Pension fund plan sponsors tend to measure a portfolio’s performance against a benchmark. The proportion of defined contribution (DC) assets in a mutual fund is used to proxy the importance of a manager beating a benchmark. The authors find that a portfolio’s DC fraction is positively correlated with either future fund-level beta or future holding-level beta. This observation indicates that the benchmarking pressure can lead managers to alter their market risk exposure.

High-DC funds are observed to hold 3.8% more of their portfolio in high-beta stocks and 2.8% less of their portfolio in low-beta stocks. The increased (decreased) demand for high-beta (low-beta) stocks incentivized by a growing number of benchmarked funds could reinforce the observed low (high) risk-adjusted returns on these stocks (i.e., the high-beta/low-alpha pricing anomaly).

A one-standard-deviation increase (0.261) in fund beta is found to improve relative performance by 0.57% annually, whereas little effect is observed on alpha. A high-beta strategy appears to be effective in improving future fund performance, relative to a style benchmark, without hurting risk-adjusted returns.

A 0.57% increase in relative performance is found to increase DC fund flow by 0.90% to 1.31%, but neither the impact of the level nor the change in beta to DC fund flows is statistically significant. Plan sponsors seem to rate relative performance as an important factor for fund selection and pay little attention to the beta rank.

With the incentive to minimize tracking error (TE), high-DC funds are observed to have significantly narrower cross-sectional distribution of fund betas. As their assets increase, funds are observed to have a higher passiveness measure (R²), lower activeness measures (active share and active weight), and no significant change in future TE. High-DC funds appear to achieve lower volatility around the benchmark by more precisely targeting beta to lie just above 1 (or the index risk level).

The fraction of DC assets in mutual funds is used as a proxy for benchmarking pressure from a plan sponsor. DC asset positions are collected from the Pensions & Investments annual survey. Return and volatility measures of fund categories and style benchmarks are obtained from the Morningstar database. Holding prices and index returns are collected from the CRSP Survivor-Bias-Free Mutual Fund Database and Thomson Reuters. The final sample is restricted to equity funds and contains 4,603 fund-year observations, which covers 1,093 distinct funds over a 10-year period (2003–2013).

Fund betas are calculated using market model regression on CRSP value-weighted return over three-month T-bills. Holding-level betas are computed by taking the value-weighted average of underlying stock betas held by each fund. The holding-level beta segregates the beta impact of manager stock selection from such other elements as changes in cash, leverage, or trading costs.

High-DC funds are compared with low-DC funds to assess whether the index-linked pressure alters fund beta preference. Regressions on performance (relative return and market alpha) and DC flows are used to test determinant factors and measure the impact of a high-beta strategy. In this way, the authors link academic findings to incentives of market players and the effectiveness of the strategy.

The authors also examine managerial activeness of each fund through such deviation measures as tracking error, active share, active weight, R², and the cross-sectional standard deviation of beta under each quintile.

Benchmarking is a popular way for investors, especially pension plan sponsors, to assess portfolio managers and decide which funds to invest in. Mutual funds with strong incentives to compete for pension money tend to be sensitive to index-relative selecting criteria.

While aiming to explore the effect of benchmarking pressure, the authors actually study the impact on fund manager behavior of the presence of pension assets. The results reveal an oversight from plan sponsors and the unintentional consequence of motivating portfolio managers to lean toward high-beta stocks, which should be helpful for plan sponsors seeking to improve measuring methods while selecting appropriate funds.

Investors in the same mutual funds may have different objectives. The Employee Retirement Income Security Act of 1974 (ERISA) influenced plan sponsors to seek anchoring points to reflect plan objectives. For other investors, without being required to disclose asset composition of retirement and nonretirement, it is difficult to be aware of potential agency conflicts and to avoid them in advance. This research could be of interest to regulators as well as non-retirement investors.

Another thought is that there is potentially larger systematic risk embedded in (supposedly) conservative pension investments. Leverage limits are placed by some regulators to avoid excessive market exposure, but the implicit leverage involved in using high-beta strategies would raise systematic risk exposure without being noticed.

Whether DC assets could be used to represent benchmarking pressure is arguable; the authors restrict their study’s scope to US equities, and thus, it may not be applicable to other markets. Return data studied are net of fees, perhaps because of data availability, so the impact of management fees has not been considered.