On Mental Model Mistakes in the Middle of the Modern Money Management Industry
In recent times, we observe that too many investors believe that the “alpha” of an investment program can be attached like an appendage to such a program’s “beta.” We believe that this wire-crossing belief stems from a conflation of sustainable recurring earnings power from an operating business and mark-to-market returns from an investment portfolio.
- The sustainable recurring earnings power of an operating business is the earnings that such a business can reasonably expect to earn if it keeps doing what it is doing, year after year. For example, if Costco continues to maintain its stores, warehouses, logistics network, and operational teams, it should reasonably expect to make a certain level of earnings each year, adjusted for changes in economic conditions, competitive dynamics, etc.
- In contrast, the mark-to-market returns of an investment portfolio for a particular time period (e.g., a year) simply reflect the change in expectations of the aggregate investment community about that portfolio over such a period. Almost by definition, such a change in expectations is not expected to keep recurring each year thereafter; otherwise, all such future changes of expectations would themselves have been capitalized and impounded in the mark-to-market return in the first year. Capitalized returns are, by their very nature, non-recurring.
The one circumstance in which annual mark-to-market returns of an investment program could be reflective of its annual sustainable earnings power is where the investment program involves a lot of rapid trading over the course of the year, employing a process that is repeatable year after year. For such a program, generating “alpha” on an annual basis can be likened to manufacturing widgets on a standardized assembly line.
However, many investment programs — including Discerene’s — do not involve the rapid trading of securities, but instead involve the buying and holding of investments over multiyear periods. In such cases, the mark-to-market return of an investment portfolio in any given year generally reflects the front-loading of portfolio returns from future years. Likewise, the mark-to-market loss of an investment portfolio in any given year generally reflects the storing-up of portfolio returns for future years. Accordingly, for buy-and-hold investment programs, mark-to-market returns are inherently mean-reverting and “alpha” is inherently lumpy.
We believe that applying the “sustainable recurring earnings” mental model to “alpha” and pursuing trend-following instead of mean-reverting asset-allocation heuristics when attempting to acquire it often amplifies investment errors, especially when confounding variables1 drive returns for stretches of time.
The amiss and artificial “alpha-as-appendage” view is particularly dangerous because returns compound geometrically, not arithmetically. At the limit, if Investor X is up +100% in Year 1 and down -100% in Year 2 in a two-year flat market, Investor X’s “alpha” over the period isn’t zero; it’s -100%.
Thus, when companies with little or no intrinsic value are the ones that run up the most, it is especially exigent for long-term investors to maintain their discipline and resist the pressure to keep up with the Joneses.
For example, in 2021 — when there was an unsubtle inverse correlation between the intrinsic values and market prices of businesses — just about the only way for an investor to “beat the market” was to be more irrationally exuberant than other investors.2 The financial-markets rally in that year was narrow and concentrated on “high-growth” companies (both public and private) with growth expectations that investors took to burlesque and gelastic extremes. Piling into these investments in 2021 required a suspension of disbelief and a shushing of common-sense judgment. But not piling into these investments meant being willing to be “left behind” as investors propelled one another to shell out ever more gaudy prices for emperor’s-new-clothes assets based on ever more frenzied and fever-pitched narratives.
*****
We also observe that money printing, the demand for more sources of investment returns, and the ascendance of Modern Portfolio Theory3 as the dominant mental model have driven many investors to search for new, uncorrelated “asset classes.” When one such “asset class” is “discovered,” a flood of money tends to flow into it. Sometimes, these “asset classes” are perceived to have “low beta,” i.e., they are not highly correlated to the world public equity index (e.g., the MSCI ACWI) and therefore worthy of larger allocations. There are often knotty methodological assumptions underpinning this conclusion.
First, Modern Portfolio Theory itself posits that the “world index” against which all assets should be regressed is the index of all possible assets in the world, public and private. This true and complete “world index” is not readily observable, so in practice, some investors simply run regressions on the returns of each “asset class” against the MSCI ACWI. This use of the MSCI ACWI as the “world index” is not problematic so long as most of the world’s financial assets are publicly traded, but, in an example of the reflexive nature of investing, this has become less and less true over the years as more and more assets stay private.
Crucially, using the MSCWI ACWI as the “world index” means that many “asset classes” are perceived to be lower beta than public equities. Some investors then conclude that a larger part of their model portfolios should be allocated to such “asset classes” but, of course, this reasoning is viciously circular. If one defines the “world index” as the “world public equities index,” then all other “asset classes” will naturally be less correlated to the “world public equities index” than the “world public equities index” itself — but the observation is trivially true and wholly unhelpful for the task of assessing actual risk and reward.
Second, many “asset classes” have less readily observable market valuations than public equities. However, the mere fact that an “asset class” isn’t readily marked to market — and mark-to-model valuations (often self-reported) seem remarkably stable — doesn’t mean that it isn’t risky or that its cash flows are not correlated to those of other asset classes. We continue to believe that it is cash flows, not mark-to-market prices, that matter when assessing risk.4
Third, marginal inflows into certain “asset classes” may be heavily driven by asset-allocation models, such that the capital gains observed in the “asset classes” may simply be caused by increases in allocations to such “asset classes” and hence self-fulfilling prophecies. Naturally, when the engine for such Ponzi-like phenomena sputters, e.g., due to the reversing of free-money policy, the correlation among many of these “asset classes” goes to one.
Despite (or perhaps because of) these asset-allocation mental-model issues, some investors end up (consciously or not) making lusty bets in exceedingly niche “asset classes” (e.g., Sand Hill Road-flavored venture capital) that account for wee fractions of the global economy, while large swaths of the real economy (with real population, real consumers, and real producers) are ignored.
1. https://www.scribbr.com/methodology/confounding-variables/.
2. “I’ll see your Sea Ltd. and raise you a Rivian!”
3. Thanks in large part to the towering work of Harry Markowitz, who sadly passed away on June 22, 2023.
4. For example, we would imagine that the cash flows of private companies that are, on average, smaller, earlier in their life cycles, and more operationally and/or financially leveraged would be riskier than their public counterparts — and in the same direction.