Note: When I wrote this explanation of budget deficits and interest rates just a year ago as a chapter of my book, Lessons from a Road Warrior, we were in a different world where budgets mattered, at least a little. We argued about deficits of $100 billion or $200 billion, and never had to use trillion in polite conversation. Today, government spending is completely out of control. Since last fall, when Congress was railroaded into giving Treasury Secretary Paulson complete discretion over spending $700 billion for the TARP bailout plan all semblance of fiscal discipline has been destroyed. Obama’s team quickly used the financial crisis to add another $1.2 trillion of spending and has added or expanded countless spending programs in his first budget. As a result the deficit for Obama’s first full budget year is likely to approach an incredible $2 trillion–not including the cost of his proposed national healthcare program. Trillions more have been provided by the Federal Reserve, without congressional approval or oversight, by taking on the authority to add all sorts of assets to their once-pristine balance sheet.

I decided to leave the numbers and the charts unchanged from the original examples in part to highlight the difference between then and now. The analysis remains unchanged. Balance sheets are still huge compared with private saving and private and government borrowing. It still takes a huge amount of government borrowing to move the needle on interest rates. And the most important question for interest rates is still whether anything is going on to change people’s willingness to hold the securities that are already outstanding, not how many new ones the Treasury is selling. But today the numbers are big enough that people all around the world are taking notice and reevaluating whether they want to hold dollar assets. This is the most important economic question of our time.


Textbooks have it all wrong when they teach students that interest rates are determined by the supply and demand for credit. For example, a leading finance textbook (Investments, by Bodie, Kane and Marcus) states:

Forecasting interest rates is one of the most notoriously difficult parts of applied macroeconomics Nonetheless, we do have a good understanding of the fundamental factors that determine the level of interest rates.

  1. The supply of funds from savers, primarily households.
  2. The demand for funds from businesses to be used to finance investments in plant, equipment, and inventories (real assets or capital formation).
  3. The government’s net supply of or demand for funds as modified by the actions of the Fed.

These statements are wrong, wrong, wrong. Had you acted on them to structure your investments over the last 30 years you would have lost all your money. Interest rates are simply a calculation we make from asset prices, as I explained in a recent post. Asset prices are determined in asset markets by existing asset supplies and the demand to hold existing assets. In spite of what the textbooks tell you, throughout history, the correlation between interest rates and deficits is actually negative; i.e., higher deficits are associated with lower interest rates. The drawings below explain why.

Interest rates are not determined by savings rates and are not determined by the demand for credit. As a logical matter, it is debt, not deficits, along with people’s demand to hold different types of assets, that determine interest rates.

Budget deficits in the ranges we have historically seen them—a few hundred billion–don’t matter much for the economy. Not for interest rates. Not for growth. The multi-trillion dollar bond markets don’t care at all whether the government is a net seller or a net buyer of $100 billion in new Treasury securities in a given year. They care whether the people who own the old paper today are still going to want to own it tomorrow. And that will depend on whether something happens to change people’s minds about future after-tax returns on bonds relative to other assets–period. The rest is all rounding errors. Nothing else matters.

People's Total Assets

People's Total Assets

When the Treasury holds an auction to finance a deficit, they print and sell new T-bills or other Treasury IOUs (notes and bonds) to investors like Aunt Tilly. This isn’t the first time the government has borrowed money from private citizens. Aunt Tilly and other investors already own a huge stock of similar old T-bills—$5.1 trillion worth at the end of 2007—that we call the national debt. At the end of last year, there were $9.2 trillion of old T-bills outstanding—the government’s accumulated borrowing since the time of George Washington. Of those, $5.1 trillion, or 55%, was held by private investors like you, Aunt Tilly and me.

New treasury paper and old treasury paper are perfect substitutes to investors. Not almost perfect substitutes—perfect substitutes. In practice, they are indistinguishable in the market. T-bills, notes and bonds are also very good substitutes for many other securities owned by investors, including agency securities, corporate securities, municipal securities and securities issued by financial institutions. These, in turn, are substitutes to some degree for equities, foreign securities and tangible assets in the minds of investors. When you buy a bond, you shop for its issuer, its maturity date, its call provisions, its tax features, and its yield—not its model year. This means that new bonds and old bonds that are the same in every other way must sell at the same price. Arbitrageurs make sure they do.

It’s like the commercial where the dad asks his teenage son to drive to the local gas station to put gas in the family car. Hours later, the dad is still standing in the driveway when his son returns with the explanation, “But Dad, you didn’t want me to mix the new gas with the old gas, did you?”

Bond investors mix the new bonds with the old bonds all the time.

Flow-of-funds is a theory to explain the price of new bonds. Portfolio balance is a theory to explain the price of old bonds. But in the real world, there can only be one price for all bonds. Which theory wins? Portfolio balance, because almost all bonds are old bonds. The important question is not whether the government will borrow money this year. It is what price it will take to make Aunt Tilly—and all the other investors who owned T-bills yesterday—still want to own the stock of T-bills tomorrow.

But balance sheets do not sit still; they grow over time. The stock of tangible assets grows as a result of building houses, factories, shopping centers and new cars faster than they wear out. The stock of private financial assets grows as people issue mortgages to finance home purchases, and companies issue new stocks and bonds to finance capital spending, home construction and durable goods production—at a faster rate than the old ones mature. We create new government securities to finance the budget deficit, or we destroy them by using a budget surplus to buy back debt.

As our net worth grows, so does our appetite to hold all assets. This growth is represented by drawing a second pie chart showing the larger stock of stuff (tangible assets) and financial assets. Historically, U.S. balance sheets have grown at about 7% per year.

Next Year's Total Asset Growth

Next Year's Total Asset Growth

Higher net worth gives investors like Aunt Tilly an appetite to own more T-Bills too, as illustrated in the drawing. You can think of this appetite as thousands of Aunt Tillys standing in line in front of the Treasury building, waiting to place orders for the new T-bills that the government will sell next year (to add to the ones they already own).

That appetite is larger than most people think. If net worth grows 7% next year, as it has done historically, investors will want to increase their holdings of T-bills by $357 billion, from $5.1 trillion to just under $5.5 trillion.

If the Treasury issues just enough new T-bills to satisfy Aunt Tilly’s appetite, then everyone will go away happy. There will be just enough T-bills to go around. Prices will remain the same as they were last year. Interest rates will remain unchanged.

But what would happen if the treasury issued too few or too many T-bills to satisfy investors’ appetites? That’s easy. There would be a scuffle. If they were too few T-Bills for sale due to a smaller government deficit, then Aunt Tilly would wrestle with the other investors for the limited supply. This would drive T-Bill prices up and interest rates down.

Too Few New T-Bills

Too Few New T-Bills

If the government ran up the deficit and there were too many T-Bills for sale, then the Treasury would be forced to lower their asking price to sell their inventory. T-Bill prices would fall and interest rates would rise.

Too Many New T-Bills

Too Many New T-Bills

It is only to the extent that budget deficits exceed Aunt Tilly’s appetite for new securities that they can be said to push interest rates up at all. In our example, we can think of the $357 billion from the previous example as the interest rate neutral budget deficit—one that will put no pressure on interest rates. The balanced budget that we all wish for would actually exert downward pressure on interest rates every year. Over time, government debt would gradually become extinct like the dodo bird; it would be irrelevant for investors.

This does not mean that I love budget deficits. It does not mean that deficits are good or bad. And it does not blunt the fact that higher government spending uses tons of resources and creates all sorts of incentive and resource allocation problems. It just means that budget deficits are unlikely to be a factor in determining interest rates in the range in which we have historically seen them.

If deficits don’t determine interest rates, then what does? The answer is any factor that could drive a wedge between the relative returns on different assets. Higher inflation does that by increasing the return on tangible assets (i.e., the gain on owning a home raises its return) relative to returns on securities. That makes people sell securities to increase their holdings of real assets, like we saw during the 1970s, which drives interest rates higher. Lower income tax rates increase the after-tax returns on taxable securities relative to real assets, for example, because people are forced to report all income from securities on their tax forms but are not required to report the value of living in their home as taxable income. This will drive interest rates lower. Rising productivity growth can quicken net worth growth, which will increase Aunt Tilly’s appetite for all assets and drive interest rates lower. And the Fed can influence the rate of net worth growth by making it easier or more difficult to finance investments.


Join the forum discussion on this post - (1) Posts

As you know from my recent posts, I spend a lot more time thinking about balance sheets and a lot less thinking about GDP, than most economists. In this post I want to look at the balance sheet of the household sector. The most recent data from the Fed’s Flow of Funds report show that at the end of Q4/2008 (12/31/08) households (including nonprofit organizations) owned the following mix of assets.

Household Balance Sheet Q4/2008

Household Balance Sheet Q4/2008

Households hold 62% of their assets ($40.8 trillion) as financial assets like deposits, T-bills, bonds, stocks and mutual funds. They keep 31% ($20.5 trillion) in real estate assets, and they hold 7% ($4.4 trillion) in the form of consumer durable goods like used cars, old washing machines and computers (in all 38%, or $24.9 trillion in tangible assets). As I have written for years, these percentages represent portfolio choice decisions for people based on their perceptions of return and risk for each asset class. These portfolio decisions are exquisitely sensitive to changes in tax rates and monetary policy.

HouseholdBalance Sheet #s Q4/2008

HouseholdBalance Sheet #s Q4/2008

The table above shows the history of household balance sheet composition. The first thing to notice is the household balance sheet numbers are huge when compared with GDP (roughly $14 trillion per year) or its components like consumption, investment, net exports and government spending. At the end of 2008, people owned $65.7 trillion worth of total assets, made up of $24.9 trillion of tangible assets and $40.8 trillion in financial assets.

The tangible assets, in turn, can be divided into $20.5 trillion in real estate assets and $4.4 trillion in consumer durable goods. It makes sense that the assets on our balance sheet are so big. They represent all the economic activity that has ever happened–all the buildings we have built, all the cars and washing machines we have ever made–less the ones that are no longer in service. In the case of autos, for example, there are more than 15 used cars and trucks on the road and in the driveways for every one that will be produced (i.e., that will appear in the GDP accounts) in the U.S. this year.

In spite of what you read in the headlines, total household liabilities, including mortgages, installment credit and credit cards, add up to just $14.2 trillion. Net worth is a whopping $51.5 trillion–more than 3.6 times total debt and almost five times disposable personal income of $10.6 trillion.

So where is the financial crisis we read about? It is in the behavior of asset values over time. Our net worth of $51.5 trillion is $12.9 trillion (20%) lower than it was just 6 quarters earlier in Q2/2007. That puts our net worth roughly where it was at the end of 2004 ($51.9 trillion) but still much higher than it was a decade ago in 1997 when it was $33.3 trillion.

You can also see from the figures when the trouble started. Tangible asset values peaked in Q1/2007 at $28.4 trillion; since then they have declined by $3.5 trillion or 12.3%. Financial assets peaked 2 quarters later in Q3/2007 (when the leveraged loan and asset-backed securities markets froze up) at $50.5 trillion but have declined by $9.7 trillion or 19.2% since then. Essentially all the adjustment in both types of assets was due to price decline; the physical stocks of tangible and financial assets did not materially change during this period.

When asset values change abruptly, as they did over the past two years, it is always a demand story. That’s because over a short period asset supplies can’t change by much because new asset creation and retirement are small compared with the stock of outstanding assets. In this case it was the sudden drop in demand when investors pushed away from the asset-backed securities market a year and a half ago.


I recently wrote about the fact that the forces impacting the U.S. economy’s balance sheet, at about $200 trillion, dominate those affecting GDP (just over $14 trillion) when thinking about interest rates and stock prices. A blog reader wrote to ask me where the $200 trillion figure comes from.

First, I want to point out that it is revealing that we have to ask the question. Why is it that people know so much about something so small (GDP) but so little about something so big (total assets)? I think it is because since the 1930’s macroeconomics has developed into a discipline concerned almost exclusively with who is spending how much money. Very little attention is paid to the capital base, or balance sheet, that makes it possible to produce the goods and services measured as GDP. A glance at a newspaper or any list of data produced by the government will convince you this is the case.

The best source of asset market, or balance sheet, information we have today is the document Z1: Flow of Funds of the United States produced after the end of each quarter by the army of economists working at the Federal Reserve Board.

The most recent (116 page!) flow of funds document, publish March 12, contains information about the balance sheet of the U.S. Economy on 12/31/08. I will warn you that you will have to dig for it–most of the 116 pages are devoted to measuring “flows of funds”, roughly the amount added and subtracted from balance sheets during the quarter. But you can find most of what you need if you hunt for it.

Total Assets by Sector Q4 2009

Total Assets by Sector Q4 2009

So what about the $200 trillion? I have constructed the table, above, by pulling figures from the report. The report reports balance sheets for some sectors of the economy but not others (which I find a little strange). They report balance sheets for 1) Households and Nonprofit Organizations, 2) Nonfarm Corporate Business (big companies), and 3) Nonfarm Noncorporate Business (small companies). These balance sheets show that at the end of 2008 housseholds and nonprofits owned $40,814 billion in financial assets like stocks and bonds and $24,905 billion in tangible assets like houses and cars, which adds up to $65,719 billion in total assets. Against that total, households and nonprofits owed debts, or liabilities, of $14,242 billion, which means they had net worth of $51,477. (These last numbers are in the document on p. 102 but not in the chart.)

Adding the three sectors together (Subtotal in row 4) produces a balance sheet with $104,049 in total assets divided between $58,639 in financial assets, and $46,301 in tangible assets.

Now it gets trickier. The Fed does not report complete balance sheets for the other sectors (farms, financial sectors, federal government, state & local governments, or rest of world (foreign owners). Instead, they report statements of financial assets, financial assets and financial liabilities. In other words, they leave out the fact that all these other sectors own tangible stuff like land, buildings, cars and computers, in addition to securities. I think that is a big mistake, reflecting the analytical bias in the macroeconomics community that somehow people consciously manage their portfolios of stocks and bonds but are passive owners of more than $46 trillion of real stuff.

We can use the Fed’s measures of financial assets held by all the sectors to get a pretty good figure for total financial assets in the balance sheet. Adding in farms, financials, governments and foreign owners brings the total financial asset figure up to $141,512 billion, which is reported on p. 115. (I say a pretty good figure because the document reports a $4,922 billion statistical discrepancy in getting to that figure themselves.) They do not report figures for tangible assets held by those “other” sectors, which is unfortunate because the “other” sectors are actually bigger than the ones they report.


That leaves us in an awkward position in trying to derive a total asset figure than makes sense for the overall U.S. economy’s balance sheet. One way to do it is to add up the numbers that we do know. I have done so in line 13. We know there are $141,512 billion in financial assets. We know that just three of those sectors own $46,301 billion in tangible assets. Adding those two numbers together produces a (reported) total asset number of $187,813 billion, pretty close to the $200 trillion number I wrote about at the top of the story. (The number would have been much closer 2 years ago before the recent drop in asset values.) Unfortunately, I have no idea what to call this number because it leaves out so many huge question marks.

If I weren’t so lazy I could dig up numbers to at least approximate the values of some of the question marks in the table. Farms own land and tractors, banks own buildings and ATM machines, governments own all sorts of crap including nearly a billion acres of land and all those cars you see on the highway that don’t have to buy license plates like you and me. And foreigners own a ton of stuff too. For today’s purposes all we have to know is that these things would add up to a very big number. And plugging these figures into the missing cells in the table would produce a total assets number far in excess of $200 trillion.

OK, that’s enough arithmetic. Why does this matter? It is to show you that the balance sheets are so big that almost any analysis of the economy that focuses on spending or saving or budget deficits alone, to the exclusion of the balance sheet, is almost certain to be wrong because balance sheet changes are so big. For example, household financial asset holdings fell from $50.5 trillion in Q3/07 to $40.8 trillion on 12/31/08 due to the collapse of stock and bond prices. And the value of their tangible assets fell by another $3.5 trillion due to falling home prices. Does anyone really think that the impact of this roughly $13 trillion drop in household net worth can be fixed by sending people checks for $700?

The most relevant application of this thinking today is how to understand the impact of the massive bailout programs on the economy and to say something meaningful about the impact of government borrowing on interest rates and stock prices. I will write more on these questions later.

You can read an analysis of budget deficits and interest rates using this approach in Chapter 4 of my new book, Lessons from a Road Warrior. You can get it from Amazon or get a signed copy directly from the John’s Book section of our website.