Valuations Aren't Great for Timing Investments
There are other variables at work.
There are other variables at work.
A version of this article previously appeared in the March 2019 issue of Morningstar ETFInvestor.
Value investing is probably the most intuitive investment strategy there is: Buy what's cheap and avoid or sell expensive alternatives. Valuations have an undeniable impact on investment returns. The higher they are, the lower future returns tend to be. Yet, valuations don't appear to be very helpful for tactical adjustments across regions, sectors, and factors, or for timing exposure to credit risk.
Valuations Matter, but …
Valuations are only moderately predictive of performance. For example, from January 1970 through January 2019, a 1-point increase in the MSCI USA Index's price/earnings ratio was associated with a 0.72% decrease in returns over the next year, while lower valuations had the opposite effect. However, valuations could explain only a small part of the variation in stock returns over this period--6% to be exact. So, the market's current valuation says little about what its return over the next year will likely be.
It can take valuations a long time to revert to the mean, so it's not surprising that they appear to have greater explanatory power of returns over longer holding periods--though it's still low. For example, with a three-year holding period, starting P/E ratios could explain 15% of the variation in the MSCI USA Index's returns. The explanatory power was slightly higher over a five-year holding period, as shown in Exhibit 1.
So, why aren't valuations a better predictor of returns? They aren't the only variable that matters. Differences in expected growth rates can justify differences in valuations. As investors' growth expectations increase, so do current valuations and stock returns. If they are realized, higher valuations don't necessarily hurt returns going forward. And there are lots of surprises along the way (both good and bad), as business conditions change, that weaken the relationship between valuations and future returns.
It's also more challenging for value investing to work for tactical adjustments across regions, sectors, and factors than it is for stock selection because portfolios aren't static. So, portfolio valuations are less comparable over time.
Stock Valuation Strategies
To test the efficacy of value-driven tactical adjustments, I created a strategy that compared the P/E ratios of the MSCI USA and MSCI World ex USA indexes once every three years (as it can take a long time for valuations to rebound). Whichever index had the lower valuation would receive a 60% weighting in the portfolio for the start of the three-year holding period, while the other would receive 40%. I chose to limit these tilts because it is always important to be diversified across both U.S. and foreign stocks, regardless of valuations.
This strategy didn't help much. From the end of December 1974 through January 2019, it returned 11.15% annualized, while a static 50/50% split between the two indexes would have returned 11.04%. (The MSCI World Index returned 10.74% over this time.) This weak performance likely stems from the tenuous relationship between valuations and future returns.
The results of valuation timing were even worse when applied to sectors and factors, though there is less data here. Certain sectors (and factors) persistently trade at lower valuations than others, so without any adjustments, using valuations to select sectors would lead to long-term sector biases. However, Morningstar research shows that value-driven sector tilts are a form of active risk that historically hasn't been well-compensated (1).
To mitigate persistent sector and factor tilts, I modified the strategy to measure the attractiveness of each sector and factor index based on how its current P/E compared with its average over the past five years, favoring those trading at the lowest levels relative to their own history. The sector strategy ranked the 10 sector indexes listed in Exhibit 2 on this metric and selected the three with the lowest values. It assigned an equal weighting to the indexes that made the cut and held them for three years before rebalancing. The factor strategy followed this same approach using the indexes listed in Exhibit 3. However, it selected the two indexes with the lowest valuations relative to their history.
The results for the sector and factor strategies are shown in Exhibits 4 and 5. The performance measurement periods start in November 2004 and December 2003, respectively, and run through January 2019.
In both cases, the results were disappointing. The sector strategy lagged a static equal sector allocation by 1.19 percentage points annually. Similarly, the factor strategy lagged an equal allocation across the factor indexes by 43 basis points annually (though it beat the MSCI USA Index by 18 basis points). As with the regional indexes, this largely owes to the weak relationship between relative valuations at the portfolio level and returns. However, it's worth noting the value investment style was out of favor during much of this time.
Bond Valuation Strategies
If valuations are helpful for timing anything, I would expect it to be exposure to different areas of the bond market. Unlike stocks, bonds have a set maturity date when investors will likely receive face value, which should make yield to maturity a better predictor of bond returns than P/E ratios are for stocks. Yet, it turns out that yields aren't helpful for timing credit risk and are only a little better for timing interest-rate risk.
To test the former, I used the option-adjusted spread, or OAS, for the ICE BofAML U.S. Corporate BBB and ICE BofAML U.S. High Yield Master II indexes from the Federal Reserve Economic Data from January 2002 through January 2019. OAS measures the amount of compensation a bond is offering for its credit risk by comparing its yield to maturity (adjusted for prepayment options) against duration-matched Treasuries.
Once every three years, the strategy compared the corporate BBB index's OAS against its average over the past five years. If it exceeded that average, it held that index for the next three years, otherwise, it held the Bloomberg Barclays U.S. Treasury Index. I then replicated this strategy using the high-yield index in place of the BBB index. The results are shown in Exhibits 6 and 7.
The investment-grade version of the strategy slightly beat a static equal allocation (reblanced monthly) between the corporate and Treasury indexes by 8 basis points, while the high-yield version lagged by 20 basis points. However, in both cases, the timing strategy was more volatile than the static allocation.
It's a bit surprising that this strategy didn't work better, as OAS is a good predictor of bond returns--better than P/E for stocks. However, widening spreads can always widen further, so higher yields don't always lead to higher returns, particularly over a holding period that doesn't match the underlying securities' time to maturity.
Valuation timing worked better for timing exposure on the yield curve. To test this strategy, I used the 10-2 Year Treasury Constant Maturity series from FRED, which measures the steepness of the yield curve, based on the yield spread between 10- and two-year Treasuries. When this spread exceeded its average over the past five years, the strategy would hold Vanguard Long-Term Treasury (VUSUX) over the next three years, otherwise it held Vanguard Short-Term Treasury (VFIRX), using data from March 1987 through January 2019.
This strategy beat a static equal allocation across the two Vanguard funds by 73 basis points annually, but with greater volatility, as shown in Exhibit 8. So, this may work, but it's not a free lunch.
Valuations are helpful for gauging expected returns, so it isn't prudent to completely ignore them. If they're unusually high, future returns will likely be lower than normal, and vice versa. However, it probably isn't a good idea to use them to make big tactical adjustments among fund investments. The benefit will likely be modest at best and can easily be outweighed by lost diversification and tax efficiency.
1) Bryan, A., & McCullough, A. 2017. Morningstar. "The Impact of Industry Tilts on Factor Performance."
Disclosure: Morningstar, Inc. licenses indexes to financial institutions as the tracking indexes for investable products, such as exchange-traded funds, sponsored by the financial institution. The license fee for such use is paid by the sponsoring financial institution based mainly on the total assets of the investable product. Please click here for a list of investable products that track or have tracked a Morningstar index. Neither Morningstar, Inc. nor its investment management division markets, sells, or makes any representations regarding the advisability of investing in any investable product that tracks a Morningstar index.
Alex Bryan does not own (actual or beneficial) shares in any of the securities mentioned above. Find out about Morningstar’s editorial policies.