The issue of social inclusion in Turkey is a controversial one. In this blog, I want to present some data that suggest Turkey experienced inclusive growth over the past decade or so. My colleagues and I have shared this basic story with a number of audiences in Turkey and often the reaction is disbelief. So what does the data say?
The bottom 40 percent can look up
I use three pieces of evidence to make my case. The first is based on recent work by Joao Pedro Azevedo and Aziz Atamanov of the World Bank on shared prosperity. Joao Pedro and Aziz’s work is ongoing and much richer than what I want to present here. So let me just focus on the following chart, which shows the growth of consumption of the bottom 40 percent in Turkey between 2006-2011 and in a number of other countries during roughly the same period. Turkey looks reasonably good albeit not exceptional. The rate of consumption growth of the bottom 40 percent was just over 5 percent, around 0.2 points below the rate of growth for the average. What this means is that during this period of significant global economic turbulence the average welfare of the bottom 40 percent improved by more than one quarter. This was better than India, Indonesia, or Mexico, albeit worse than Brazil, China and Russia.
Further analysis of the Turkish data shows that he two main drivers of income gains in Turkey are the labor market and pensions. For the bottom 40 percent, labor earnings and increases in employment together account for almost half of all income gains even during the crisis period. Pensions account for another fifth. This can be related to government policies, as minimum wages were raised during the crisis, jobs were protected through short-term work schemes and reduced social security contributions for women and labor force entrants, and pension coverage was extended. But unlike in much of Latin America, redistribution through social transfers does not account for a large share of improvements in the welfare of the bottom 40 percent. This is consistent with inequality failing to improve during this period. Growth can be inclusive without redistribution.
Closing the gap in health and education
The second piece of evidence I presented was the improvement in health and education coverage and resulting reductions in inequalities of health and education outcomes across socio-economic groups. This is well documented in other reports even if not always well known (see for instance: Universal health coverage in Turkey: enhancement of equity; and Promoting Excellence in Turkish Schools). Just a couple of statistics to make the point: infant mortality rates for children whose mother had no education were almost three times higher than for children whose mother had at least secondary education in 2003 – by 2008 the difference had vanished. Infant mortality in the lowest asset quintile was 31.1 per 1000 live births higher than in the highest quintile in 2003, but this narrowed to just 11 by 2008. It should be noted that these improvements extended gains made since 1990, but the reduction in inequalities in access and health coverage is nonetheless remarkable.
In the case of education, between 2003 and 2009 PISA results for the bottom quintile improved by 20 points on average across reading, maths and science and by 25 points in the second lowest quintile, but only by 3 points in the top quintile. The bottom 1 percent of achievers, which are predominantly drawn from the lower socio-economic strata, even saw improvements of between 25-33 percent in their PISA results. This is equivalent to around three quarters of a full academic year of additional schooling.
More women are getting to work
The third piece of evidence, and here it becomes most contentious, is the fact that the share of women in in Turkey who work has been increasing since a low point in 2005. This is because around that time the number of women in salaried work started to increase faster than the number of women in agricultural and informal employment declined. In fact, there were some increases in informal and agricultural employment of women during the crisis, but the share of women in salaried work has risen continuously since the 1990s. Together with evidence that employment rates among more educated women are higher and the fact that education levels among women are increasing, I argued that we may be seeing a trend towards gradually increasing female employment.
The evidence is not in dispute. So why does the suggestion that Turkey had experienced inclusive economic development often meet with incredulity? I offer two possibilities. First, while I have looked at improvements over time, others may compare Turkey with the standards of advanced economies – for me the glass is filling, for them it is still less than half full. For instance, Turkey’s income inequality with a Gini of 0.43 in 2011 is higher than in all EU countries. Turkey’s average PISA ratings are still around one full year, or 40 points, behind the OECD average, and of course Turkey’s female labor force participation rate is much below the level of even other middle income countries. Second, perhaps the concept of inclusion is broader than what economic and social statistics convey. It may have something to do with a sense of being included in a national dialogue, from which some groups may feel excluded.
We are currently engaged in an exercise to draw lessons from Turkey’s experience. The fact that economic development seems to have been inclusive is one of the possible emerging messages. I’d be curious to hear your views whether this is correct and how if at all it should be qualified.