Betting on where interest rates will be in a year or two does not offer very attractive rewards compared to the risk.


In the present yield curve, one year rates are at 2% and 10-year rates are at 5%. If we make the simplifying assumption that interest rates can only wind up at two equally probable levels a year from now, then the present yield curve implies that there's a 50% chance that the 10-year (or more precisely, the 9-year) interest rate will drop to 3.6750%, and a 50% chance that it rise to 7.2833%.


Your bet is to either buy or sell a 10-year zero coupon bond.


[In terms of your ability to predict where interest rates will be, other bets, including for example the S&P 500, are effectively leveraged versions of this wager, with some extra noise. They offer the same ratio of expected profit to risk. Actually, more often than not, they offer a somewhat poorer ratio due to the extra noise.]


If you predict that interest rates will go down (to 3.6750%), you'll buy the 10-year; if you predict that they'll go up (to 7.2833%), you'll short the 10-year (and park your principal plus the proceeds of your short sale in a one year zero).


With either bet, if your prediction is correct, you'll make 18.4%; if your prediction is wrong, you'll lose 14.4%.


If you flipped a coin, then in terms of arithmetic average, your expected profit is 2%, which equals the return on a one year note. However, in the long run, you'd be better off sticking with the one year note if you're going to flip a coin, because your expected compounded rate of return will be 0.69%. This is because a given percentage loss hurts your rate of compounding more than the same size percentage gain helps it.


But let's suppose you're good at predicting interest rates. How good? Well, let's suppose that in the long run you can expect to win 60% of your interest rate bets. My guess is that it's very unlikely that anyone is this good: the yield curve represents the consensus of an ongoing debate between many highly capable arbitrageurs. It figures to be tougher than a point spread sports bet. ... But let's suppose you're this good.


Then your expected compounded rate of return is a whopping 3.93% per year. Not too exciting, eh? And don't think you can improve this by leveraging up. Leveraging up will will actually reduce your expected rate of compounding (if the one year and ten year zeros are your only investment options, you should actually leverage *down* in order to maximize your expected rate of compounding).


If you're looking for an edge in the financial markets, you really should look someplace besides betting on where interest rates are going. Furthermore, you should try to *avoid* betting on interest rates: if you're unwittingly exposed to rates (and most investors and traders are), you're tossing a large amount of volatility into your results with no compensation. And even if you happen to be very skillful at predicting rates, you really only get to make statistically independent bets every couple of years. In a lifetime of investing, you might make 15 to 20 bets. ... Many investment managers with stellar track records just made one or two lucky interest rate bets, even though they and most of the financial world didn't recognize it. And of course, there's no reason to believe that winning one or two coin flips is a strong indication of enduring skill.


One way to avoid betting on where interest rates are going is to have a very long investment horizon (and holding period) and invest primarily in dividend paying securities. Warren Buffett comes to mind (he doesn't like dividend paying companies because of the tax disadvantages, so he actually prefers to buy positive cash-flow companies outright. But if not for the tax penalty, he'd much prefer that the companies he invested in pay dividends rather than retain profits).


Another way to avoid betting on interest rates is called "immunization" in the fixed income world. A typical textbook example is buying 3 and 5 year bonds and shorting 4 year bonds. The resulting portfolio doesn't fluctuate as interest rates change, at least in the short run. But you don't have to completely flatten your results, and it isn't necessary to take short positions. Many bonds have implicit options (e.g., mortgaged-back bonds) which give them "interesting" total return profiles vs. interest rates ... for example, it's possible to create bond portfolios that will outperform Treasuries regardless of where interest rates wind up a few years from now. Working stocks into an "immunized" investment strategy is more difficult (unless you want to change less than 10% of your portfolio annually), but it can be done.