Chinese debt has risen over the past ten to fifteen years at perhaps the fastest pace in history, leaving China with one of the highest debt-to-GDP ratios in the world—second only to Japan among the major economies. What is more, surging debt has always been the warning for investment-driven “miracle” economies that their growth model has become unsustainable.

That is why it is not surprising that analysts have become increasingly concerned about how China will ultimately resolve its debt burden. Some look back to the cleanup of China’s banking system in the early 2000s, following the explosion of bad lending during the 1990s, as a possible model for dealing with today’s debt overhang.

Unfortunately, they often draw the wrong lesson, seeing the experience of the 2000s as suggesting that China can resolve a banking and debt problem at little cost to the economy. But in fact, the cleanup was extraordinarily costly, and shaped the Chinese economy in ways that are discussed almost daily today in the Chinese press. The way Beijing chose to resolve its debt burden laid the foundations for many of the deepest structural imbalances in China’s economy over the subsequent two decades.

For that reason, it is worth reconsidering how China resolved its debt problems in the 2000s. The original problem emerged in the 1990s, when banks functioned primarily as quasi-fiscal institutions supporting state-owned enterprises (SOEs), local governments, and industrial policy, rather than allocating capital on commercial terms. Credit allocation was largely administrative, as it had been throughout the four decades since the founding of the People’s Republic. Loans were extended because enterprises were politically important, because local governments wanted investment, or because employment had to be maintained, and rarely because the projects they funded were expected to generate adequate returns.

The way Beijing chose to resolve its debt burden laid the foundations for many of the deepest structural imbalances in China’s economy.

Many of these loans were, in effect, fiscal transfers disguised as bank credit, and in many ways, this remains one of the primary functions of China’s banking system today. As a result, the largest source of bad debt was lending to SOEs. Many SOEs remained grossly inefficient, with excess employment maintained for reasons of social stability even as prices were liberalized, competition intensified, and profitability collapsed. Yet banks continued lending because allowing widespread bankruptcies was politically unacceptable.

By the late 1990s, China’s banking system was burdened with enormous quantities of non-performing loans. The Big Four state-owned banks—which accounted for the overwhelming majority of commercial bank lending—were technically insolvent. Official estimates suggested that roughly 25 to 30 percent of their loan portfolios were non-performing, while many independent analysts, including the World Bank, believed the true figure was closer to 40 or 50 percent. Rural credit cooperatives and many local banks were widely believed to have even higher proportions of non-performing loans.¹

As Beijing prepared for World Trade Organization (WTO) accession and sought both to modernize and partially to privatize its banking system, policymakers recognized that banks burdened with such large volumes of bad loans would be extremely difficult to sell, especially to the foreign strategic investors they hoped to attract. The decision was therefore made to recapitalize the banks before listing them on the Hong Kong and Shanghai stock exchanges.

As part of this process, Beijing transferred bad loans from the Big Four banks to newly created asset-management companies (AMCs), injected fresh capital into the banks, strengthened regulation and accounting standards, and brought in foreign strategic investors.² At the time, many economists assumed the cleanup would be largely technical. Bad loans would be transferred, banks recapitalized, accounting standards improved, and the banks successfully listed. That is broadly what happened. What was much less widely understood was who ultimately paid for the most important—and by far the most expensive—part of the reform: the bad loans themselves.

Every Banking Crisis Is Ultimately About Loss Allocation

What is often forgotten in discussions of banking crises is that every bad-debt resolution is fundamentally a question of loss allocation. When the value of an entity’s assets falls below the value of its liabilities, someone must absorb the resulting loss. Normally, the first losses fall on the owners through lower equity values. If those losses exceed the owners’ equity, creditors typically absorb the remainder.

But when the owners or creditors are politically powerful—or when imposing losses on them threatens broader financial stability—governments often intervene to socialize the losses instead. There are many ways to do this, but in almost every case the losses are ultimately transferred, often indirectly, to the household sector. This can occur through explicit fiscal bailouts financed by taxes, through inflation, or through financial repression that transfers income from depositors and other holders of monetary assets to borrowers.

Moving bad loans from one balance sheet to another does not eliminate them.

This leads to two fundamental points that are essential for understanding China’s banking cleanup. First, losses never disappear, even when they appear to. They are always absorbed by some sector of the economy. Second, loans that finance projects whose economic value ultimately proves to be less than their cost represent real economic losses, whether or not those losses are immediately recognized. Moving bad loans from one balance sheet to another does not eliminate them. It merely changes the way in which the losses will be allocated.

The only meaningful question, then, is whether those losses are to be borne primarily by borrowers, banks, governments, businesses, creditors, or households. Contrary to widespread perception, Beijing did not primarily ask the government to absorb those losses. Instead, it relied overwhelmingly on financial repression. That is probably the single most important feature of China’s banking cleanup—and also the most widely misunderstood.

The Role of the AMCs

The mechanics of the cleanup were relatively straightforward. In 1999, Beijing established four state-owned AMCs—Cinda, Huarong, Orient, and Great Wall—to purchase non-performing loans from each of the Big Four banks: China Construction Bank, Industrial and Commercial Bank of China, Bank of China, and Agricultural Bank of China, respectively. The AMCs were partially capitalized by government entities, but between 1999 and 2000 they acquired roughly RMB 1.4 trillion of bad loans—equivalent to about 20 percent of GDP—at full face value, even though the market value of those loans was far lower. Later transfers took place at prices ranging from roughly 40 to 60 percent of face value, still far above the ultimate recuperable value of the loans.

The AMCs financed these purchases primarily by issuing ten-year bonds to the very banks from which they acquired the bad loans. These bonds were supported by letters of comfort from the Ministry of Finance and accommodated by liquidity from the People’s Bank of China. In effect, the AMCs purchased bad loans with the proceeds of bonds they sold to the banks, while assuming responsibility for restructuring, selling, or liquidating the underlying assets.

It is important to recognize what these transactions did and did not accomplish. They did not eliminate the losses. They merely transferred them. The Big Four no longer held loans to borrowers that could not repay. Instead, they held bonds issued by AMCs whose principal assets were those same non-performing loans. Bad loans on bank balance sheets had simply been replaced by bonds whose value ultimately depended on recovering value from the same insolvent borrowers.

The difference between the face value of the loans and their true economic value still had to be absorbed by someone. The purpose of the AMCs, in other words, was not to resolve the losses but to manage the long process by which those losses would eventually be allocated elsewhere.

Financial Repression Was the Real Bailout

That “elsewhere” turned out to be primarily the Chinese household sector. This is where financial repression became critical. China already operated within one of the world’s most tightly controlled financial systems. Deposit rates, lending rates, exchange rates, and capital flows were all subject to administrative control. During the banking cleanup, these controls became powerful mechanisms for transferring income from one part of the economy to another.

Deposit rates were kept well below nominal GDP growth and, for extended periods, below inflation itself. Because households constituted the overwhelming majority of net savers, they earned extraordinarily low, and often negative, real returns on their bank deposits. Businesses, state-owned enterprises, and local governments, by contrast, borrowed at artificially low interest rates.

Although these policies were often presented as monetary measures designed to support growth and maintain financial stability, they also represented an enormous redistribution of income. Every year, households transferred purchasing power to banks and borrowers through artificially suppressed returns on their savings. Depending on how one measures the transfer, it may have amounted to as much as five percentage points of GDP annually during the most repressive years, between 2001 and 2010.

The banking cleanup succeeded because households, largely without realizing it, recapitalized the financial system.

Households therefore did much more than simply earn low returns on their deposits. Because deposit rates were frequently negative in real terms, households effectively paid a hidden tax on their savings. Over roughly a decade they quietly absorbed losses that otherwise would have remained within the banking system or been transferred directly to the government’s balance sheet.

The banking cleanup succeeded because households, largely without realizing it, recapitalized the financial system. This is why the common claim that China resolved its banking crisis at little cost is so misleading. China certainly paid the cost. It simply allocated that cost primarily to households through negative real deposit rates rather than explicitly to banks or to the fiscal authorities. Because the transfer occurred gradually and indirectly, it appeared as though the losses had simply disappeared.

Why Household Consumption Collapsed

Once viewed this way, one of the most puzzling developments in modern Chinese economic history becomes much easier to understand. In the early 1980s, total consumption accounted for roughly 65 to 68 percent of GDP. While somewhat below the global average, this was not unusual for a rapidly industrializing economy. Consumption declined during the early 1990s but partially recovered by the end of the decade. As late as 2000, it still represented 63.9 percent of GDP.

Instead of rising thereafter, as many economists expected once China had largely completed its initial phase of industrialization, consumption collapsed (see figure 1). By 2010, it had fallen to just 49.4 percent of GDP, by far the lowest share ever recorded in any major economy.

At almost exactly the same time, China’s current-account surplus surged, eventually exceeding 10 percent of GDP. Most economists attributed this primarily to China’s accession to the WTO in 2001. That explanation has never been entirely convincing. Joining the WTO does not require a country to run trade surpluses. It simply allows both exports and imports to expand more rapidly under a more predictable trading regime. Nor does WTO accession explain why China’s consumption share of GDP collapsed so dramatically during precisely the same period.

The more convincing explanation lies in the banking cleanup itself. Because Beijing chose to recapitalize the banking system primarily through financial repression, it transferred a substantial share of national income from households (overwhelmingly net savers) to businesses and governments (overwhelmingly net borrowers). The consequences were exactly what standard macroeconomic theory would predict. Household income declined as a share of GDP, and household consumption fell with it. At the same time, businesses and local governments gained access to extraordinarily cheap capital. Investment accelerated even as household purchasing power weakened.

Because Beijing chose to recapitalize the banking system primarily through financial repression, it transferred a substantial share of national income from households to businesses and governments.

The effects extended far beyond the banking system. By sharply reducing the household share of GDP, Beijing also automatically increased the producer share. Manufacturers benefited from lower financing costs, cheaper capital, and higher investment, while households lost purchasing power. Manufacturing competitiveness improved dramatically, but domestic demand weakened relative to production.

Initially, because China was still substantially underinvested, much of the resulting increase in production could be absorbed through additional investment. As productive investment opportunities gradually diminished over the following decade, however, sustaining rapid growth increasingly required either rising debt or rising trade surpluses. Seen in this light, the banking cleanup and China’s subsequent debt buildup are not separate stories. They are different manifestations of the same underlying process.

The Legacy of the Cleanup

The irony is that the policies that successfully recapitalized the banking system also made future rebalancing extraordinarily difficult. Once household income had been suppressed, maintaining rapid growth required either ever-higher investment or ever-larger trade surpluses. As productive investment opportunities gradually diminished, investment could be sustained only by rapidly rising debt. What had initially been a mechanism for recapitalizing the banking system eventually became a mechanism for generating another debt problem.

The lesson of the early 2000s, therefore, is not that China found a painless way to resolve bad debt. It is that every debt resolution requires losses to be allocated, whether explicitly or implicitly. Beijing chose to allocate those losses primarily to households through financial repression. That decision successfully recapitalized the banking system, but it also depressed household income, reduced consumption, raised the national saving rate, increased manufacturing competitiveness, and created many of the structural imbalances that continue to shape the Chinese economy today.

The problem today is that the politically easiest solution—forcing households indirectly to absorb the losses—is far less available than it was twenty years ago.

This history matters because China once again confronts an enormous debt overhang. If policymakers attempt to resolve today’s debt burden in the same way they resolved the last one—by once again transferring resources from households to producers and governments—they may stabilize the financial system temporarily, but they will also deepen the very structural imbalances that have made China’s adjustment so difficult and contributed to the rapid accumulation of debt over the past fifteen years.

The problem today is that the politically easiest solution—forcing households indirectly to absorb the losses—is far less available than it was twenty years ago. China’s household share of GDP, along with its consumption share, was already unusually low when the banking cleanup began. Today it is lower still, among the lowest ever recorded for a major economy. Further transferring income from households to producers and governments would only deepen the structural imbalances that lie behind China’s excessive reliance on debt and external demand.

If Beijing wants to resolve today’s debt problem without further suppressing household income, the losses must instead be allocated elsewhere. In practice, that means they must fall primarily on either businesses or the government. If businesses absorb a larger share of the losses—through higher wages, higher interest rates, currency appreciation, the removal of implicit subsidies, or other policies that transfer income toward households—manufacturing profitability and investment will almost certainly decline. Growth is likely to slow as the economy rebalances, and China’s extraordinary manufacturing competitiveness will inevitably weaken.

If, instead, the government absorbs a larger share of the losses through explicit fiscal transfers, debt assumption, or the sale of public assets (the last of these being by far the most economically efficient way to resolve China’s bad-debt problem), the economic adjustment may prove less disruptive in the short run. But it would raise difficult political questions about how losses are allocated among different levels of government and among the powerful business, financial, and political interests that have grown up around China’s investment-led growth model over the past three decades. 

There is no escaping this arithmetic. Every bad-debt resolution requires that someone absorb the losses. Those losses cannot simply be transferred from one balance sheet to another indefinitely, nor can they disappear through accounting or financial engineering. They must ultimately be borne by households, businesses, or the government. The difference between the early 2000s and today is that the option Beijing chose then—quietly transferring the losses to households through financial repression—is no longer available on anything like the same scale. That leaves policymakers facing a much more difficult choice than they confronted in the early 2000s. They can ask businesses to absorb a greater share of the losses, with the resulting slowdown in growth and decline in manufacturing competitiveness, or they can ask the state to absorb them, with all the political consequences that implies.

There is no painless solution. The remarkable achievement of the early-2000s banking cleanup was not that China found a way to eliminate the costs of resolving bad debt. It was that it found a politically acceptable way to hide those costs. The challenge today is that the costs are likely to be much harder to hide.