Nearly half of American families have individual retirement accounts. Several other millions participate in public pensions or in workplace “defined benefit” pension plans.

If you belong to one of them, chances are your retirement contributions are allocated with a particular target in mind such as 60% equities (like stocks) and 40% fixed-income securities (like bonds).

But as asset values change, the 60/40 portfolios may drift out of balance — say to 65/35 — just because of the shifting market value of their underlying securities.

When this happens, the pensions need to ‘rebalance’ their portfolios, which means buying or selling equity, said Professor Campbell Harvey, of Duke University’s Fuqua School of Business, and most funds have “mechanical rules” in place for this rebalancing.

However, “automatic” rebalancing may lead to important investment losses for the pension, as shown in new research Harvey co-authored with Alessandro Melone of Ohio State University and Capital Group’s Michele Mazzoleni.

In the new working paper, “The Unintended Consequences of Rebalancing,” the researchers explain that most institutional investors rebalance either at regular times — monthly (when pensions need liquidity to pay member benefits) or quarterly — or whenever the portfolios drift from their target ratio.

These “calendar” and “threshold” adjustments, they say, offer predictable “signals” to other investors, for example hedge funds.

“Suppose the pension has drifted to 65-35. They need to sell some equity. Sophisticated traders know this and they start selling ahead of the pension – they ‘front run’ the trade – driving equity prices down,” Harvey said. “This means the pension has to sell for less, and the losers are the pensioners.”

Estimating the costs of ‘rebalancing’

The researchers calculated the deviations from a 60/40 portfolio using market data between 1997-2023.

Assuming both “threshold” and “calendar” rebalancing (both “easy to compute, available in real time,” they write), they found that these adjustments significantly reduce portfolio returns.

 “The cost is economically important given that $20 trillion in U.S. retirement assets are subject to rebalancing,” Harvey said. “The average U.S. annual salary is about $63,700, and employees save roughly 7.4% of their pay for retirement — or about $400 monthly. We estimate the total losses as $16 billion per year which means that the average pensioner loses $200 per year due to the mechanical rebalancing – and our estimates are conservative.”

“For many pensioners, a month’s worth of contributions is lost and over a 24-year horizon, this is like losing two years of contributions,” he said.

A confidential roundtable of pension fund managers

To validate their findings, the researchers convened a roundtable discussion with a “group representing a global network of public pensions,” managing approximately $2 trillion in assets, they write. 

The participants, which included senior managers like Chief Investment Officers, confirmed they are aware of the costs associated with predictable rebalancing policies. 

One source also noted the difficulty of trying to convince their “investment committee” to change rebalancing policy, the researchers write.

One fund mentioned that it is easier for them to let their “alpha desks” (internal teams responsible for generating excess returns) exploit the predictability of automatic rebalance. 

“It is extraordinary that a pension would effectively ‘front run’ its own trading – as well as other pensions’ trading,” Harvey said. 

Making rebalancing less predictable

The researchers say rebalancing is important for maintaining portfolio diversification and managing risk. “Without rebalancing, a balanced 60/40 equity/bond portfolio would drift to 80/20 within approximately 10 years,” they write.

However, the authors say that mechanical rebalancing is “unwise” and lead to unnecessary costs. 

“The problem is the trades are known in advance,” Harvey said.

“These funds manage trillions, and their actions can move the market”, he said. “It’s never a good idea to announce your trade in advance. Certain investors know the pension will be buying or selling and they will front-run or get ahead of the trade”.

Harvey said their paper is about “uncovering the problem,” and more research is needed to develop solutions.

But “it is not that difficult to solve the problem,” he said.

“Why not just make the rebalancing less predictable? This end of month or quarter stuff needs to go.”