An ominous sign for stock returns?

Goldman Sachs recently reported that the allocation to equities as a percentage of household assets had risen to levels last seen at the height of the NASDAQ Bubble. Is this an ominous sign of a crowded trade? Are investors in a crowded long that stocks are about to enter a painful 2000–2002-style bear market?
 

 

 

A less alarming view

The Goldman study, which is based on Federal Reserve data, calculates equity exposure as a percentage of household assets. But the demographics of the household change over time and investors go through a savings and investing cycle as they age.
 

In 2011, a team of academics led by John Geanakoplos wrote a paper entitled Demography and the Long-Run Predictability of the Stock Market. Geanakoplos et al related demography to long-term stock returns. They found that P/E ratios were correlated to the ratio of middle-aged people to young adults, otherwise known as the MY ratio. When MY rises, the market P/E will tend to rise and when it falls, P/Es tend to fall.
 

In 2012, a San Francisco Fed study, Boomer Retirement: Headwinds for U.S. Equity Markets?, used a slightly different methodology than the paper by Geanakoplos et al, and they postulated a market bottom in 2021.
 

 

A San Francisco Fed follow-up study in 2018 found that the forecast P/E ratios had deviated from the population ratio and attributed the deviation to the pronounced effect of Baby Boomers’ retirement.
 

 

The Gen X demographic, which is the age cohort behind the Baby Boomers who are just entering retirement, is now in their prime earnings and savings years. The stock market boom may be attributable to fund flows from Gen Xers.
 

 

Investors can see a demographic-adjusted view of portfolio allocation by the AAII Asset Allocation Survey. Unlike the weekly survey, which asks AAII members how they feel about the stock market, the AAII monthly survey asks members how they are invested. That is to say, what they are doing with their money instead of how they feel.

 

The AAII Asset Allocation Survey shows that equity allocations (black line) is indeed elevated relative to its own history, but levels are nowhere near the NASDAQ Bubble years. Similar elevated readings have seen the S&P 500 either stage minor pullbacks or consolidate sideways. In all cases, excessive equity allocation is not an actionable intermediate-term sell signal.

 

 

 

Valuation challenges

Even then, U.S. equities face valuation challenges. John Authers at Bloomberg observed that, for the first time in 22 years, bonds are yielding more than stocks, though that’s also not an actionable investing signal. Nevertheless, it is a warning that the equity risk premium is becoming overly compressed.
 

 

 

A productivity debate

Even then, these valuation warnings are not an automatic sell signal. Much depends on the growth rate of earnings and, more importantly, the future trajectory of productivity gains from adoption of artificial intelligence.
 

 

AI boosters, like NVIDIA CEO Jensen Huang, have characterized AI as the start of another industrial revolution. On the other hand, labour economist and newly minted Nobel laureate Daron Acemoglu argued in a paper that productivity gains from AI will be limited to “no more than 0.66% increase in total factor productivity over 10 years” and predicts gains of “less than 0.53%”. In a NY Times interview, Acemoglu explained that AI “can automate only about 5 percent of an office worker’s tasks” by automating routine work and “free workers to tackle more brainy challenges like developing a business strategy for a new product launch”.

 

Aswath Damodarn, finance professor at NYU Stern School of Business who is considered the dean of company valuation, wrote about the challenges of AI in a Financial Times article when a friend put all of Damodarn’s work through an AI LLM to absorb everything Damodarn knew about company valuation. He concluded that, in today’s environment of AI adoption, analysts have to be better and more creative at what they do. AI has to be faster and perform better than humans at financial modeling using historical data, but will have difficulty with “valuation built around a business story, enriched with soft data”. Learn to avoid the “Google curse” or looking up answers instead of “reasoning answers on our own”. It’s no surprise that companies are seeking employees with soft skills instead of just hard modeling skills.

 

In conclusion, U.S. household equity allocations are becoming extended relative to their own history, which warns of a challenging long-term return outlook. Demographically adjusted allocations are less extended and similar episodes have resolved in either minor pullbacks or sideways consolidations. Equity valuations are stretched compared to bonds and the long-term outlook will depend on the future advances in productivity.

 

While excessive equity allocation could be a warning if a subdued long-term equity return outlook, perhaps the best short-term contrarian indicator of U.S. returns is this cover from the Economist.

 

 

1 thought on “An ominous sign for stock returns?

  1. Indeed AI models are very capable assistants to experienced workers. Once people learned how to prompt with these models the productivity goes up substantially. This is especially true for very specialized technical fields. This is a big boon to corporations. It saves a ton of money and time in R&D compared with the old days. The first order effect would be higher profit margin. Then there is this virtualization which further narrows the target area. So there will be a group of very highly compensated specialists and the rest of the population. As the technical progress is accelerating, the gap is getting bigger and bigger. Let’s see how our political leaders solve this problem. If you pay attention to the Nobel Prize winners this year you see the harbinger of the trend. Mr. Acemoglu is mainly addressing those mundane work. But if you look at all these technical fields the advance is breath-taking. It is just so overwhelming.

Comments are closed.