Modern monetary theory (MMT) isn’t discussed much in Chinese universities as far as I can tell, except by a few of my Chinese friends and former students who mostly work in banking and finance. Analysts abroad often claim that China is somehow proof that MMT “works,” but to me and most of my Chinese friends, this statement is either unintelligible or just wrong.

Every Sunday afternoon, a few of the brightest math, finance, and economics students from Peking University converge on my courtyard house for an informal seminar on debt-related issues. For our first meeting of the new school year, I had them read up on MMT and research the MMT debate so that we could work out how it might apply to their own thinking.

From these readings (including a 1943 piece on “functional finance” by economist Abba Lerner), we assume that the key insight of MMT is not that debt doesn’t matter, as mistaken stereotypical views seem to assume, but rather that governments that issue their own fiat currency have no funding constraints insofar as they do not need to issue debt or raise taxes to spend money. They simply budget the expenditure, and then go ahead and spend the money.

There is a lot of confusion about this. A January 2019 Financial Times article had this to say:

Advocates for modern monetary theory argue that, for a sovereign country with its own currency, there is no inherently unacceptable level of government debt—that country does not automatically begin to collapse when debt reaches 90 per cent of GDP, or even 200 per cent of GDP. The country appropriates what it believes is necessary for domestic programs, regardless of revenue.

This, however, is almost certainly an unfair caricature of MMT. It assumes that if a country increases debt to fund spending until the economy is at capacity, it will cause the country’s debt-to-GDP ratio to rise. But if government spending directly or indirectly causes productive investment to rise in line with the debt, this kind of spending increases both debt and GDP, so neither the debt ratio nor the debt burden changes.

It is only when money is borrowed (or created) and spent in ways that do not cause GDP to rise that the debt burden rises. For that reason, a rising debt-to-GDP ratio over the medium term is almost prima facie evidence that the government should cut back its spending (over the short-term there can be timing differences between when an investment is made and when it starts to pay off). This is also true of inflation: if the government “prints money” to spend on projects that reduce excess capacity or employ workers productively, the result will not be inflationary to the extent that the increase in demand caused by printing the money will be matched by an increase in supply.

What MMT actually does say is that governments must issue debt or raise taxes if either is necessary to control the potentially adverse economic impact of the additional demand created by spending the additional money. As Lerner puts it:

The first financial responsibility of the government (since nobody else can undertake that responsibility) is to keep the total rate of spending in the country on goods and services neither greater nor less than that rate which at the current prices would buy all the goods that it is possible to produce. If total spending is allowed to go above this there will be inflation, and if it is allowed to go below this there will be unemployment. The government can increase total spending by spending more itself or by reducing taxes so that the taxpayers have more money left to spend. It can reduce total spending by spending less itself or by raising taxes so that taxpayers have less money left to spend. By these means total spending can be kept at the required level, where it will be enough to buy the goods that can be produced by all who want to work, and yet not enough to bring inflation by demanding (at current prices) more than can be produced.

Lerner points out that this idea, “like almost every important discovery,” is very simple (albeit counterintuitive), to the point that when academics understand it, they often dismiss it as “merely logical.” He goes on to explain:

An interesting, and to many a shocking, corollary is that taxing is never to be undertaken merely because the government needs to make money payments. According to the principles of Functional Finance, taxation must be judged only by its effects. Its main effects are two: the taxpayer has less money left to spend and the government has more money. The second effect can be brought about so much more easily by printing the money that only the first effect is significant. Taxation should therefore be imposed only when it is desirable that the taxpayers shall have less money to spend, for example, when they would otherwise spend enough to bring about inflation.

This seems to be where much of the confusion lies. One of the main criticisms of MMT is that it seems to imply to some people that governments can spend on any project they like without worrying about the consequences. By extension, these critics also assume that for this reason there are no effective limits to government borrowing: debt can always be serviced by creating additional fiat money for the sole purpose of servicing the debt.

To sort through these various claims, our seminar decided to concentrate on the conditions under which there are no intrinsic constraints on government spending, a state that can be called MMT heaven. We also considered the conditions under which there are constraints, in which case governments would have to raise taxes to balance the government expenditures. To simplify matters, we decided that the government could basically spend the money in three ways:

  1. Give to the rich: The government can fund projects that effectively give money to businesses or the rich, who, for simplicity’s sake, can be defined as entities who consume little or none of any incremental increase in income and who therefore save all or most of it. The government can do this by cutting taxes on businesses and the rich, or by engaging in polices that directly or indirectly increase the income or wealth of the rich, such as certain kinds of quantitative easing, which tend to cause a rise in the prices of assets, most of which are owned by the rich.
  2. Give to the poor: The government can fund projects that effectively give money to average or poor households, who can be defined as entities who consume most of any incremental increase in income and who therefore save little of it. The government can do so by cutting taxes, funding social safety nets, creating jobs, or setting minimum basic income policies, and so on.
  3. Build infrastructure: The government can also spend more on infrastructure.

In all three of these cases, the bottom line is this: if the government can spend these additional funds in ways that make GDP grow faster than debt, politicians don’t have to worry about runaway inflation or the piling up of debt. But if this money isn’t used productively, the opposite is true. The flowchart below shows how each of these forms of government spending might or might not affect the economy.

Figure 1

Give More Money to the Rich

Taking the first case, assume that the government implements central bank or fiscal policies that effectively deliver additional income to the rich, who (by definition) save all or most of these proceeds. These policies could be enacted under two different sets of circumstances. In the first scenario, assume that there are supply-side constraints on investment and that interest rates are quite high. Businesses have many profitable investment opportunities that would increase productivity, in other words, but they are unable to invest in many or most of them because the cost of capital exceeds the risk-adjusted expected return. To put things another way, desired investment in this economy exceeds actual investment.

Under these conditions, creating money and passing it on to the rich may lead at first to inflation, depending on how tight the resources and labor markets are, but this effect should be temporary. These policies increase savings and consequently lower the cost of capital. This causes businesses to invest more and labor productivity to increase, so the total value of goods and services produced in the economy (GDP) also rises. Because GDP goes up by more than the amount of additional money created, in the end there is no inflation and the additional value created is more than enough to service the debt.

Essentially, the government can create money or debt, so to speak, to fund transfers to businesses or to the rich without worrying about funding the expenditures through taxes or borrowing. This is what is meant by MMT heaven.

But now assume that there are no supply-side constraints on investment: this second scenario looks a lot like today’s world, with low interest rates, weak demand, and businesses hoarding huge amounts of cash for which they can find little purpose beyond speculating on stock buybacks or acquisitions. The reason businesses don’t invest, in other words, is not because they cannot access reasonably priced capital but rather because demand for their products is too weak to justify investing in additional capacity.

If that is the case, what happens when the government borrows or creates money and passes it on to the rich? As I discussed in a June 2017 blog post, when such policies are enacted under such conditions, there is no corresponding increase in investment because the inability to access savings was never a constraint on investment. This also means that, in a closed economy (or in an open economy whose capital account, like that of the United States, is determined by foreign conditions), there can be no increase in savings within the economy. This might sound strange, given that the savings of the rich definitely would have increased, but this only means that the increase in the savings of the rich had to result in a decrease in savings elsewhere in the economy. (In an open economy, of course, the increase in the savings of the rich would flow out in the form of a capital account deficit.)

There are many ways that savings in another part of the economy could decline. If the government borrowed the money, for instance, the corresponding decrease in savings could be explained by a rise in government debt (which is just negative savings). If instead the central bank created the money and used it to buy assets, the resulting rise in asset prices could create a wealth effect that would cause households to consume a larger share of their income (and therefore save less), or it could cause lenders to lower their lending standards and encourage a rise in household debt. Or if the increase in savings were to flow abroad, it would be matched by a current account surplus as some of the goods produced at home were exported. The larger point is that, however it happens, because the increase in domestic or foreign demand is not met by an increase in the value of goods and services produced, ultimately these policies enacted under these conditions must result in higher inflation, which in turn would reduce real household income to balance out the increase in the wealth of the rich.

A second adverse effect could easily occur too. Because this policy effectively would transfer wealth from the poor to the rich, there would be a net reduction in consumption. In a closed economy (or in an economy that has little or no control over its capital account, as I explain here) that net reduction would result in a further decrease in business investment and with it, by necessity, a further decrease in national savings, driven by a rise in unemployment, household debt, or the fiscal deficit.

Again, creating or borrowing money does not increase a country’s wealth unless doing so results directly or indirectly in an increase in productive investment. In this second scenario, creating or borrowing money and passing it on to businesses or the rich only redistributes wealth from those who did not benefit from government policies to those who did. That requires that the government raise taxes to reduce demand and prevent unwanted inflation.

Give More Money to the Poor

In the second case, the government enacts policies that directly or indirectly give more money to average or poor households. Again, these policies could be enacted in two different scenarios. For the first scenario, assume that the economy operates under demand-side constraints. This mean that the economy suffers from weak demand—consumption and investment—and that there is resource and labor slack, or inefficiencies in the economy.

Under these circumstances, delivering money to the poor directly boosts consumption (because, by definition, they consume most additional income) and indirectly boosts investment as businesses increase capacity to meet this additional demand. There are various iterations in which the increase in household income would be partly consumed and partly saved, with the saved portion funding new investment to serve the increase in consumption, which further increases household income and, in doing so, creates additional consumption and savings, and so on. The basic arithmetic (found in any introductory economics textbook) isn’t worth getting into, but it is easy to show that as long as there is slack in the economy, both investment and savings will rise as consumption is boosted.

Once again, the result is MMT heaven. The government can create money or debt, so to speak, to fund transfers to the poor, but because the consequence is that investment grows by at least as much as the amount of money or debt, there is neither inflation nor an increase in the country’s debt burden (as the debt grows by less than GDP does, which is a proxy for debt-servicing capacity).

In the second scenario, however, assume that there are no demand side constraints. In such a world, growth is constrained by high costs of capital, regulations, or other supply-side constraints. Under these conditions, delivering money to the poor boosts consumption, but because the supply-side constraints aren’t addressed there is no equivalent boost in investment or GDP. This means that, at current prices, total demand for goods and services has risen with no equivalent rise in the supply, so the country must suffer from inflation or, if unlike the United States it has control over its current account, it must run a trade deficit, which is balanced by equivalent capital outflows. The way to prevent either outcome is for the government to raise taxes enough to reduce aggregate demand to its original level.

Spend Money on Infrastructure

For the third case, a government could pour money into building new infrastructure. Again, assume that this policy could be enacted under one of two scenarios: either there is substantial room for productive infrastructure spending or there isn’t. In the first scenario, if the government creates debt or money to fund productive infrastructure spending, the result is MMT heaven once again because the total value of GDP—that is, the goods and services produced by the economy—rises faster than the money created, so there is no inflation, and rises faster than the debt created, so the country’s debt burden doesn’t increase.

In the second scenario, government spending on infrastructure is largely nonproductive and creates little value for the economy—as is the case in China today. This time, the consequences are a little more complicated because they can involve wealth transfers from one group to another. But ultimately, either money increases faster than the value of goods and services produced, in which case there is inflation, or debt rises faster than the country’s overall debt-servicing capacity.

For an example of MMT working under the first scenario, consider China in the 1990s. Money creation and debt went mostly to fund businesses and the rich (as ordinary households were forced to heavily subsidize borrowing costs), thus forcing up Chinese savings rates; the proceeds were then poured into productive investment, generating enough growth to trickle down to the poor at a substantial rate. Perhaps that is what people mean when they say that China supposedly proves the validity of MMT. There seemed to be no constraints on the country’s ability to expand its money supply directly or via increases in debt because real GDP grew at least as quickly as the money supply and the debt, so there was little inflation and no increase in the country’s debt burden (since debt grew no faster than GDP).

More recently, however, China represents the second scenario. Money creation or debt still go mostly to fund businesses and the rich, but there are few productive investment opportunities left. Although investment still continues to rise quickly, much of this investment is nonproductive and only shows up as increases in GDP because the entities responsible for the investment are not subject to hard budget constraints, so they never write down bad investments on the books. That being the case, debt creation causes a surge in the country’s debt burden as debt rises much faster than the country’s debt-servicing capacity. Because of the distribution of income, money creation shows up more as asset price inflation than as consumer price inflation.

When Do Increases in Money or Debt Matter?

As always in economics, it depends on the circumstances. It is easy to consider cases in which rising debt or money creation results in an equivalent or greater increase in total GDP or debt-servicing capacity. In such cases, in the aggregate, an increase in debt or money mostly doesn’t matter insofar as it results in greater wealth and a declining debt burden. It is also easy to consider cases in which rising debt or money creation doesn’t result in an equivalent increase in total debt-servicing capacity. In these scenarios, in the aggregate, it does matter because rising debt or money results in less wealth due to inflation and/or a rising debt burden.

For simplicity’s sake, it can be assumed that rising debt or money creation doesn’t matter under the following conditions that characterize MMT heaven:

  • When it involves a transfer of wealth to the rich or to businesses—in a way that boosts domestic savings—in a supply-constrained economy with high capital costs (typically, developing countries);
  • When it involves a transfer of wealth to the poor in a demand-constrained economy with a great deal of slack or inefficiency (most advanced economies today);
  • When it involves the building of needed infrastructure.

All that being said, I don’t want to be too cavalier about rising nominal debt levels. Balance sheet conditions and constraints matter to a country’s future economic growth in at least three ways not discussed above.

First, an economy driven by rising debt tends to adjust in ways that systematically incorporate strong positive feedback mechanisms (as I explain here), pushing domestic balance sheets away from their optimal structures. These balance sheet mismatches can lead to suboptimal resource allocation and can cause future growth to slow.

Second, high levels of debt directly increase the volatility of earnings and income, so a country with more debt has less room to withstand temporary shocks, even if more debt comes with more assets that can service the debt over time.

Third, when debt levels are high enough and balance sheets sufficiently fragile that there is uncertainty about how future debt-servicing costs will eventually be allocated, the different sectors of the economy that might be forced to absorb the debt-servicing costs will change their behavior in ways designed to protect themselves. As financial distress theory shows, the ways in which they change their behavior almost always reduce growth, further increase balance sheet fragility, or both.

It is not clear to me that pure money creation under the positive circumstances listed above comes with comparable problems, so perhaps this means that governments should fund wealth-enhancing income transfers or productive investment mainly by creating money, not by borrowing. This suggests that hard-core MMT proponents are right when they say that governments don’t borrow or raise taxes to fund spending. Instead, they simply spend. The purpose of borrowing or raising taxes in those circumstances is to counter the impact that MMT can have in some cases, but not in all.

There are, on the other hand, cases in which governments can simply create money or borrow with no ill effects, that is to say, with no inflationary impact and no increase in the debt burden. As always in economics, the outcome depends on the underlying conditions.

So how do these insights apply to the world today? If they so choose, the U.S. and European governments should be able to create money or debt with no ill effects if the proceeds were used to fund needed infrastructure or to reverse income inequality by increasing the incomes of the poor and middle classes. Either way, productive investment would rise faster than debt or the money supply, as would the total value of goods and service produced.

As for China, money or debt can no longer be used to fund infrastructure because the resulting increases (in money or debt) will not be matched by increases in real GDP. Beijing should, however, be able to create money or debt with no ill effects if the proceeds were used to reverse income inequality by increasing the incomes of the poor and middle classes.

My twitter account is @michaelxpettis. Aside from this blog I write a monthly newsletter that covers some of the same topics covered on this blog. Those who are interested in receiving the newsletter should write to me at chinfinpettis@yahoo.com, stating their affiliations.