Bayesian Investor Blog

Book review: Investing Amid Low Expected Returns: Making the Most When Markets Offer the Least, by Antti Ilmanen.

This book is a follow-up to Ilmanen’s prior book, Expected Returns. Ilmanen has gotten nerdier in the decade between the two books. This book is for professional investors who want more extensive analysis than what Expected Returns provided. This review is also written for professional investors. Skip this review if you don’t aspire to be one.

The Specter of Lower Investment Returns

Real interest rates have been unusually low the past decade, and reached an extreme in 2021.

Another way to phrase that is that discount rates have been unusually low.

There’s been more savings than can find a way to be safely invested at at positive expected return. Investors kept some assets in bonds and money market funds even when those investments were expected to lose a bit of value, and increasingly put money into riskier startups that aren’t expected to become profitable for years (Rivian, Luminar, early stage drug development companies).

Note that Ilmanen finished writing the book in 2021. Some of the low returns that he predicted happened while the book was being printed. This is especially true of bonds – Ilmanen’s warning would have been very valuable a year ago. At roughly the same time as he finished writing the book, bonds started a one-year decline of more than 30%. So his advice about bonds, which was great a year ago, has become possibly obsolete today, due to bond yields that have returned somewhat close to normal. (I warned my readers about the coming bond market decline in mid-2020.)

Pension Problems

Most of the book’s advice applies to a wide range of market conditions. The main recommendation that is fairly specific to recent conditions is to save more for retirement than you would in an environment of high expected returns.

Some money managers, particularly those running pension funds, are pressured to target fairly specific returns.

They usually base their target returns on realized returns for the past few decades. But the past 40 years have produced returns that were partly due to rising price/earnings ratios for stocks, and declining yields on bonds. I.e. price appreciation that is unlikely to be sustained, because the appreciation causes other sources of returns on those investments to approach zero.

Ilmanen shows that the expected long-run returns on a typical stock/bond portfolio dropped below 4% in 2021. Yet US state and local pension funds only reduced their target to 7.25%.

The most tempting strategies to earn a 7.25% return when standard strategies are expected to 4% involve taking more risk, usually with increased leverage and/or investing in illiquid assets. This has some chance of working, at the cost of making failure more dramatic. The UK recently had some sort of pension fund crisis which likely resulted from this kind of strategy. And the FTX meltdown was likely in part an extreme version of that – sometimes the only alternative to lowered expectations is increasingly desperate gambles.

Most of what investors need is the willingness to accept that sometimes it isn’t possible to earn the returns that are available in normal times.

Diversify

We’ve all heard that diversification is the only free lunch in investing. But did you know that well-executed diversification is indistinguishable from magic?

The best parts of the book describe strategies for better diversification. He recommends using a diverse set of criteria across which we should diversify, including:

• industries
• countries
• asset classes: stock, bonds, commodities, currencies
• strategies: value, momentum, defensive, carry
• multiple signals, e.g. for value, use a mix of dividends, free cash flow, book value, etc.

“Let every man divide his money into three parts, and invest a third in land, a third in business and a third let him keep by him in reserve.” – Talmud.

Patience

Most importantly, the fact that so many investors capitulate from almost any factor after a few years of underperformance justifies the word “risk” in alternative risk premia strategies. Wise men have said “no pain, no premium” – it is precisely the painful times that will sustain the premium and prevent it from being arbitraged away.

How would you balance disappointing performance over the past three years against positive evidence over the past century? … Guided by … “the law of small numbers“, investors tend to expect any long-run edge to manifest itself within a short period.

Investors persist in evaluating strategies using time horizons of 1 to 5 years. That creates a modest cyclical tendency for a strategy to outperform for a few years, as investors become more bullish on the strategy based on its recent good performance. Then when the strategy becomes about as crowded as it can get, that effect stops fueling the outperformance, and several years of disappointing performance cause many investors to abandon the strategy.

I still feel some temptation to follow trends that have been going on for a couple of years. My best strategy for breaking that temptation is to remind myself of the much longer-term historical evidence, by reading sources such as Ilmanen’s books.

Market timing mostly doesn’t work, at least for the S&P500. It keeps looking like some strategy based on Shiller’s CAPE should tell us to buy at major lows and sell at major highs. Ilmanen did some rigorous testing of such strategies, using buy and sell thresholds that avoided lookahead bias, and showed that market timing did work well over the 1900-2020 period, but approximately all of the benefit came from a 30-year period that ended shortly after 1950. So a little bit of market timing is likely to be a good idea, but don’t count on it working during your lifetime.

I made some good money by timing the market between September 2008 and March 2009, based on signs of macroeconomic problems and on a comparison to 1937-8. But the conditions under which I can repeat that kind of success are rare.

For most such successes, there’s an offsetting mistake, such as selling when Alan Greenspan expresses concern about irrational exuberance.

Unorganized Tidbits

Why did the expected returns of US treasury bonds drop to -1% in 2021? Some of the answer is high savings rates causing plenty of capital to be available for all investments. But there’s also the effect of negative bond-stock correlations in making bonds a valuable hedge for stock market risk.

Ilmanen suggests that the bond-stock correlation is negative when there’s little uncertainty about inflation. “As long as inflation uncertainty remains mild, stock-bond correlation is likely to remain negative.”
That provided an excellent prediction about what would trigger the kind of bond crash we saw in 2022. It’s embarrassing that it was combined with a faulty view of inflation (I presume much of the book was written before the 2021 burst?? of inflation was visible).

How reliable is the equity premium? Newly compiled historical evidence say there was no equity premium in the 19th century.

The size premium is an illusion. There’s an illiquidity premium, which maybe causes a slight tendency for small illiquid stocks to outperform.

The shift to passive, as well as to ETFs and factor investing, could lead to higher correlations between single stocks and thus higher systematic risk.

Summarizing The Impact of Market Conditions on Active Equity Management:

There is also some evidence that active stock pickers tend to outperform during recessions, at times of high dispersion between stock-specific returns, and especially during “differentiated declines” – when weak markets and wide dispersion coincide.

The book has only one brief mention of cryptocurrencies, in a section about gold, implying that he thinks of cryptocurrencies as a minor subset of commodities.

Ilmanen sounds a bit strange when he strays into macroeconomics: “As long as economies are weakened by the virus and the lockdown effects or by precautionary saving, inflation pressures should be muted.”

The old story of “the Fed taking away the punch bowl when the party gets going” sounds quaint in this century when the only question seems to be how much the Fed is spiking the bowl.

(See Scott Sumner for a better view of the past 20 years.)

Could an AI-related acceleration of economic growth invalidate Ilmanen’s forecast of low returns? Alas, the evidence is weak and confusing as to whether faster growth causes higher stock market returns. Cross-country comparisons show a negative correlation over the 20th century.