Originally Published: February 2020
U.S. Treasury bonds (and bills and notes) are often thought of as risk-free investments.
This is true in a sense; the U.S. Treasury has always fully paid back its debts in dollar-denominated nominal terms. However, what many investors miss is the fact that the purchasing power of those bonds is not guaranteed. They can lose real purchasing power during their holding period, and sometimes rather dramatically.
This chart, for example, shows the historical yield of the 10-year Treasury note (blue line) along with the forward 10-year annualized inflation-adjusted return from the start of that year (orange bars):
As you can see, there were long stretches of negative real returns.
In addition, ETFs such as the iShares 20+ year Treasury Bond ETF (TLT) don’t hold their bonds to maturity, so they can experience a capital loss on their investment if the bond yield goes up (and thus bond price goes down) during the investor’s holding period.
It’s important for investors to understand the risks associated with this defensive asset class, and to ensure that they have a risk-management process for their holdings, just like with any other asset class. I like Treasuries as a tactical holding, but it’s a position that needs to be managed just like any other.
This article provides a description of three times in the past century that U.S. Treasury bonds lost real value, looks at one similar international example, and then dives into why the current environment is similar to those times.
A Useful Abstraction
In modern times, money is an abstraction that represents purchasing power, but is not valuable in and of itself. We don’t walk around with silver coins in our pockets, paying each other like pirates.
Instead of being made of a rare and valuable substance, modern money is just a placeholder for value, an abstraction of value. From day to day and year to year, that abstraction doesn’t typically matter in daily life, because the purchasing power remains relatively consistent. We pay with paper, cards, and digital platforms, exchanging abstract units of currency, knowing that we can buy all sorts of things with those units.
Every few decades though, that abstraction suddenly matters, because policymakers use that abstraction to reset debt burdens as a percentage of GDP by devaluing those units of currency. In addition to happening in emerging markets frequently, it happens in developed markets occasionally.
Example 1) USA, 1930’s
During the Great Depression, total debt as a percentage of GDP in the United States reached what at the time was record highs. Government debt as a percentage of GDP wasn’t very high, but the total amount of private debt and government debt combined was skyrocketing relative to the size of the (then shrinking) economy. It was an extreme, self-reinforcing, deflationary contraction.
However, suddenly in 1933, debt as a percentage of GDP dropped like a rock:
Chart Source: Hoisington Investment Management Co
Back then, the dollar was on a gold standard. However, in April 1933, President Roosevelt signed Executive Order 6102, which made it a criminal offense to own meaningful amounts of gold. Everyone had to turn in his or her gold for a payment of $20.67/ounce, which was what the dollar was backed by at the time.
A year later in 1934, once they had effectively confiscated the nation’s gold, Roosevelt devalued the dollar from being worth 1/20.67 an ounce of gold to 1/35 of an ounce of gold. This allowed them to print a ton of money, which they started doing around 1932 (front-running their official devaluation):
Chart Source: Bridgewater Associates
U.S. nominal GDP grew at a double-digit average rate for the next four years from 1934 to 1937 and inflation heated up moderately (bouncing back from a low of -10% to as high as about +5%), until they encountered the next recession. This reduced the total debt burden as a percentage of GDP, and they also reduced wealth and income concentration.
The United States wasn’t the only country to do this. All major currencies were devalued relative to gold, with some being devalued more than others:
Chart Source: Bridgewater Associates
If you bought $10,000 worth of 10-year Treasury notes in 1931, you were expecting to be paid interest as well as your principle back. At the time of the purchase, the dollar was set at 1/20.67 of an ounce of gold (0.0484 ounces per dollar), meaning that you lent the equivalent of about 484 ounces of gold to the government, and expected to be paid back the equivalent of about 484 ounces of gold plus interest.
However, by the time the bond matured, the dollar was only set at 1/35 of an ounce of gold (0.0286 ounces per dollar), meaning that when you received your $10,000 principle back, it only translated into 286 ounces of gold. It was, in essence, a default. Every dollar was paid back in nominal terms, but the definition of the dollar was reduced vs gold, and the number of dollars in existence was greatly increased. Plus, by then it was illegal to own gold in the United States. Thanks for playing, try again next time.
Imagine, for example, that you and I were to both sign a contract, written by me, wherein you will give me your car, and I will give you twenty thousand dollars. However, after you give me the keys, I give you a briefcase filled with twenty thousand Australian dollars, which are worth significantly less. “The contract never said what kind of dollar!” I shout gleefully, driving away.
I didn’t break the letter of the contract, although I clearly broke the intention of it. That’s what the U.S. government did for bond holders. All debt holders that were owed dollars in one form or another, public and private, were essentially defaulted on due to the redefinition of the dollar even though all dollars were paid back in full.
Very high levels of inflation, as measured by the consumer price index, never materialized during this 1930’s period, because the inflationary potential was offset by so much deflationary credit destruction. In other words, this was an example of a largely non-inflationary devaluation, although the devaluation did mark the bottom of the deflationary spiral and a shift back towards a moderately inflationary environment (which later became very inflationary during World War II).
Data Source: Robert Shiller
Financial assets did reasonably well from their lows. In addition to the price of gold going up by definition, stocks soared and real estate went up moderately. The stock market bottomed in mid-1932, and more than doubled a year later by mid-1933. The stock market then chopped along in a volatile sideways pattern for the next decade, paying an average dividend yield of about 5% along the way, before starting the next true bull market.
Example 2) USA, 1940’s
By the time the Second World War was ramping up, total U.S. debt was relatively low as a percentage of GDP. However, the war effort required massive government deficit spending, so government debt skyrocketed, eventually reaching a level even higher than it is today as a percentage of GDP.
Chart Source: Congressional Budget Office
In 1942, in order to fund the war effort, the Fed capped US Treasury bond rates at 2.5%, and T-bill rates at 0.38%, thus creating an artificial low and steep yield curve, regardless of inflation.
A 2016 research paper called “The Fed’s Yield-Curve-Control Policy” by Owen Humpage at the Cleveland Fed described this well:
In early 1942, shortly after the United States declared war, the Fed effectively abdicated its responsibility for monetary policy despite its concern about inflation and focused instead on helping the Treasury finance the conflict. After a series of negotiations with the Treasury, the Fed agreed to peg the Treasury-bill yield at 0.375 percent, to cap the critical long-term government bond yield at 2.5 percent, and to limit all other government securities’ yields in a consistent manner. The Fed would maintain a yield curve that was both low and relatively steep. The low rates kept the Treasury’s borrowing costs down, while the firmly harnessed term structure convinced investors that waiting for higher yields was pointless and that the risk of capital loss from holding longer-term securities was small. Setting interest rates in this manner, however, allowed the Treasury to expand bank reserves by issuing more securities than the public wished to hold when yields reached their caps, because the Fed then had to purchase them.
The Fed’s balance sheet increased dramatically due to buying a portion of all the newly-issued debt to maintain their target rates:
Chart Source: Cleveland Fed
During the later stages of this period, the inflation rate ran very high, briefly into the double-digits. However, the Fed ignored market forces and kept bond yields pegged at 2.5% with the power of their balance sheet. Neither bond prices nor bond yields moved much during that decade due to this peg.
Investors that bought 10-year Treasury notes in the beginning of 1942 made modest gains in nominal terms over the next decade, and of course had every dollar paid back with interest, but lost a substantial amount of purchasing power as inflation greatly exceeded the total nominal bond returns. Specifically, investors in Treasury notes lost nearly 28% of their CPI purchasing power over a decade. Thank you for your patriotism; your sacrifice has not been in vain.
Data Sources: Robert Shiller, Aswath Damodaran
This was an example of an inflationary devaluation. The U.S. rapidly paid down its war debt as a percentage of GDP because the value of most things- the stock market, home prices, nominal GDP, wages, and so forth- all went up quickly, while most debts including government debt were at a fixed nominal amount and yields locked below the inflation rate.
Lael Brainard, a member of the Fed’s Board of Governors, publicly proposed in 2019 capping long-term rates again in the future. The Treasury market is a free market unless the Fed doesn’t like the number, at which point they will buy enough to set the number however they want it.
As one example, when the interbank lending market ran into a problem and the overnight repo rate spiked to 7% in September 2019, the Fed didn’t like that number, so they effectively nationalized the repo market and set the rates lower by supplying liquidity.
As a second example, the long-duration Treasury market had a big bull move during the initial phase of the February/March 2020 sell-off, but then the long-duration Treasury market briefly crashed in mid-March and gave up all its gains, as foreigners started selling Treasuries to get dollars amidst the global dollar shortage, and risk parity funds unwound positions. Seeing this, the Federal Reserve cited severe dysfunction in the Treasury market, and then stepped in and ramped up Treasury security purchases to an astounding $75 billion per day until the market had enough liquidity, and Treasury bond yields were back down and Treasury bond prices were back up to where they had been. In the six weeks from mid-March through April, the Fed bought $1.4 trillion in Treasury securities across the duration spectrum.
Data Source for Treasury Purchases: NY Federal Reserve
So, we’ve seen brief echoes from the 1940’s in today’s Treasury market, both in terms of what Fed officials have stated and in ways they’ve decided to intervene.
Example 3) USA, 1970’s
In the 1960’s, Americans still couldn’t own gold or exchange dollars for gold, but the dollar was still officially backed by 1/35 of an ounce of gold for international settlements. However, due to persistent deficits, gold was flowing out of the United States. International creditors were calling the bluff, saying they want the actual gold, not the dollars that were supposedly backed by gold.
After drawing down a considerable portion of the nation’s gold reserves from over 20,000 tons in the 1950’s down to just over 8,000 tons, President Nixon closed the gold window in 1971. At the time, this move was claimed to be temporary, but looking back on it nearly five decades later we can see that it was permanent. The dollar was once again redefined and devalued.
Similar to the 1930’s, anyone who happened to be a Treasury bond investor at the time received every single dollar owed to them. However, the definition of the dollar changed during the holding period from 1/35th of an ounce of gold to nothing at all.
If you bought 10-year Treasury notes in the beginning of 1969, you had strong nominal gains going forward. However, inflation in many years exceeded the interest rates and total bond returns. After a 14-year period, 10-year Treasury notes (including reinvestments into more 10-year notes), went up 78% nominally but were down nearly 33% in CPI purchasing power.
Data Sources: Robert Shiller, Aswath Damodaran
In addition, if you were an international creditor and invested $10,000 into U.S. Treasuries in January 1969 when the dollar was backed by 1/35 of an ounce of gold, you basically lent 286 ounces of gold-equivalent to the U.S. government. When the note matured in January 1979, a dollar was worth 1/233 an ounce of gold, so you could buy only 43 ounces of gold with your returned principle. Thanks again for playing, may the odds be ever in your favor.
This was another inflationary devaluation.
Example 4) Japan, 2010’s
For many decades, Japan had a positive trade balance and positive current account. Their disciplined and highly-educated culture has consistently produced more value than they consumed, even though they have historically had to import a lot of their energy and commodities.
However, they also experienced a declining workforce and an aging population. Their government deficit as a percentage of GDP became very high.
Starting in late 2010, Japan developed a trade deficit as well, which eventually reached a low point of about -2% of GDP in 2013 and 2014. Their current account balance dropped to as low as zero during this time; relatively flat.
In order to fund large fiscal deficits and attempt to stimulate the economy, the Bank of Japan began a massive quantitative easing program in 2012. Over the next three years, they more than doubled the size of their already-large central bank balance sheet, and today it is just over 100% of GDP (compared to about 20% for the U.S. Federal Reserve and about 40% for the European Central Bank). In other words, they created a ton of new yen, used that yen to buy Japanese government bonds, and stored those bonds on the Bank of Japan’s balance sheet.
Chart Source: St. Louis Fed
In addition, the U.S. Federal Reserve stopped its quantitative easing program by the end of 2014, which resulted in tighter U.S. monetary conditions and a strong dollar.
In January 2012, one dollar was equal to about 78 yen. By January 2015, the yen had devalued such that one dollar was equal to about 120 yen. In other words, the yen devalued by about 35% relative to the dollar. It also devalued against the euro and other major currencies, but not by quite as much.
During that three-year period, Japanese investors in 10-year Japanese government bonds made modestly positive nominal gains in yen terms, as the interest rate dropped from about 1% to 0%. Domestic inflation for Japan was low. However, the international purchasing power of those bonds decreased substantially. The ability of that same amount of yen to buy foreign goods diminished.
Japan’s stock market doubled during that period in yen terms, outperforming bonds. Their corporate earnings recovered from the global recession, and valuations remained relatively flat.
This was an example of a non-inflationary devaluation; a loss of real global purchasing power for Japanese government bonds, but without consumer price inflation.
The best investment that Japanese investors could have made at the time was foreign assets- stocks, real estate, and/or bonds of regions with stronger currency fundamentals.
By 2015, the weakening yen had made Japan’s exports more competitive, and made imports more expensive, and thus it helped fix their trade balance back up to even (zero), and their current account balance to being positive once again. At that point, even though they continued massive quantitative easing, their currency stabilized against other major currencies. In fact, the yen has slowly strengthened against the dollar since then from its weakest point, although it is still over 100 yen to the dollar.
Parallels to Today
The main point of this article is to highlight tangible examples of why Treasury bonds and other “riskless” sovereign bonds do have real risks. They may have little or no nominal risk because they are first in line for bailouts, but they do have the ability to lose purchasing power in various ways.
There are certain points in history where debts, deficits, or trade balances reach such levels that a currency devaluation is almost mathematically unavoidable during the decade. Every few decades, that layer of abstraction between money and real value that normally doesn’t matter, suddenly matters again. Policymakers use that abstraction to reset things, much to the detriment of fixed income investors or holders of currency.
Today’s situation with the dollar is an amalgam of many of these past scenarios.
Much like the 1930’s, U.S. total debt is extraordinarily high as a percentage of GDP. Much like the 1940’s, specifically U.S. federal debt and deficits as percentages of GDP are very high. Much like Japan in the 2010’s, the United States has a huge twin deficit (fiscal deficit and trade deficit) and is just beginning, since September 2019 to ramp up its central bank balance sheet to print all the money needed for its obligations.
Debt is deflationary. Fiscal and monetary responses to that deflationary environment can be extremely inflationary, or can still be non-inflationary domestically but devaluing for the currency compared to commodities or other currencies.
All of these environments were bad for bonds and cash in real purchasing power terms, although they did fine nominally. Some environments were good and some were bad for stocks in real terms depending on the details of the situation. Most of them were particularly good for precious metals.
If the goal of investing is to protect and grow your purchasing power, then cash and “riskless” sovereign bonds do have a risk, and can go a decade or two with negative purchasing power returns. Cash is a position. Bonds are a position. Both have benefits and risks that are worth measuring and managing.
I am using cash and Treasuries tactically in this environment, to temporarily store value to buy into things cheaper during sell-offs. Within my various accounts, I had larger Treasury positions before the sell-off, and shifted some Treasuries into equities in tranches during the sell-off as we moved lower. They are useful deflation hedges for periods of deflation or disinflation, and particularly during market crashes.
However, I would never personally go all to cash or bonds, and instead seek to protect purchasing power via a globally diversified multi-asset portfolio that includes equities, real estate, cash, short-term bonds, and alternative asset classes such as precious metals. I personally like a combination of gold and Treasuries as the defensive section in my portfolios, rather than Treasuries alone.
Traders can still make money with bonds going forward, and investors can use them as tactical placeholders, but make sure you have a specific process and risk-management approach, just like with any other asset class.
Check out these articles for further reading: