|Shares Out. (in M):||1||P/E||0.0x||0.0x|
|Market Cap (in $M):||1||P/FCF||0.0x||0.0x|
|Net Debt (in $M):||0||EBIT||0||0|
“When the financial history of this decade is written, it will surely speak of the Internet bubble of the late 1990s and the housing bubble of the early 2000s. But the U.S. Treasury bond bubble of late 2008 may be regarded as almost equally extraordinary.”
- Warren Buffett, Shareholder letter dated 2/27/2009 (2008 intra-day low on 10yr yields was 2.04%)
I don’t expect this idea to win any awards for originality as several variations of this thesis have been posted to VIC over the last several years. All I can say is that sometimes the consensus is correct and I view this as a true “not if, but when” situation that allows long-term investors to take advantage of a genuine time-horizon arbitrage.
Investment thesis - U.S. Treasury bonds are a compelling short because:
¦ Real and nominal treasury yields are at the lowest levels ever and will eventually normalize
¦ Low yields limit the possibility of capital impairment
¦ Economic data and Fed signaling have made deflation a highly remote possibility
¦ The U.S. fiscal position is unlikely to materially improve
¦ Shorting levered ETFs provides additional upside
Shorting U.S. treasury securities offers multiple ways to win, including eventual interest rate normalization, inflation and further deterioration in the U.S. government fiscal situation. Nominal losses on an unlevered position are capped at <2%/year; the only way to lose more than this in real terms is under a regime of sustained deflation – a highly unlikely scenario not seen in the U.S. since the 1930s. Shorting levered ETFs provides additional upside via the inherent decay of these instruments.
Treasury yields are near the lowest levels ever - Yields on U.S. treasury securities are currently near the lowest levels ever on both a real and nominal basis. Nominal 10-year yields hit an inter-day low of 1.70% in September 2011 (vs. current yield of 1.78%), but prior to 2011 yields had not fallen below 2% since WWII. Ex-ante real interest rates are difficult to calculate prior to the development of an active TIPS market, but the best data I’ve seen suggests that long-term (20 yr +) real yields in the U.S. had not fallen below 1.2% over the last 100 years prior to 2008 vs. current real yields of approx. 0.75%.
Low yields limit the possibility of capital impairment
By definition, the most you can lose (in nominal terms) by shorting an unlevered 10-year treasury would be 1.78% /year if held to maturity. On a 5% position this would be less than a 0.1% annual loss to the portfolio. However, if yields went to 4% - an historically very low level (where bonds traded as recently as mid-2010) the short would make approx 18%. Figure 1 shows the range of outcomes for shorting a 10 yr bond.
Figure 1: Yield / price relationship for an available 10- year treasury bond
|Last date yields hit this level||Yield||Price Change||
|1.78%||0.0%||- Current yield|
Interest rates will eventually normalize
Real “risk free” interest rates represent the price of current consumption vs. the price of future consumption and are determined in the market by the intersection of the supply of money (i.e., savings) and the demand for money (i.e., investment). Currently, real 10-year interest rates are at ≈ 0%, implying that the marginal economic actor is indifferent between current consumption and future consumption due to a lack of projects that generate a positive rate of return. Real interest rates vary widely over time and estimates of a “normalized” rate are difficult to project accurately. However, in the absence of a savings imbalance, normalized rates are always well in excess of 0% for even a slowly growing developed economy, because even slow growth would provide investment opportunities that generated a positive rate of return. I’ve asked several experts and no one I’ve spoken with was able to offer an example of sustainable real interest rates at or below 0% in a growing developed economy.
Furthermore, in order for rates to go negative for a sustained period, not only would there have to be a lack of investments with a positive rate of return, but also savings demand would have to be greater than the amount of output that can be stored (i.e., you would be better off buying a case of toothpaste than saving your money because in the future your money will buy less toothpaste). Thus, it is highly unlikely that long-term real interest rates will remain below 0% over an extended period of time.
Economic data and Fed signaling have made deflation a highly remote possibility
While our nominal losses on an unlevered short of a 10-year bond are capped at <2%/year, real losses could be much greater if the U.S. experiences a sustained period of deflation. This is because we must repay our short position with nominal dollars in the future, and if the price level in the future is lower than it is today, these dollars will be more valuable when it comes time to repay the short.
In the modern era, only three major economies have experienced deflation and currently deflation appears highly unlikely in the U.S. In Q1 real GDP rose 2.1%, and inflation is steadily above 2%, signaling that the U.S. is not close to a deflationary environment. Furthermore, in January the Federal Reserve released the first-ever inflation target of 2%. This target is significant when one considers how aggressive the Fed has shown it is willing to be. During the current financial crisis, the U.S. has not experienced deflation; however, the mere threat of falling prices has spurred the Federal Reserve to increase its balance sheet to >19% of GDP today, from about 6% in 2007.
Of course, the world economy is volatile, and a worsening of the situation in Europe or Asia could cause deflationary pressure in the U.S. If Fed actions are not timed properly, this could indeed cause temporary deflation. However, given the economic problems deflation causes and the range of policy responses available to U.S. Congress and the Fed, after reaching out to several experts on the subject I am unable to construct a realistic scenario in which the U.S. would experience sustained deflation (for the record, I don’t consider Ron Paul winning the presidency as ‘realistic’).
The U.S. fiscal situation could deteriorate further, causing rates to rise
The fiscal strength of the U.S. government depends on numerous macroeconomic and political factors that are impossible to forecast. However, it’s worth noting that U.S. government debt has risen 66% in only the last 3yrs and now stands at >100% of GDP for the first time since the 1940s. Thus according to the Government Accountability Office, in order for the debt-to-GDP ratio not to expand further, the U.S. needs to run a “primary” surplus (i.e., before interest) of 1.1% of GDP (or 4.5% of the 2011 budget) vs. a 2011 primary deficit of 5.5% of GDP (or 23% of the 2011 budget). Looked at another way, based on numbers from the Congressional Budget Office, even if one (rather generously) assumes that over the next 10 years government revenue as a percent of GDP rises 25%, borrowing costs rise 3% and U.S. real GDP grows at 2.5% annually, total debt will climb to 120% of GDP – a level on par with Italy and greater than Ireland and Portugal.
I would emphasize that even over the long-term I don’t view U.S. fiscal insolvency as a “likely” scenario, but given the difficult political choices needed to fix the situation I don’t view it as impossible either. Currently, treasury yields are priced as if there is no credit-risk to US government obligations. Any material change in the perception of the solvency of the U.S. government would not only cause bond yields to rise dramatically, but could also trigger a complex series of events with a potentially dire impact to global capital markets. A short position in treasuries would provide a hedge in this scenario.
Shorting levered ETFs provides additional upside to this view
One way to express a view on Treasury yields is by shorting levered ETFs. For example, the Direxion Daily 20+ Year Treasury Bull 3X Shares (NYSEArca: TMF) is a levered ETF that holds a combination of securities (mostly index swaps) to produce “daily investment results, before fees and expenses, of 300% of the performance of the NYSE 20 Year Plus Treasury Bond Index.” Since TMF is designed to replicate 3x the daily change in the index, the ETF must be rebalanced daily in order to increase the position when prices rise and decrease when prices fall. As such, the volatility of the underlying index will cause the value of this ETF to decrease over time even if the value of the index is unchanged.
For a simple example of how this works, supposed the underlying index rises 5pts on day 1 and falls 5pts on day two. Even though the index is unchanged, the levered ETF will lose value.
Figure 2: How volatility destroys value of a levered ETF
|Day||Index Value||Index Performance||3x Index Performance||Value of Investment|
Mathematically, the before –fee return formula for a 3x levered ETF is:
Annual return +1 = (1+ index return)^3 * exp(-3*(annual volatility^2))
The average volatility of this index is about 17%, implying that TMF would be expected to lose over 8%/yr even if the index is flat.
The NYSE 20 Year Plus Treasury Bond Index has an approximate yield of 2.85%, an approximate average maturity of 27.5 years and a duration of 17.47 years. Using this data we are able to calculate the relationship between long-term treasury yields and a short position in TMF as shown in figure 3.
Figure 3: Relationship between yields and return for a short TMF position
|Change in Long-term Rates||Estimated Index Return||3x Index Return||10%||12%||15%||17%||20%||22%||25%|
Assumes annual fees of 1.14% and 2.96% cost to borrow
Interest rate normalization
Continued gov’t dysfunction