For my friends, anything; for my enemies, the law.
Oscar R Benavides, President of Peru, 1914-1915 and 1933-1940
Policy. Policy. Policy. One of the most widely used words in applied economics. But what if – at least in Africa – we have no idea what we are talking about. What if it we cannot really specify what “policy” is in terms of its actual effects on how firms do business and interact with governments?
Economists mostly talk about public policy in the common sense use of the word, i.e. policy as a mapping from states of the world to actions by an agent. An insurance policy is mapping from states of the world (“did a tree fall on your car?”) to actions by the agent (payment of a claim). Public policy just means the policies are implemented by agents of the state. Central banks control monetary policy – a mapping from certain indicators to actions over the instruments of their control.
This view often takes the details of implementation for granted. Outside of the consideration of the inter-temporal consistency of policy, the economics literature mostly assumes that if there is a sales tax of 5% sales then each firm pays 5% of sales (or the exemptions to that rule are also rules, like “food items are exempted.”).
An experimental study (Bertrand et al. 2007) of a very simple administrative process implemented in every country – getting a driver’s license – in the city of Delhi casts doubt on the whole notion of policy as rules, or at least as the published rules. The study approached people entering the building to get their driver’s license and asked them to participate in a study of the process. After the potential license seekers got past the experimental crew, they encountered the touts who, for a fee, mediated the process. Most of the control group hired a tout. Of the control group who hired a tout only 12% had to do the driver’s test that assessed the applicant’s ability to drive. Of the control group that did not hire a tout 94% had to take the driver’s test. The study tested the control group individuals who received a license by hiring a tout and without taking the test – the vast majority could not drive.
What is the policy for getting a driver’s license in Delhi?
The policy on paper is identical to nearly every other place. Prove your age, identity and residence, prove you are a knowledgeable and competent driver, and the agent of the state issues a license.
So, what is the real
policy for getting a driver’s license in Delhi?
Clearly the policy in practice is “hire a tout and the part about a driving test is nearly always waived.” The real policy includes the availability of “deals” – individual specific deviations from the implementation of the legal policy based on states of the world nowhere specified – but often widely known. The flip side is, “If you don’t hire a tout, you must take your chances with the paper implementation”.
What has this got to do with firms in Africa?
The World Bank has supported interviews of firms around the world, including more than thirty countries in Africa (some multiple times). A standard section of the interview asks firms to name, from a prompted list, what they believe are the barriers to their growth. From these lists emerge the usual complaints of firms. Firms often complain access to and cost of finance, and there is now a massive literature both macro and micro about how much firm growth is inhibited by access to finance. Interestingly, in the early rounds of the surveys, the worldwide number one obstacle identified by firms was “policy uncertainty” (Smith and Hallward-Driemeier 2005).
In our work on the Africa data we find that it is a strong number two (see Figure 1), right after electricity but ahead of the usual litany: tax rates, macroeconomic instability, access to finance, corruption, etc… (Hallward-Driemeier et al. 2010).
Figure 1. “Policy Uncertainty” is the second most frequently cited obstacle to firm growth in Africa – what does that mean?
But that just raises the question, “what is this ‘policy uncertainty’ firms hate?” Is it unpredictable inter-temporal changes in the rule? Or, they could be taking a firm level view and regard policy, not at the high level mapping, but policy is what happens to them. With weak capability for implementation even if the policy doesn’t change, there could be a great deal of uncertainty for individual firms about what the agents of the state are going to do. They may ask themselves: what are the deals available to me?
Our paper explores this interpretation of policy implementation uncertainty as what firms hate. The first point is that when firms are asked about how policies affected them, like how long it took them to get an operating license or construction permit or get goods through customs what was striking was the variability within countries compared to that across countries. Figure 2 shows the days firms reported it took them to get an operating license. The range of the median firm response across countries was 20 days – the median firm reported 1 day in Rwanda and 20 days in South Africa. But the spread within countries, the difference between the 90th and 10th percentile of responses, was over 20 days for each country. So, knowing the average “policy” for a country explains almost none of the firm variance.
Figure 2. Typically the variability in implemented policy outcomes across firms in the same country is much larger than the entire range across countries
We then use the variability across firm responses within in the same sector, in the same country, in the same size city, of the same size, to proxy for “policy uncertainty” to show that these measures of policy uncertainty are as predictive of actual firm growth in employment as are the levels of “policy.” We also use “differences in differences” to show that greater policy uncertainty has a differentially large effect on government intensive sectors (benchmarked for Germany, a country with strong implementation) in firms that are particularly susceptible to uncertainty.
Bringing the role of policy implementation into consideration, particularly in weak environments for implementation, can potentially change how we think about the implications of “policy” and help think about, though by no means single-handedly resolve, several puzzles.
One puzzle is the apparently quite different overall growth responses to apparently similar policy changes. Why is it that announced, but initially quite modest actual steps towards “free markets” produced massive growth responses in China, in Vietnam, and in India while many other countries received nothing from their “policy reform.”
Back to the driving license example, suppose that the “policy” shifted to change the content of the driving exam – adding “demonstrate U-turns” to the list of skills to be displayed by drivers taking the test. What impact would this have on actual skills of those getting licenses? In a deals environment, it is hard to say. There is a chance that very little would change, only the basis for deals and hence the equilibrium price for touts. This view of policy implementation uncertainty facing firms in “deals” environments brings the problem of the credibility of policy reform out of the inter-temporal problems of time-consistency into the micro level. Does “policy reform” really change what firms expect to happen?
A second puzzle is the collection of “institutions rule” papers that show that controlling for institutional quality drives “policy” measures out of growth regression (Acemoglu et al. 2003, Rodrik et al. 2004, and Easterly and Levine 2004). Once policy implementation is taken into consideration it is obvious that one cannot sensibly talk about the impact of “policy” without discussing how the policy mapping translates into actual actions. But this is, at least in part, what is meant by “quality” institutions of governance – that the rules are followed and the deals are not wide open.
A third puzzle is the Harberger triangle or growth impacts puzzle. It is hard to get the standard measures of the welfare losses from specific policy distortions up to much more than 5% of GDP yet, as Hausmann et al. (2007) show, growth accelerations in which output increases by more than 50% over a decade or less are common. As Ed Glaeser (1996) showed for rent control, if allocations are person specific – i.e., a specific person benefits from rent-controlled housing – and the price mechanism is not fully functional then the usual triangle representation of the distortion is completely wrong. In a “deals” environment, policy actions are firm specific, not rule specific. The consequences of policy reform in the deals world can be much lower or much higher than standard triangles would suggest.
Acemoglu, D., S. Johnson, J. Robinson and Y. Thaicharoen (2003), “Institutional Causes, Macroeconomic symptoms: Volatility, Crises, and Growth.” Journal of Monetary Economics. 50(1), pp 49-123.
Bertrand, M, S. Djankov, R. Hanna, S. Mullainathan (2007), “Obtaining a Driver’s License in India: An Experimental Approach to Studying Corruption.” Quarterly
Journal of Economics. 122(4). Pp. 1639-1676
Easterly, W. and R. Levine (2003), “Tropics, Germs and Crops: How Endowments Influence Economic Development,” Journal of Monetary Economics. 50(1), pp. 3-39
Edward L. Glaeser, 1996. "The Social Costs of Rent Control Revisited," NBER Working Papers 5441
Smith, Warrcik and Mary Hallward-Driemeier (2005) “Understanding the Investment Climate.” Finance and Development
Hallward-Driemeier, Mary, Gita Khun-Jush, Lant Pritchett (2010), “Deals versus Rules: Policy Implementation Uncertainty and Why Firms Hate It”, NBER Working Paper 16001.
Hausmann, R., L. Pritchett, and D. Rodrik (2005), “Growth Accelerations.” Journal of Economic Growth. 10(4), pp. 303-329
Rodrik, Dani, A. Subramanian, and F. Trebbi (2004), “Institutions Rule: The Primacy of Institutions over Integration and Geography in Economic Development.” Journal of Economic Growth. 9(2), pp. 131-165