To illustrate the thoughtlessness of this allocation “bunching,” we compared the median balanced manager to two alternative approaches—a “religious rebalance” and a “diehard drifter.” Our “religious rebalancer” maintains a near-continuous 60% equity allocation by rebalancing the portfolio back to 60/40 every month. In contrast, our “diehard drifter” never rebalances, allowing his portfolio mix to drift with the whims of the market.
The black line in Figure 1 plots the mid-point between these two allocators. Virtually all of the allocation movement of the peer group is captured by a combination of price drift and continuous rebalancing!
The other interesting tidbit is the performance of stocks versus bonds over this stretch. From June 2004 through September 2011, the S&P 500 underperformed the BarCap US Aggregate Bond Index by a significant margin (2.0% versus 5.8% compound annual return). This underperformance shouldn’t come as a shock as during virtually the entire time horizon—save for a couple of months at the depths of the Global Financial Crisis—stocks were expensive, trading at Shiller P/E ratios well above the historical average. Asset allocation managers evidently don’t put a whole lot of independent thought into these asset mixes.
The Method Behind the Madness—Benchmarks!
The late economic historian and consultant Peter Bernstein wrote a wonderful piece lamenting that the days of astounding active manager performance, like baseball’s .400 hitters of yesteryear, were a thing of the past.2 While much of this was due to markets becoming more efficient, Peter also noted that the focus on benchmarks was increasingly to blame. He cited Mark Kritzman’s contribution to one of his Economics and Portfolio Strategy newsletters:
Failing unconventionally was never a happy event in this business, but clients’ love affair with benchmarks has made large tracking errors extremely perilous for managers. As Mark Kritzman recently pointed out, active managers are reluctant to form portfolios on the basis of unconstrained optimization of risk and return because the recommended allocations typically deviate too far from the allocations in the benchmark.3
Returning to the eVestment Alliance US Balanced/TAA peer group, we find 75% of the managers with a stated benchmark are tied to a 50/50 or 60/40 equity–debt combination. No wonder their allocations are so tightly concentrated! Peer pressure attributable to short-term benchmarking has transformed asset allocation—arguably the largest determinant of future portfolio returns4—from an independent and informed exercise of risk and return to an automated process getting us to an approximate normal allocation.5 Certainly adjustments are made—a tweak in favor of equities here, a small-cap bias there—but these differentiations tend to be trivial. The pattern still seems to be one size fits all.
The Tonic—Outcome-Oriented Investing
Solving this problem requires us to take a step back. What are these portfolio strategies trying to achieve? What is the desired outcome? Individual, long-only asset class mandates in equities or fixed income naturally should be expected to “beat the benchmark.” Such a relative comparison is sensible given the huge swings in absolute performance over even 5- to 10-year stretches, particularly in equities. But active asset allocation, between basic diversification and shifting the mix, can presumably at least partially offset big declines.
We assert that the success of an asset allocation fund should be measured versus a full market cycle outcome rather than an intermediate-term benchmark. Since 1900, a 60/40 blend of U.S. equities and bonds have produced a total return of 8.1% according to our own research, not far from the 7–8% embedded in actuarial return assumptions and most 401(k) calculators. Of course, that is a nominal result, which is only of use to those whose liabilities are not subject to inflation—that is, very few if any! In the real world, retirees buy consumer products whose prices generally increase and in specific cases, like health care, rise very fast. Thus, we need to translate this ultra-long-term historical experience into real (post-inflation) returns. With inflation running at an annualized clip of 3% since 1900, we get to a 5% per annum real return. If history repeats, this figure is the long-term outcome investors are looking for from basic asset allocation management.
We further assert that it is reasonable to use such a “CPI-plus”6 as our targeted outcome and build a long-term asset allocation program to meet it, benchmarks be damned. On this outcome-oriented approach, the construction of an asset allocation or balanced mandate allows for several important advantages versus the constraints of an asset allocation benchmark:
- Greater Diversification. Using our oft-cited “Equally Weighted 16 Asset Class” portfolio,7 Figure 2 shows that this expanded opportunity set produces returns commensurate with a 60% S&P 500 / 40% BarCap Aggregate blend at a lower risk level. The Sharpe ratio of this approach is 0.61 versus 0.50 for the traditional 60/40 mix, an increase of over 20% in risk-adjusted results. However, this greater efficiency comes at a cost to benchmark-sensitive investors in the form of nearly 5% tracking error. An easy interpretation of this number is that, on average, once every six years the portfolio will underperform 60/40 by more than 5%, well outside the comfort range of many fiduciaries on a total portfolio basis.
- Tyranny of Asset Class Benchmarks.8 The focus on 60/40 or 50/50 mixes also creates intra-asset class dilemmas at odds with producing a reasonable risk-adjusted outcome. As we’ve claimed before, cap weighting in equities structurally places more of the portfolio in overpriced stocks and less in underpriced. Ask 1999 era S&P 500 investors how Cisco, Oracle, and Lucent—the so-called axis of wealth destruction—treated them. As counterintuitive as cap weighting equities appears, bond cap-weighting may be even more egregious. Why on earth do we want to lend more to those companies or countries who are the biggest debtors? Given our 3-D prognosis,9 where over-indebted nations will try to reflate away their obligations, short-term comparisons to a capitalization-weighted global sovereign debt yardstick may be incredibly burdensome to a forward-looking asset allocator.
- Dynamic Risk Posture. Asset allocation benchmarks are more or less static risk portfolios as seen in Figure 3, which displays rolling 10-year returns and volatility for a 60% S&P 500/40% Ibbotson Long-term Government Bond blend in the post-World War II era. The 60/40 portfolio hovers around 10% volatility over most 10-year stretches, though it occasionally dips below 8% or rises above 12%. While the risk is more or less static, the returns swing from barely positive to 18% per annum, indicating regimes where risk bearing is rewarded and penalized. By focusing on an outcome, we allow ourselves to be in line with Warren Buffett’s quote: “Be fearful when others are greedy and greedy when others are fearful.” Using a benchmark with more or less constant risk exposure is inconsistent with a Buffett-like contrarian approach.