“If you don’t believe you can estimate the expected return, you shouldn’t be in the investment business.” — Bernd Scherer
While the Capital Asset Pricing Model (CAPM) has weaknesses, it provides a useful place to begin for forecasting returns. “It links expected returns to an objective measure of risk and current interest rates,” says Page.
In theory, the market portfolio at the guts of CAPM calculations includes all assets, private and non-private. In practice, listed stocks and bonds are suitable substitutes for many investors. The global market was about 60% stocks and 40% bonds in 2000. Today, on account of share buybacks, privatizations, fewer IPOs, and the issuance of huge amounts of presidency bonds, the ratio is about 40% stocks and 60% bonds. Investors can calculate expected returns for the wide selection of assets in multi-asset portfolios by combining the weighted estimates for stocks and bonds after which multiplying by the beta of every asset.
A straightforward inversion of a stock market’s price-to-earnings (P/E) ratio gives an affordable rough estimate of stock returns. Which P/E ratio? The Shiller CAPE (cyclically adjusted P/E) provides a cyclically adjusted measure for the United States. The low return implied by today’s high levels could also be too pessimistic if the rise in profitability over the past decade will be maintained. Higher profits could also be everlasting on account of the quasi-monopoly position of the large technology firms. In addition, recent profits could also be understated on account of accounting problems. In contrast, measurements based simply on today’s earnings could also be too optimistic. The writer finds that combining the historical and current earnings approaches leads to forecasts which can be near the estimates of a variety of industry peers.
Forecasting local currency government bond yields is easy and comparatively reliable. The current yield to maturity provides a superb estimate of long-term yields. Yield shocks can drive bond prices down (or up), but are offset by higher (or lower) reinvestment rates in the longer term.
The CAPM is a valuation-agnostic model. However, equity valuations exhibit a powerful mean-reversion effect over the long run, so investors can improve their estimates by incorporating valuation forecasts. Equity returns will be decomposed into three components, with income and growth alongside the change in valuation. Dividend payouts are constant, so income forecasts based on current returns are reliable. Earnings growth ought to be anchored to economic growth because earnings as a share of economic output must mean-revert over the very long run.
Page explores various methods for fine-tuning forecasts, including analyzing institutional investor flows and dynamics across asset classes. The sheer volume of macro data makes it difficult to differentiate signal from noise. Color-coded dashboards are a superb method to display data on relationships where macro aspects matter for asset prices.
A review of 93 academic studies by Ser-Huang Poon and Clive Granger noted that “there is no clear winner in the great race to predict risk.” Investment risk is complex. But increasing the complexity of risk models doesn’t necessarily improve their predictability. So what should investors do? Page suggests using a variety of various models – and using judgement.
The simplest approach is to assume that the volatility of every asset class next month will likely be the identical because the previous month. This approach can be hard to beat, because volatility stays the identical from month to month. But over the long run, the other is true. Five years of calm markets usually tend to be followed by five years of turbulence, and vice versa.
Models based on normal distributions underestimate the probability and magnitude of downside risks. However, Page has not found persistent patterns that may help us predict skewness and kurtosis, the statistical measures of those extremes. Instead, he suggests other approaches to modeling tail risks.
Modeling risk-on and risk-off environments individually can provide a more realistic view of potential downside risks by incorporating stressed betas and correlations. Scenario evaluation – using each historical events and forward-looking scenarios – can add one other layer of understanding. However, investors need to think about how markets have modified since those historical events. For example, emerging markets are less sensitive to commodity price changes today than they were in 2008, while bonds, as measured by the Barclays Aggregate Index, are more sensitive to rate of interest changes as the common maturity has increased (from 4.5 years in 2005 to 6 years in 2019).
Once investors have forecasts for returns, risks, and correlations, they’ll input them into an optimizer to calculate the really helpful asset mix. Most optimizers recommend concentrated portfolios and are sensitive to small changes in inputs. Investors can overcome these limitations using five methods:
- Limit the weighting to individual asset classes.
- Apply group constraints, akin to holding in alternative assets. (This isn’t a random alternative. Many alternative asset forecasts overestimate expected returns and underestimate risk, resulting in recommendations for big holdings.)
- Use resampling methods developed by Richard Michaud that have in mind the uncertainty of forecasts.
- Apply the Black-Litterman approach, which mixes the forecasts of lively investors with forecasts from the CAPM, making an allowance for the reliability of those forecasts.
- Optimize in three dimensions: risk, return and tracking error in comparison with peer group weights.
The stock-bond mix is a very powerful decision multi-asset investors make, but this mix doesn’t reliably reduce risk. The diversification advantages of Treasuries often show up during stock sales, but stocks haven’t protected investors from bond sales. Stock-bond correlations were positive within the Seventies and Nineteen Eighties when inflation and rates of interest drove volatility. This was also true throughout the 2013 “tantrum” when the Federal Reserve signaled monetary tightening and in 2018 when benchmark rates of interest rose.
Fixed income investors usually tend to achieve their retirement goals with bonds, especially inflation-indexed bonds. However, most investors haven’t saved enough for retirement. They usually tend to achieve their retirement goals with stocks.
Are carbon-based energy firms a essential hedge against inflation or future stranded assets? How do social and governance issues affect the sustainability of sovereign debt in emerging markets? Asset managers have necessary decisions to make on these issues, but surprisingly the book doesn’t address environmental, social and governance evaluation.
There isn’t any one right approach to asset allocation. Page quotes his father, a now-retired finance professor: “We don’t know the outcomes in advance. The information we use is always incomplete and we can’t control the variables. Nevertheless, we have to make decisions, because often no decision is worse.” Investors must use their judgment to pick the fitting tools for the job. The range of tools Page presents on this book can assist investors make higher decisions.
If you liked this post, don’t forget to subscribe.
Photo credit: ©Getty Images / Ioannis Tsotras
