Missing the cause for the symptoms

The stock of non-performing loans (NPLs) is increasing in both public and private banks. This is raising the threat to financial stability, impairing financial intermediation and damaging the resilience of the banking sector to shocks, thus increasing systemic risk. NPLs are also associated with higher funding costs and a lower supply of credit. However, the recent hot debate in Bangladesh has centred on whether high NPLs are a cause or a consequence of high lending rates.

Those identifying NPLs as the cause argue that the link between NPLs and lending rates is driven by higher risk premia with higher NPL stocks passed on to borrowers. Default risk is an essential component of loan pricing. Those identifying NPLs as the consequence argue that high lending interest rates enlarge the debt burden of borrowers eventually causing loan defaults.

The focus on the lending rate-NPL nexus has crowded out attention to the impact of NPL on credit volume. NPLs affect banks’ lending volume through losses caused by loan provisioning against the NPL stock. Increases in provisioning leads to a reduction in capital via the profit and loss statement. NPLs also erode bank’s liquidity leading to depletion in their capacity to lend. Banks thus react to a high stock of NPLs by restricting loan supply.

In an economic system where everything depends on everything else, this debate may sound a bit like the blind men and an elephant parable. Each blind man feels a different part of the elephant’s body and describes the elephant based on their limited touch of the animal. The parable shows how our individual perceptions or “mental models” can lead to far less than full truth. 

The NPL, lending rate and loan volume issue can be settled just the same way the Rajah settles it in the parable. “The elephant is a big animal,” he said. “Each man touched only one part. You must put all the parts together to find out what an elephant is like.”

To discover the whole truth about the relationship between NPLs, interest rate and credit growth, we must put all these together by looking at what the data says.

What do we look for in the data? Granger predictability is a way to investigate “causality” between two variables in a time series. The method seeks to find patterns of correlation in empirical data. Technically, it is a method for determining whether one time series is useful in forecasting another. What one can uncover is whether a particular variable consistently comes before another in the time series.

Since predictability is a central feature of “causal” attribution, causality could be tested by measuring the ability to predict the future values of a time series (lending rates or credit growth) using prior values of another time series (NPLs) or vice versa. This does not necessarily imply true causality, only conditional dependence, that is, the probability of one event (say rise in NPL) given the occurrence of another event (rise in lending rate).

Both quarterly (from December 2013 to September 2019) and annual (from 1997 to 2018) data on NPLs and the bank lending rates suggest there is no clear-cut relation between gross NPLs and lending rates. There is no evidence of Granger predictability either way—from lending rate to gross NPL or from gross NPL to lending rate. However, there is fairly strong relationship between gross NPL and growth of credit to the private sector with predictability running from NPL to credit growth and not vice versa. 

The impact of NPLs on credit growth is fairly large. Simple regression analysis based on quarterly data suggests that a 1 percent increase in gross NPLs reduces private sector credit growth by 0.47 percentage points. Similar regression further suggests that one percentage point decrease in credit growth increases lending rate by 0.34 percentage points.

Deterioration of the credit risk caused by rising NPLs reduces growth in the supply of loanable funds. High default risk and the difficulty of assessing the soundness of each debtor in a default-friendly system generate adverse selection and loss aversion (preferences to avoid losing compared to gaining the equivalent amount) among banks. NPLs, as a source of anticipated future losses, have the same impact on lending policies as actual capital shortfall. There is an abundance of evidence in the related literature suggesting capital corrosion entail cuts in bank lending. Not only that NPLs decrease the supply of loanable funds and banks may even shift the composition of their borrower portfolio towards riskier borrowers who can be charged higher risk premia. Both contribute to increasing lending rates.

There is no shortcut to getting to the root of the problem. Measures to reduce interest rate by administrative fiat cannot dent the NPL problem when interest rate has no predictable effect on the NPLs. The reason is simple. It only touches the symptom, not the root cause. High NPL is also a symptom of deeper malaise in the market for bank loans. The issue is not whether NPL causes high interest rate or high interest rate causes NPLs. The issue is what corporate and regulatory failures in banking governance keeps both interest rates and NPLs high? 

The legal and judicial framework and banks’ governance structures determine the circumstances under which macro-prudential policy can be effectively enforced. When bankers expect the enforcement of collateral and the outcome of insolvency proceedings to be lengthy, costly and variable from case to case, they incorporate this into their pricing decision. This keeps lending rates high and credit availability restrained. With borrowers aware that the collateral will not be easily and quickly enforced, the incentive to pay their loans in time is feeble. This keeps NPLs high. The absence of resolute action in these two areas is antithetical to ensuring the resilience of banks and preserving their capacity to finance the economy.

The author is an economist.