It was only a few months ago that debt was readily available to all who asked. Actually, you didn’t even have to ask. Banks fought to extend loans in highly leveraged takeovers and paid high commissions to agents for digging up deadbeat US mortgage borrowers. It has all changed completely. There is talk of a credit squeeze as banks and brokers scramble to strengthen their balance sheets.

What went wrong? The subprime mess started the ball rolling. But much of the real blame belongs to rapid growth of a very dangerous practice—“pro-cyclical" leverage.

The idea is explained in a research paper by Tobias Adrian and Hyun Song Shin, published by the New York Federal Reserve Bank. The researchers studied 43 years of leverage history in the US. The patterns vary widely between different economic groups: households, industrial companies, commercial banks and brokers.

Households are strongly “anti-cyclical" in their leverage. When asset values increase quickly (the up part of the cycle) leverage, measured as the ratio of debt to equity, generally declines. Conversely, shrinking asset values go with rising leverage. This is because Mr and Mrs Homeowner pretty much keep on paying down their mortgage, without paying much attention to house or share prices. Imagine the Homeowners take out a $300,000 (nearly Rs118.20 lakh) mortgage on a $400,000 house. That gives them equity of $100,000 and a leverage ratio of three. Suppose that house prices increase sharply, so their humble abode is now worth $500,000. Their equity doubles to $200,000, bringing the leverage ratio down to 1.5. Now look at industrial companies and ordinary commercial banks. They are neither strongly anti or pro-cyclical. Instead, they seem to target a fixed leverage ratio, whatever is happening to the value of assets.

Finally, there are brokers and investment banks. They are strongly pro-cyclical. The more asset prices rise, the more their leverage ratio goes up. This is largely because brokers use specialized risk-based accounting, which allows them to borrow more when asset prices are rising and forces them to borrow less when asset prices are falling.

Brokers might attempt to deny this pro-cyclicality. That’s because the tier I capital ratio, the ratio of so-called “risk-weighted assets" to equity doesn’t change much whatever happens to asset prices. But dig a little deeper and one discovers that the risk weightings, which determine whether a particular asset is high or low risk, are highly pro-cyclical.

There’s nothing wrong with adjusting for risk. But risk is a slippery concept. A non-specialist might think that risks in a portfolio increase when the prices of the investments in it are high by historical standards, especially when prices are rising simultaneously for many different types of investments. The amateur would also be wary of calm markets, because storms tend to come up sooner or later.

But the professionals reverse those judgements. So-called “value at risk" (VAR) calculations which are embedded in brokers’ assessment of riskiness, rely on market values when they are available, no matter how unreasonable the market. For unmarketable securities, models are used to provide proxy prices, with little attention paid to the plausibility of the model. In the calculation of the VAR of complex portfolios, only quite recent market history is used. So, a few years of past calm is taken to indicate many more years of future market tranquillity. In effect, when the real risk of a market collapse is high, the ratio of VAR to the value of the underlying assets is low. Finance directors look at the low ratio of VAR to assets and say: “We’re not using our balancesheet efficiently. Leverage it up."

From 2001 to early 2007, apparent risk in the markets kept declining, brokers kept expanding their portfolios and reported tier I ratios generally stayed pretty low. Adrian and Song show that all but one of the brokers studied (Bear Stearns & Co. Inc., Goldman Sachs Group Inc., Lehman Brothers Inc., Merrill Lynch & Co. Inc. and Morgan Stanley) kept the ratio of VAR to equity fairly constant.

The exception was Goldman, which rapidly increased its leverage, even by this generous measure. That increase might help explain some of the firm’s superior performance in the period. But then Goldman has also done well in the last few months, when other firms have stumbled. So the firm might genuinely be better at managing risk.

Even Goldman will have to deal with the effect of more difficult markets on VAR. The study doesn’t cover the last few months, but it’s safe to say that the VAR calculations have changed direction. Just as the signal was to buy when prices were high and rising fairly steadily, it is now to sell when markets for risky assets are turbulent. The VAR signal won’t change when markets get cheap. It will only turn to buy when markets get stable. Unfortunately, the brokers’ pro-cyclical leverage makes that market stability harder to achieve. The VAR-guided sector of the market is now big enough to have a significant influence on prices. The operations of a big part of the financial world now amplify market trends. They provide additional demand when prices are rising and additional supply when prices are falling. No wonder the central banks have their hands full battling a crunch.