Parliamentary election cycles and the Turkish banking sector

Abstract

This paper analyzes the effects of parliamentary election cycles on the Turkish banking system. Using annual bank-level data representing all banks in Turkey during 1963–2007, we present evidence of meaningful differences in the structure of bank assets, liabilities and financial performance across different stages of the parliamentary election cycle. However, we find that government-owned banks’ behavior does not meaningfully differ from that of either domestic and foreign-owned private-sector banks before, during or after elections. Our estimates also show that government-owned banks underperform both domestic and foreign-owned private-sector counterparts.

Introduction

Despite the observation that the role of commercial banks in market-oriented economies is shrinking, it is an undeniable fact that they retain a pivotal role in financial markets. However, when we investigate the structure of banking systems across different countries, we find substantial differences. For instance, in pure market-oriented economies such as the US and the UK, governments have traditionally only played a regulatory role in the banking sector,¹ while in many other countries governments directly control financial resources through ownership of one or more banks in addition to their regulatory functions.² Nations that contain government-owned or state banks base the existence of such institutions on the ability of state-owned banks to finance projects that would not pass scrutiny at private-sector banks. This reasoning is quite controversial as it harbors the seeds of corruption and inefficiency. For instance, state-owned banks can be misused by the governing party, who may direct state banks to channel funds to projects which will benefit those who support the government rather than those which serve the greater public interest. In that sense, actions of state-owned banks in financial markets are scrutinized very closely by the public and by international organizations such as the IMF and the World Bank.

Given the potential controversies embedded in a mixed financial system in which private-sector and state-owned banks coexist, researchers are keen to document the behavioral differences between state-owned and private-sector banks. The questions that researchers investigate vary from the significance of bank ownership in the promotion of economic growth to politicians’ ability to influence election results through state-owned banks’ actions. At the heart of the matter, researchers seek to determine whether state-owned banks fulfill their functions efficiently and effectively or if they promote abuse and economic inefficiency through misallocation of capital.

Recent research points out that state-owned banks in developing countries seem to have lower profitability and higher costs. An earlier study by Molyneux and Thornton (1992), who use a sample of European banks during 1986–1989, finds that government ownership has a positive impact on bank profitability. Yet Barth et al., 2000, La Porta et al., 2002 provide evidence that state-owned banks are associated with lower economic growth. Similarly, Bongini et al. (2002) suggest that political connections may determine governments’ intervention to rescue failing banks. Dinç (2005) investigates the effects of politicians’ influence on state-owned banks, concentrating on bank lending behavior. Using a large cross-country dataset, he finds evidence that state-owned banks increase their lending in election years relative to private-sector banks. In a similar vein, Brown and Dinç (2005) show that prior to elections, failing banks are less likely to be taken into administration. A recent study by Micco et al. (2007) points out that the performance of state-owned banks worsens during election years as they are influenced by political concerns. A broad cross-country study by Hauner (2008) provides evidence of a significant negative effect of credit to government in developing economies.

Studies that concentrate on country-specific data on the behavior of lenders and borrowers also reveal that firms that have political ties with politicians are favored by state-owned banks. Sapienza (2004) shows the effect of political connections on state-owned banks in Italy. Concentrating on Italian firms, he provides evidence that state-owned banks charge lower interest rates than do private-sector banks. He also shows that firms pay lower interest rates in areas where the head of the local state-owned bank has the same party affiliation as the ruling political party. Using data from Pakistan, Khwaja and Mian (2005) provide evidence that low-quality borrowers with political connections can borrow from state-owned banks. Fraser et al. (2006) suggest that Malaysian banks’ leverage is affected by the share of government ownership, informal ties to politicians, and the ownership share held by “institutional investors”, de facto controlled by the government or government-sponsored agencies. Baum et al. (2008) find that politically affiliated banks in Ukraine have significantly lower interest rate margins and that the level of activity of affiliated deputies in parliament has a positive impact on linked banks’ capitalization ratios. Cole (2009), using data from India, shows that state bank credit is 5–10% higher in election years and even more so in heavily contested electoral districts. He also indicates that state bank loans are less likely to be repaid. Other researchers have shown that Chinese state-owned banks are less profitable, less efficient and have worse asset quality than other types of banks (Lin and Zhang, 2009, Berger et al., 2009).

Research on Turkish political business cycles also highlights the possibility of state-owned banks being used for the advancement of governments’ ambitions. Tutar and Tansel (2001) point out that during election periods both the budget deficit and line items of the budget are significantly negatively affected. Alper and Öniş (2004) claim that these ‘deficits reflect the existence of powerful pressures for rent distribution as a major element of Turkey’s political economy in recent decades.’ Similarly, Akalin and Erkişi (2007) find that government expenditures increase during election years over the 1950–2006 period.

We study the behavior of private-sector and state-owned banks in Turkey throughout parliamentary election cycles. Our dataset covers the period between 1963 and 2007, during which Turkey conducted 11 parliamentary elections. Our sample consists of 2158 bank-year observations pertaining to 86 banks. Since the establishment of the republic in the 1920s, Turkish financial markets benefited from state-owned banks as well as domestic and foreign private-sector banks. Hence, the data, collected and made available by the Banks Association of Turkey, are rich and consistent throughout the period of interest. We carry out our empirical analysis using the system dynamic panel data (DPD) estimator which combines equations in differences of the variables with equations in levels of the variables (see Blundell and Bond, 1998).

Our study differs from the earlier literature on several dimensions. First, our study provides a thorough investigation of all types of banks in Turkey over election cycles rather than a knife-edge comparison of banks’ behavior between election versus non-election years. Next, we also concentrate on the behavior of a number of bank finance ratios, their growth rates and performance measures rather than a single factor. Finally, by focusing on a single country’s entire banking sector, we scrutinize a consistent dataset which is not prone to sample selection bias or accounting problems that might have affected the results presented in earlier research.

Several findings emerge in a setting where confounding factors are taken into account within the framework of a dynamic model. Although Turkish election cycles significantly affect bank behavior, these effects do not differ meaningfully across state, domestic and foreign-owned private sector banks. In particular, we find no evidence that state-owned banks increase their lending in comparison to other bank categories before, during, or after elections. This is an interesting finding, and stands in contrast to earlier research that has used cross-country or country-specific data which points out that state-owned banks behave differently during election years. Furthermore, our regression results also show that election cycles do not lead to a differential impact on other key financial ratios across banks. The second set of results, in line with earlier research, indicate that state-owned banks are less efficient in comparison to both domestic and foreign-owned private sector banks.

The rest of this paper is organized as follows. In the next section we describe parliamentary elections in Turkey followed by a short summary of developments in the Turkish banking sector. Section 3 presents the data. Section 4 lays out the econometric model and estimation results. Finally, Section 5 concludes.