A recent study reported that DC plans underperformed DB plans by 0.70% per year over the 1990–2012 period.¹ That’s a substantial gap given the long time horizon of retirement investing. And it comes after adjusting for the differences (in size and equity allocation) among plans that might explain the performance gap. In the end, the researchers concluded the difference was due to fees, because DC plans are more likely to invest in higher-cost mutual funds. In many ways, the study reinforced conventional wisdom on this topic.
Yet, if we dig a bit deeper into the data, it’s not clear to me that we can conclude that DB plans are winning the investment horse race. Nor can we conclude that higher DC fees account for most of the difference. In fact, exactly the opposite may be true: DC plans may be outperforming DB plans.
Consider the following, using data from the published study. Imagine investing $100 in the typical DB plan and the typical DC plan in January 1990. And assume, at the end of each calendar year through 2012, you are credited with the average DB or average DC return reported in the study.² Think of this as an investment competition between the aggregate DB and DC systems.
By 2012, the $100 investment in the typical DB plan would have grown to $374. But the $100 investment in the average DC plan would have reached $429. In other words, the typical DC system investor would have had 15% more in aggregate wealth than the typical DB investor. The average annual compound returns are 6.53% for DC plans and 5.90% for DB plans. That’s a difference of 0.63% in favor of DC plans.
So how can the same data yield exactly opposite conclusions? One reason is that the original study pooled individual plan year returns, while my calculation is based on compound average returns. In particular, my calculation considers the sequence of average returns over time. In full disclosure, in my calculation, both DB and DC systems actually ran neck and neck from 1990 to 2001. It was during the 2001–2012 period that DC plans pulled ahead.
Those who think regularly about investment performance will immediately see a broader issue. There are a host of factors influencing DB and DC plan performance. The most critical is asset allocation policy. Over the study period, there were major changes in the portfolios of many DB and DC plans. DB plans reduced equity exposure, shifted to longer-duration fixed income, and added alternative assets. DC plans increased equity exposure, and generally held mostly large-cap U.S. stocks and short-duration fixed income assets. Also, in the latter period, DC plans shifted to a more professionally diversified approach with the growing use of target-date funds and advice services. Unfortunately, none of these factors are tracked with this study’s data.
But there’s more. Rebalancing policy stands out as a major difference. DB plans rebalance regularly, whereas DC plans typically do not, although this is changing due to target-date funds and advice. Another factor is contribution timing—with DB plans having more control over the timing of contributions, while DC contributions are linked to payroll cycles. And finally, there are fee differences. DB plans, which are generally larger, are more likely to use lower-priced trust vehicles. But as DC plans have grown, they have expanded their use of trusts and institutionally priced mutual funds.
A paradox in the data
Three other points highlight the paradoxes within the data.
First, in the study’s actual results, the DB/DC difference fell from 0.80% in 1990–2002 to 0.30% in 2003–2012. So the data also suggests that, using its preferred metric for performance, DC plans substantially narrowed the gap with DB plans. This is consistent with a view that the Department of Labor’s fee policy, plus market competition, began to drive down DC investment fees in the latter period. My own sense is this is part of the explanation, but more than one factor is at work.
A second anomaly in the data is that DC plans took more stock market risk—yet they underperformed when the stock market rallied, and outperformed when the stock market fell! As one example, during the global financial crisis, DC plans outperformed DB plans (by 760 basis points) in 2008, when the market slumped. And then, when the market rallied in 2009, DB plans outperformed DC plans (by the same amount). This was despite the fact that DC plans held a larger position in equities. This type of variation has nothing to do with fees.
A final paradox in the data is this. The largest DC plans not only underperformed large DB plans—but they also underperformed smaller DC plans. Again, fee differences seem an unlikely explanation. After all, the most intensive fee competition occurs for the largest DC mandates.
What do I take away from all of this? It seems like we all have a good understanding of what factors will influence the relative performance of DB versus DC plans. But we certainly don’t have the right dataset to answer the question fully, or attribute the results to a given factor like fees.
The very fact that, with the same dataset, you can conclude that DB plans outperformed (in the original study) or that DC plans outperform (by my alternative metric)—all of this suggests that calling the horse race between DB and DC plans awaits better data and a deeper analysis of the competition.
¹ Munnell, Alicia H. Jean-Pierre Aubry, and Caroline V. Crawford, “Investment Returns: Defined Benefit vs. Defined Contribution Plans,” Center for Retirement Research at Boston College, Number 15-21, December 2015.
² In this calculation, in each year, I used the average plan return published by the study in its appendix. Using median plan returns yields about the same result.
- All investing is subject to risk, including the possible loss of the money you invest.
- Past performance is no guarantee of future results.