FIVE years after the Asian financial crisis, people are still arguing about what exactly went wrong, and how it should have been put right. But most can agree on the background conditions that led up to the debacle. For many years the countries concerned had been pursuing a growth strategy centred on the exports of manufactures, following the then highly successful Japanese example. Financial resources were marshalled by keeping a lid on domestic consumption, and by encouraging the financial sector to channel savings to exporters. In return, governments shielded the financial sector from competition, along with other domestic services and natural resources.

During the early 1990s, exports boomed, and both foreign direct investment and portfolio money poured in as financial markets were liberalised. Companies assumed ever higher levels of debt, notably in the form of short-term foreign-currency borrowings, while a rush of foreign money fed stockmarket and property booms. Current-account deficits ballooned as national currencies tied to the dollar continued to give the impression of immutably fixed exchange rates.

Yet from the mid-1990s export growth softened and money, both foreign and domestic, started to leave these hot economies. On July 2nd 1997, Thailand's government, after futile attempts to stop the capital flight by raising interest rates and selling dollar reserves, allowed the baht to sink. Over the next few months other countries followed. At the height of the panic the Indonesian rupiah lost over four-fifths of its value, rendering the entire banking system insolvent. The panic spread beyond the region, to Russia and Brazil, and to American hedge funds. In Asia itself, financial mayhem led to political chaos.

Five years on, a degree of calm has settled upon the region, though the political tremors still reverberate. Indonesia is on its fourth leader in five years. And in the Philippines a corrupt president was, with questionable constitutionality, jettisoned in mid-term, to be replaced by his vice-president who now says she has too insecure a mandate to run in elections in 2004. On the other hand, in Thailand the crisis engendered a new constitution and stronger, more stable government.

Macroeconomically, the crisis countries are in a far sounder state. External finances have been restored to health, thanks to balance-of-payments surpluses on the back of renewed exports and much-reduced domestic demand. The short-term foreign-currency borrowings that were the catalyst for the whole mess have largely been rescheduled or paid off. Governments have rebuilt robust—perhaps unnecessarily robust—levels of reserves, and are paying back the multilateral and bilateral loans that bailed them out during the crisis. Currencies are no longer volatile, and interest rates and inflation in most of the region are low.

Yet despite these external strengths, Asia has a vast and unresolved legacy to deal with: the region's non-performing loans (NPLs), exposed as the crisis receded. The extent of these loans can be seen as a ready-reckoner of the cost of Asia's old, unbalanced economic model, a kind of marking-to-market of the whole region.

A proper accounting for the cost has yet to take place, for a big proportion of these bad loans has still not been disposed of. And for as long as an overhang of NPLs persists, the companies responsible for them will put off the needed restructuring. They are sometimes called the region's zombie companies, kept alive by unnaturally low interest rates, or by overdue interest payments being rolled over into the loan's principal.

Accurate estimates of the NPLs are hard to come by. Ernst & Young calculates its total figure for the region of $800 billion, or $2 trillion including Japan, by applying best international practice to the assessment of loans. One reason why the figures are hard to pin down is that Asia is still on the way to adopting uniform standards for assessing delinquent loans. Official figures invariably understate the problem, sometimes by a huge margin (see chart 3).

Whatever their exact size, the NPLs represent a misallocation of resources on a giant scale. And the greater the delay in tackling the zombie companies—by breaking up and selling off corporate bits with any value, putting seized collateral (usually property) on the market, and liquidating companies that are not worth anything—the longer the region will take to make a full economic recovery.

Many of these loans, admittedly, have been shifted away from the banks that made them in the first place. All the crisis economies except the Philippines have set up some sort of government-sponsored body, usually called an asset-management company (AMC), to take distressed loans off banks' books. Philippine banks actually have a higher proportion of dud loans on their books now (19% of all loans) than during the worst of the financial crisis (15%), but elsewhere the burden of bad loans has been passed, at least in part, from the financial sector to the public. Still, the record is mixed.

The Korean Asset Management Corporation (KAMCO) has been so successful that it now advises other countries, from Russia and China to Indonesia, on tackling bad debts. Choi Beom, an executive director, says that some of his 1,250 colleagues favour eventual privatisation: KAMCO's right to use fresh public funds to buy non-performing loans ran out last November, and its public remit expires in 2005.

Immediately after the crisis, KAMCO took on the task of buying from banks (at a discount to face value of around 60%) some $58 billion of bad loans, equivalent to nearly four-fifths of banks' total loans, or 30% of GDP. Since late 1998 it has sold over 70% of these loans, and the debt of a single giant borrower, Daewoo, which was rescued in 2000, accounts for a large proportion of the remainder.

The Korean authorities accepted early on that governments make poor asset managers, and that the more quickly distressed assets are disposed of, the lower the cost to the taxpayer. KAMCO took an innovative approach. For advice, it sought out former regulators from the Resolution Trust Corporation, charged in the 1980s with cleaning up America's savings-and-loan (thrift) crisis, and from Securitas, which handled Sweden's banking crisis in the early 1990s. It brought in western investment banks who know how to bundle up and securitise property assets for sale. It sold some distressed assets in bulk, by auction or tender, and others to American buy-out funds run by companies such as GE Capital, Lone Star, Goldman Sachs and Morgan Stanley. Several of these had learnt specialist skills in so-called loan workouts during the savings-and-loan crisis and, later, when leveraged buy-outs were all the rage in America. Sometimes KAMCO formed joint ventures with such firms, profiting from any value created. KAMCO's efforts, in turn, have helped establish a secondary market in distressed debt, with a face value of $22 billion, which makes it easier for banks to offload bad loans still on their books.

In September 1998 Malaysia drew widespread criticism overseas for slapping on currency controls in response to the crisis, which the prime minster, Mahathir Mohamad, blamed on a western conspiracy. That did not help Malaysia's case; nor, later, did the curious arrest and imprisonment of Dr Mahathir's finance minister. Still, Malaysia used the relative calm created by the controls to get to work on problem loans.

For a start, it set up a restructuring body to deal with the biggest corporate debtors. This was wound down last year, having finished its business. Meanwhile, the national asset-management company, Danaharta, has done a good job helping to clean up the banks' non-performing loans. Danaharta has been lucky in avoiding the moral hazard (or worse) to which AMCs in some other Asian countries are exposed by their conflicting duties, to the taxpayer, to banks that need recapitalising and to companies that the government wants to keep afloat. Danaharta's job is simply to resolve non-performing loans at the lowest cost to the taxpayer consistent with lightening the banks' NPL burdens enough to get them lending again. It is under no pressure to buy loans from banks which, in turn, are under no pressure to sell.

Instead, independent valuers estimate what the loans that banks want to offload are worth. If the loans look viable, Danaharta restructures or reschedules the debt; if they do not, the collateral is sold and the business liquidated, or a special administrator is appointed to sell the business by tender. As payment for NPLs, the offloading banks get zero-coupon bonds issued by Danaharta. If the recovery value for an NPL is higher than Danaharta's cost of acquisition, the selling bank gets four-fifths of the difference—an incentive for banks to offload such loans. Danaharta says that it acquired half of its portfolio of non-performing loans for an average of 46% of their face value; the other half were bad loans already owned by the government, and given to Danaharta to sort out. It expects an overall average loan-recovery rate of 57% before it winds itself down in 2005, turning a small loss overall. In all, though, the government expects the bank clean-up to have cost a mere 5% of GDP, far less than many other countries are having to spend.

However, a spanner has recently been thrown in the works. The most effective national asset-management companies everywhere have powers to bypass the courts when foreclosing on property assets, the most common form of loan collateral in Asia. Yet in December a local court ruled that Danaharta's special exemption from legal challenges to its restructuring plans was unconstitutional. Danaharta is appealing against the ruling.

KAMCO and Danaharta were fortunate in being handed clear and powerful mandates, which some other asset-management companies lack. In Thailand, for instance, the Thai Asset Management Corporation (TAMC), set up only in 2001 and controlling some two-fifths of all the distressed loans in the financial system, is charged not with minimising the claims on taxpayers but, in the words of an international adviser, “with helping debtors stay in the game”. It is attempting to restructure the assets on its books rather than to sell them off quickly, partly for fear of being seen to be selling assets to foreigners at knockdown prices. It is a strategy that will turn out expensive for the taxpayer.

Moreover, the TAMC acquired only a minority of private debts—those with many creditors. The bulk of its assets, on the other hand, are all the bad loans from the state banks. At the height of the crisis, 90% of the loans extended by Krung Thai Bank, which has the largest number of branches and employees in the country, were reckoned to be dud. Cleaning up the state banks risks merely setting them up afresh to make yet more politically directed loans, rather than ones based on commercial merits. It simply prolongs the financial system's agony.

Indonesia's asset-management company, the Indonesian Bank Restructuring Agency (IBRA), acquired—at their hideously inflated face value—some $31 billion of distressed assets. As a result, it has come to dominate Indonesia's economic landscape. As well as owning great swathes of industry, it also has a controlling stake in most of Indonesia's banks. Its activities are therefore central to Indonesia's economic recovery.

Yet so far it has sold only one-sixth of the loans it acquired, at 5% of their face value; and in the long run it will be lucky to recover 20% of the face value of all loans. During its short, politicised and often corrupt life, IBRA has had half a dozen chiefs. In 1999 it was implicated in kickbacks from Bank Bali, which dealt the reform process a blow. One former chief, who unwillingly answered the summons to lead the agency, points out that IBRA has an impossible number of stakeholders, including the finance ministry, the central bank, powerful parliamentary committees and an independent review committee.

During the savings-and-loan crisis, America's government spent $80 billion buying dud loans from the thrifts, but got four-fifths of it back. That mattered to congressmen, who had more to lose from angering voters by wasting taxpayers' money than to gain from helping a friend with an empty shopping mall or office building. In America, it helped that every meeting between the Resolution Trust Corporation and a member of Congress was automatically followed by a press release. In Asia, says a Thai banker, there is no taxpayers' culture, at least not yet; and there are precious few press releases to explain what national asset managers are up to.