• Keine Ergebnisse gefunden

Decomposing by sector and skill

The Labor Income Share in Developing Countries: A Review and Analysis of

3.5 The components of the labor share

3.5.3 Decomposing by sector and skill

The magnitude and development of labor shares vary considerably between sectors and workers of different skills. Unfortunately, SNA data on compensation of employees and mixed income is usually only available on the aggregate level but the WIOD data allows us to decompose by sector and skill. This is only possible, however, for middle income countries.

Standard economic propositions suggest that the globalization process especially favors the low-skilled workers of developing countries because they are the relatively abundant

Figure 3.7:Mean labor shares of the wage and self-employment sector, 1990-2011

01020304050Labor Share in % of GDP

1990 1995 2000 2005 2010

Wage Employment (SNA) Self−Employment (SNA G1) Self−Employment (SNA G2) Self−Employment (PWT) Self−Employment (Trapp)

01020304050Labor Share in % of GDP

1990 1995 2000 2005 2010

Wage Employment (WIOD) Wage Employment (ILO) Self−Employment (WIOD) Self−Employment (ILO)

Note: Data is interpolated and kept constant beyond start and end points for this illustration to deal with unbalanced panel data sets.

production factor. Figure 3.9 decomposes the labor share by skill and reflects the contrary.

Although the low-skilled make up almost two-thirds of the labor force, they only earn about one third of all labor incomes and about 16.5 % of national income in 1995. Until 2009, their portion in total labor and national income has further declined to 12.2 % and 23 %, respectively, while their number of working hours remained more or less unchanged. The finding for high-skilled workers is exactly the opposite: In 1995, only a very small share of the workforce was high-skilled (6.6 %), but their share in the overall labor share was disproportionately high with 23.4 % and they further increased their share in labor income to 33 % and 17.3 % (from 12.7 %) in national income over time. Their representation in the workforce has also slightly increased to 10 %. The findings for the middle-skilled workers are somewhere in between: They are the second largest group in the workforce and their working hours increase over time, while their share in overall labor and national income remains more or less on the same level during the observed period. The gap between overall productivity and the average labor income of the middle-skilled thus has widened over time.

This indicates that the low levels in aggregate labor share are primarily driven by the low labor incomes of the low-skilled workers, which is in line with the findings for the self-employment sector above as it typically accommodates the low-skilled. Also the decline in the labor share was entirely at their expense. The literature discusses various reasons for these trends, such as technological change, which is biased towards high-skilled activities (Dabla-Norris, 2015, Goldberg and Pavcnik, 2007). These dynamics have direct implica-tions for the income inequality in developing countries: As the decline in the labor share

3.5 The components of the labor share 85 Figure 3.8:Percentage change of labor share components, 1995-2009

0.8

−2.1

14.1 28.2 713.9

507.3 904.8

648.5 807.5

903.5

02004006008001,000change in %

LS (Wage Empl.) LS (Self−Empl.)

Working Hrs. (Wage Empl.) Working Hrs. (Self−Empl.) Mean Labor Inc. (Wage Empl.) Mean Labor Income (Self−Empl.) Tot. Labor Inc. (Wage Empl.) Tot. Labor Income (Self−Empl.)

Tot. Labor Inc. Tot. Value Added

Source: WIOD (2013).

affected the low-skilled and hence the bottom of the income distribution, it is likely that inequalitywithinthe labor share, i.e. within workers of different skills, has increased. By contrast, high-skilled laborers – who earn the highest incomes – could further increase their income shares but still only make up a small proportion of the total work force (see also Francese and Mulas-Granados, 2015).

Finally, I decompose by economic sector. This gives a mixed picture of the labor share (see figure 3.10): The labor share is highest in the sectors of agriculture and transport (over 80 %), followed by the textile and leather industry as well as public administration, educa-tion, health and social work with labor shares of around 70 %. The labor share is lowest in the real estate business (13 %). Most of the industries, however, exhibit a labor share of around 50 %. The low level of the aggregate labor share hence results from the generally rather low labor shares of 50 % of most sectors but also from the fact that sectors with high labor shares only play a limited role in terms of value added. For example, the value added share of labor-intensive agriculture is only 8.3 %, on average, whereas the value added share of the capital-intensive real estate sector amounts to 7.7 %, on average. About half of the industries experienced a decline in their labor share (especially the sectors mining and

quar-Figure 3.9: Mean labor shares and working hours by skill, 1995-2009

0204060Labor Share in % of GDP

1995 2000 2005 2010

year

Low−Skilled Middle−Skilled High−Skilled

0500100015002000Working Hours (in thousands)

1995 2000 2005 2010

Low−Skilled Middle−Skilled High−Skilled

Note: Data is interpolated and kept constant beyond start and end points for this illustration to deal with unbalanced panel data sets.

Source: WIOD Data.

rying as well as postal services and telecommunications) between 1995 and 2009, whereas the other half shows constant or increasing labor shares. A similar picture emerges when only looking at the different branches of manufacturing. Overall, there is no clear trend within sectors regarding factor shares that could explain the decline of the aggregate labor share. However, on average, there is a tendency that sectors with low labor shares increased their share in total value added (mainly stemming from renting of machines and equipment and postal services and telecommunication), while sectors with high labor shares (especially agriculture but also transport) became less important in terms of value added shares (see fig-ure 3.11 ). It appears that there has been a shift away from the agricultural sector towards capital-intensive sectors such as real estate and the renting of machines and equipment. The observed decline in the overall labor share therefore seems to result from a shift away from sectors with high labor shares towards sectors with higher capital shares, rather than changes within industries.

3.6 Synthesis of results and recommendations for data