Global debt, according to a recent report by the Institute for International Finance, amounted to nearly $300 trillion in 2021, equal to 356 percent of global GDP. This extraordinarily high debt level represents a 30 percentage-point rise in the global debt-to-GDP ratio in the past five years. No wonder analysts increasingly worry about the possible adverse consequences of excessive debt levels.
Unfortunately, few economists have a clear understanding of why too much debt is a bad thing, let alone how much debt is too much. That makes it hard to know what to worry about and why. Because economists tend to assume that the extent of a country’s debt burden is measured by its national debt-to GDP ratio, they often fail to distinguish between types of debt, instead treating a rise in one country’s debt-to-GDP ratio as equivalent to the same rise in another country’s ratio, even though the two cases may have very different implications.
So when is debt a burden for the economy and why? Crucially, different kinds of rising debt can have very different effects on an economy. Moreover, even in countries that are seen as having too much debt, the adjustment costs can vary significantly.
Dispelling a Myth About Modern Monetary Theory
Before going into the different consequences of debt, it is necessary to discuss briefly some of the confusion that has arisen in recent years from a naïve understanding of modern monetary theory (MMT). Many proponents of the theory have argued that MMT teaches that there are no spending constraints on a government that is monetarily sovereign (one that produces its own credible fiat currency) and that such a government can spend an unlimited amount without worrying about the consequences until there is a surge in inflation.
This simply isn’t true. What MMT really shows is that there are no direct spending constraints on a government that is monetarily sovereign. It can always create money or debt to fund its spending needs without first needing to obtain the funding. This doesn’t mean, however, that there are no indirect constraints. In fact, there are always economic constraints. This is because an economy cannot consume and invest more than it produces and imports, and any ex ante imbalance between the two must be resolved by implicit or explicit transfers that reduce the purchasing power of some sector of the economy by enough to bring the two back into balance.
How Demand Adjusts to Supply
This is a simple idea, even if it isn’t always well understood. Government spending, whether financed by money creation or debt creation, automatically increases the purchasing power of some sector of the economy, in effect increasing demand. If government spending simultaneously increases supply by an equivalent amount—either directly (by funding investment in infrastructure, for example) or indirectly (by funding increases in consumption that in turn cause business investment to rise)—there is indeed no meaningful constraint on government spending.
But any government spending that increases demand without directly or indirectly increasing supply by the same amount creates an imbalance in ex ante supply and ex ante demand, an imbalance that must be resolved by implicit or explicit transfers. These transfers must reduce the purchasing power of one or more sectors of the economy by enough so that the ex ante gap between demand and supply is reduced to zero. Demand and supply must always balance at all times, and the ability to create credible money at will doesn’t change this.
What is more, and contrary to what many MMT proponents seem to think, taxes and inflation are not the only mechanisms that can resolve the gap between demand and supply. There many such mechanisms, including the following ones.
- Inflation erodes the value of fixed incomes and financial assets, which in turn reduces demand by curbing consumption or investment through income and wealth effects.
- Higher income taxes can reduce consumption by cutting disposable income.
- Wealth and property taxes shift the adjustment cost onto the wealthy mainly through wealth effects, which means taxes must be very high if they are going to reduce demand materially.
- Tariffs shift the adjustment cost onto importers, including all household consumers, and works by forcing up domestic savings. It does this by reducing the real value of household income, and hence consumption, relative to total production.
- Currency depreciation also shifts the adjustment cost mainly onto domestic importers in much the same way that tariffs do.
- Trade deficits, in countries in which investment is not constrained by scarce savings (including all advanced economies), force down domestic savings, shifting the adjustment cost onto workers.
- A shortage of consumer goods, as occurred in the former Soviet Union, can shift the cost onto ordinary households by forcing up involuntary savings.
- Financial repression, often in the form of distorted deposit rates, works a lot like inflation but typically shifts the adjustment cost mainly onto middle-class and working-class savers. This is especially the case in economies in which the banking system is highly controlled and in which there are limited investment alternatives.
- Countries that centralize the purchase of locally produced agricultural commodities can shift the adjustment cost onto farmers by reducing purchase prices. This has often happened in economies like the Soviet Union and China in which the agricultural sector was forced to fund rapid industrialization.
- Any government-controlled monopoly buyer or seller of goods and services does largely the same thing. For instance, a monopoly energy supplier can shift income between producers and consumers of energy by raising or lowering energy prices.
- Higher unemployment transfers the adjustment cost onto workers that lose their jobs. This usually happens as part of a larger adjustment process that causes domestic businesses to cut back on production (through widening trade deficits, for instance).
- Wage repression also effectivity transfers the adjustment cost onto workers. This is especially likely in a highly globalized system in which government spending is matched by an increase in the trade deficit, or in which businesses try to keep wages down to increase international competitiveness. It can also happen when governments use their control of labor unions and wage negotiations to force workers to fund rapid industrialization (under general secretary Nicolae Ceaușescu, this method was used in Romania to fund the repayment of external debt, among other things).
- Reduced business profits shift the adjustment costs onto businesses and may result in lower business investment.
- Falling government spending on the military, the social safety net, and other public goods shifts the cost of adjustment onto the affected sector.
The point is that there are many ways to resolve ex ante imbalances between demand and supply created by government spending, and all these ways involve the same underlying process: some mechanism, whether intended or not, implicitly or explicitly allocates the adjustment cost onto some sector of the economy. Understanding this is important for understanding some of the ways in which excessive debt can undermine an economy.
When Is Debt a Problem?
There are at least four separate consequences of rising debt that can adversely affect the current and subsequent performance of an economy. These include transfers, financial distress, bezzle (or fictional wealth), and additional spillover adjustment costs termed hysteresis.
- Transfers: When rising government debt creates an ex ante disparity between total demand and total supply in an economy, as discussed above, there must be some automatic adjustment mechanism that restores an equilibrium between the two by transferring income from one sector of the economy to another. This transfer mechanism can itself distort the economy in ways that directly undermine growth, though this is not always necessarily the case. This seems to be the main and perhaps only adjustment mechanism that naïve MMT proponents acknowledge.
- Financial distress: Rising government debt can also indirectly undermine economic growth. When there is enough uncertainty about how real debt-servicing costs will be allocated through the explicit or implicit transfers described above, the debt can cause various sectors of the economy to change their behavior in ways that protect themselves from being forced to absorb the real cost of the debt. These behavioral changes either undermine growth, increase financial fragility, or both. What is more, this behavior tends to be highly self-reinforcing.
- Bezzle: As I explained in this August 2021 Carnegie piece, under certain circumstances, the rapid rise of debt can lead to the systematic creation of fictional growth and bezzle. This fictional wealth, the creation of which distorts economic activity, can consist either of inflated asset prices or of the capitalization of expenditures that should more properly be expensed. But bezzle is only temporary, and it is always eventually amortized, although its creation can persist for many years.
- Hysteresis: Finally, as adjustments occur, rising debt can create a kind of hysteresis in which the equilibrating adjustment mechanism creates additional future adjustment costs. The most obvious example is when rising government debt triggers a financial crisis, which in turn either locks the economy into debt-driven deflation or leads to a political crisis. Most economists, when asked to explain why too much debt is a problem, will argue that debt becomes a problem to the extent that it leads to a financial crisis. Besides being totally circular, this argument is wrong for other reasons. A crisis is simply one of the ways—and not even the most costly way economically—in which an economy can adjust from excessively high debt levels.
Debt is a problem when it sets off one or more of these four reactions, which in turn causes economic growth to slow. Each of these mechanisms works in different ways, and while the last one is largely self-explanatory, it is useful to consider the other three ways in greater detail.
Rising Debt Can Distort the Economy Directly by Creating Suboptimal Transfers
It is important again to recognize that rising debt is not a problem when it causes the supply of goods and services to rise along with the demand it creates. When that happens, the debt is effectively self-liquidating, with debt rising no faster than the real debt-servicing capacity of the economy (for which GDP is usually a proxy). This doesn’t mean that rising debt in such cases is irrelevant. To the extent that it causes shifts in the distribution of income, it may benefit individual sectors of the economy at the expense of others, but this is a problem of income distribution, not of debt.
The problem is more concerning when debt rises faster than the country’s real debt-servicing capacity. This occurs when debt boosts the demand for goods and services without directly or indirectly causing an equivalent rise in the production of goods and services. This can happen for a variety of reasons. For example, rising government debt in an economy without excess labor and capacity may fund additional consumption (through welfare payments, for instance), be poured into defense spending, or go toward nonproductive investment in wasted infrastructure (a particular problem in China); ballooning debt could also encourage speculative spikes in prices in the property sector, the stock market, or other assets as demand is boosted through the wealth effect.
In these cases, rising debt increases ex ante demand for goods and services relative to ex ante supply. But because demand must always equal supply, this requires some adjustment mechanism to equilibrate the two (some of which I list above). The first of the three problems with rising debt, then, is if and when the implicit or explicit transfers associated with the adjustment undermine growth directly.
Debt is conventionally viewed as consisting of transfers of resources from the future to the present, but that isn’t quite right. On the contrary, borrowing consists of a current transfer of resources from the lender to the borrower, followed by a future transfer that closes out the loan. This future transfer is normally expected to reverse the original transaction—as resources are transferred from the borrower back to the lender—but this occurs mainly in cases where increases in demand created by the debt are matched by increases in supply. In other cases, depending on a variety of circumstances, the transfers may end up being to or from other sectors of the economy, as described above, that end up bearing the cost of the debt.
One way that excessive debt becomes a problem is when these transfers adversely affect the economy. For example, if the transfers take the form of high levels of inflation or financial repression, they can raise business uncertainty and otherwise distort economic activity. If they take the form of higher taxes, they can undermine what economist John Meynard Keynes called “animal spirits” and reduce investment in risky but productive sectors of the economy. Or if they result in trade deficits, they can force up unemployment or household debt and so on.
Rising Debt Can Distort the Economy Indirectly by Setting Off Financial Distress Costs
A related, and far more important, potential problem is the extent to which these implicit or explicit transfers undermine growth indirectly. Economic agents aren’t stupid or incapable of learning. As uncertainty rises about how real debt-servicing costs are to be allocated, there is also increasing uncertainty about which sector will be forced to absorb the cost and how, so economic agents are likely to alter their behavior in ways that protect themselves.
Historical precedents show the many ways in which various sectors of the economy do this. As rising debt causes a growing gap between ex ante demand and supply, economic agents understand that this gap will be resolved by some combination of means including inflation, higher taxes, rising unemployment, wage suppression, financial repression, capital controls, and currency depreciation. As this happens, households—especially wealthy ones—shift their wealth into movable assets or into foreign currency (capital flight), consumers cut back on spending, home buyers and equipment buyers delay purchases, manufacturers move operations abroad, farmers hoard production or cut back on land development, and workers, if they are allowed to do so, will unionize and become more militant or, if not, they will work less efficiently. In countries where foreigners might be seen as acceptable political targets, foreign businesses in particular are likely to react to uncertainty over debt repayment by liquidating assets and moving abroad.
These actions in turn can cause a host of spillover effects. Business owners may generally disinvest or shorten their investment time horizons, real estate developers may cut back on development projects, and factories may postpone expansion plans, and all these actions reinforce similar behavior in other economic sectors. By reducing economic activity, these behavioral shifts can also reduce the value of existing infrastructure and manufacturing facilities, which in turn makes it harder for local authorities and businesses to service the associated debt. In addition, creditors will raise lending costs and shorten maturities; suppliers who get paid in IOUs will raise their prices, demand cash payments, or refuse to deliver; and policymakers are likely to shorten their policy time horizons as they strive to resolve short-term problems as quickly as possible, even if these policies are not optimal for the medium and long term. If any of these behaviors by households, businesses, foreigners, and policymakers affect government revenues, governments may respond by cutting back on spending.
The point is this: to the extent rising debt must lead to some sort of adjustment in which certain economic sectors will lose purchasing power to equilibrate demand and supply, each sector recognizes the risk and changes its behavior to protect itself from being forced to absorb the cost. These behavioral changes cause growth to slow, exacerbating the adjustment costs that must be borne by vulnerable economic agents as financial distress automatically spreads through the economy.
This process has been on display most recently with Beijing’s clampdown on China’s overleveraged property sector. Until last year, as long as everyone assumed the debts of the property sector would be resolved through implicit government guarantees, no one cared much about how the debt payments would be allocated.
Once the market believed, however, that debts were no longer guaranteed, every part of the real estate sector—property developers, contractors, suppliers, creditors, homebuyers, and local governments—worried about how the costs of insolvency would be implicitly or explicitly distributed, so each sector altered its behavior to protect itself. Lenders withheld credit, suppliers halted deliveries, contractors stopped construction projects, homeowners postponed their purchases, local governments suffered revenue declines and cut back on spending, and so on. And as they did so, the problems within the property sector deepened and spread to other, less-leveraged developers as well as to upstream suppliers and downstream clients.
Once that happens, only credible guarantees of the debt—which do not solve the debt problems but effectively transfer them to the guarantor—can prevent things from spiraling out of control. The past few months in China have illustrated very clearly how debt-related uncertainty can spread through the real economy.
Finance specialists will note that a lot of this resembles corporate-finance discussions about the behavior of stakeholders in a business whose probability of default is rising. In fact, what I describe above is the macroeconomic equivalent of something that is well understood in corporate finance circles. Governments are different from businesses because they do not run the same kind of default risk as businesses do, but their debt-servicing capacity is still constrained by the production of goods and services in the real economy, and as uncertainty over the allocation of their debt-servicing costs rises, it has a very similar impact on stakeholders.
Unfortunately, while this behavior is well-documented and well-understood in corporate finance, its macroeconomic counterpart seems to be much less familiar to economists, despite many historical precedents. The United States in the early 1930s illustrated an especially virulent form of the automatic spread—once it was set off—of financial distress and debt deflation, as did low-income countries like Brazil and Mexico during the debt crisis of the 1980s, Japan in the 1990s, and so on.
In either the corporate case or the macroeconomic case, by changing the behavior of economic agents, behaviors associated with financial distress cause growth to slow, uncertainty to rise, and balance sheets to become more fragile, with the latter two effects causing growth to further slow. This behavior is especially destructive because it is so highly self-reinforcing: rising levels of debt create increasing uncertainty about how the associated costs will be allocated, which sets off financial distress behavior that undermines growth, further increasing the gap between debt-servicing needs and debt-servicing capacity and driving even more financial distress behavior.
Debt Can Create Bezzle
The third of the economic problems associated with rising debt is that it can encourage and accommodate a rise in fictitious wealth or bezzle. This is often the most damaging consequence of rising debt because this fictitious wealth creates distortions in economic behavior both as it is created and, much more importantly, as it is destroyed. When this fictitious wealth is eventually destroyed, the process can occur either quickly, in the form of a financial crisis, or slowly in the form of lost decades of stagnation and low growth.
There are two main ways in which the impact of fictitious wealth can affect economic activity, and although the link between rising debt and the creation of fictitious wealth isn’t always causal, even when it isn’t they are usually both symptoms of the same set of underlying monetary distortions.
First, rising debt can be associated with soaring prices for stocks, real estate, and other assets, prices that exceed what is justified by their future contribution to the production of goods and services. When this happens, as Berkshire Hathaway vice chairman Charlie Munger has explained, the perception of wealth can exceed the reality of wealth. The linking factor between this form of fictitious wealth and debt is probably rapid expansion in underlying liquidity—whether this is created by direct money creation, by an increase in asset liquidity, or by advances in financial technology—that feeds into speculative asset purchases as prices increasingly diverge from fundamental expectations.
Second, rising debt in certain economies can also be the result of formal or informal pressure on banks to lend into nonproductive investment that, because of soft budget constraints, isn’t written down for many years. (This has occurred most notoriously China in recent years, Japan in the 1980s when investment was subject to window guidance, and the Soviet Union and Brazil in the 1970s.) When that happens, expenditures that should be expensed are in fact capitalized, causing income-statement expenses to be understated and balance-sheet assets to be overstated, so that both net income and wealth are artificially boosted to higher levels than they otherwise would have been if only productive activity were recorded.
Both forms of fictitious wealth creation feed into economic activity, but in different ways. In the former case, when excessive liquidity drives up the prices of assets, the main impact on the economy is indirect, through the wealth effect. Higher asset prices make the households who own the assets feel richer, and as a result they may spend more and save less out of current income than they would normally choose. Higher asset prices also increase the value of assets that can be used to collateralize debt, making it easier (and cheaper) to borrow to fund spending.
In the second instance, nonproductive spending also affects economic activity directly, through an income effect. Money spent on nonproductive investment boosts income and economic activity in the same way as money spent on productive investment, except that in the latter case the increase in income and economic activity is justified by a real increase in the future value of goods and services produced by the economy, whereas in the former case, it isn’t.
Both forms of fictitious wealth creation boost domestic demand and economic activity (usually when debt is rising), but these artificial increases in demand and economic activity must eventually be reversed, and usually much more sharply. This happens when the factors creating fictitious wealth are eventually reversed once debt stops rising and stabilizes, or when it begins to decline. At that point, households and businesses feel collectively poorer than they had, and as the value of collateral declines, households and businesses come under increasing pressure to pay down debt.1
When pressure on banks leads to substantial amounts of nonproductive investment and surging leverage ratios, the impact of this nonproductive investment on economic activity is much greater. As it is being created, economic activity is boosted directly by the investment spending, and this activity overstates the real value created for the economy. It is as if workers, contractors, and suppliers were paid $100 first to build a house and then another $100 to knock it down again, except that rather than recording the combined $200 as wasted spending or as a rise in costs, they record it as a $200 rise in the value of assets they own. In this example, the economic agents involved would collectively record a $200 rise in their wealth even though there has been no real wealth creation. This increase in wealth shows up typically in the form of loans on a bank’s balance sheet against which there are $200 of income-earning assets.2
More importantly, while this activity can continue for many years, it immediately begins to depress future growth by contributing less to future economic activity than its value implies. In the same way, just as the creation of fictitious wealth leads to more spending through wealth effects (and thus more growth), the amortization of fictitious wealth depresses spending and growth. In many cases, this downward pressure on growth may encourage monetary authorities to expand the domestic money supply even faster, so that the net creation of fictitious wealth (new bezzle minus the growing amortization of old bezzle) keeps the growth in economic activity constant. Bezzle creation, in such cases, must accelerate to prevent economic activity from slowing. What is more, because the resulting (lower) amount of growth is measured against a higher base GDP level, the GDP growth rate is further depressed.
The Effect of Nonproductive Investment on the Lending Banks
Some economists and analysts argue that countries like China, in which there are substantial amounts of nonproductive spending, don’t face significant adjustment costs because both the borrowers and the banks are state-owned and state-controlled. Local governments, they say, can simply write off the debt.
This Kramer form of analysis suggests just how poorly most analysts understand the consequences of debt. Consider a bank with $10 of capital and $90 of deposits that makes a $100 loan into a project that ultimately generates only $50 dollars of economic value. Whether or not the bad loan is written off, the bank must still come up with the resources to service the deposits. To say the investment project generates only $50 dollars of economic value is also to say it generates enough resources to service only $50 dollars of the bank’s deposits.
There are many ways a banking system can resolve this mismatch between the earning power of its assets and the obligations on its liabilities. The bank can default on its depositors and force them immediately to take the loss. Alternatively, bank regulators, worried about the political consequences of a bank default, may use some form of financial repression to force down the deposit interest rate. This is effectively a hidden tax on savings, and it spreads the recognition of the loss out over many years. This is what Chinese regulators did in the 2000s to clean up the pre-IPO banks, during which time the household share of GDP dropped sharply, largely because of this hidden tax. Finally, regulators can recapitalize the bank at the expense of taxpayers or holders of monetary assets.
Whichever method the banking system uses, either the depositors must take a loss up front, or they must take the loss over a longer period of time, or some other sector of the economy must absorb the loss. Writing off bad debt isn’t a mysterious process that makes losses disappear. It is simply the process of explicitly allocating a previously unacknowledged loss.
Duck Soup Economics
One way to understand how the bezzle associated with capitalized, nonproductive investment will ultimately affect future growth is through a thought experiment. Assume that there are two countries—Fredonia and Sylvania—that are economically identical in every way except that Sylvania, unlike Fredonia, begins to use its underutilized resources to engage systematically in investment that has no impact on increasing the future value of goods and services. Assume further that in Sylvania, this nonproductive investment is not expensed but instead is capitalized as an asset.
One result will be that for some time, Sylvania’s recorded GDP and its recorded wealth will grow faster than Fredonia’s. Even if they do not create economic value, bridges to nowhere nonetheless increase economic activity in Sylvania in ways that show up in its GDP growth data, and this increase is matched by an increase in the reported value of assets and loans held by local entities and banks. Sylvania would continue to record higher GDP growth and higher wealth accumulation than Fredonia until Sylvania cannot continue increasing debt any longer—usually because soaring debt levels convince, or force, Sylvanian officials to cut back on nonproductive investment.
When that happens, the tides begin to turn. Whereas the two countries’ GDPs and measures of collective wealth diverged in Sylvania’s favor for some period of time, over some subsequent period of time these measures should once again converge. The two countries, after all, are identical in terms of the real value of goods and services they produce.
But for them to converge, Sylvania’s nominal GDP growth must decline in the future relative to that of Fredonia’s and relative to what it would have been without seemingly inflated past growth. If the two economies produce the same real value of goods and services, then clearly any period during which one economy’s GDP growth outpaces the other’s (for reasons that have to do mainly with bad accounting) must be followed by a period in which these distortions are reversed. This is the only way that their reported GDP can converge to reflect the equality between their real economies.
But while the two economies should in principle converge as the bezzle created by nonproductive investment is wrung out of the system, in practice, there is likely to be additional costs that will reduce the value of Sylvania’s real economy to below that of Fredonia’s. These are the incremental or frictional costs associated with the economic distortions created during Sylvania’s many years of nonproductive economic activity. As Sylvania’s GDP growth and wealth is first boosted by the creation of bezzle, and later reduced by the amortization of this bezzle, this trend will have distorted the behavior of businesses and households in ways that ultimately leave the country poorer than if it had never engaged in bezzle creation. This does not even consider the effects of the potential financial distress associated with Sylvania’s higher debt.
This is just a way of pointing out what should be obvious. Anything that for some period of time artificially inflates reported growth, relative to some unbiased proxy for real economic value, must result eventually in lower reported growth, as reported growth ultimately converges with—or, more likely, drops below—that of the proxy.
Mapping the Impact of Excessive Debt onto the United States and China
Because the United States and China are the world’s two largest economies and the two most commonly compared—and because the structures of their economies, their financial systems, and their debt are so different—it might make sense to compare the two on these points to show just how different the adverse impact of too much debt in different countries can be.
Before making such a comparison, it is worth noting that there are many reasons, which are usually ignored, unfortunately—that can make direct comparisons of debt levels between two countries of very limited use. These include income levels, debt structure, and underlying economic volatility.
On the subject of income levels, if a group of countries are graphed by their debt levels (debt-to-GDP ratios) versus their income levels (GDP per capita), it becomes obvious that the relationship has a positive slope. This means that the richer a country is, the higher the debt level it is able to sustain. That being the case, comparing the debt level of a rich country with that of a middle-income or poor country understates the riskiness of the latter.
As for debt structure, like I explain in my 2001 book, The Volatility Machine, debt can be structured in such a way that debt-servicing costs are either positively correlated or inversely correlated with debt-servicing capacity. In the latter case, which I call “inverted” debt—which includes funding long-term assets with short-term debt, funding domestic assets with external debt, collateralizing debt with assets whose prices are boosted by rising debt—the structure of the debt automatically increases the volatility in the relationship between debt-servicing costs and underlying economic performance, thus increasing the riskiness of each unit of debt. If two countries have similar debt levels, in other words, the debt in the country with a more inverted balance sheet structure is more likely to be excessive.
Finally, the more volatile a country’s underlying economy is—perhaps because it is less diversified, more correlated to global commodity cycles, or more poorly managed—the less efficiently it can absorb any unit of debt. Because the financial distress cost of debt to an economy is a function of debt-servicing uncertainty, the more volatile an economy, the greater the financial distress costs of any fixed measure of debt.
These various conditions all work by exacerbating the four consequences of excessive debt discussed in this essay. It is instructive to compare China and the United States according to each of the consequences: transfers, financial distress, bezzle, and hysteresis.
Because of the very different structures of their economies and their balance sheets, it is hard to compare the United States and China on the basis of the costs associated with the future transfers that will ultimately equilibrate supply and demand (real debt-servicing costs and real debt-servicing capacity). This problem is compounded by the fact that no one knows as of yet how the costs will be distributed.
In the past, however, it has almost always been ordinary households who are forced to absorb the cost of bad debt—through taxes, inflation, financial repression, and unemployment— probably because they are politically the most vulnerable sectors of the economy. If that is also the case in the future, China’s unbalanced economy compared to that of the United States, with its extremely low household share of GDP, means that the economic distortions created by transfers from the household sector are likely to be more costly for China than for the United States.3
In such an event, China would once again be forced to choose between much slower growth (as the consumption share of GDP declines) or a much more rapid increase in the country’s debt burden, as growth would rely increasingly on nonproductive investment. To keep this from happening, China must find ways of allocating the debt-servicing costs to sectors other than households, a task that has always proven highly difficult politically.
On the other hand, rising debt in the United States has been associated with large trade deficits, unlike in China where large surpluses have occurred.4 For that reason, part of the U.S. adjustment costs will consist of transfers to foreigners, whereas in China these costs will consist of transfers from foreigners. This dynamic could lower the relative cost to China of a unit of debt compared with the United States.5
The ways in which financial distress costs work though an economy are likely to be extremely different in countries as different as the United States and China. I would argue that there are at least three major relevant differences between the two.
First, the U.S. financial system is among the most flexible in the world and has more moving parts, unlike the fairly rigid Chinese financial system, which is dominated by government-supported banks. Both banking systems are underpinned to some extent by moral hazard, and moral hazard complicates the allocation of insolvency costs, but moral hazard is almost certainly a much bigger issue for Chinese banks than for U.S. banks. What is more, in China the banks completely dominate the financing of households, businesses, and the government, unlike in the United States where bond and equity financing are roughly twice as important.
Second, governance and the legal and bankruptcy systems are more transparent and institutionalized in the United States than in China, and the U.S. system has clearer procedures that are less susceptible to political interference. This creates useful precedents for how insolvent borrowers are to be liquidated, how assets are to be divided among creditors, and how insolvency costs are to be allocated and absorbed. All of these factors reduce uncertainty about the extent to which economic agents will be forced to absorb insolvency costs.
Third, the Chinese government is a much more powerful economic player within China than the U.S. government is within the United States. To the extent that Beijing is able to identify the least damaging way in which to allocate debt-servicing costs, and proves able to implement this allocation of costs quickly and efficiently, it may be able to reduce financial distress effects more effectively than the United States can. Of course, the opposite would be true if Beijing is not able to identify the least damaging way in which to allocate debt-servicing costs, or if political objectives interfere with the government’s economic objectives.
I would argue that these differences imply that, for any unit of uncertainty about the gap between debt-servicing costs and debt-servicing capacity, because there is likely to be more uncertainty in China about how the costs of insolvency will be allocated, financial distress costs in China are likely to be larger than in the United States, although in theory Beijing could allocate costs more efficiently and in doing so reduce financial distress costs.
One way of demonstrating this is in observing the behavior of capital flight, business disinvestment, and dollar hoarding by the middle classes. These have often been major issues in China but are almost nonissues in the United States. Among other things, this suggests that economic agents in China are more uncertain about their ability to protect their domestic assets than those in the United States are, and for that reason Chinese economic actors are more likely to engage in financial distress behavior.
The next year or two may offer a chance to see how insolvency is resolved in the Chinese property sector as a guide to how China would handle more widespread insolvency generally. As insolvency among property developers evolves, Beijing either will opt for a transparent and rules-based bankruptcy process, or it will choose instead to resolve each case as a function of local politics and the relative power of different agents. A recent announcement by American fund Oaktree Capital that it had seized as collateral a major Evergrande development near Shanghai may be a step in the right direction. If Oaktree is able to manage, dispose of, or liquidate the asset in a way that maximizes its recovery value, and if this becomes a replicable model for both Chinese and foreign creditors, it will help reduce the costs of other corporate insolvencies.
This will matter in the future as bankruptcies spread. The decision about whether insolvencies are resolved according to a well-understood set of rules, or according to specific political circumstances, will necessarily affect uncertainty about how the costs of insolvency will be allocated and whether future loans are made on commercial grounds or for political purposes. Such decisionmaking will also help determine whether lenders put resources into increasing political access or into understanding their projects’ commercial merits.
Financial distress behavior and costs are not just driven by how transparent the cost allocation process is but, perhaps more importantly, how large the insolvency gap between debt-servicing costs and debt-servicing capacity is. Here is where the difference between the two economies is probably greatest. How big are the bezzle effects likely to be in either economy?
Stock Market. At roughly $50–60 trillion in estimated value, the U.S. stock markets are equal in size to roughly 250 percent of the U.S. economy and are about four times the size of the Chinese stock markets (about $10–15 trillion), which are in turn equal in size to roughly 70 percent of the Chinese economy. Obviously, it is extremely difficult to calculate the extent of overvaluation in either market. but it seems that if both markets are considered to be overvalued to the same extent, the U.S. stock market is likely to contain roughly three to four times as much bezzle as the Chinese stock market, even adjusted for the relative sizes of the economy. Based on those calculations, for every 10 percent overvaluation in the market, in other words, bezzle is equal to 25 percentage points of GDP in the United States and 7 percentage points in China. Other forms of equity ownership in both countries that are subject to overvaluation (such as private equity or venture capital) may also involve systematic overvaluation, but they are much smaller.
Real Estate Market. It is hard to get precise numbers for either country’s real estate markets, but according to one source, U.S. residential and commercial real estate is worth $30–35 trillion while Chinese residential and commercial real estate is worth $80–85 trillion.6 Other sources place the value of Chinese residential real estate at $55 trillion and U.S. residential real estate at less than half that. Given that the U.S. economy is larger—about $21 trillion in reported GDP in 2020 versus approximately $15 trillion—it seems reasonable to assume that the value of U.S. real estate should exceed that of China. Assuming that is the case, and using the lower valuation numbers, the overvaluation of Chinese real estate might exceed any overvaluation in the United States by a minimum of $27 trillion, and perhaps as much as twice that. This implies that fictitious wealth in Chinese real estate could easily exceed that of the United States by 150-200 percentage points of China’s GDP. This clearly is a potentially enormous problem, especially given that real estate represents around 60 percent of the household wealth in China.
Capitalized Spending on Nonproductive Investment. It is very hard to get estimates here, but in economies with hard budget constraints, like in the United States, there is relatively little systematic misallocation of investment, and when it does occur, it is quickly written down. In an economy like that of China, where governments, state-owned entities, and some private entities are able to operate under soft budget constraints, nonproductive investment can be a much larger and longer-lasting problem. In China, investment makes up roughly 45 percent of GDP, and the property sector and infrastructure spending – the two sectors underpinned by moral hazard – have accounted for roughly 60 percent of this total. That means that the two sectors most susceptible to systematic nonproductive investment and the creation of bezzle comprise about 25 to 30 percent of GDP. For every 10 percent misallocation, in other words, there has been bezzle creation of 2 to 3 percentage points of GDP, perhaps since as early as 2008 or 2009, when Chinese debt first began to soar relative to GDP.
Hysteresis effects are impossible to predict, but given the nature of China’s closed capital markets and powerful regulators, it is probably safe to say this: on the one hand, China is less likely to suffer from a financial crisis then the United States, but on the other hand, the U.S. system is more able to absorb the social and political costs of a financial crisis. What is more, given the greater premium the Chinese political system places on stability, a sharp, unexpected economic slowdown is more likely to be politically disruptive in China than in the United States.
A McKinsey Report on Global Wealth
A November 2021 McKinsey report on global wealth reached broadly similar conclusions. Worryingly, it concluded that measures of global wealth, of which real estate accounts for 68 percent, have risen in the past two decades to levels that are roughly 40–50 percent higher than normal. The report reinforces the idea that there may be a great deal of bezzle in the world’s major economies, underscoring that this is especially likely to be a problem in China.
China, according to the report, accounted for 50 percent of the growth in global wealth from 2000 to 2020, followed by the United States at 22 percent. During that same two decades, China accounted for 26 percent of the growth in global GDP, while the United States contributed 21 percent.7 The McKinsey study lists the measure they use for national wealth as a multiple of national GDP (see table 1).
|Table 1. A Proxy for Measuring Major Countries’ National GDP|
|Source: Jonathan Woetzel, Jan Mischke, Anu Madgavkar, Eckart Windhagen, Sven Smit, et al., “The Rise and Rise of the Global Balance Sheet: How Productively Are We Using Our Wealth?” McKinsey Global Institute, November 15, 2021, https://www.mckinsey.com/industries/financial-services/our-insights/the-rise-and-rise-of-the-global-balance-sheet-how-productively-are-we-using-our-wealth.|
According to their data, the only major economy that has ever had a GDP multiple above 8 was, perhaps not surprisingly, Japan in 1990, although at the time global wealth was 4.5 times global GDP, where it stands at roughly 6.1 times today. It is probably not a coincidence that the Japanese multiple more than doubled in the twenty years before 1990, during a time of extremely high Japanese GDP growth rates; nor is it a coincide that in the subsequent fifteen years, as GDP growth collapsed to near zero, its multiple dropped by 25 percent even as the global multiple rose by 20 percent. This is consistent with the idea that the creation of bezzle boosts economic activity while its amortization reduces economic activity.
I suspect that most readers who have gotten this far will have hoped for me to provide some quantitative comparison of debt in the two countries by the end, but I hope that I have shown how difficult, even impossible, such a task would be. The only thing we can say for certain is that it makes little sense to compare debt-to-GDP ratios for countries that have very different financial systems, economies, and hard budget constraints.
The important point is to understand how high levels of debt are likely to affect the subsequent performance of an economy. Excessive debt can undermine economic performance when it is followed by transfers that are economically suboptimal. More importantly, these transfers can set off financial distress behavior that undermines subsequent growth, in many cases substantially. In addition, to the extent that excessive debt creates fictitious wealth, it can boost current growth in suboptimal ways and harm future growth as the trend reverses. And finally, excessive debt can set off unpredictable hysteresis effects.
What is often forgotten (until it becomes obvious, by which time it is too late) is how highly self-reinforcing these effects can be. As the impact of a high debt burden eventually causes growth to slow, this slower growth in turn raises the existing debt burden and causes new debt problems to emerge in sectors once though relatively safe. These dynamics in turn reinforce factors that caused growth to slow. It is notable that in almost every case in history when a country’s rapid growth has been associated with even more rapid growth in its debt burden, the subsequent adjustment has always turned out to far more difficult than even pessimists had predicted. People have always systematically underestimated the positive impact on economic activity of rising debt and the negative (and asymmetrical) impact on economic activity of the subsequent adjustment. There is little reason to believe that the future will be much different.
Aside from this blog, I write a monthly newsletter that focuses especially on global imbalances and the Chinese economy. Those who would like a subscription to the newsletter should write to me at firstname.lastname@example.org, stating their affiliations. My Twitter handle is @michaelxpettis.
1 The extent of this wealth effect probably depends to some degree on the extent of income inequality. When the bulk of the assets are owned by a small elite, the wealth effect on spending is likely to be lower than when the ownership of assets is widely distributed. In addition, the wealth effect is probably concave in the sense that a unit increase in wealth tends to boost spending by less than the reduction in spending caused by a unit decrease in wealth.
2 Although this fictitious wealth is associated with rising debt—because it is usually the banking system that funds soft budget entities—the economic impact lies not in the debt itself but rather in the fictitious wealth. One way to understand this is to consider what would happen if this nonproductive investment were funded by equity rather than debt. This would slightly improve matters by potentially reducing financial distress costs associated with debt, but ultimately the fictitious wealth still creates economic activity that is recorded as an increase in income and wealth rather than an increase in expenses.
3 Observers saw how this works, for example, in the cleanup of Chinese banks in the 2000s through high levels of financial repression. This caused China’s already very low household share of GDP in 2000 to drop by more than 10 percentage between 2000 and 2010, to the lowest level ever recorded, a problem that still hasn’t been resolved more than a decade later.
4 Much of the rise in U.S. debt has been driven by consumer debt, whereas in China it has been driven by producer debt.
5 Unless the United States chooses inflation as its main adjustment mechanism, in which case foreign holders of U.S. debt will bear a disproportionate cost.
6 According to this source, residential real estate is worth roughly four times the value of commercial real estate in the United States and roughly eighteen times as much in China.
7 The report also notes that Japan held 11 percent of global wealth in 2000, down substantially from the 31 percent it held in 2000.