(This post assumes you’re familiar with the concepts outlined in the preceding mental model post on how to make money investing)
There are no shortage of people in this world selling promises. Financial advisors sell you promises. Banks sell you promises. If you’re an allocator, asset managers, consultants, third party marketers and cap intro groups all line up to sell you promises, too. Such-and-such returns over such-and-such a time period with such-and-such volatility.
Only rarely are these promises derived through anything resembling deductive logic. They’re almost always based on storytelling and data-mining.
When your financial advisor tells you she can get you 8% (god forbid, 10%) annualized on your US-biased public equity portfolio, what is that number based on? It’s almost always just a historical average. Same with your Fancy Consultant pitching private equity or middle market lending or crypto-cannabis venture capital or whatever other magical strategy happens to be selling well at the moment.
He who builds on historical averages, builds on sand.
There’s no natural law requiring US equities to return somewhere between 8% and 10% on average over 20-year rolling periods. Same with your private equity and middle market lending and crypto-cannabis venture funds. As with everything, you should build up your return expectations from first principles.
For bonds, your expected nominal return* over the bond’s tenor is equal to the starting yield.**
For stocks, your expected nominal return is equal to the starting dividend yield, plus expected growth in earnings, plus any change in valuation (price).
For the remainder of this post, we’ll focus on stock returns.
Remember the two ways to make money investing? You’ve got cash distributions and changes in investor preferences. Dividend yields and expected growth in earnings are the fundamentals of your cash distributions. Multiple expansion, as we’ve noted before, is always and everywhere a function of changes in investor preferences for different cash flow profiles.
Where people get themselves into trouble investing is extrapolating too much multiple expansion too far into the future. When you do this, you’re implicitly assuming people will pay more and more and more for a given cash flow stream over time. This kind of naive extrapolation is the foundation of all investment bubbles and manias.
If you want to be as conservative as possible when underwriting an investment strategy, you should exclude multiple expansion from the calculation all together. This is prudent but a bit draconian, even for a curmudgeon like me. I prefer a mean reversion methodology. If assets are especially cheap relative to historical averages, we can move them back up toward the average over a period of, say, 10 years. If assets are especially expensive relative to historical averages, we can do the reverse.***
Below are a couple of stylized examples to illustrate just how impactful changes in valuation can be for realized returns. Each assumes an investment is purchased for an initial price of $500, with starting cash flow of $25, equivalent to a 5% annual yield. Cash flows are assumed to grow at 5% per year, and the investment is assumed sold at the end of Year 5. Only the multiple received at exit changes.
In the Base Case, you simply get your money back at exit.
In the Upside Case, you get 2x your money back at exit.
In the Downside Case, you only get 0.25x your money back at exit.
Despite the exact same cash flow profile, your compound annual return ranges from -12% to 17.9%. I hope this conveys how important your entry price is when you invest. Because price matters. It matters a lot. At the extremes, it’s all that matters.
If you’d like to get more into the weeds on this, Research Affiliates has a fantastic primer on forecasting expected returns. Research Affiliates also offers a free, professional grade interactive asset allocation tool. It covers a wide range of asset classes in both public and private markets.
In the meantime, what I hope you take away from this post is that there are straightforward models you can use to evaluate any investment story you’re being told from a first-principles perspective. Often, you’ll find you’re being sold a bill of goods built on little more than fuzzy logic and a slick looking slide deck.
Notes
*Real returns for bonds can vary significantly depending on inflation rates. This is a significant concern for fixed income investors with long investment horizons, but lies beyond the scope of this post. Really, it’s something that needs to be addressed at the level of strategic asset allocation.
**Technically, we need to adjust this with an expected loss rate to account for defaults and recoveries. This doesn’t matter so much for government bonds and investment grade corporate issues, but it’s absolutely critical for high yield investments.
***Why is it okay to use historical averages for valuation multiples while the use of historical averages for return assumptions deserves withering snark? The former is entirely backward looking. The latter at least aspires to be forward looking.
Also, if you can establish return hurdles based on your investment objectives, you can back into the multiple you can afford to pay for a given cash flow stream. That’s a more objective point of comparison. Incidentally, the inputs for that calculation underscore the fact that valuation multiples are behaviorally driven. Mathematically, they’re inversely related to an investor’s return hurdle or assumed discount rate.