عملکرد اطلاعات اعتباری، تثبیت و اعتبار بازار: شواهد در سطح بانک از کشورهای در حال توسعه

Paying particular attention to the degree of banking market concentration in developing countries, this paper examines the effect of credit information sharing on bank lending. Using bank-level data from African countries over the period 2004 to 2009 and a dynamic two-step system generalised method of moments (GMM) estimation, it is found that credit information sharing increases bank lending. The degree of banking market concentration moderates the effect of credit information sharing on bank lending. The results are robust to controlling for possible interactions between credit information sharing and governance.

Information asymmetry and poor contract enforcement lead to suboptimal credit market equilibrium (Stiglitz & Weiss, 1981). To the extent that these problems are endemic in underdeveloped countries, financial sector underdevelopment in these countries could be attributed to poor credit information about borrowers. Credit information sharing is therefore expected to facilitate lending decisions (Bennardo et al., 2010 and Pagano and Jappelli, 1993), reduce loan default by increasing borrowers' incentive to repay (Padilla and Pagano, 1997 and Padilla and Pagano, 2000), and increase competition which in turn leads to higher lending (Pagano & Jappelli, 1993). The benefits of information sharing are hypothesised to be particularly helpful in less consolidated or more competitive banking markets, where borrower credit information is dispersed (Marquez, 2002). Although recent empirical interest has been drawn to the potential benefits of credit information sharing on lending decisions, the moderating effect of banking sector consolidation has been largely ignored. In this paper I examine the effect of credit information sharing on bank lending in African countries. I further condition this effect on the extent of banking sector consolidation. This paper focuses on African countries for a number of reasons. The region exhibits record high levels of default. This, coupled with inadequate credit information and poor creditor rights protection, makes lending decisions within African banking markets a difficult task. Unsurprisingly, therefore, African banking markets remain dramatically underdeveloped, even compared to other developing countries (Honohan and Beck, 2007 and Mylenko, 2007). Bank credit to the private sector in the region lags behind that of other regions. The region records the lowest credit penetration in the world (Mylenko, 2007) with less than 20% of households having access to formal banking services (Beck, Demirguc-Kunt, & Levine, 2009). A key feature to which Africa's financial sector under-development may be attributed is weak contract enforcement. With rule of law, regulatory quality, and control of corruption well below the world average, it is unsurprising that it takes an extremely lengthy process to recover bad loans (Sacerdoti, 2005). The high credit risk translates into high interest spreads and margins (Beck et al., 2009). With low banking depth and breadth, as well as high credit risk, the potential benefits of credit information have been appreciated in a few African countries. A few years ago, public credit registries and private credit bureaus were virtually non-existent. In recent times, significant efforts have been made to have operational information sharing systems in a number of African countries. In many of these countries, however, information sharing systems are in their infancy (e.g., Zambia, Nigeria and Ethiopia) and have low coverage. Several other countries are also in the process of establishing operational credit information sharing (e.g., Ghana, Tanzania and Uganda). The effort to establish functional credit information sharing schemes in Africa is consistent with several years of financial sector reforms that have promoted banking competition in the region. With significant reforms across the African financial sectors over the past two decades,1 the region has witnessed significant financial deepening and broadening in recent times (see Allen et al., 2012 and Beck et al., 2009). Compared to developing countries in other regions, however, the pace of improvement is much slower (Allen et al., 2012). The years of reforms have also led to a downward trend in banking sector concentration, which has been characteristically high for the region (Fosu, 2013). Whilst the downward trend in concentration does not necessarily indicate improved competition (Boone, 2008, Boone et al., 2005, Boone et al., 2007 and Demsetz, 1973), it does suggest that credit information is becoming more dispersed as the pool of borrowers per bank becomes smaller (Marquez, 2002). In view of the above-mentioned features, this paper seeks to answer the following questions: first, how does credit information sharing affect lending in developing countries? Second, to what extent does the depth (or the characteristics) of credit information affect lending decisions? Third, to what extent is the effect of credit information sharing conditional on the degree of banking market concentration? The results suggest that credit information sharing improves bank lending. It is also found that the depth of credit information is similarly important in increasing bank lending. Furthermore, it is found that the effect of credit information sharing is higher in less concentrated banking markets. The findings are robust to controlling for several measures of institutional quality and their possible interactions with credit information. The paper contributes to the existing literature in several ways: first, the paper provides the first bank-level (supply side) evidence of the effect of credit information on credit allocation. Bank-level data ensures that individual banks' reactions to credit information sharing are not confounded by aggregate variation in credit allocation. In particular, bank-level data helps to isolate variations in credit allocation arising from (unobserved) heterogeneity of banks. Using aggregated credit data makes it impossible to isolate lending behaviour of specialised banks, especially those that are there to serve government motives. Second, this paper is the first to provide empirical evidence about the moderating effect of banking sector consolidation on the benefits of credit information sharing. Third, the paper further investigates the extent to which a wider range of institutional factors interacts with credit information sharing to impact on credit allocation. Finally, this is the first paper to attempt a comprehensive study of credit information sharing and bank lending in African countries. The rest of this paper is organised as follows. Section 2 provides a review of the theoretical literature and empirical evidence that motivates this study. Section 3 outlines the research hypotheses. The data and variables used for the study are described in Section 5, whilst the empirical estimation methods are provided in Section 4. The findings of the study are discussed in Section 6. Section 7 concludes the study.

Using bank-level data, the results from this paper suggest that credit information sharing increases bank lending. Moreover, this study finds that the increases in bank lending arising from credit information sharing decrease with banking market concentration. The results are robust to alternative measures of credit information sharing and banking market concentration. Whilst banking market concentration may signal less dispersion of credit information, the evidence in this paper suggests that this informational advantage does not outweigh the distortion of optimal credit market performance caused by banking market concentration. Given the wave of regulatory reforms across many banking markets in developing countries, which have already opened up the domestic banking markets for entry of new and foreign banks, the evidence suggests that embracing or deepening credit information sharing will help boost financial development in these countries. The evidence further suggests that policy makers cannot necessarily view quality governance as a perfect substitute for ensuring better access to credit information. Even though the benefits of credit information sharing decrease with the quality of governance, some positive benefits still accrue from information sharing even at very high levels of governance. This is consistent with the fact that, even in developed countries where rule of law, for example, is robust, credit information sharing is advanced. Hence, the findings of this paper implore developing countries to strive to achieve effective and efficient credit information sharing schemes alongside the ongoing regulatory reforms and the promotion of quality governance.