Bearish on Bonds: The Long-Run Costs of Owning TBT

Includes: TBT, TLT
by: Sebastien Buttet

by Sebastien Buttet

The poor financial performance of leveraged ETFs, when held over long periods of time, is well known by Wall Street professionals and retail investors alike (perhaps to a lesser extent for the latter group). In fact, the financial powerhouses that offer leveraged ETFs make very clear to potential investors in their prospectus that performance should only be evaluated on a daily basis.

In a related article entitled Leveraged ETFs: A Look at a Play on the S&P 500, we compared the relative performance of the Proshares S&P 500 ETF (NYSEARCA:SPY) and the Proshares Ultra Short S&P 500 (NYSEARCA:SDS) when funds are held for different time periods. The results were astonishing: if held for only one day, tracking was almost perfect; over a month period, slippage was about 1%; over a three-year period between April 12th, 2008 and April 11th, 2011, the S&P 500 Index lost 1.5% of its value; the twice bearish SDS fund was down 66% implying a loss of 69%.

We propose a comparative analysis of the relative performance of two currently popular Treasury bond ETFs:

  1. the iShares Barclays 20+ year Treasury bond fund TLT - The Fund seeks investment results that correspond generally to the price and yield performance of the long-term sector of the U.S. Treasury market as defined by the Barclays Capital 20+ Year Treasury Index. Average daily volume for TLT is equal to 7 million shares.
  2. the Proshares Ultrashort Treasury bond fund TBT - The Fund seeks daily investment results, before fees and expenses that correspond to twice (200%) the inverse (opposite) of the daily performance of the Barclays Capital 20+ Year U.S. Treasury Index. Average daily volume for TBT is equal to 10 million shares.

We suspect that daily volume for both ETFs will pick up in the next couple of months as investors position themselves for potential gains or losses in the Treasury market once the Fed's QE2 program ends in June. In fact, even professional investors are utterly divided about the fate of the bond market:

1. Bill Gross, the co-head of PIMCO fund who supervises more than 1.2 trillion dollars in assets is bearish on Treasuries and has initiated short positions. His thesis is quite intuitive:

  • Interest rates on Treasuries will rise in the coming months (implying a fall in Treasury bond prices) as demand for Treasury bonds gradually fades;
  • The Fed's QE2 program is scheduled to end in June;
  • Japan might slow down on buying U.S. Treasury bonds as it needs to repatriate some of its yens home to start rebuilding;
  • China needs to change its overall economic model by moving away from being an export-oriented country with high saving rates to becoming a consumer-led nation.

2. On the other hand, Rick Rieder, chief investment officer of BlackRock with 1.6 trillion dollars under supervision, has been buying. His thesis also relies on intuitive arguments:

  • Markets have already discounted the end of QE2 as investors have known for a long time the end date of the Fed's bond buying program;
  • Very few fixed-income assets offer the quality of the U.S. debt, even after last week's warning by S&P. Potential investment substitutes include the debt of strong European countries such as Germany but the fate of the euro currency is uncertain because countries at the periphery (the PIIGS) have dire economies with anemic (or negative) growth and large budget deficits. The other option would be buying the Japanese debt, but the economic future for the country of the rising sun is equally uncertain following the tragic earthquake and tsunami, and the country's debt to GDP ratio is greater than 200%;
  • If QE2 is inflationary, then the end of QE2 should be deflationary, usually a good environment to hold fixed-income securities. Treasury bonds actually rallied when the Fed's first QE ended, on the premise that the U.S. economy was not enough to grow on its own without stimulus provided from the Fed and the U.S. government. True, the economic situation in the U.S. and other countries around the world has improved, but the minority of investors who believe that QE3 is well on its way is growing every day.
  • Finally, if history is any guidance, interest rates in Japan in the early 2000s were quite stable even after the country's sovereign debt was placed on a debt watch by S&P.

In Table 1 below, we compare the returns for TLT and TBT for a hypothetical investment over different time periods that would have been liquidated on 04/26/2011.

Tab 1. Comparison of TLT and TBT for different holding periods
Returns for TLT Returns for TBT Relative Performance: 2*TLT versus TBT
Intraday - 04/26/2011 0.97% -1.95% -0.01%
1 month 1.82 -4.57 -0.93
3 months 1.54 -6.43 -3.35
6 months -7.69 7.63 -7.75
1 year 4.15 -24.61 -16.31
3 years 2.21 -48.25 -43.83

Results are as expected and as described in the TBT prospectus. For short periods of time (a few days and even up to one month), returns on TBT are roughly equal to twice the inverse return on TLT, the 20+-year Treasury bond index. For example, TLT surged by 0.97% on April 26th, 2011 while TBT that same day lost 1.95%. Pretty close!

However, for holding periods longer than 3 months, returns on TLT and TBT differ markedly. Between April 26, 2010 and April 26, 2011, prices for the 20+-year Treasury index increased by 4.2%. Adding interest payments of about 4.5%, the total 1-year return on TLT was equal to 8.7%. Should TLT and TBT track each other closely over periods of time longer than a day, you would expect TBT to be down 17.4 % in the one-year time period. However, the Ultra Short ETF performed much worse and lost 24.6% of its value.

Things get even dicier when looking at a three-year time period. Between April 26, 2008 and April 26, 2011, TLT gained 2.21%. Assuming that the average interest rates on the 20+-year Treasury bond was no greater than 4%, TLT's total return in the last three years was equal to 14.21%. If returns on TLT and TBT were as described, you would expect TBT to be down 28.42% over the same time period. How did TBT perform? It lost about 48% of its value, a 20% underperformance!

Since a graph is worth 1000 words, we illustrate the 3-year performance for TBT and TLT below.

TLT: iShares Lehman 20+ Year Treasury Bond Fund

Our message is simple. Investors who are bearish on Treasuries and wish to benefit from a decline in bond prices would be better served by shorting TLT directly rather than going long TBT shares. Those who still decide in favor of owning TBT should be prepared to see their principal reduced by 7% per year. A rather unfortunate situation could materialize where investors are correct about the declining Treasury prices but could still lose money if they hold shares of TBT!

Finally, for aggressive investors with deep enough pockets to answer margin calls, we offer a trade to benefit from the poor performance of TBT when held over long periods of time: sell short say $1,000 worth of TBT and sell short $2,000 worth of TLT (the amounts $1,000 and $2,000 are for illustration only. The idea is to short twice as much TLT compared to TBT). Although we caution that we did not perform full due diligence, we believe that the proposed trade (short TLT - short TBT) is profitable in the long run.

First, let us explain the rationale behind our trade. TBT does a poor job of tracking the 20+-year Treasury bond index over the long run, so it should be sold, which in essence is equivalent to go long Treasuries. To make the trade profitable regardless of the market's direction, TLT must be sold to hedge the short position on SDS. And since TBT is supposed to return twice as much as TLT, we need to short twice as much TLT compared to TBT.

Next, we examine the returns for that particular trade over a three year time period. Between April 26, 2008 and April 26, 2011:

  • TLT gained 2.21%, implying that the short $2,000 position on TLT would have declined by about $44. In addition, short-sellers need to make interest payments. Assuming that rates on 20+-year Treasuries averaged 4% in the last three years, investors would have to disburse another $240. The total loss on the TLT position over a three-year time period is equal to $284.
  • TBT declined by 48% over the same time period. A short position on $1,000 would be up by $520.

The total payoff for the proposed trade is $236.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.