Growth of labour value added per worker

Labour value added per worker is computed as the sum of all wages and net profits divided by the number of persons employed.⁹ Table 2.2 shows the growth in labour value added per worker in Uganda and Peru between 2005 and 2009. The estimates are well in line with the overall economic performance of both countries. In Peru, labour value added per worker (in constant LCU¹⁰) grew by about 9 per cent annually between 2005 and 2009 (from a 2005 baseline of Int. $3,424 in PPP), while the annual growth rate of gross national income (GNI) per capita was 6.3 per cent (World Bank 2016). In Uganda, with a per capita income of only Int. $353 in 2005, the estimated annual growth of labour value added per worker (in constant

9 We count all smallholders in both Uganda and Peru as employed. While this may partially conflict with the new labour statistics practices, it is in line with the definitions used when the surveys were carried out. Yet, it should be noted that – even in Uganda – the majority of smallholders produce for the market.

10 Local currency units, i.e. Peruvian nuevo soles and, for Uganda, Ugandan shillings.

31

LCU) was 14.1 per cent over the period – considerably higher than the economy’s per capita GNI growth rate of only 4.8 per cent (ibid.). Closer inspection of the Ugandan data reveals that this strong growth was driven by labour value added in agriculture, while labour value added per worker in non-agricultural self-employment actually declined.

Table 2.2. Labour value added per worker in Uganda and Peru, 2005 and 2009

Country 2005 2009 Annual growth

(%)

Uganda

In constant 2005 Int. $ 353 451 6.3

In constant LCU 263,009 446,320 14.1

Peru

In constant 2005 Int. $ 3,424 4,504 7.9

In constant LCU 5,650 7,656 8.9

Note: LCU refers to local currency units (Ugandan shillings and Peruvian nuevo soles).

Source: Authors’ calculations based on UNHS 2005/06 and 2009/10 and ENAHO 2005 and 2009.

Technically, the differences between the PPP-adjusted values and the values in local currency for Uganda are noteworthy. The much lower growth rate of labour value added in constant Int. $ can be explained by major changes in the PPP conversion factors. While the figures in Int. $ are useful for cross-country comparisons, the analysis of within-country changes over time should be complemented by indicators measured in local currency. Finally, this exercise also illustrates that a meaningful aspiration level for this indicator might be the projected growth rate of GNI per capita.

2.5.2 Working poverty rate

In the next step, we compute the incidence of working poverty in Uganda and Peru. Table 2.3 shows the headcount poverty ratios and the WPRs for both countries in 2005 and 2009, using different approaches. The results of calculations based on national conventions are contrasted with those obtained from the World Bank’s approach to computing the internationally comparable poverty rates (applying the 1.25 Int. $/day poverty line to monitor MDG 1). For

32

the purposes of its official national poverty statistics, Uganda uses a consumption aggregate per adult equivalent and regional poverty lines that further distinguish between rural and urban areas. Peru bases its estimates on consumption per capita and also uses regional poverty lines. The regional poverty lines can differ considerably: in Peru, the highest is 52 per cent higher than the lowest, and in Uganda it is four times higher. Depending on the context, the consideration of equivalence scales and regional price differences is known to matter for measuring poverty and, accordingly, working poverty. While these are only two of the problems raised by the calculation of internationally comparable (working) poverty rates, a more extensive discussion of the issues goes beyond the scope of this article. Yet, we consider the advantage of having internationally comparable indicators to outweigh those disadvantages.

Table 2.3. Working poverty rates in Uganda and Peru, 2005 and 2009 (percentages)

Uganda Peru

2005 2009 2005 2009

Poor (headcount – national poverty line) 29.3 24.5 55.6 33.5

Poor (headcount – Int. $1.25) 44.6 35.7 5.6 2.2

Poor (headcount – Int. $2) 70.1 64.4 16.8 8.7

LFPR 77.0 77.7 71.7 76.9

Total number of persons employed 9,799,816 11,432,223 13,107,577 15,418,822

WPR (national poverty line) 27.9 22.5 55.7 32.0

WPR (Int. $1.25) 45.9 29.7 5.5 2.0

WPR (Int. $2) 73.3 60.2 17.0 8.2

Note: LFPR and WPR refer to the labour force participation rate and the working poverty rate, respectively.

Source: Authors’ calculations based on UNHS 2005/06 and 2009/10 and ENAHO 2005 and 2009.

Table 2.3 shows that poverty decreased in both countries, and the results illustrate the importance of the choice of applied methods and poverty lines. Uganda’s national poverty statistics suggest a moderate decline in the headcount ratio between 2005 and 2009. Using the international poverty lines, however, the reduction of extreme poverty (less than Int.

33

$1.25 per capita per day) appears to have been much more pronounced than that of poverty according to the higher poverty line (less than Int. $2 per capita per day). In Peru, the national poverty line is much higher than the international poverty lines. Albeit at different levels, however, all of the country’s poverty indicators show a drastic decline over the period under consideration.

Some very interesting patterns emerge with regard to the WPR, illustrating the possible value of this indicator as a complement to poverty measures. For Peru, the WPR is very close to the headcount ratio, i.e. the share of poor workers among all workers is similar to the share of poor people in the overall population. This holds for all indicators and both years, implying that progress in poverty reduction correlates with progress in workers’ incomes. In Uganda, in contrast, this only applies to 2005; in 2009, the WPR was between two and six percentage points lower than the headcount ratio. This means that the working population fared better than the non-working population and that income from work enables people to escape poverty more effectively than do other sources of income. This is consistent with the above finding of a considerable increase in labour value added per worker in Uganda; it is also likely to reflect the country’s higher labour force participation rate, with more household members contributing to household income with their labour. In general, the fact that overall poverty rates in Uganda are higher than the country’s working poverty rates – while being similar in Peru – is likely to be related to differences in the composition of household income sources between the two countries. In particular, non-labour incomes (e.g. old-age pensions and other transfers), which partially sustain households without employed individuals, are more common in Peru than in Uganda. In other words, in economies without social support programmes people are more dependent on decent labour income.

2.5.3 Workers earning less than absolute and relative minimum labour incomes