IMF Survey: What are the relationships between land inequality, on the one hand, and institutions, financial development, and public provision of education, on the other?
Erickson: The literature has broadly seen institutions as crucial in determining economic growth or, at least, it has identified a powerful relationship between the two. But, to our surprise, we were unable to identify any robust relationship between political institutions broadly defined and land inequality. Nor did we see any strong correlation between financial development and land inequality. We were able to find some negative relationship between public provision of education and land inequality, but only if we measured inequality across the entire agricultural population.
We are of the view that, all other things being equal, if a relationship that holds in a micro setting is important, then it should also appear in the macro data. If it does not, this raises some important questions. We do not answer these questions in our paper, but, in addition to bringing new data to the question, we argue that, if other researchers are going to claim that these relationships are important, the onus is on them to explain why we are not seeing this in the cross-country data.
We think this research is part of a productive move in the macroeconomic literature toward trying to build more bridges with areas that have traditionally been considered the province of microeconomics. There is strong resistance by many microeconomists to thinking about these questions at all in a cross-country setting, but these are the sorts of intellectual debates that make our work as economists interesting.
IMF Survey: In what ways were your research approach and methodology unique?
Erickson: As a first step, we came up with a measure of land inequality that included the landless. This is the paper’s main contribution. Previous measures had only looked at households that owned land—largely because of the lack of data on the number of landless households.
IMF Survey: How did you select the countries in your sample?
Erickson: We were able to obtain data from the UN Food and Agricultural Organization (FAO) for 82 countries, but for most of our analysis we used data for 45 or so of these countries. The period under study is post-World War II, mostly from the early 1960s through the late 1980s. We excluded the more developed countries, with a few exceptions, because we had strong suspicions that agricultural inequality in western Europe and in Japan during the postwar era had been driven by other factors, like agricultural subsidy policy.
IMF Survey: How were you actually able to measure land inequality across agricultural populations to include the landless?
Erickson: The standard way of measuring inequality is to construct a Gini coefficient. The larger the Gini coefficient, the greater the inequality measured within the particular population. The previous literature had used FAO data to determine the inequality of landholdings as measured by a Gini coefficient within the group of landowners. But a major shortcoming of this methodology was that it did not measure inequalities across landholders and landless in a country. Taking an oversimplified example, imagine a country with two landowners and each has a farm with 5 million acres and thousands of peasant farmers working the land. If you were to measure only inequality among the landowners, the data would show a country that has perfect land equality. This is certainly at variance with common-sense conceptions of inequality.
To address this, we construct what is probably the simplest measure possible. We don’t have information on the number of landless households, but we do have information on overall agricultural populations. So we take the size of the overall agricultural population and divide it by the number of landholdings. This improves substantially on the previous measure in that it does include the landless, but it is nonetheless very crude. We hope that in future work we can come up with a more sophisticated measure, but this is a first step and an improvement on what has been done before.
IMF Survey: William Easterly found that income inequality is linked to poor institutions. Aren’t your findings—which indicate that there is no relationship between land inequality and institutions—inconsistent with this? Aren’t land inequality and income inequality very closely related?
Erickson: It is inconsistent, and we were quite surprised by our findings. One possibility is that somehow income inequality has effects on institutions, yet agricultural inequality, for some reason, does not. However, this seems counterintuitive to us. We do not think that we are overturning Easterly’s results. Rather, we are raising questions about what the causes could be. This inconsistency in our findings could also indicate that we have our own data problems. For example, we may have looked at the wrong time period.
IMF Survey: You find scant evidence that land inequality and financial development are linked. But wouldn’t there be a link given that those who do not own land lack an obvious source of collateral for obtaining credits to finance business ventures, which, in turn, hinders private sector development and, by implication, further financial sector development?
Erickson: The paper discusses cases at the micro level where high land inequality could lead to low financial development, and it also discusses cases where low financial development could lead to high land inequality. We find all of these arguments to be quite plausible, and we agree that causality could run both ways. Although we are not claiming to disprove these arguments, we are raising significant questions based on the premise that if these relationships exist at the micro level and they are important, they also should show up in the macro data. But they don’t.
IMF Survey: Your findings seem to show that whatever strong effects land inequality has on education are mitigated over time. What might explain this?
Erickson: Over the course of the time period we looked at, the provision of education improved dramatically for all countries. We try to correct for this effect, but we may not have completely succeeded. Beyond this, we are not quite sure why the strong effects that land inequality has on education are mitigated over time. In the postwar era, the public provision of education skyrocketed, and by the end of the 20th century it had, thankfully, become more difficult to find countries where large numbers of children did not attend school. This goes beyond the scope of our work, but one of the important ways to move forward on this topic is to develop better measures of cross-country land inequality historically. Getting meaningful land inequality data going back before World War II and even into the 19th century would, however, entail comprehensive archival work.
Copies of IMF Working Paper No. 04/158, “Dimensions of Land Inequality and Economic Development” are available for $15.00 each from IMF Publication Services. See page 16 for ordering information. The full text of the paper is also available on the IMF’s website (http://www.imf.org).
Try Finance & Development
the quarterly magazine from the IMF that examines economic policy, globalization, poverty reduction, trade, and development issues.
Published in English, French, Spanish, and Russian online at www.imf.org/fandd
Print versions also published in Arabic and Chinese.