Tyranny of numbers

There have been reports of rising inequality under the Ahmadinejad’s administration (for example, in the Persian sites of Rastak and Aftab), which, unlike their claims for rising poverty, are grounded in facts.   Survey data show convincingly that inequality has increased in the last few years, but what has caused it is uncertain and subject to dispute.  The popular explanation (popular among reformists) for the rise in inequality in recent years is, of course, President Ahmadinejad policies.  But there is a deeper, somewhat related, explanation which should not be overlooked– the oil boom itself.  Deciding which explanation is more important goes to the heart of political economy questions that have occupied many minds in Iran in recent years.

If it is policy, then changing administrations can make a difference and improving inequality is a legitimate issue for presidential elections.  If it is oil, then inequality is more structural and harder to change with policy, at least with one election.  Remedies would require deeper and more slow changes in the distribution of access to oil rents.  Some of the factors that would increase equity in access are things that also improve productivity and access to jobs, such as access to education, health, and basic infrastructure (which now includes the Internet, besides electricity and roads).  There has been much improvements in equity of access to these factors in the last thirty years, but the results have not shown.  Most of the relative improvements have come in rural health and education (see my recent paper), with little to show for in rural incomes, as I showed in my previous post.   This could be because, as Keyvan pointed out, the better educated leave the rural areas so the fruits of rural development all show up as higher incomes (greater equity?) in urban areas.

Understanding the causes of inequality is complicated.  The most famous generalization about inequality is the Kuznets inverted U curve, which suggests that in the long run inequality rises and then falls with economic development, is not helpful for understanding changes in inequality in the short run.   One generalization for Iran, and perhaps most oil exporting countries, to which I have alluded in a recent paper, is that inequality increases during oil booms and falls in downturns.   If true, such a correlation probably has something to do with the way in which oil revenues trickle down the economic chain–a complex mechanism that includes the system of subsidies and the criteria for allocation of government contracts and credit, among others.  Policy matters here because it influences how much and how fast the oil money flows into the country, but the mechanisms themselves are not easily manipulated by policy makers.

Let us take a look at facts.  How did inequality change during the three different administrations since the end of the war with Iraq? I will postpone doing the same with respect to the price of oil.   The figure below shows a very interesting picture:

Source: Author’s calculations from Expenditures and Income Surveys, the Statistical Center of Iran.

The figure shows the average growth rate of per capita household expenditures by decile of expenditures.   For example, during the 1989-1996 period (the Rafsanjani period), average expenditures for all  deciles grew by 2.1-2.3 percent per year.  As a result, inequality stayed constant.  The same happened during the eight years of the Khatami administration, but the growth rates were three times higher.  The third period is different because it shows a distinct rise in inequality: the bottom two deciles grew by about 2 percent while the higher deciles enjoyed more than 3 percent growth in their per capita expenditures.  The result is increase in inequality, which is an increase of about one Gini point (from 0.43 to 0.44).

The temptation to attribute the deteriorating distribution of income to government policies, especially in these election times, is understandably high, and to some extent justified, but caution is advised.  First, one should not ignore the role of rising oil income in increasing the level of inequality.  In general, as more oil revenues trickle down the social and economic pyramid, more may stick at the top than reach the bottom.   Second, once an oil boom is underway, it is difficult for the government to deny the public the right to enjoy its fruits.  You could say that Mr. Ahmadinejad, who had won the 2005 election in part because of his promise to take the oil money to people’s dinner tables, was in a particuarly difficult position in this respect.   This is not an excuse but it is not so unusual for the government to ingore the ex-post austere budget and spend the riches on public sector salaries, subsidies, easy credit and more lucrative government contracts.  The same temptations did not exist under Khatami most whose time was spent oil poor, and during the small oil boom of the Rafsanjani era, 1990-92, the nation did such an overdose on imports and expenditures that it spend the next five years paing for it.  So, the

One final points to take us back to an earlier discussion about poverty.  Notice that in all three periods the expenditures of the lowest deciles rise, meaning that if we measured the incidence of poverty with a constant poverty line (in real terms, of course), despite the rise in inequality, we would arrive at a declining trend in poverty.  So, the controversy about changes in poverty is really about whether or not to use a constant poverty line in different years. As I have argued here before, I am not for recalculating the poverty line every year, especially when using the calorie method, which I believe is very imprecise.   There is less disagreement regarding changes in inequality because its calculation does not involve such judgements on the part of the researcher and is more mechanical–put the same data into any computer, get the same level of inequality every time.