Determinants of bank credit in pakistan

The growth in bank credit to the private sector is used as dependent variable whereas growth of liabilities from abroad, growth in domestic deposits, money market rate, M2 as percentage

DETERMINANTS OF BANK CREDIT IN PAKISTAN

“Determinants of Bank Credit in Pakistan: A Supply Side Approach”

Kashif Imran University of Karachi, Karachi Mohammed Nishat, PhD Institute of Business Administration (IBA), Karachi

Abstract

This study empirically identifies the factors which explain the bank credit to the businesses in varying financial environments and emerging global challenges The growth in bank credit to the private sector is used as dependent variable whereas growth of liabilities from abroad, growth in domestic deposits, money market rate, M2 as percentage of GDP, real economic growth, inflation and the exchange rate are identified as major explanatory variable to explain the behavior of bank credit With the major focus on the supply side this study uses the ARDL econometric approach using annual data from the period 1971 to 2008 for Pakistan The empirical results indicate that the foreign liabilities, domestic deposits, economic growth, exchange rate, and the monetary conditions have significant impact on banks credit to the private sector in Pakistan, particularly in long run Whereas the inflation and money market rate do not affect the private credit Moreover, in short run the domestic deposit does not influence private credit The reason may be that the banks do not issue immediate loan from currently deposited amount by account holders The results also infer that the financial health and liquidity of the banks play a significant and vital role in the determination of loan A strong economic condition measured by GDP, as motivating factor to banks has statistically significant impact on issuance of more private credit to businesses Results also indicates that the long run relationship is stable and any disequilibrium formed in the short run will be temporary and get corrected over a period of time with a high speed of 53.5 percent per year This study does not statistically distinguish the behavior of bank credit during non-financial (1971-1989) and financial reforms periods (1990-2008) in Pakistan

JEL Classification: G21, E44, E51

Key Words: Bank Credit, Reforms, Pakistani

A strong financial system is essential for economic growth, whereas a developed economy support to sturdy financial system, thus it becomes a two way process A well developed businesses as well as economic growth raise demand for credit and lead to credit growth The mild monetary conditions and a vigorous banking sector tend to enhance more credit The underlying financial market imperfections create borrowing constraints, hence lower economic and credit growth In the case of Pakistan the domestic credit by banking sector has declined from 51.1 percent of GDP in 1971 to 46.8 percent in 2010 (World Development Index, 2011) As shown in figure 1a (the trend line exhibits a declining pose) The various factors influence on banking decision to allocate the credit in the economy beside of investors’ own characteristics e.g unstable political environment of the country, unstable government economic policies, and

the legal risk The figure 1b show that before the financial liberalization in Pakistan (before 1990) the credit by banks was increased (as trend line show upward slope), however, in post liberalization period the credit ratio was declined (downward slope can be seen in figure 1c) So,

it can be concluded that the financial liberalization/ reforms has negative association with credit growth in the case of Pakistan

Figure 1a: Domestic credit provided by banking sector (% of GDP) from 1971-2010

Figure 1b: Domestic credit provided by banking sector (% of GDP) from 1971-1989

Figure 1c: Domestic credit provided by banking sector (% of GDP) from 1990-2010

In general, the bank credit can be viewed from two perspectives, the demand side (firms or individual's access to credit) and the supply side (financial intermediaries like banks) The present study identified the supply side factors that affect the credit growth

To assess the growing trend of bank credit before and after boom and bust cycles observed in most economies it is crucial to identify the factors which determine the bank credit especially from the bank side Most of the studies, done for Pakistan considering the demand side approach [Qayyum (2002); Afzal and Mirza (2010); Awan (2009); Khawaja (2007); and Ali et al (2011)] Not any study is undertaken considering the supply side variables of bank credit in Pakistan The main objective of present study is to identify the determinants of bank credit from supply side The study also empirically identifies if the bank credit behavior is different during financial reform and non-reform period The rest of the paper is organized that section 2 presents review

of the existing literature, followed by data and methodology in section 3 The, estimation and discussion of results is presented in section 4 The summary and concluding remarks are presented in section

Literature Review

In a recent study Guo and Stepanyan (2011) indicated that domestic and foreign funding are positively associated with the credit growth The stronger economic growth leads to higher credit growth, whereas higher inflation lessens the real credit growth in the economy Monetary policy has also a significant impact on the credit growth, the soft monetary conditions of the country as well as global, lead a more credit, and finally the strong banking sector positively influence on the credit growth Chernykh and Theodossiou (2011), by using a sample of Russian banks, found that the median banks assign only 0.5 percent of its total assets in terms of long-term loans to business and there is large cross-sectional disparity in this ratio among banks They argued that the bank’s capacity to expand long-term business loans depends on various factors including its capitalization, size and the availability of long-term liabilities, however, the ownership of banks did not matter They also concluded that the banks hesitated to issue business loans with more than three years maturity Their results exhibit that the banks with lower level of capital, the banks having lower funding for long term loans and banks in most competitive areas are reluctant to supply long term loans They considered weak creditor rights protection, enforcement and the low creditworthiness of risky borrowers as other hurdles in providing long-term loans to firms

According to Aisen and Franken (2010) prior to financial crisis the bank credit growth was larger

as compared to post crisis period Using a sample over eighty countries they also concluded that the countercyclical monetary policy and liquidity position of the banks played a crucial role and lessened the bank credit reduction in the post crisis era These findings advocate that the countries should follow the appropriate institutional and macroeconomic structure favorable to countercyclical monetary policies.They also found that the countries responding differently in

various regions of the world, due to diversity in countries’ structural characteristics e.g financial

depth and integration etc

Takats (2010) studied the bank lending behavior and empirically found that during the financial crisis the cross-border bank lending declined sharply Using the data of twenty-one emerging markets, he concluded that during the financial crises the demand and supply factors contributed

to the fall in bank lending, but the impact of supply factors were dominated However, both the factors appear to have more balanced effects in pre-crisis period Furthermore, supply shock was the key determinant of slowdown in cross border lending to emerging economies The credit growth before the crisis was vastly different across countries and regions

During the post crisis period the emerging markets experienced a considerable slowdown in credit growth (Guo and Stepanyan 2011) compared to pre crisis period The Pakistan also face slowdown in post crises period (as shown in figure 2)

Figure 2 Domestic credit provided by banking sector (% of GDP) in post crises period

Theoretically there is a long term relationship between bank health and the foreign bank credit growth in emerging markets The negative shocks to bank health created slowdowns in credit growth Moreover, the financial crises badly hit the banks health which leads to a lower credit

growth Empirical studies indicated that the reduction in bank credit during financial crisis left a negative impact on the real side of those emerging economies, which relied heavily on bank financing (McGuire and Tarashev 2008) Over the last two decades, most of the fastest growing economies of the developing nations have experienced lending booms and financial crises (Ranciere et al 2003) Kamil and Rai (2010) found that the countries rely more on external finance suffered the most during the crisis era Bakker and Gulde (2010) considered external factors as key reasons for credit booms and busts in new European Union members The pre-crisis boom and slowdown in partner economies were the fundamental determinants of credit growth during the crisis It is argued that the banks which faced ultimate liquidity stress lost their ability to lend more (Aisen and Franken 2010)

3 Data And Estimation Techniques The present study investigates the factors determine the bank credit in the case of an emerging economy like Pakistan The annual data used for econometric analysis, spanning the period of forty years from 1971 to 2010 The data obtained from various sources e.g Banking Statistics of Pakistan, World Development Index (WDI) and the International Financial Statistics (IFS)

The model used in the study is as follows

All variables are taken in natural logarithm form

By concluding the existing literature and according to the theory the liabilities from abroad, deposits by the domestic businesses and individuals, inflation rate, economic growth, exchange rate and the monetary conditions of the economy create a positive impact on the growth of private credit from the supply side, whereas the money market rate decreases the private credit

To find the long run relationship between variables, in terms of methodology this study used the robust econometric technique Autoregressive Distributed Lag (ARDL) by Pesaran (1997), Pesaran and Shin (1995, 1999) and Pesaran et al (1996) The ARDL has several advantages upon other conventional methods of cointegration like Johansen (1998) and Johansen and Julius (1990) The ARDL method can make a peculiarity between dependent and independent variables One major advantage of ARDL approach is that the estimation is possible even the explanatory variables are endogenous (Pesaran and Shin 1999; Pesaran et al., 2001) Another advantage is that it can be applied whether the variables are integrated at level or at order one or fractionally co integrated (Pesaran and Pesaran 1997) The empirical results are usually very sensitive to the method and diverse alternative choices are available in the estimation procedure (Pesaran and Smith 1995) Moreover, it is possible with ARDL that different variables can have different number of lags So this study also used ARDL approach for cointegration analysis and the follow-on error correction mechanism (ECM)

Firstly to find the long run relationship estimates the following equation:

+ 𝜏𝑖∆𝑀𝑀𝑅𝑡−𝑖𝑝

The lag orders are selected on the basis of Akaike Information Criterion (AIC) For short run dynamics the ECM has constructed The ECM version of ARDL approach can be written as

4 Discussion Of Empirical Results

Descriptive statistics is highlighted in table 1

Although, it is not necessary for the variables to be integrated at the same level for the application of ARDL but the use of unit root test will prove us whether or not the ARDL model should be used In order to check the level of integration of variables, this study used Augmented Dickey Fuller (ADF) and Phillips-Perron (PP) unit root tests Table 2 shows a mixture of level I(0) and I(1) of underlying variables, hence we can proceed the ARDL methodology

Table 1: Descriptive Statistics of the Variables

PC FL DD CPI GDP ER MMR M2

Mean

419070.9

Table 2:

Unit Root Test Results Variables

Level 1st Diff Level 1st Diff

PC 3.616 -4.796* 12.671 -7.380*

ADF and PP statistics with trend and intercept

* and ** are statistically significant at 1 and 5 percent level of significance respectively

The first step after checking the level of integration is to find the long run relationship among variables In order to decide the number of optimal lags Akaike Information Criterion (AIC) is used The results of long run relationship are presented in table 3 The growth of liabilities from foreign (FL), domestic deposit (DD), real GDP (GDP), exchange rate (ER) and the M2 as percentage of GDP (M2) are significantly affecting the private credit (PC) Whereas consumer price index (CPI) and money market rate (MMR) does not affect the private credit As banks get loan from foreign financial institutions their assets as well as their liquidity goes up, as a result they can lend more at domestic level Similarly the same impact imposed by the domestic deposits Increase in real GDP boost up the manufacturing sector’s income as well as the general peoples earning, which leads to higher domestic deposits, hence increase the liquidity of banks and they can lend more for investment needs, so the GDP has a positive association with private credit Exchange rate also positively associated with private credit in the case of Pakistan, which confirms that private credit in terms of domestic currency appear to pick up some valuation

effect of foreign currency credit Inflation level does not significantly impact the private credit but a positive sign indicates that private credit also increases with the inflation level The M2 considers as an alternative gauge of monetary conditions As monetary conditions of the country going up, the credit to private sector also enhances The insignificant impact of money market rate on private credit does not seem strange because the money market rate does not show a noteworthy value of standard deviation (as can be shown in table 1), it shows a smooth pattern throughout the whole study period Most of our results support to Guo and Stepanyan (2011) Another factor which can affect the private loan is financial liberalization reforms Pakistan adopted these reforms in 1990 to promote his financial sector In case of Pakistan these reforms

do not pose a significant impact on the private credit1

1

To capture the impact of financial liberalization reforms on private credit, we estimated several models consisting on a dummy variable having value 1 from 1990 to 2010 and 0 otherwise, and the interaction term of some independent variables with financial liberalization dummy The results can be seen in table 1 and 2 in appendix The results do not exhibit some different pattern

M2

26967.33**

(2.330) R-squared 0 999 Durbin-Watson stat 1.696 t-statistics are in parenthesis *, ** and *** are statistically significant at 1, 5 and 10 percent respectively

The results of error correction model are given in table 4 Most of the results are similar in both long run and short run However, little difference exists like in short run domestic deposits have not significant relationship with private credit The reason could be that banks do not issue loan immediately from the currently deposited amount by account holders The money market rate

has positive impact on private credit in the short run However, the empirical evidences do support significant effect of money market on the private credit in the long run Results of remaining variables are same in short run as in the case of long run The coefficient of ECt-1 (-0.535) is indicates that the above said long run relationship is stable and any disequilibrium formed in the short run will be temporary and get corrected over a period of time with a high speed of 53.5 percent per year A number of diagnostic tests are also applied in current study The results of those diagnostic tests evidenced that there is not serial correlation, heteroskedasticity and the Autoregressive Conditional Heteroskedasticity (ARCH) effect in the disturbances