Corruption is a pervasive phenomenon around the world, and there is increasing recognition that it imposes substantial costs on the economy. Yet, despite considerable research examining the causes and consequences of corruption, there is little empirical evidence on the specific channels through which it operates. Detail has been lacking on how corruption is carried out, what factors promote or hinder it, or how the private gains from corruption compare to its aggregate costs.

The contributions of this paper are that it addresses these questions in detail and in a manner that allows one to make causal inferences and better interpret the results. It uses a loan-level dataset that comprises the universe of corporate lending in an emerging economy in Pakistan, to first establish the presence of political rents in banking, and argue that these are indeed rents and not socially motivated lending.

“Our results highlight how significant a problem corruption may be in emerging markets not just in terms of its prevalence but also the aggregate cost on an economy and its future development.”

It then goes a step further by identifying the means of rent provision focusing on the role of the public sector, and hinting at the checks imposed by the electorate. Finally, the paper is able to use this micro-level examination to provide macro-level estimates of the costs imposed by such rent-seeking behavior on the economy.

Moreover, the methodology used is relatively straightforward and can be replicated in other contexts to examine the role political and other avenues of corruption play in the economies of both developed and developing nations.

The scope and depth of the data used in this study provide several advantages. First, instead of relying on subjective proxies, we have direct measures of a firm’s political connections, defined as the firm having a politician on its board. We can therefore test at the individual firm level if political status obtains preferential lending.

Second, by using firm fixed-effects and hence only exploiting variation within the same firm over time or across lenders, we can account for unobserved firm-specific factors that do not vary over time or across lender types. This allows cleaner identification of the impact of political status on rent provision.

Third, using measures of political strength and electoral participation, we can examine the extent to which rents are affected by the local political environment. Finally, given that we have the universe of corporate lending in the country, we can use our micro-level estimates to back out tentative economy-wide costs of political corruption.

Our results highlight how significant a problem corruption may be in emerging markets, not just in terms of its prevalence, but also the aggregate cost on an economy and its future development.

Next, by identifying a specific means through which such rents are provided and limited, it better speaks to policy reform and offers a deeper understanding of how political corruption operates and may be curtailed in practice.

In terms of prevalence, we find that politically connected firms receive substantial preferential treatment. Not only do such firms receive 45% larger loans, but they also have 50% higher default rates on these loans. This preferential treatment is entirely driven by loans from government banks. Private banks show no such political bias. Moreover, such preference come at a substantial cost to the economy, with even conservative estimates as high as 1.9% of GDP lost each year!

We go beyond these estimates by highlighting a particular means of rent provision, and suggesting the means to check it. Politically powerful firms obtain rents from government banks by exercising their political influence on bank employees. The more powerful and successful a politician is, the greater is his ability to influence government banks. This influence stems from the organizational design of government banks that enables politicians to threaten bank officers with transfers and removals, or reward them with appointments and promotions.

Government banks survive such high levels of corruption because of the soft-budget constraints that often characterize state institutions. However, firms whose politicians run from constituencies with greater voter turnout receive lower preferential treatment, hinting at checks imposed by electoral participation and political accountability.

The paper is a direct consequence of the value of interactions between academics and policy makers. As part of a conference we had organized in the Lahore University of Management Sciences (LUMS) in Pakistan, we invited academics and policy makers to attend the conference. Not only was the conference extremely valuable in allowing policy makers to pose important questions to academics, but, by allowing the two to interact, it provided for opportunities for mutual research.

One such opportunity was offered by Dr. Ishrat Hussain, the then-governor of the central bank, the State Bank of Pakistan (SBP). Having been at the World Bank before, Dr. Hussain was very open to academic collaborations and, over the course of the next year or so, both of us not only spent time interacting with researchers at the SBP, but also worked out the legal arrangement to get access to the detailed data used in this paper.

As we often joke in presenting the paper, "no we were not politically connected in order to obtain such access." Instead, in what we hope to be an increasingly common occurrence in studying developing markets, policy makers are often very open to sharing insights and data with reputable academics, as long as one assures confidentiality and value-added results through such exchanges.

In terms of problems, it definitely takes time in emerging markets to build such relationships, since the standard collaborative arrangements between academics and policy makers are just beginning to be developed. Moreover, the data, while not as bad as one may first suppose, do require substantial consistency checks and detailed cleaning.

While this has been made easier with the computation power and various matching and cleaning algorithms available these days, just as an example, it took us around a year to get the relevant data compiled, matched, and cleaned, in order to be in reasonable shape for the analysis required in this paper.

However, with sufficient patience and a willingness to add value, we believe that not only are such problems overcome, but that this offers a promising direction in empirical work on developing economies. In fact, since then, not only have other researchers obtained access to the data we had, but moreover, researchers, at times even graduate students, have been able to get such access to equally detailed datasets in other economies and produce extremely interesting work.

This paper elaborates on the nature and consequences of political corruption in the form of rents in financial markets by carrying out a detailed micro-level analysis. The techniques used are relatively straightforward and can be replicated in other contexts to examine the role political and other avenues of corruption play in the economies of both developed and developing nations. As such, the policy implications of not only this work, but related papers that we hope to increasingly see, are substantial.

Beyond the immediate results of this paper and their policy import—that political rents are pervasive and costly, yet may be checked through direct policy measures that curtail the rents provided by the public sector, and a more active electorate—there are, even further, less obvious policy implications.

For example, the rents identified in this paper are likely to have an impact on the structure of industry. Differential access and subsidized credit to politically connected firms are likely to affect entry and exit of firms and their competitive strategies in general. Firms may devote resources to seek such rents and build political links. Exploring such effects offers promising areas for further research and policy work.

Another question that arises is how these rents affect the decision to enter politics and the actions chosen by, and success of, politicians. If greater wealth has an impact on political entry and strength, then our results imply a feedback mechanism where influential individuals, particularly the most corrupt, may progressively increase their wealth and influence.

Our results also hint at the importance and robustness of political networks as politicians are able to obtain rents even when not directly in power. They also raise questions about the extent to which political competition imposes checks on rents. Are the excessively corrupt penalized and do rents have to be distributed to retain power? How the nature and extent of rents affect the political and institutional environment presents another interesting direction of future enquiry.

Finally, a positive policy interpretation of our results is that private banks do not provide any political rents and their low default rates suggest they lack such concerns in general. Moreover, they show little evidence of related lending. This lends credence to the government’s current push for privatization, with three government banks privatized since 1990.

However, we should caution that our results do not suggest that full privatization will eliminate rent provision. If government lending is reduced significantly, those with influence may choose other avenues to seek rents. To the extent that constraining the political rents identified in this paper leads to alternative sources of rent extraction, the country may not recover the full cost of corruption identified in this paper.

Understanding the importance and costs of alternative sources of rent-seeking when more common channels are shut down, is an interesting area for future work, especially given that emerging economies are increasingly carrying out such reforms.