A fundamental problem with target-date funds
July 15, 2020
Target-date funds lead to defined contribution plan members taking more risk and becoming more diversified than they would be if left to their own devices. But despite these benefits, TDFs fail to provide what members really want and need — a reliable income replacement for life managed by somebody else.
My colleague Ioulia Tretiakova and I were puzzled by a fundamental problem with DC plan design and by TDF’s blunt approach to risk and inability to accommodate personal circumstances when we wrote a paper for the Rotman International Journal of Pension Management back in 2011 titled “What DC Plan Members Really Want.”
Specifically, we found that target-date funds are not appropriate for most members. Few DC plan members can fully fund their retirements upon joining a pension plan. Most members face a contingent liability problem, a requirement to plan, save and invest for an ultimate liability based upon unknown future events. TDFs, lifestyle funds and other accumulation strategies invest as though plans are fully funded using mean-variance optimization or modern portfolio theory to preserve and grow capital. Yet, MPT/mean variance optimization cannot handle a contingent liability. Further, TDFs try to preserve and grow capital to a retirement date (some through retirement), but members need income replacement through retirement as well.
TDFs are also poor vehicles for replacing income like defined benefit plans do. In our research, we looked at data for various TDFs, collectively representing 70 per cent of U.S. TDF assets, using a nine per cent contribution rate, reasonable assumptions and historical returns. We found these replaced 70 per cent of income 52 per cent of the time. While better than a balanced portfolio, which has a probability of delivering 70 per cent replacement income 37 per cent of the time, TDFs offer no better than a flip of a coin chance of achieving this goal.
We decided to look for a better solution with the belief that we shouldn’t expose investors to periods of excessive volatility when it is the risk of an asset allocation rather than the allocation itself that delivers the portfolio to its goal. So we used a dynamic risk approach that applied goals-based research developed for high net worth investors using risk rather than asset allocation as the foundation.
The dynamic constant risk approach changes the TDF model in two ways. It targets capital at age 65 to fund 70 per cent replacement income – or whatever target replacement rate is appropriate – and uses an algorithm to dynamically adjust portfolio risk to target capital, like a GPS on your phone would. It also uses constant portfolio risk (standard deviation) rather than fixed asset allocation to provide a stable model input.
In our study, a dynamic constant risk approach delivered 70 per cent of replacement income 97 per cent of the time in historical testing with less risk than bonds in the critical five years before retirement date.
This research drew the attention of QSuper, the public service superannuation plan for Queensland, Australia. QSuper’s members, like the balanced funds in our research, were falling short of adequate pension outcomes, so it implemented a plan incorporating many key principles of our approach in 2014 and has been very successful to date.
Overall, while taking more risk and broader diversification have benefited TDF investors, TDFs were never designed to be pension plans. Rather, harnessing technology to provide mass-customized goals-based solutions is the real trend to watch.
Mark Yamada is president and chief executive officer of PUR Investing Inc. These views are those of the author and not necessarily those of the Canadian Investment Review.