Herding in Pension Fund Asset Allocation

Pension funds don’t allocate capital in isolation. They follow benchmarks, hire consultants who recommend the same allocations, and monitor their performance against peers. The result is herding: nearly all large pension funds hold nearly identical portfolios—60–70% equities, 20–30% bonds, small alternatives—regardless of their liabilities or time horizons. When markets stress, this crowding amplifies losses and creates systemic risk.

Why Pension Funds Converge: The Mechanism

Pension fund asset allocation is not a free-market outcome. Three forces drive herding:

1. Benchmarking and policy liability targets

Most defined-benefit (DB) pension funds operate under a liability-driven investment framework. A fund’s “policy” allocation is set to match the duration and inflation exposure of its liabilities (promised benefits to retirees). Because most DB liabilities look similar—indexed partially to wages, partially to inflation, with a long but finite tail—target allocations converge naturally. A pension fund with 20 years of liabilities to pay needs growth assets (stocks) to meet return targets, but not as much as a younger fund. Two funds with similar liability profiles will adopt similar allocations.

2. Consultant recommendations

Pension fund trustees—often non-expert board members—hire consultants to design their asset allocation. A few large consulting firms dominate the space (Russell Investments, Aon, Mercer, etc.). These firms codify their recommendations into standard “glide paths” and asset-class mixes. A consultant might recommend “U.S. equity 40%, Int’l equity 18%, Fixed Income 22%, Alternatives 20%” to any fund matching certain liability profiles. Hundreds of funds receive this same recommendation and adopt it. The consultant doesn’t need to control any capital; the recommendations propagate through delegation and trust.

3. Career and reputational risk

A pension fund manager who deviates wildly from peers risks career damage. If allocations diverge and underperformance results, the manager can be blamed. But if an allocation is standard across the industry and underperforms, responsibility diffuses; “everyone faced the same headwind.” Career risk incentivizes conformity. A manager who holds 35% equities while peers hold 65% faces scrutiny if markets rally—they’ll lag by multiples. It’s safer to herd.

The Aggregate Effect: Crowded Markets and Synchrony

By 2020, the top 100 U.S. pension funds held roughly $5 trillion in assets. Despite different names and governance, they held nearly identical allocations:

• Equities: 60–65%
• Fixed income: 25–30%
• Alternatives (private equity, real estate, infrastructure): 5–10%

Within equities, the same pattern repeats: 60–65% U.S., 25–30% developed international, 10–15% emerging markets. Within alternatives, the same preference for private equity over direct real estate or credit.

This homogeneity is the crowding. When markets move, all three decisions ripple through the same positions simultaneously.

The Mechanism of Crisis Amplification

Here’s what happens during a shock:

1. Asset prices fall suddenly (e.g., March 2020, post-Lehman 2008). Equity allocations fall from 65% to 55% of portfolio value (because equities are now worth less).

2. Liability-driven rebalancing kicks in. A fund’s policy says “hold 65% equities, 30% bonds.” Now that equities are underweight at 55%, the fund is supposed to buy equities to restore 65%. But first, to maintain liquidity for rebalancing, many funds must cover margin calls on derivatives or redemptions from pension contributors.

3. Pension funds sell what’s liquid first: public equities, bonds, or alternative assets that permit early exit. They don’t sell illiquid private equity; instead, they raise cash by dumping public markets.

4. Crowding magnifies: Hundreds of funds facing the same rebalancing problem all try to buy equities (or sell alternatives) at the same moment. Bids for credit and illiquid alternatives evaporate. Selling pressure intensifies.

5. Feedback loop: Forced selling by pensions pushes prices lower, triggering more pension losses, forcing more rebalancing sales. This is a fire-sale dynamic.

Historical Examples: Herding Under Stress

August 2015: A sharp equity correction forced rebalancing sales. Pension funds, endowments, and other large institutions all raised cash by selling illiquid alternatives (private equity, hedge funds) at distressed valuations. The “dash for cash” made headlines.

March 2020: The COVID shock forced one of the largest simultaneous pension rebalancing events. Funds faced margin calls on hedges and couldn’t meet pension payments. They dumped public equities and alternatives to raise cash, even at losses. The coordinated selling exacerbated the selloff for one of the worst weeks in market history.

March 2023: Regional bank failures triggered a flight to safety. Pension funds with exposure to commercial real estate or emerging-market debt all moved for the exits. Illiquid assets (a major pension holding) faced no bids; funds had to accept fire-sale prices or hold to maturity.

In each case, the crisis was deepened not just by panic but by the synchronized rebalancing of a sector that thinks alike.

The Alternative Hypothesis: Liability-Driven is Rational

Defenders of pension convergence argue that herding is not irrational but optimal. A pension fund with fixed liabilities should hold allocations that match those liabilities—immunization is prudent. If two funds have similar liabilities, similar allocations are correct, not a bug.

This is true in theory. But in practice, the convergence goes beyond liability matching. Pension funds differ in their level of funding (some are 100% funded, others 70%), their payout horizons (some have 30 years, others 10), their wage-inflation sensitivity (public vs. private), and their ability to adjust benefits. Yet allocations are remarkably uniform—suggesting that liability matching alone is not the only driver. Benchmarking, peer pressure, and consultant groupthink add a layer of synchrony beyond what funding ratios would predict.

Systemic Risk and Policy Concerns

Regulators and policymakers have noted the concentration risk. If a shock forces pension funds to delever simultaneously, and they all own the same assets, the result is severe illiquidity for those assets—credit spreads widen, real estate bids evaporate, even public-equity markets suffer from crowded exits. A shock that might be manageable if responses were dispersed becomes dangerous when responses are synchronized.

Some proposals to reduce herding risk include:

• Diversification mandates: Require pension funds to hold heterogeneous allocations; this is hard to enforce without forcing suboptimal liability matching.
• Liquidity requirements: Ensure funds hold sufficient liquid assets to meet obligations without forced sales; this increases costs.
• Stress-testing: Model synchronized deleveraging scenarios and adjust allocations; many funds now do this, but it’s backward-looking.
• Regulatory friction on alternatives: Make it harder to exit alternatives during crises; this reduces panic selling but increases illiquidity.

None is a panacea. The fundamental tension is that optimal portfolio construction (matching liabilities, minimizing cost) naturally converges across similar funds. The price of eliminating that convergence is either inefficiency or policy complexity.

The Moral: Crowding and Hidden Risk

Pension fund herding illustrates a broader principle: consensus allocations hide tail risk. When an allocation is held by 80% of the market’s largest institutions, that allocation is the market. There’s no other side to take. In normal times, this matters little. In a crisis, it’s dangerous.

A pension manager who holds an unusual allocation—say, 50% equities instead of 65%, or overweight alternatives—faces performance drag in a bull market but protection in a crash. The hidden cost of herding is not visible until the herd panics.

See also