More FAQS on deficits and debt: Where is the money coming from?

Former Federal Reserve Chair Alan Greenspan told CNBC last week that his “biggest concern” in regard to the economic outlook included too-high budget deficits. This concern is both completely misplaced and widely expressed.

We have already addressed a number of FAQs about deficits and debt. Besides those FAQs, however, another common question people ask is, “Where is the money the government borrows coming from?” Related to this fear is concern that the source of borrowing (whatever it is) might dry up at any time, and the result could be spiking interest rates that force firms to cut back on investment in productive private-sector investments, which in turn would slow economic growth in the future.

The short answers to these concerns are pretty simple: The money the government is borrowing comes mostly from ourselves, and that’s a key reason why we can be sure that interest rates won’t start rising unless and until we have reached a full recovery.

Below, we’ll explain what it means that federal budget deficits mostly amount to U.S. households borrowing from themselves.

In March and April this year, whole industries in which U.S. consumers spend money were shut down (restaurants, hotels, air travel, gyms, etc.). While there are some substitutes for some of this spending (when restaurants close, people tend to spend more money on groceries, for example), these offsets are nowhere near one-for-one. This means that as the places where people spend lots of money closed, households spent less. Because one person’s spending is another person’s income, as household consumption spending collapsed, market-based incomes collapsed in turn (i.e., workers in COVID-19-shutdown sectors stopped getting paychecks and business owners stopped earning profits).

Absent any policy response, a horrific vicious cycle would have started. Laid-off restaurant and airline workers would have cut back spending on everything—including spending in sectors that COVID-19 had not directly shut down. At that point, workers in non-coronavirus-affected sectors would have seen cuts to their incomes and reduced their own spending—and the downward cycle would continue. The few months of significant economic support provided by the CARES Act starting in April didn’t just provide vital income support to laid-off workers—it also broke this vicious cycle and put up a firewall between the coronavirus-driven shutdowns and the rest of the economy.

The CARES Act was forecast to increase the federal budget deficit by just under $2 trillion. Concretely, this means that the U.S. Treasury had to issue bonds in this amount. When people express worries about where the money comes from for the government to borrow, or if rising deficits will lead to spiking interest rates, they are concerned that there will be no willing buyers for U.S. Treasury bonds at going interest rates, and that the Treasury will have to raise rates to find buyers for new debt. But since March, even as the supply of bonds exploded, interest rates have fallen sharply.

How did this happen? Apparently a huge source of new demand to buy Treasury bonds emerged starting in March. Where did this new demand for Treasury bonds come from? From the huge increase in household savings that was the flip side of the collapse in household spending. Households not experiencing income losses in March and April had far fewer opportunities to spend money as whole swaths of the economy shut down, so their savings increased dramatically.

Figure A shows the personal savings rate up to the second quarter of this year—April, May, and June.

Figure A

U.S. households are supplying the money borrowed by the federal government: Personal savings rate, 1979–2020

Personal savings rate
1979-Q1 11.1%
1979-Q2 10.4
1979-Q3 9.9
1979-Q4 9.8
1980-Q1 10.1
1980-Q2 11.3
1980-Q3 11.4
1980-Q4 11.4
1981-Q1 10.9
1981-Q2 10.9
1981-Q3 12.3
1981-Q4 12.9
1982-Q1 12.3
1982-Q2 12.5
1982-Q3 12.3
1982-Q4 11
1983-Q1 11
1983-Q2 9.8
1983-Q3 9.5
1983-Q4 10
1984-Q1 11.1
1984-Q2 11.3
1984-Q3 11.7
1984-Q4 11.3
1985-Q1 9.3
1985-Q2 10.2
1985-Q3 8.2
1985-Q4 8.9
1986-Q1 9.3
1986-Q2 9.5
1986-Q3 8.5
1986-Q4 8.1
1987-Q1 8.9
1987-Q2 6.8
1987-Q3 7.5
1987-Q4 8.5
1988-Q1 8.3
1988-Q2 8.6
1988-Q3 8.6
1988-Q4 8.5
1989-Q1 9
1989-Q2 8.3
1989-Q3 8
1989-Q4 8.3
1990-Q1 8.3
1990-Q2 8.7
1990-Q3 8.3
1990-Q4 8.2
1991-Q1 8.7
1991-Q2 8.6
1991-Q3 8.6
1991-Q4 9.4
1992-Q1 9.6
1992-Q2 9.9
1992-Q3 9.3
1992-Q4 8.9
1993-Q1 8.8
1993-Q2 8.3
1993-Q3 7.4
1993-Q4 7.3
1994-Q1 6.8
1994-Q2 7
1994-Q3 6.8
1994-Q4 7.1
1995-Q1 7.6
1995-Q2 6.9
1995-Q3 6.8
1995-Q4 6.6
1996-Q1 6.7
1996-Q2 6.5
1996-Q3 6.6
1996-Q4 6.4
1997-Q1 6.3
1997-Q2 6.6
1997-Q3 6.1
1997-Q4 6.3
1998-Q1 7.4
1998-Q2 7
1998-Q3 6.7
1998-Q4 6.1
1999-Q1 6.1
1999-Q2 5
1999-Q3 4.5
1999-Q4 4.5
2000-Q1 4.9
2000-Q2 4.9
2000-Q3 5
2000-Q4 4.4
2001-Q1 5
2001-Q2 4.6
2001-Q3 6.5
2001-Q4 4
2002-Q1 5.9
2002-Q2 6.2
2002-Q3 5.5
2002-Q4 5.6
2003-Q1 5.4
2003-Q2 5.5
2003-Q3 5.8
2003-Q4 5.3
2004-Q1 5
2004-Q2 5.4
2004-Q3 5
2004-Q4 5.1
2005-Q1 3.4
2005-Q2 3.1
2005-Q3 2.5
2005-Q4 3.4
2006-Q1 4.2
2006-Q2 3.9
2006-Q3 3.5
2006-Q4 3.7
2007-Q1 4
2007-Q2 3.9
2007-Q3 3.5
2007-Q4 3.3
2008-Q1 3.9
2008-Q2 5.6
2008-Q3 4.3
2008-Q4 6.1
2009-Q1 5.9
2009-Q2 7.2
2009-Q3 5.6
2009-Q4 5.7
2010-Q1 5.9
2010-Q2 6.7
2010-Q3 6.8
2010-Q4 6.8
2011-Q1 7.4
2011-Q2 7
2011-Q3 7.1
2011-Q4 7.2
2012-Q1 8.2
2012-Q2 8.9
2012-Q3 8.1
2012-Q4 10.2
2013-Q1 6
2013-Q2 6.6
2013-Q3 6.7
2013-Q4 6.3
2014-Q1 7.1
2014-Q2 7.4
2014-Q3 7.4
2014-Q4 7.5
2015-Q1 8
2015-Q2 7.5
2015-Q3 7.3
2015-Q4 7.4
2016-Q1 7.4
2016-Q2 6.8
2016-Q3 6.6
2016-Q4 6.6
2017-Q1 6.9
2017-Q2 7.4
2017-Q3 7.5
2017-Q4 7
2018-Q1 7.7
2018-Q2 7.8
2018-Q3 7.9
2018-Q4 8.1
2019-Q1 8.4
2019-Q2 7.3
2019-Q3 7.2
2019-Q4 7.3
2020-Q1 9.6
2020-Q2 25.7
ChartData Download data

The data below can be saved or copied directly into Excel.

Source: Bureau of Economic Analysis (BEA) National Income and Products Accounts, Table 2.1.

Copy the code below to embed this chart on your website.

The spike in savings in the most recent quarter is obvious—and that spike mostly answers the question of where the money the government borrowed came from—it came from the increased savings of U.S. households. Basically, the increase in debt that financed the CARES Act’s few months of income support for laid-off workers recycled this increased household savings into somebody else’s income, and kept this household saving (or nonspending) from acting as a pure anchor on the economy’s growth. Yes, the entire story is more complicated than that, but that’s the essence of it.

It’s a pretty familiar story as well—federal budget deficits swell as recessions hit and household savings increase to finance lots of these increases. Figure B shows data from the four years before and after the beginning of the Great Recession (January 2008). Annual borrowing by the federal government rose from $375 billion on average in the four years before the Great Recession to $1,250 billion in the four years after—an increase of $875 billion. But households went from being net borrowers of $40 billion in the four years before the recession to lending $630 billion annually in the four years after, a swing of $670 billion, or enough to account for more than three-quarters of the rise in federal borrowing (the rest can be more than accounted for by an increase in business-sector savings).

Figure B

During recessions, federal borrowing usually comes from us: Net lending (+) or borrowing (-) by household sector and federal government (billions), 2004–2007 and 2008–2011

2004–2007 2008–2011 Change
Households -38.2 628.7 666.9
Federal government -376.9 -1,251.1 -874.3
ChartData Download data

The data below can be saved or copied directly into Excel.

Source: Bureau of Economic Analysis (BEA) National Income and Products Accounts, Table 5.1.

Copy the code below to embed this chart on your website.

In some economic models, we don’t need the public sector to translate increased savings into somebody else’s incomes to avoid a slowdown in economic growth—it’s supposed to be done by private financial markets. In these models, a big influx of household savings puts downward pressure on interest rates, and the decline in interest rates induces firms to borrow more to finance increases in investment. (Note: The only things that count as business “investment” in macroeconomic terms are spending money to build plants, purchase equipment, add to physical inventory, or paying people to engage in research and development. Exchanging ownership of already-existing assets— say through mergers and acquisitions—doesn’t count.)

These models that assume interest rates are the essentially the “price” of “loanable funds,” which are supplied by household savings and demanded by firms looking to undertake investment, are pretty misleading for a whole host of reasons. But even if they were sometimes true, they’re obviously not true in a world where the interest rate needed to balance planned household savings and planned business investment is negative. The interest rates relevant to business investment decisions generally can’t go negative for any real length of time, and yet—particularly in times of economic crisis—there is no guarantee that positive interest rates will indeed suffice to balance planned investment and planned savings.

In short, the hypothetical private financial system that nimbly translates an increase in household savings into increased business investment that exists in many economic models is a fantasy and cannot be relied on during crisis. But in the real world we can instead have the public sector take the increased private savings, put it into a useful asset for households to store wealth in (Treasury bonds), and make sure it is spent out into the economy to support incomes.

In this case, the money the federal government borrowed to finance relief from the most recent economic crisis comes from us. This fact is key to explaining why there’s no need to worry about some sudden spike in interest rates that would make servicing the increase in debt ruinous. The exact same influence—the COVID-19-driven collapse in household spending and concomitant increase in planned savings—drove both the rise in public debt and the collapse of interest rates. As the economy recovers, household savings will be cut back, but so will lots of the public spending that was driving the increased federal budget deficits. As unemployment falls, for example, there would be steady decline in unemployment insurance payments even if we kept the post-COVID-19 enhanced benefit levels turned on (and we certainly should do that).

Could economic recovery happen so quickly—with households spending like mad and businesses eager to balloon their investments—that interest rates really would see some upward pressure as planned savings fell well short of planned investments and the Federal Reserve needed to ramp up policy interest rates to control inflation? This is theoretically possible, but it is supremely unlikely. The main macroeconomic problem in the U.S. economy for the past two decades has been trying to get spending to grow fast enough to keep unemployment at acceptably low levels, not trying to keep it muffled. And an economy growing so fast that it threatens to spark inflation and interest rates is a challenge for sure, but it would be replacing a much worse challenge—an economy growing so slowly (or even contracting) that mass unemployment was crushing American families’ incomes. No state of the world is perfect, but I’d much rather policymakers err in overshooting in trying to get us out of the current state rather than go timid and let us stay mired in the one we’re in.