The credit crunch has thrown a spotlight on the inadequacies of the Basel system of bank regulation. The system loosens regulations when they should be tightened, and contains several glaring and dangerous loopholes.

Three significant banks—the UK’s Northern Rock Plc. and IKB Deutsche Industriebank and Sachsen LB in Germany—nearly collapsed after funding themselves in foolish, but legal ways. Regulators should tear up the Basel system and start again—paying special attention to liquidity, off-balance sheet vehicles and flawed risk models.

The Basel II system was sold as a necessary modernization of the original 1989 Basel regulations, under which banks had to keep a minimum of 8% of certain assets in capital. But, in fact, it has gone in the wrong direction. It has introduced two new features that make the banking world more vulnerable, while failing to address several of the weaknesses of the old system.

The first of the new features is a two-tier risk management system depending on the size and sophistication of the bank being regulated: big banks get more freedom to leverage themselves than small ones.

As market worries about large banks have escalated, governments have shown themselves willing to guarantee deposits or otherwise bail out large institutions. Since size is no protection against foolish activity, but merely makes it more likely that the taxpayer will be called in to bail out the results of failure, large institutions should in principle face more stringent capital requirements than small ones.

The second new feature is the principle of “self-regulation" by which large banks with sophisticated risk management systems can monitor and control their own risk. But neither regulators nor the banks themselves truly understand the risk profile of complex and sophisticated trading positions, as has been demonstrated by the losses sustained in recent months.

The popular value at risk (VAR) system, for example, carries an underlying assumption that by analysis of past data, a “99% VAR" can be determined, which provides a reasonable assessment of a position’s risk, and is likely to be inaccurate by only a moderate amount.

Basel II allows banks to assess their risk levels based on their 30-day VAR, which is supposed to be the maximum loss that can arise over a 30-day period—and provide capital accordingly.

There are three problems that make this method a trap, likely to fail spectacularly in periods of market turbulence.

First, even if VAR were truly a level of loss likely to be exceeded only 1% of the time, 100 months is only eight years.

Second, when volatility increases suddenly, VAR calculations demand that risk positions be cut back correspondingly. However, if the whole market cuts back risk simultaneously, liquidity disappears. VAR accentuates cycles, as banks are encouraged to take on additional risk positions in periods of market calm, and then forced to cut them back when it becomes turbulent.

Third, VAR assumes that losses much greater than the “99% VAR" level are unlikely, and that prices behave in a “near-normal" pattern. The fallacy of that assumption was exposed last month by Goldman Sachs Group Inc.’s Global Alpha Fund, which announced that it had been subjected to a “25-standard deviation" event. Had the VAR model been valid, such an event could have occurred only once in the history of the universe; reality demonstrates that something like it happens every few years.

As if these new problems with Basel II are not bad enough, two big errors in the old Basel system haven’t been dealt with: liquidity and off-balance sheet vehicles. Both have been exposed by the crunch.

The Basel II regulations do not directly address liquidity risk. As demonstrated by the shrinkage of the asset-backed commercial paper (ABCP) market and the run on Northern Rock, illiquidity can itself be a cause of greatly increased cost and possible bank failure.

Indeed, it is something of a truism that the reason banks fail is because they run out of money.

Additional capital requirements address this problem only indirectly. Bank leverage is sufficiently high that funding requirements can be several times available capital. Finding the right answer will require thought. But there are several possible solutions.

One would be to focus on the maturity gap between the assets and liabilities of banks and to require them to fund themselves so that that they have cash for a specified minimum period. Another would be to require those banks that rely on short-term funding to hold more capital. That would impose a cost on them, giving them a countervailing incentive to raise long-term funds that are more expensive.

The Basel regulations have demonstrated a further inadequacy in recent weeks: loopholes. ABCP vehicles, financing medium-term assets, can be set up by a bank without requiring substantial additional capital under both the original Basel and Basel II. The banks have no direct legal obligation to the vehicles that themselves are not banks, and can thus be thinly capitalized.

However, in the event of market turbulence, ABCP markets may close for such vehicles, requiring backstop finance from the parent bank, increasing its risk and draining its liquidity. Thus Sachsen LB required a bailout of $23 billion (Rs91,540 crore) because of illiquidity in its Irish investment conduits.

Loopholes in bank capital requirements are particularly pernicious because profit-maximizing banks will always exploit them. The ABCP market reached a volume of $1.2 trillion, when almost its only advantage was to allow banks to earn additional returns while using less capital. When combined with self-regulation and a lengthy period of high liquidity such as 1995-2007, the regulations become effectively nugatory, as banks invent new ways to evade them, increasing the boldness and volume of their evasions as the high-liquidity period rolls on.

The original Basel regulations offered the advantage of relative simplicity and reduced the scope and incentives for gaming the system. The main changes needed are to eliminate the loopholes and to introduce limitations on liquidity risk. All banks should be treated equally and self-regulation should not be permitted because of its potential for abuse. Banks will complain that such a system is too restrictive, forcing them to keep too much capital compared with their asset base. However, the cost to society of overcapitalizing banks is trivial compared with the cost of credit bubbles followed by crunches and bailouts.