We received a phone call recently from a rather large bank's investment officer asking us this very question.
Total return uses three components: beginning price, income & reinvested cash flow, and ending price (market value) at a horizon date.
The question really focused on ending price, for if we never sell, then our ending price does not matter until the security matures. In essence, the only thing that matters is the current income a security provides. When the security matures, then income will be addressed again (a typical buy-and-hold strategy).
The question of market value arises because investors focus on income. Since we know all bonds someday mature at par, market value at times is overlooked if we assume the bond will not be sold. We also assume we will lock into whatever interest rate is available at the time our bond either matures or is called away.
But bonds purchased in a higher-rate environment have higher market values because they provide higher income. Market value quite simply is the present value of future income. To ignore market value is to ignore future income. And to ignore market value on a callable security is to surrender yourself to a portfolio comprised of securities purchased at interest-rate lows.
Market value is also the only way to compare securities with different maturities. As one bond matures, market value captures the difference between purchase yield and current yield. To ignore market value is to state that income today is paramount; income tomorrow insignificant.
Let's explore some examples illustrating these points:
Example 1
Typical buy-and-hold investors, because they only focus on income, could utilize a "break-even" method to compare two bullet securities. For example, suppose that we were trying to compare the risk/reward between a 1-year treasury at 5.58% and a 2-year treasury at 5.98%. If we ignore market value, then we would have to compare total income on both securities for two years (because that is the first date both securities will be free from market- value risk). Because the 2-year is locked in for both years, then the way we would solve for income is as follows:
1-year 2-year ?=6.38, i.e., in 1-year we must be able to 5.58 5.98 reinvest in a 1-year treasury at 6.38%, ? 5.98 which means that the 1-year must rise 11.96 11.96 from 5.58% to 6.38%, or 80 basis points.
Clearly in this example neither security is sold, but rather is held to the same ultimate maturity. The 2-year will provide more overall income if rates, one year later, do not rise by more than 80 basis points, and conversely the 1-year will provide more income if rates rise by more than 80 basis points 1 year later.
This method is a perfectly acceptable way to measure maturity/income risk/reward between two bullet securities with different maturities. It answers the question, "What has to happen for me to win and to lose?"
Total return effectively does the same thing by using a math short cut: present value. Employing market value does nothing more than present-value future income. But ultimately, it is future income--income you as an investor will enjoy or hate in the future as you "hold" the security. Let's solve the same 1-year vs. 2-year using total return over a 1-year horizon.
Using total return, we arrive at virtually the same answer as the break-even method of buy and hold, that one year later a 1-year treasury could be 6.41% (vs. the 6.38% using break-even analysis).
The reason that the answers are not exactly the same has to do with the time value of money...a dollar today is worth more than a dollar tomorrow. Break-even analysis does not incorporate that rule of finance. The point is that total return measures overall income vs. maturity risk and gives investors a valid math method to measure the effect future interest rates will have on total overall income. Selling at market is not required, measuring the market is.
Example 2
Our second example concerns "bragging" rights. This particular institutional investor had recently looked up its IDC Financial Publishing's total return ranking and noticed that it was in the top 90th percentile. Upper management, however, was looking at the Uniform Bank Performance Report and noticed that their institution's portfolio yield was "lower" than some of their peers. Yet, both were looking at what seemed to be the same portfolio. What gives?
Uniform bank performance reports only cover current yields, not future ones. Just because an institution currently has a higher portfolio "yield" does not mean it will stay that way. What if Bank A had an entire portfolio consisting of 3-year treasuries purchased in 1993 at 4.25% whereas Bank B had the entire portfolio consisting of 30-year treasuries, also purchased in 1993, at a yield of 5.75%? The UBPR in 1994 would say that Bank B did "better" than Bank A because it had a higher "yield" by 150 basis points.
Likewise, in 1995 and 1996 the results were the same. Bank B's portfolio looked 150 basis points better than Bank A's. At this point, the portfolio manager of Bank A is trying to tell upper management that even though their portfolio "yield" was 150 basis points lower on the UBPR, in actuality their total returns are 275 basis points better than Bank B!
What element missing in the UBPR could have shed some light on the apparent conflict? How about market value?
Sure, Bank B had bragging rights for the first three years; but if you understand market value and its implication regarding future income it was clear Bank B's "bragging rights" would be short lived when you included market value.
In this case a long bond at 5.75% was way below market and Bank A would expose this in future UBPR reports once their 3-year that was purchased in 1993 matured.
Now let's take a look at future UBPRs. Suppose Bank A reinvests its maturing 3-year treasuries into the same treasury as Bank B, only at a yield at a 1996 market of 6.95%. Just think, Bank A will now outperform Bank B by 120 basis points (Bank A now owns 6.95% whereas Bank B is stuck at 5.75%) for the next 27 years...guaranteed!
Who had the ultimate bragging rights, Bank A or Bank B? How could you have determined this fact long before the maturity of 1996? Perhaps by including market value as well as income (total return) rather than only focusing on current portfolio yields.
The opposite might also be true. Perhaps Bank A purchased a portfolio of high-quality 5-year PAC CMOs (planned amortization class collateralized mortgage obligations) at a yield of 7.3% earlier this year whereas Bank B purchased 5-year callable agencies at a yield of 7.5%. In the first year, whether rates rise or fall, Bank B will show a higher portfolio yield than Bank A by 20 basis points. Since those purchases, rates have fallen, and the callable agency that Bank B owns get called, and now the best that Bank B can get is 6.9% on a similar maturity callable agency. Market value captured the pain of reinvesting at lower rates. Current yield does not address future yields. Bank A still owns the 7.30% PAC CMO.
Clearly Bank A will hold the long-term bragging rights. Attention bank board members! If you're going to reward a portfolio manager on performance, make sure that performance is based on long-term shareholder value--TOTAL RETURN, not just yield.
Example 3
Total return is the only valid measuring device we know of when comparing three or more different bullet securities simultaneously. Under buy-and-hold (break-even analysis), how would you arrive at a single interest rate measurement if you were to simultaneously compare 1-year, 2-year, and 3-year treasuries? Since market value would not be used, you would be forced to use the longest of the three maturities.
1-year Treasury 5.58%
2-year Treasury 5.98%
3-year Treasury 6.15%
In one year, only the 1-year treasury would mature, so you would have to solve that break-even...in one year you would have to reinvest in a 2-year treasury at 6.435% to equal the yield of the 3-year (5.58+6.435+6.435=3 years of 6.15(18.45)). So, in one year the yield of the 2-year treasury would have to rise by 45.5 basis points (5.98 to 6.435). The problem is that in one year, the 2-year at 5.98% has yet to mature. Its break-even with the 3-year could not yet be measured.
Only after two years could you measure the break-even, and hence the corresponding future rate environment. The break-even would be in two years reinvesting in a 1-year treasury at 6.49% to equal the yield of the 3-year (5.98+5.98+6.49=3 years of 6.15 (again, 18.45)), or a rate rise of 91 basis points (5.58% to 6.49%).
What does this analysis tell one regarding all three simultaneously? Nothing! We have two different time periods to deal with, one year from now and two years from now.
If, however, we used a total return methodology over a one-year horizon, we could simultaneously compare the three different maturities if we solved all three for the total return of the 1-year treasury. It would turn out that in one year the 1-year could rise to 6.4% (up 82 basis points) and the 2-year could rise to 6.46% (up 48 basis points). Notice that all three are now reduced to a quantified rate environment one year from now, no selling required.
Broader applicability
Total return is the only valid methodology to evaluate any non-bullet security because none of these securities have guaranteed income to a guaranteed date. Using other methods like yield, spread to treasuries, etc., falls short.
 |
 |