector experienced a decline in education intensity (table 1), and the same is true for the education intensity of the most productive part of services (table 3a). Education intensity increased in vulnerable employment in services (but at much lower levels than in non-vulnerable employment), which is part of the explanation of the polarized pattern of educational attainment in this sector. Table 3a. Structural change and education intensification – non-vulnerable employment Employed (%)

Education intensity (%)

Between sector

Within sector

Contribution by sector

Contribution by sector (%)

2001

2006

Change

2001

2006

Change

Agriculture

12.4

9.7

-2.7

6.1

5.1

-1.0

-0.2

-0.1

-0.3

97.8

Industry

16.4

19.7

3.3

18.0

17.1

-0.9

0.6

-0.2

0.4

-155.7

8.6

10.0

1.4

20.1

16.2

-3.9

0.3

-0.4

-0.1

40.6

Manufacturing Mining and construction Services Aggregate

7.7

9.7

2.0

15.7

18.1

2.4

0.3

0.2

0.5

-204.3

71.3

70.6

-0.7

33.2

32.9

-0.3

-0.2

-0.2

-0.4

165.9

100.0

100.0

0.0

27.3

27.1

-0.2

0.2

-0.5

-0.3

100.0

-74.1

174.1

Percentage of intensification due to between and within sector effects

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: Industry includes manufacturing, mining and construction.

Table 3b. Structural change and education intensification – vulnerable employment Employed (%)

Agriculture

Education intensity (%)

2001

2006

Change

2001

2006

Change

Between sector

Within sector

Contribution by sector

Contribution by sector (%)

88.6

82.3

-6.3

1.5

1.5

0.0

-0.1

0.0

-0.1

-17.3

Industry

1.7

3.2

1.5

4.4

6.4

2.0

0.1

0.1

0.1

23.9

Manufacturing

1.0

2.3

1.3

6.6

6.9

0.3

0.1

0.0

0.1

17.0

Mining and construction

0.7

0.9

0.2

1.2

5.2

4.0

0.0

0.0

0.0

7.0

9.7

14.6

4.9

6.8

8.0

1.2

0.3

0.2

0.5

93.3

100.0

100.0

0.0

2.1

2.6

0.5

0.3

0.2

0.5

100.0

60.0

40.0

Services Aggregate

Percentage of intensification due to between and within sector effects

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: Industry includes manufacturing, mining and construction.

At the same time, UNIDO (2012) also indicates significant skills gaps – differences between required and actual shares – for tertiary educated workers. 6

16

4.3 Occupational change and qualifications mismatch The occupational distribution provides insights into sets of jobs whose main tasks and duties are characterized by a high degree of similarity. Structural change in developing economies will be visible in this distribution as the share of skilled agricultural and fishery workers decreases, and industrialization results in larger shares of plant and machine operators and assemblers or craft and related trades workers. Development and rising educational attainment can be expected to result in increasing shares of highskilled occupational groups such as professionals and technicians. In fact, the rising shares of professionals and technicians and associate professionals (major group 3) is a major trend in both developed and developing countries (ILO, 2011). In Tanzania, the share of skilled agricultural and fishery workers decreased by 9.6 percentage points between 2001 and 2006 (see table 4, first three columns). Much of the growth occurred in the major group service and sales workers (5.4 percentage points) and in craft and related trades workers (2.2 points), and plant and machine operators and assemblers increased by 0.5 percentage points in this period. On the other hand, structural change was accompanied by a slight decrease in high-skilled jobs which require tertiary education (in particular technicians and associate professionals). This decrease is consistent with the decrease in education intensity of non-vulnerable employment which was discussed before, to the extent that high skilled jobs are likely to be more prevalent in paid employment. Part of the decrease may also be due to horizontal mismatch, that is, difficulties in finding workers with appropriate skills, given the imbalances in science relative to liberal arts in tertiary training institutions that was noted in section 2. Skills mismatch may exist in several forms (see e.g. Johansen and Gatelli, 2012; ILO, 2013b), including in the form of workers employed in occupations that underutilize their skills set (overqualified workers) or in occupations normally requiring skills that they do not possess (underqualified workers). In both cases, this (vertical) mismatch affects job satisfaction and wages of individual workers, as well as the productivity of firms, and may lead to increases in turnover of staff (Quintini, 2011). The occupational distribution can be used to assess underqualification and overqualification of workers, as major occupational groups are linked to education and skills levels. Major groups 2 (professionals) and 3 (technicians and associate professionals) are linked to advanced levels of education (above secondary education), while major groups 4 to 8 are linked to secondary education. The rationale is that for these occupations the ability to read information such as safety instructions, to make written records of work completed, and to accurately perform simple arithmetical calculations, is essential and many occupations require relatively advanced literacy and numeracy skills and good interpersonal communication skills (ILO, 2012). In Tanzania, this rationale is reinforced by the concerns over the quality of primary education, to the extent that additional years of secondary education are needed to achieve the objectives of primary schooling. Furthermore, lower secondary education is considered vital in the development of foundation and core employability skills (UNESCO, 2012b). Major group 9 (elementary occupations) is linked to primary education (see annex B).

17

Table 4. Occupational distribution and qualification mismatch 2001 (%) 2001 (%)

2006 (%)

Change (percentage points)

2006 (%)

Change (p.p.) A

U

O

A

U

O

A

U

O

Managers

0.2

0.2

0.0

10.4

89.6

0.0

10.8

89.2

0.0

0.4

-0.4

0.0

Professionals Technicians and associate professionals Clerical support workers Service and sales workers Skilled agricultural, forestry and fishery workers Craft and related trades workers Plant and machine operators, and assemblers Elementary occupations

0.4

0.7

0.3

25.4

74.6

0.0

12.0

88.0

0.0

-13.4

13.4

0.0

2.3

1.9

-0.5

1.6

98.4

0.0

7.0

93.0

0.0

5.4

-5.4

0.0

0.4

0.4

0.0

61.1

38.6

0.3

48.6

48.8

2.5

-12.5

10.3

2.2

5.0

10.4

5.4

10.9

89.0

0.1

11.4

88.3

0.3

0.5

-0.7

0.2

81.6

72.0

-9.6

1.5

98.5

0.0

1.5

98.5

0.1

-0.1

0.0

0.0

2.8

5.0

2.2

9.9

90.0

0.1

11.0

88.6

0.5

1.1

-1.4

0.4

0.9

1.4

0.5

10.0

90.0

0.0

16.4

83.6

0.0

6.4

-6.4

0.0

6.4

8.1

1.7

57.0

39.8

3.2

64.7

29.3

6.0

7.7

-10.5

2.8

100.0

100.0

6.2

93.5

0.2

8.7

90.7

0.6

2.5

-2.8

0.4

Total

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Notes: See annex B for methodological details. p.p. = percentage points A = Adequately qualified U = Underqualified O = Overqualified

Table 5. Occupational distribution and qualification mismatch by level of education All workers Major groups 1-3 (tertiary) 4-8 (secondary 9 (primary) Total

2001 (%)

2006 (%)

Adequately qualified workers Change (p.p.)

2001 (%)

2006 (%)

Change (p.p.)

Underqualified workers 2001 (%)

2006 (%)

Change (p.p.)

Overqualified workers 2001 (%)

2006 (%)

Change (p.p.)

2.9

2.7

-0.2

2.4

2.6

0.2

2.9

2.7

-0.2

0.0

0.0

0.0

90.7

89.2

-1.5

38.8

37.0

-1.7

94.4

94.7

0.3

10.6

17.1

6.5

6.4

8.1

1.7

58.8

60.3

1.5

2.7

2.6

-0.1

89.4

82.9

-6.5

100.0

100.0

100.0

100.0

100.0

100.0

100.0

100.0

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007).

Confrontation of the L-shaped education structure in Tanzania with the occupational structure in which the majority of workers are classified in major groups 4 to 8 inevitably results in substantial mismatch. Indeed, a strikingly low proportion of workers is adequately qualified, 6.2 per cent in 2001 and 8.7 per cent in 2006 (table 4). 18

The remainder of the workforce is mostly underqualified, with overqualification accounting for very small proportions of workers. Apart from elementary occupations, only in the major group clerical support workers the proportion of workers with adequate qualifications is close to (in 2006) or above (in 2001) half of all workers in this group. Among associate professionals and technicians only 1.6 per cent was adequately qualified in 2001, rising to a still very low 7.0 per cent in 2006. Comparing 2001 and 2006 shows that the share of adequately qualified workers increased by 2.5 percentage points. Analysis of the pattern of mismatch by level of education shows that jobs in major groups 4-8 account for the lion’s share of underqualification (as most workers are in these groups), and also suggests that workers with secondary education are particularly in demand, despite the fact that the share of jobs in these major groups slightly decreased (table 5).7 Contrary to major groups 1-3, which are linked to tertiary education, and major group 9 (linked to primary education), both the share of adequately qualified workers in groups 4-8 declined and the share of underqualified workers increased. Furthermore, the share of overqualified workers in these groups also increased, which means that relatively more workers with a tertiary education are classified in groups 4-8. This may reflect the lack of growth of jobs in groups 1-3, but given the very low proportions of adequately matched workers in these groups demand for secondary qualifications in major groups 4-8 is likely to be important as well. Similarly, overqualification decreased in the group of elementary occupations, which means that fewer workers with a secondary education work in these occupations. Most of the adequately qualified workers are in this major group, and the increase from 2001 to 2006 reflects the increasing number of workers with primary education.

4.4 Returns to education Rates of return are often higher in low-income countries and in sub-Saharan Africa (Psacharopoulos and Patrinos, 2004b; World Bank, 2012b), which is perhaps not surprising in view of the generally low levels of educational attainment and the high level of (vertical) mismatch. Additional insights derived from rate of return analysis concern the patterns of returns across sectors, levels of education and over time. Rates of return to schooling in Tanzania have been calculated based on Mincerian earnings functions (see annex B). Across all sectors, the average private rate of return to years of schooling is close to 20 per cent in 2001 (figure 3 and table A2), which is much higher than the global mean rate of return of 9.7 per cent reported in Psacharopoulos and Patrinos (2004b, Table 3). Note that social rates of return, which include private and external costs, are usually lower than the private rates discussed here. Returns to investment in education by level, latest year. Mincerian returns and mean years of schooling. Social returns to investment in education by income level. Private returns to investment in education by income.

Across all sectors, the average private rate of return to years of schooling decreased from 2001 to 2006, which is consistent with the decrease in education intensity in paid employment discussed before. As more than two thirds of employees are in service sectors, national returns are similar to those in the broad service sector and the decrease in returns is also consistent with the decrease in education intensity shown in One reason for a decrease in the share of major groups 4-8 may be difficulties in the accurate classification of workers in elementary occupations versus skilled agricultural, forestry and fishery workers (ILO, 2011).

7

19

table 3a. Similarly, the sharp fall in returns in manufacturing and rise in returns in mining and construction are consistent with the pattern of change in education intensity in (non-vulnerable) employment between 2001 and 2006. Rates of return in Tanzania are lowest in agriculture, and contrary to what might have been expected on the basis of the decrease in education intensity, returns to education marginally increased (by 0.5 percentage points). Between 2001 and 2006 the returns to education decreased for primary education and tertiary education, and increased for secondary education (figure 4 and table A3). The relatively low and decreasing returns to primary education appears to be a consequence of the rapid rise in the number of graduates at this level, and is in line with recent international trends. Although the evidence for many years indicated higher returns at the primary level compared with secondary and tertiary levels, more recent studies suggest a lower return at the primary level as well as a decreasing long-term trend in returns to primary education, pointing at a convex earnings function (Colclough et al., 2010; Teal, 2011). Colclough et al. (2010) highlight that the average return to primary education in a sample of studies with similar methodologies shows a decrease by two percentage points in studies undertaken since the year 2000. [Figures 3 and 4 around here] The strong increase in the wage returns for secondary graduates and the decrease for tertiary education appears to be consistent with the analysis of mismatch (which covers all workers), and in particular the demand for secondary graduates. The decrease in returns to tertiary education may also be due to some of the quality issues in tertiary education that were noted in section 2, as well as possible (horizontal) skills mismatch. According to UNESCO (2012a), differences between wages of secondary and those of tertiary graduates are small in the public sector, which is the main employer of tertiary graduates. Returns to education are generally higher in urban areas than in rural areas, while women face a ‘penalty’ on their returns (tables A2 and A3). Disturbingly, the penalty for women increased between 2001 and 2006, but the premium for urban workers decreased during this period.

20

5. Conclusions This paper explores patterns of structural transformation in Tanzania and how changes in educational attainment are related to productivity and the quality of employment. The analysis is using a framework which is building on a sectoral approach and is taking the quality of employment explicitly into account. The framework is based on the premise that the level of educational intensity of jobs and sectors is an important determinant of productivity and the quality of jobs. The analysis of education intensity in vulnerable and non-vulnerable employment suggests that this premise is justified. Furthermore, the framework suggests to measure changes in education intensity by two distinct effects - the between sector and the within sector effect. Hence, by using education intensity as an indicator for the quality of jobs, this paper determines which sectors contribute most to changes in educated employment and to decent employment as well as the extent to which changes in job quality and productivity are accounted for by the within sector effects and between sector effects. We apply this framework to explain the growth of productivity in Tanzania during the period 2001-2006, and in particular the decrease in productivity in industry and service sectors, despite the steady increase in levels of educational attainment of the labour force (albeit still at low levels). Decreases in sectoral productivity in industry and services appear to be partly due to the decrease in education intensity in the highproductivity parts of these sectors and productive transformation within sectors was largely absent. The analysis based on this framework also shows that the pace of increases in educational attainment was barely sufficient to keep up with the sectoral reallocation of workers in the economy. Furthermore, the analysis of qualifications of workers, benchmarked against occupations and the education and skills defining these occupations, indicates continuing high levels of (vertical) mismatch, which is principally due to underqualification. Qualifications mismatch not only affects the high-skilled occupations in major groups such as professionals and technicians, but is an issue in virtually the full range of occupations. Mismatch is decreasing slowly, in line with the expansion of the educational system. The sectoral pattern of rates of return to education in Tanzania appears consistent with the analysis of structural change and the education intensity, while increasing returns to secondary education are in accordance with the high levels of mismatch between levels of education and occupations, and the demand for workers with this level of education. However, the decreasing returns to tertiary education raises important questions and challenges further research to better understand patterns of structural transformation that are able to provide jobs and occupations for tertiary educated workers. Another important area for research is the complementarity of primary, secondary and tertiary educated workers in enhancing the options for diversification of production structures (across and within sectors) and upgrading technologies (see Nübler, 2013). The findings provide some important policy conclusions. Education policies in lowincome countries aiming at supporting productive transformation and decent work should not be limited to enhancing investment in high quality basic education as the basis for increased effective investment in secondary and higher education. Policies 21

should also aim to expand secondary and tertiary education to provide the options for diversification into economic activities that create more and better jobs, and prevent a situation in which the slow pace of increase in educational attainment limits the diffusion of decent work. Training policies need to ensure the supply of qualifications and skills to avoid skills mismatches. Finally, industrial development policies need to support structural transformation patterns that create more and good jobs and these policies need to be aligned with education and training policies to ensure the supply of a labour force that accommodates such industrial development patterns.

22

References ADB. 2007. Asian Development Outlook 2007, Asian Development Bank (Manila). Baumol, W.J.; Blackman, S.A.B; Wolff, E. 1985. ‘Unbalanced Growth Revisited: Asymptotic Stagnancy and New Evidence’, American Economic Review, Vol. 75, No. 4, pp. 806-817. Colclough, C.; Kingdon, G.; Patrinos, H. 2010. ‘The changing Pattern of Wage Returns to Education and its Implications’, Development Policy Review, Vol. 28, No. 6, pp. 733-747. Fares, J.; Guarcello, L.; Manacora, M.; Rosati, F.C.; Lyon, S.; Valdivia, C.A. 2005. ‘School to work transition in Sub-Saharan Africa: An overview’, Working Paper No. 15, Understanding Children’s Work (UCW) Project (Rome). Ghose, A.K.; Majid, N.; Ernst, C. 2008. The Global Employment Challenge, International Labour Office (Geneva). Hausmann, R.; Hidalgo, C.A.; Bustos; S.; Doscis, M.; Chung, S.; Jimenez, J. 2011. The atlas of economic complexity, Center for International Development at Harvard University (Cambridge). ILO. 2010. Decent Work Country Profile Tanzania (mainland), International Labour Office (Dar es Salaam and Geneva). ILO. 2011. Key Indicators of the Labour Market, 7th edition, International Labour Office (Geneva). ILO. 2012. International Standard Classification of Occupations: ISCO-08, International Labour Office (Geneva). ILO. 2013a. Global Employment Trends 2013. Recovering from a second jobs dip, International Labour Office (Geneva). ILO. 2013b. Global Employment Trends for Youth 2013. A generation at risk, International Labour Office (Geneva). Imbs, J.; Wacziarg, R. 2003. ‘Stages of Diversification’, American Economic Review, Vol. 93, No. 1, pp. 63-86. Johansen, J.; Gatelli, D. 2012. Measuring mismatch in ETF partner countries. A methodological note, European Training Foundation (Turin). Klinger, B.; Lederman, D. 2004. Discovery and development: An empirical exploration of new products, World Bank (Washington, D.C.). Kucera, D.; Roncolato, L. 2012. ‘Sectoral drivers of growth and the labour productivityemployment relationship’, ILO Research Paper, No. 3, International Labour Office (Geneva).

23

Kucera, D.; Tejani, S. Forthcoming 2013. Feminization, Defeminization, and Structural Transformation, International Labour Office (Geneva). McMillan, M.; Rodrik, D. 2011. ‘Globalization, structural change, and productivity growth’, NBER Working Paper, No. 17143 (Cambridge). MOEC. 1995. Education and Training Policy, Ministry of Education and Culture (Dar es Salaam). MOEVT. 1999. National Higher Education Policy, Ministry of Education and Culture, Ministry of Education and Vocational Training (Dar es Salaam). MOEVT. 2006. Primary Education Development Programme II (2007-2011), Ministry of Education and Vocational Training (Dar es Salaam). MOEVT. 2008. Education Sector Development Programme (2008-2017), Ministry of Education and Vocational Training (Dar es Salaam). MOEVT 2010. Higher Education Development Programme (2010-2015), Ministry of Education and Vocational Training (Dar es Salaam). MOEVT, 2012. Basic Education Statistics, Ministry of Education and Vocational Training online database (Dar es Salaam). MOFEA. 2010. National Strategy for Growth and Reduction of Poverty II, Ministry of Finance and Economic Affairs (Dar es Salaam). Nübler, I. Forthcoming 2013. Capabilities for productive transformation and development: a new perspective on industrial policies, Palgrave McMillan/International Labour Office (Geneva). Ocampo, J.A.; Rada, C.; Taylor, L. 2009. Growth and Policy in Developing Countries: A Structuralist Approach, Columbia University Press (New York). Psacharopoulos, G.; Patrinos, H.A. 2004a. ‘Human capital and rates of return’, Chapter 1 in Johnes, G.; Johnes, J. (eds.), International Handbook on the Economics of Education, Edward Elgar (Cheltenham). Psacharopoulos, G.; Patrinos, H.A. 2004b. ‘Returns to Investment in Education: A Further Update’, Education Economics, Vol. 12, No. 2, pp. 111-134. Quintini, G. 2011. ‘Over-qualified or under-skilled: a review of existing literature’, OECD Social, Employment and Migration Working Papers, No. 121 (Paris). Sparreboom, T.; Albee, A. (eds.) 2011. Towards Decent Work in sub-Saharan Africa. Monitoring MDG Employment Indicators, International Labour Office (Geneva).

24

Teal, F. 2011. ‘Higher Education and Economic Development in Africa: a Review of Channels and Interactions’, Journal of African Economies, Vol. 20 (supplement 3), pp. 5079. UNESCO. 2012a. Tanzania, Education Sector Analysis (Dakar). UNESCO. 2012b. Youth and skills: Putting education to work. EFA Global Monitoring Report (Paris). UNIDO. 2012. Tanzania Industrial Competitiveness Report 2012 (Dar es Salaam). United Republic of Tanzania. 2000. Poverty Reduction Strategy Paper (Dar es Salaam). United Republic of Tanzania. 2001. Education Sector Development Programme (Dar es Salaam). United Republic of Tanzania. 2002. Integrated Labour Force Survey 2000/01 - Analytical Report, National Bureau of Statistics (Dar es Salaam). United Republic of Tanzania. 2005. National Strategy for Growth and Reduction of Poverty, Vice President’s Office (Dar es Salaam). United Republic of Tanzania. 2007. Analytical report for Integrated Labour Force Survey (ILFS) 2006, National Bureau of statistics (Dar es Salaam). Walker, I.; Zhu, Y. 2001. ‘The Returns to Education: Evidence from the Labour Force Surveys’, Research Report RR313, University of Warwick (Coventry). Wedgwood, R. 2007. ‘Education and poverty reduction in Tanzania’, International Journal of Educational Development, Vol. 27, No. 4, pp. 383-396. World Bank. 2012a. World Development Indicators & Global Development Finance (Washington, D.C.). World Bank. 2012b. Jobs. World Development Report 2013 (Washington, D.C.).

25

Annex A. Tables A1. Enrolment in education (thousands) 2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

4,382.4

4,881.6

5,981.3

6,562.8

554.8 7,083.1

638.6 7,541.2

669.1 7,959.9

795.0 8,316.9

874.0 8,410.1

896.1 8,441.6

261.9

289.7

323.3

345.4

432.6

524.3

675.7

1,020.5

1,222.4

1,515.9

O-Level

238.2

264.9

296.5

319.5

401.6

489.9

630.2

967.1

1,164.3

1,401.6

A-Level

23.7

24.8

26.8

26.0

31.0

34.4

45.4

53.4

58.2

64.8

Open distance learning Teacher training*

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

n.a.

49.5

11.4

16.0

13.2

28.5

32.7

28.1

30.5

23.4

21.9

35.4

1,063.1

1,073.3

0.0

0.0

466.0

357.5

1,890.0

1,473.9

1,170.5

1,040.3

n.a.

n.a.

n.a.

n.a.

233.9

192.8

150.7

109.5

69.2

53.1

n.a.

n.a.

232.1

164.7

Pre-primary Primary Secondary

Adult education/ non-formal education Cohort I Cohort II

n.a.

n.a.

ICBAE

1,063.1

1,073.3

Technical education ** Non-higher technical Higher technical Vocational education and training VTC Long Courses FDC Long Courses Higher education ***

1.6

4.1

3.8

70.7

75.7

42.2

29.9

1,668.5

1,288.7

1,059.1

957.3

36.6

48.6

49.6

20.1

31.1

28.8

16.5

17.5

20.8

57.1

67.0

95.8

68.5

72.9

52.8

62.7

90.8

63.4

68.2

9.6

2.0

4.3

4.3

4.9

5.1

4.7

30.8

36.6

40.1

48.1

80.0

97.2

115.3

Source: UNESCO (2012a). Notes: * Teacher training refers to teacher training colleges only. ** Technical education refers to NACTE-registered institutions; whereas ‘non-higher’ refers to certificates and ordinary diplomas, ‘higher’ includes advanced diplomas, degrees and beyond; technical education figures for 2009 are based on UNESCO estimates. *** Higher education includes universities and university colleges.

26

A2. Rate of return analysis: returns to years of schooling 2001 Education years

Age

All sectors

All sectors

0.118 7.3

0.211 12.0

0.101 4.7

0.195 34.8

0.100 21.2 -0.001 -12.5

0.067 4.6 -0.001 -2.9

0.035 2.1 0.000 -0.1

0.009 0.5 0.000 0.3

0.1 22.0 -0.001 -13.2

2.490 14.7

2.850 13.1

4.207 17.6

1.980 30.6

0.22 319

0.37 284

0.12 207

0.54 2318

Agriculture Manucturing

Mining and construction

Constant

2.097 36.8

R squared N

0.49 3128

0.51 3128

Urban

Education years

Age

All sectors

All sectors

Services

0.186 41.4

0.181 40.1

0.123 8.3

0.148 9.3

0.173 10.1

0.184 33.5

0.084 17.0 -0.001 -9.4

0.093 5.8 -0.002 -5.0

0.061 4.2 0.061 -2.5

0.051 3.0 -0.001 -1.7

0.085 14.7 -0.001 -7.2

2.967 15.3

3.551 18.4

3.480 14.9

2.786 38.4

0.23 332

0.27 320

0.28 316

0.43 2719

Constant

2.824 46.2

0.081 16.5 -0.001 -9.0 -0.165 -5.1 0.205 6.4 2.799 43.1

R squared N

0.42 3687

0.43 3687

Age squared

Services

0.187 38.5

Female

2006

Mining and construction

0.196 40.0

0.101 21.8 -0.001 -13.0 -0.097 -2.9 0.351 11.5 1.964 32.7

Age squared

Agriculture Manucturing

Female Urban

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: All coefficients are significant at 1 per cent, except those that are underlined (not significant) or in italics (significant at 10 percent); the rows below the coefficients are t-values.

27

A3. Rate of return analysis: returns to levels of education 2001 Primary

All sectors 1.095 22.2 1.960 18.3 2.350 17.5 0.112 21.3 -0.001 -13.5

All sectors

Constant

2.511 39.0

1.017 20.8 1.850 17.6 2.129 16.0 0.115 22.0 -0.002 -14.1 -0.077 -2.1 0.400 11.7 2.334 34.6

R squared N

0.38 3128

0.40 3128

Secondary Tertiary Age Age squared Female Urban

Agriculture Manucturing

Mining and construction

Services

0.578 5.1 3.698 1.3 2.853 1.7 0.069 4.5 -0.001 -3.0

1.325 7.1 2.182 6.6 2.665 3.7 0.006 0.3 0.001 1.3

0.305 1.8 0.883 2.4 2.170 4.0 0.012 0.6 0.000 0.0

1.065 16.6 1.857 16.2 2.187 16.2 0.130 22.4 -0.002 -14.6

2.748 16.4

3.475 14.7

4.626 19.9

2.424 30.9

0.17 319

0.22 284

0.11 207

0.41 2318

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: All coefficients are significant at 1 per cent, except those that are underlined (not significant) or in italics (significant at 10 percent); the rows below the coefficients are t-values.

28

A4. Rate of return analysis: returns to levels of education (continued) 2006 Primary

All sectors 0.903 18.8 2.258 29.7 2.434 18.8 0.086 16.3 -0.001 -9.5

All sectors

Constant

3.381 49.8

0.845 17.4 2.168 28.3 2.343 18.1 0.084 15.9 -0.001 -9.2 -0.140 -4.0 0.234 6.8 3.338 46.7

R squared N

0.33 3687

0.34 3687

Secondary Tertiary Age Age squared Female Urban

Agriculture Manucturing

0.570 5.3 2.717 7.5 2.891 2.5 0.081 5.1 -0.001 -4.4

0.492 3.7 1.752 7.5

Mining and construction

Services

0.057 3.8 -0.001 -2.5

0.410 2.4 1.835 5.5 2.369 5.6 0.048 2.6 -0.001 -1.8

0.764 11.4 2.060 22.5 2.198 15.6 0.093 15.1 -0.001 -8.3

3.337 18.0

4.235 22.3

4.436 17.3

3.452 40.3

0.25 332

0.21 320

0.19 316

0.35 2719

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007). Note: All coefficients are significant at 1 per cent, except those that are underlined (not significant) or in italics (significant at 10 percent); the rows below the coefficients are t-values.

29

Annex B. Methods B1. Qualification mismatch According to the normative method, qualifications mismatch is determined using the presumed correspondence between education levels and major occupational groups in the International Standard Classification of Occupations (ISCO-1988), as shown in the table below (see ILO, 2012). Workers with educational levels in accordance with this correspondence are considered as adequately educated, while those with a lower (higher) level are undereducated (overeducated). The national classification of occupations in Tanzania (TASCO) is similar to ISCO. For the preparation of table 4, we assumed the following correspondence between TASCO major groups and levels of education: major groups 1-3: tertiary education major groups 4-8: secondary education major group 9: primary education To improve comparability between the two household surveys used in this paper, TASCO sub-major group 13 has been excluded from the analysis (see ILO, 2010). International Standard Classifications of Occupations 1988 ISCO-1988 – Major groups 1 2 3 4 5 6 7 8 9 0

Legislators, senior officials and managers Professionals Technicians and associate professionals Clerks Service workers and shop and market sales workers Skilled agricultural and fishery workers Craft and related trades workers Plant and machine operators and assemblers Elementary occupations Armed forces

ISCO skill level* -4 3 2 2 2 2 2 1 --

* (1). The first ISCO skill level was defined with reference to ISCED category 1, comprising primary education which generally begins at the age of 5, 6 or 7 and lasts about five years. (2) The second ISCO skill level was defined with reference to ISCED categories 2 and 3, comprising first and second stages of secondary education. The first stage begins at the age of 11 or 12 and lasts about three years, while the second stage begins at the age of 14 or 15 and also lasts about three years. A period of onthe-job training and experience may be necessary, sometimes formalised in apprenticeships. This period may supplement the formal training or replace it partly or, in some cases, wholly. (3) The third ISCO skill level was defined with reference to ISCED category 5, (category 4 in ISCED has been deliberately left without content) comprising education which begins at the age of 17 or 18, lasts about four years, and leads to an award not equivalent to a first university degree. (4) The fourth ISCO skill level was defined with reference to ISCED categories 6 and 7, comprising education which also begins at the age of 17 or 18, lasts about three, four or more years, and leads to a university or postgraduate university degree, or the equivalent.

30

B2. Rate of return analysis Mincerian earnings functions have been estimated following Psacharopoulos and Patrinos (2004a) and Walker and Zhu (2001). The log of hourly wages (lnW) is regressed on years of schooling (S), years of experience in the labour market (EX) as well as its square (EX 2), using ordinary least squares. The basic Mincerian earnings function takes the form: lnWi = α + βSi + γ1EXi + γ2EX 2i+ εi In this equation, β can be interpreted as the average private rate of return to one additional year of schooling, regardless of the educational level this year of schooling refers to. This method assumes that forgone earnings represent the only cost of education, and so measures only the private rate of return, and assumes further that individuals have an infinite time horizon. The function above does not distinguish between levels of schooling. To gain insights into the returns to different levels of schooling, a series of dummy variables are substituted for S which correspond to discrete educational levels (primary, secondary and tertiary): lnWi = α + βpDp + βsDs + βtDt + γ1EXi + γ2EX2i + εi A separate regression has been performed including control variables for being female (F) and living in an urban area (U), as follows: lnWi = α + βSi + γ1EXi + γ2EX 2i+ γ3Fi + γ4Ui + εi

31

Figure 1. Average annual growth of value added, employment and productivity by broad economic sector, 2001-2006 (%)

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007) and World Bank (2012a).

Figure 2. Distribution of the employed by educational attainment and broad sector, 2006 (%)

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007).

32

Figure 3. Returns to years of schooling, 2001 and 2006 (%)

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007).

Figure 4. Returns to levels of education, 2001 and 2006 (%)

Source: Authors’ calculations based on the Tanzania National Bureau of Statistics, Integrated Labour Force Survey, 2000/01 and 2006 (United Republic of Tanzania, 2002 and 2007).

33