Lending with (nearly) zero risk

The persistent failures of banks to lend sensibly in Japan, in other East Asian countries, and in many other countries around the world has brought the question of safety in lending to the fore. Why do banks persistently lend so imprudently, and how should lending be done at minimum risk?

One essential problem with banks is the human and managerial challenge of motivating employees of banks to cater to the interests of the owners (shareholders) of the bank. The history of banking is replete with episodes of employees favouring friends and relatives with loans. In some East Asian countries, there are well--defined "market rates" for bribes for obtaining loans from banks. These problems are also present in India, though the record of India's banking system in these aspects is much better than that of many other countries.

Another aspect of the problems of banks concerns prudent levels of leverage. A bank is a financial intermediary with fairly small equity capital, which borrows money from depositors, and invests it into risky assets. This involves a high degree of leverage: many `staid' banks are more leveraged than most `aggresive' derivatives traders. This leverage implies that fluctuations in asset prices can wipe out the capital of a bank. Leverage, at the level of the bank, is dangerous regardless of the quality of credit analysis which has gone into each loan.

These problems have been exacerbated by governments, which have peculiarly accorded banks a special status amongst financial intermediaries, guaranteeing them against failure. This guarantee generates `moral hazard' in the form of incentives for bank managers to adopt extreme levels of leverage. High leverage generates high risk and high returns. If high returns are obtained, the bank takes the profits, but it is protected from high losses by the government.

Regulators have tried many policy initiatives aimed at obtaining a banking system which has controlled leverage, high quality lending, and thus a reduced risk of failure. These include capital adequacy requirements based on clumsy measurement of risk (e.g. the Basle norms), prohibition of lending against real estate, restrictions on lending against shares, rules governing collateral, etc.

We will be able to better understand the issues by asking a polar question: is there a way to lend while bearing near--zero risk?. The answer, remarkably enough, is in the affirmative. There are procedures for "riskless lending" which can be used by banks without any "capital adequacy" limits, and are easily implemented despite improperly motivated employees. While these procedures are not universally applicable, they will help us better understand the difficulties with other kinds of lending.

Lending with (nearly) zero risk. A riskless loan is one which is fully collateralised using actively traded assets. These assets should be traded objects, so that a "mark to market" can be done daily, to ensure that the collateral is always larger than the outstanding loan. The value of the asset that is measured when marking to market should be the liquidation value, thus taking into account the problems faced by the bank when selling off the assets.

A loan would commence with collateral of Rs.110 protecting a loan of Rs.100. For example, a person with Rs.10 would borrow Rs.100 and pledge the asset worth Rs.110 to the bank. If the market price of the collateral drops below 110% of the outstanding loan, then the daily mark--to--market would reveal insufficient collateral, and the borrower would have to put up additional collateral. If the borrower fails to comply, or if he fails on his repayment schedule, then the bank would instantly liquidate the asset on the market. Since the asset is traded on an active secondary market, there would be no delays in liquidating it, nor would there be any associated price risk. Since the collateral has been valued at its liquidation value, the illiquidity of the asset would not catch the bank by surprise.

This is hence a remarkably safe lending relationship. It is only if the secondary market for the asset breaks down at the same time that a borrower defaults (for example, in a war, or if the satellite used by NSE fails) that the bank suffers a loss.

This kind of fully collateralised loan requires liquidity and transparency in the assets which are accepted as collateral:

There is only one kind of market structure which supports this level of transparency about both prices and liquidity: the open electronic limit order book market. This is the market structure that is used in trading equity at NSE and BSE. All these ideas can hence be implemented for loans against shares. Share prices are observed daily, which allows a daily mark to market. Snapshots of the limit order book allow measurement of impact cost and hence liquidation value.

This suggests a remarkable fact: shares are the only form of collateral which can be accepted in India, today, to support this (nearly) riskless form of lending. In the future, if bonds trade on an electronic limit order book market, they would also satisfy these requirements, thus enlarging the universe of assets that are acceptable as collateral in such (nearly) riskless lending.

This reasoning emphasises the information revelation function of securities markets. Exchanges make information about prices and liquidity available to the economy, which enables such sophisticated contractual relationships to develop. Prior to the electronic limit order book market, lending relationships were intrinsically riskier because liquidity was not observable.

Another advantage of this style of lending is that it supports the development of well--defined, technology--intensive internal procedures in the bank. Credit analysis of borrowers is not required, and improperly motivated employees would not be able to distort lending.

This helps us understand the problems of lending against real estate. The market for real estate is non--transparent. It is not possible to recalculate the liquidation value of a property every day, and even if liquidation were to be done, the bank would need months to sell off the asset. This makes loans against real estate risky. The issue here is not the volatility of real estate prices, which is probably similar to the volatility of share prices. The problem with real estate is the inability to observe prices and liquidity.

We would hence agree with folk wisdom in frowning upon loans against real estate, but disagree with it when it prohibits loans against shares. Loans against shares, implemented using the procedures described above, are the safest form of lending which can be done in India, as of today. When projects are unable to put up collateral of this nature, lending involves risk, and credit analysis is unavoidable. In lending organisations which lack good credit analysis, the procedure above is a safe strategy to fall back upon. RBI could potentially require that "weak banks" be allowed to do no other kind of lending.

Back up to Ajay Shah's media page