Retirement Investment Strategies: A Downside Perspective

|
Includes: VWESX
by: Dennis Uyemura

Summary

Analysis of negative return periods for long-term corporate bonds versus S&P composite.

Effect of longer holding periods on total return distributions.

Rising corporate bond rates do not necessarily result in negative bond returns due to term premium.

In an earlier article ( Retirement Investment Strategies: Risk and Return), I illustrated the risk and return profiles and Sharpe ratios for the following major asset classes over several long investment periods:

  • LTCorp (MUTF: VWESX): Long-term investment grade corporate bonds. (Adjusted prices from Yahoo Finance.)
  • SPComp: S&P composite adjusted for dividends using data from Robert Shiller website. (Adjusted prices were derived from the index values and dividends.)

For most long investment periods, LTCorp outperformed SPComp on a risk-adjusted basis. However, based on the questions and discussion from that article, many still view that there are outsized risks related to investments in LTCorp given the current prospects for future rising interest rates. The purpose of this article is to analyze and discuss those downside risks.

How do I define a downside risk? For this article, I will define it simply as any defined holding period where investing in LTCorp or SPComp results in a negative total return. However, before proceeding with the analysis, it is important to describe the behavior of interest rate cycles with respect to business cycles.

The Nature of Interest Rate Cycles and Business Cycles

The first exhibit illustrates the close relationship between interest rate cycles and business cycles.

Exhibit 1: History of 3-month Libor and 10-year Treasury yields (monthly averages) from 1965 through 2015. For reference, shaded bars indicate recessions (NBER).

Notice that historically, recessions provide an excellent marker for the start of falling short-term interest rates during most interest rate cycles. There is only one major decline in rates not denoted by recession: in late 1984 and early 1985. (Interest rate cycles are far easier to define and visualize using short-term rather than long-term rates.) Typically, rates rise and fall about 400 basis points (4 percentage points) or more during a typical rate cycle.

In contrast, 10-year Treasury yields provide a clearer view of the secular trends in interest rates: a rising secular trend from 1965 through 1980 compared to a falling secular trend from 1981 through 2016. It is critical to distinguish the 7 or 8 interest rate/business cycles (normally each about 5 to 10 years in length) from the very long-term secular trends that have lasted several decades in length. In the current low rate environment, markets are anticipating either the onset of a new rising secular trend sometime in the future or an extended period of very low but stable rates.

The Frequency of Negative Returns with LTCorp and SPComp

To detect periods of negative returns, monthly adjusted prices from January 1980 through December 2015 were used. Rolling 1-year, 3-year, and 5-year total returns were calculated for LTCorp and SPComp. The histograms of total returns for each investment period are shown in the next three exhibits.

Exhibit 2: Rolling 1-year total returns for LTCorp and SPComp from 1981 through 2015 (420 observations for each series). There were 58 negative LTCorp returns and 84 negative SPComp returns.

Exhibit 3: Rolling 3-year total returns for LTCorp and SPComp from 1983 through 2015 (396 observations for each series). There were 3 negative LTCorp returns and 62 negative SPComp returns.

Exhibit 4: Rolling 5-year total returns for LTCorp and SPComp from 1985 through 2015 (372 observations for each series). There were 0 negative LTCorp returns and 53 negative SPComp returns.

The bin ranges were held constant in the last three exhibits to allow direct comparison of the distributions of returns. For the 1-year returns shown in exhibit 2, the distributions appear similar with modestly higher averages and dispersions for SPComp compared to LTCorp, which is not surprising. For the 3-year returns in exhibit 3, the SPComp returns become much more broadly dispersed. For the 5-year returns in exhibit4, SPComp returns become extremely dispersed from -30% to +200% in the last exhibit. In contrast, the LTCorp return distributions remain relatively narrow and clearly unimodal. What is most important is the frequency of negative returns, as summarized in the next exhibit.

Exhibit 5: Frequency of negative returns for LTCorp and SPComp for various rolling investment periods.

This exhibit demonstrates that negative returns with LTCorp are dramatically reduced with the length of the investment period. In contrast, the frequency of negative returns is only modestly lessened with SPComp with longer investment periods. The lesson from this exhibit is that, with LTCorp, time is your friend. That is, as long as your investment horizon is 3- or more years, it is extremely rare to ever experience a negative total return for LTCorp. (It may be of interest to readers that the only three negative LTCorp 3-year returns occurred during the worst of the financial crisis. They correspond to the 3-year periods that ended in Sep, Oct, and Nov 2008 when corporate bond prices were exceptionally depressed.)

Rising Interest Rates and Negative Returns

The prospect of rising interest rates concerns anyone considering investing in LTCorp. This section deals with that issue. First, we need to select an indicative market interest rate to use in analyzing this risk. The next exhibit displays how 1-year LTCorp returns correlate with changes in three candidate indicator market rates: 10-year Treasury yields, Aaa corporate bond yields, and Baa corporate bond yields. (All were obtained from the Federal Reserve H.15 report.)

Exhibit 6: Correlation matrix for rolling 1-year changes in 10-year Treasury yield, Aaa corporate bond yield, Baa corporate yield versus the 1-year total return on LTCorp fund.

Based on this exhibit, changes in Baa corporate bond yields had the highest absolute correlation with LTCorp total returns. Of course, the correlation is negative, indicating that total returns decline when Baa rates rise. Hence, we will use rolling changes in Baa bond yields for the analyses presented in the next three exhibits.

Exhibit 7: Rolling 1-year LTCorp total returns versus 1-year changes in Baa yields. Of 420 observations, 162 had Baa yield increases and 58 had negative LTCorp total returns.

Exhibit 8: Rolling 3-year LTCorp total returns versus 3-year changes in Baa yields. Of 396 observations, 96 had Baa yield increases and 3 had negative LTCorp total returns.

Exhibit 9: Rolling 5-year LTCorp total returns versus 5-year changes in Baa yields. Of 372 observations, 46 had Baa yield increases and 0 had negative LTCorp total returns.

What may be very surprising to some is that, most of the time, Baa rate increases are not associated with negative LTCorp total returns. With 1-year returns in exhibit 7, about one-third of the Baa yield increase periods were associated with negative returns. With the 3-year and 5-year investment periods, negative returns were miniscule in number compared to the number of rate increase periods. (And some of those rate increases were more than 100 basis points, or 1 percentage point.)

These exhibits show that rate increases are not necessarily (indeed, not usually) associated with negative LTCorp total returns. What is critical is to understand why this is true.

The Term Premium As Protection Against Negative LTCorp Returns

This is the major take-home lesson of this article. In the prior article, it was stated that the term premium embedded in the yield curve may represent an excess return because the implied forward interest rates are generally far higher than actual future interest rates. In essence, this means that the term premium is providing downside protection against future rate increases.

Here is the point: The only time that total returns can be negative are when future interest rate increases are higher than the already rising pattern of implied forward rates embedded in term rates. Historically, this does not happen very often. The result: Multi-year total returns for LTCorp are overwhelmingly positive even during periods when corporate yields rise. That is, the term premium provides an effective bias against negative multiyear returns.

Once more, we conclude that time protects against negative LTCorp returns. And what type of investing provides ample time to achieve this protection? Retirement investing, to be sure. Hence, I believe that LTCorp should be considered as a strong candidate asset class for retirement investing.

Disclosure: I am/we are long VWESX.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.