Reconsidering regime type and growth: Lies, dictatorships, and statistics

In this blog, Professor Christopher Magee and Professor John Doces tell us about their innovative research, which sheds light on the exaggerated growth rates that many dictatorships publish. The findings of their research were recently published in full in International Studies Quarterly. Christopher Magee is professor of economics at Bucknell University. John Doces is assistant professor of political science at Bucknell University.

The rapid growth of China’s economy has given new credence to an old idea that the best way to promote rapid economic growth is through heavy-handed authoritarian political systems. The argument is that China’s impressive growth has been carefully guided by the Communist Party’s dictates and policies, which have been able to efficiently provide infrastructure, quell domestic disturbances, and protect foreign investors. The fact that the three fastest growing countries between 1992 and 2006 were all Asian dictatorships (China, Myanmar, and Vietnam) supports the view that some authoritarian governments can adroitly control and manage the growth process. Many have thus concluded that an authoritarian model of growth is the way forward for other poor countries, especially those in Africa.

While we do not contest all the theoretical arguments about how dictatorships may be able to promote growth, we are skeptical of the statistics used to back up these arguments. The problem is that countries’ GDP growth rates in the World Bank and IMF data sets rely heavily on statistics that are provided by national authorities. All governments, we believe, have an incentive to exaggerate the growth rates they report to the World Bank and other institutions, but not all are equally capable of doing so. Specifically, regimes governing in a democracy face many constraints, like a free press or independent government agency, that will call out inflated reports of growth. In contrast, authoritarian regimes, such as those in China and many countries in Africa, face fewer constraints from overstating their growth rates. The incentive to misreport economic data is pervasive in authoritarian regimes: as Sabrie (2012) notes, “Officials at all levels of government are under pressure to report good economic results to Beijing as they wait for promotions, demotions and transfers to cascade down from Beijing.”

There is some evidence from prior research that dictatorships are less forthcoming about reporting economic statistics than are their democratic counterparts. Hollyer, Rosendorff, and Vreeland (2011) find that authoritarian regimes report less data to the World Bank than do democracies. Our question is slightly different, as we consider the quality rather than the quantity of the data reported by governments. Do dictatorships tend to exaggerate the growth rates that they report to international organizations? Are the actual growth rates as high as claimed by many of these governments?

In order to answer these questions, we use a data set of satellite images of nighttime lights. Henderson, Storeygard, and Weil (2012) show that growth in the average intensity of nighttime lights in a country is strongly correlated with its GDP growth. Growth in nighttime lights is an imperfect measure of economic expansion, but unlike GDP data, it is not one that authoritarian regimes would have had any reason to manipulate. Thus, we compare the growth in each country’s nighttime lights to the growth rates it reported to the World Bank. Countries that are exaggerating their GDP growth rates would have officially reported growth rates that are faster than the growth suggested by how fast nighttime lights are increasing in the country. We estimate that dictatorships overstated their GDP growth rates by between 0.5 and 1.5 percentage points per year on average when they reported data to the World Bank.

Beyond the statistical results we look at several cases that further support to our claim. One is Ethiopia, where the Ministry of Finance and Economic Development (MoFED) is the primary data source for macroeconomic variables used by the World Bank. Our colleague Berhanu Nega (2010) notes that until about 2005 Ethiopia’s economic data were thought to be reliable but that since then MoFED has been guilty of overstating growth and other key economic numbers. After 2004, the country reported a dramatic increase in its GDP growth rate and also reported rapid growth (9.5% per year) in agricultural value added despite several years of severe drought and large food price increases that necessitated food aid to about one-sixth of the population. A special report in the The Economist by Oliver August (2013:12) notes that, “Annual productivity gains in agriculture are probably not 5-6%, as the official statistics suggest, but more like 2-3%.” As with China, the report notes that, “An insider says: ‘Officials are given targets and then report back what superiors want to hear.’ International experts are suspicious of the GDP growth figures of 11% flaunted by the government. They say the actual growth rate is only half that, around 5-7%–which is still respectable.” Thus, there seems ample reason to question whether Ethiopian statistics are being reported accurately since 2004.

Our model is consistent with the claim that Ethiopia’s economic statistics were relatively reliable until around 2004 or 2005 but that economic growth has been exaggerated for political purposes since then. Prior to 2004, Ethiopia’s reported GDP growth matched the growth in its nighttime lights extremely well. The intensity of nighttime lights in Ethiopia grew on average by 8% per year between 1992 and 2004 (23^rd fastest out of 168 countries in our data set). Its reported GDP growth during this time was 5.4% (24^th fastest). Between 2004 and 2008, however, Ethiopia’s reported GDP growth more than doubled to 11.3% per year (up to 5^th fastest) while lights growth plummeted to 1.7% per year (89^th fastest).

Our results cast doubt on the economic successes of Ethiopia and other authoritarian regimes. While they may have grown economically, dictatorships have a tendency to exaggerate their growth rates, as our paper has shown. Perhaps most importantly, our results call into question claims that authoritarianism can be used to promote growth and development in Africa.