By Christopher Blattman
If corruption hurts growth, why have economists not found much evidence for it?
Should foreign donors care about corruption in developing countries? Or is corruption an Anglo-American fetish? Corruption infuriates me as much as the next person. It is despicable, pure and simple. But Westerners care about corruption far out of proportion to its impact on poverty alleviation and economic growth.
Recently, the British Prime Minister David Cameron wrote in a Wall Street Journal op-ed that eliminating corruption ought to be the prime focus on aid and development strategies worldwide, along with ending civil wars or improving property rights protection. Just last week, at a UNDP conference on anticorruption, a UNDP official argued that corruption is one of the biggest obstacles to achieving the Millennium Development Goals.
This is absurd.
To be clear, my point is not that corruption is unimportant. Of course it is. But if we are talking about where the UN ought to focus its aid energy for the next 15 years, or what ought to be the top priority of a government like Uganda’s, never would I use corruption in the same sentence (or even paragraph) as civil war or property rights.
Here is a policy litmus test I use: If you believe issue X ought to be a top priority, you need to believe two things: (A) it matters a lot for development, and (B) you can do something about it.
When it comes to corruption, I don’t know a whole lot about B. But a lot of economic research has gone into A, and corruption does not jump out as very important.
We usually think of corruption as sand in the wheels of prosperity. This is surely true. Corruption reduces the incentive for a small entrepreneur or a big firm to invest. This robs a nation of industry, and with it jobs and growth. In many ways, corruption is like a tax on business, one that seldom finds its way into providing public goods to society.
This logic is so persuasive that economists have been pretty surprised not to find much evidence. The economists Nauro Campos and Ralitza Dimova reviewed 41 different cross-country statistical studies of corruption and development in 2010. Two-thirds of the studies do not even find a negative correlation.
Cross-country studies have mostly bad data and empirics, so we should not rest here. But the economist Jacob Svensson—famous for his studies of corruption in Ugandan education—has written about the broader evidence and draws the same conclusion: there’s not much to show that corruption reduces economic growth on net.
Why might this be so?
One reason: Most of us fail to imagine that corruption can also grease the wheels of prosperity. In places where bureaucracies and organisations are inefficient (meaning entrepreneurs and big firms struggle to transport or export or comply with regulation) corruption could improve efficiency and growth. Bribes can act like a piece rate or price discrimination, and give faster or better service to the firms with the highest opportunity cost of waiting.
In theory, this could improve overall efficiency in inefficient environments. If bribes subsidise large chunks of the government, then corruption reduces the need to collect taxes and allocate government spending efficiently, which is a difficult and expensive task in poor countries. The “tax” that corruption imposes could be more efficient than the seemingly clean alternative.
Most of the economists I have read conclude that the gains from corruption are probably outweighed by the losses. But, looking at the numbers, it seems the losses do not outweigh them by much (assuming either matter much at all). No evidence, though, that it ought to be policy priority one, two or three.
A second reason corruption might not cause underdevelopment is that it is a symptom of a deeper problem and not a cause of poverty itself. If you listen to political scientist (and former Liberian President) Amos Sawyer, the real problem facing African societies is the lack of any real constraints on the President. Over-centralized power, and the absence of real checks and balances, is the root of governance failures. This is a common theme running through the history of development in Europe, China, the US and, now, Africa.
If over-powerful Presidential systems are responsible for corruption (plus all the other nasty things that impede growth), then anticorruption programs don’t tackle the fundamental problem. In fact, I think fighting corruption has become a distraction from the real issues.
So why do world leaders like David Cameron care so much about corruption? One reason, I think, is the theoretical case is so compelling. Another is the cultural fetish I mentioned—people react to bribery in the same instinctive and emotional way that they react to people jumping the queue or drivers who do not yield to pedestrians.
I am willing to bet three-quarters of my readers felt a flash of outrage and a churn in their stomach when I suggested that corruption could be efficient. Did you? I understand the rage, but I’m not sure it makes for sound policy.
It is because of this fetish that I think Cameron’s inside voice sagely says the following: “If even a little bit of British aid money is stolen, the British public will yank it all away.” I think that’s the real reason corruption is in his top few priorities.
Let’s return to my hasty policy litmus test, where policies need to: (A) matter, and (B) be fixable. On part A, we might conclude corruption matters for growth only because the uptight American and British publics will drop support for things that do help development because some dictator’s son drives a Ferrari in Cannes. Not the best case I have heard for a policy priority, but (I admit) one I would follow were I in Cameron’s shoes.
What about part B of the formula? Are we any good at stopping corruption? I don’t pretend to know. But we will need to be realistic about the pace of change. I worry most that the US or UK or UN will set unattainable goals, fail to meet them, and then punish themselves and (more importantly) the aid recipients.
Harvard scholars Lant Pritchett, Michael Woolcock, and Matt Andrews have an original take on this point. They run the following thought experiment: What if we tried to measure the quality of governance (including corruption) over time and space, and figure out the 20 fastest improvements countries have made in recorded history?
Then, using this best-case-scenario, we ask how long it will take today’s poorest and weakest states just to get to the median level of governance–not to the level of a Canada or Sweden, but to the (decidedly less ambitious) level of a Tanzania or Guatemala. The answer, depending on the country, is 15 to 35 years. And that’s if the corrupt countries experience a miracle of change.
The goal the US and UK commonly set for the Afghanistans, Ugandas, and Liberias of the world, of course, is about three to five years. Just witness the public and political angst every time the New York Times uncovers corruption in its Afghan regime. Be prepared for more angst.
If this saddens you, just remember my first point: the corruption they can’t get rid of probably does not matter much for growth, and policymakers should probably be putting their attention elsewhere.
So your despair is part of the solution.
Christopher Blattman is an Assistant Professor of Political Science and International Affairs at Columbia University. He writes at http://chrisblattman.com.