- For an active equity, passive fixed-income approach to be advisable, the performance of active equity would have to dominate that of active fixed-income. This is an update of a year-ago paper that examined whether such active-equity dominance was indeed indicated by the real-world data.
- Previous results showed active fixed-income outperforming active equity over sample periods ending in 2006, but active equity showing slight outperformance when the financial crisis (2007-08) was added into the mix.
- After adding in 2009 results in the present paper, active equity and active fixed-income results become roughly comparable, while active fixed-income retains its edge for sample periods not including the 2007-09 crisis cycle.
- All in all, the results analyzed here do not indicate any dominance of active equity over active fixed-income. If anything, for periods excluding the recent financial crisis, active fixed-income has outperformed active equity.
Some commentators are recommending that investors restrict fixed-income allocations to passive mandates, confining active management to equities or alternatives. In a 2009 white paper,1 “Active Equity and Passive Bonds?”, we pointed out that in order for such an approach to be advisable, information ratios for active equity managers should thoroughly dominate those for active fixed-income. We found that actual results did not support this. If anything, active fixed-income information ratios were superior to active equities.
However, in those previous results, inclusion of 2007-08 results improved the performance of active equity relative to that of active fixed-income. For sample periods ending in 2006, active fixed-income performance was generally superior to that of active equity, but for sample periods ending in 2008, active equity performance was generally on par with or a bit better than that of active fixed-income.
As most investors are no doubt painfully aware, 2007 and 2008 subjected both fixed-income and equity markets to trauma unlike anything seen since the 1930s. However, over the last 10 months of 2009, equity and spread-sector fixed-income markets recovered so dramatically that they reversed much of the preceding declines. It is natural to ask how the active management performance results look with the 2009 experience folded in, and that is the focus of this paper.
The next section updates the results presented a year ago to include 2009. For a full discussion of why information ratios (and analysis of active alpha) should be the final criterion for a passive fixed-income approach, the reader can consult our earlier paper.2
Active Management Results Through 2009
We analyze manager returns as reported in the Rogerscasey Investworks database for the four manager universes: large-cap equity, long duration fixed-income, core fixed-income and full discretion core fixed-income. Manager alphas are calculated according to the respective benchmarks shown in Exhibit 1. Annual holding period returns for these are analyzed from 1989 to the present.
For each manager, we calculate average alpha and tracking error (standard deviation of alpha), and from these, information ratios,3 as well as the correlations of alphas with their respective benchmarks. Exhibit 2 shows the median level and low/high range for each of these measures for each manager universe over a range of sample periods. Also displayed is the percentage of managers within each universe and sample period with statistically significant information ratios (statistically significant alpha at 5% critical levels). Finally, to provide some perspective on the relative strength and weakness of stock and bond betas in the various sample periods, the last line of the table shows compound average return premia for large-cap equities over the Barclays Capital U.S. Long Government/Credit Bond Index for each sample.
While a large number of managers are included in the Investworks database, only a few report results all the way back to 1989, so in order both to take advantage of long sample periods and to maximize manager counts, we analyze results over a number of different sample periods. A year ago, we analyzed the sample periods 1989-2008, 1994-2008 and 1999-2008 to span 20, 15 and 10 years. To show comparable length sample periods that didn’t include the financial crisis, we also analyzed results over the 15- and 10-year periods 1992-2006 and 1997-2006. With the inclusion of 2009 results, each of the former three sample periods is now a year longer, and so, to complete the parallel, we extend the latter two samples to 1991-2006 and 1996-2006. Also, to further gauge the effects of the financial crisis, the 1989-2006, 1994-2006 and 1999-2006 samples are analyzed with the same manager sets as for the former three samples. Moving down a column shows how different manager universes compare within a particular sample, while moving across a row set shows the same manager universe over various sample periods.
In the results of a year ago, sample periods ending in 2006 generally showed information ratios for active equity below those for active fixed-income (the exception being 1999-2006), while for sample periods ending in 2008, active equity information ratios were a bit higher than those for active fixed-income. As seen in Exhibit 2, sample periods ending in 2006 again show active equity information ratios generally below those for active fixed-income (once again with the exception of 1999-2006). However, for sample periods ending in 2009, active equity information ratios are now generally about the same as or lower than those for active fixed-income. Adding in 2009 results substantially improves the performance of active fixed-income relative to active equity, whereas adding 1991 and 1996 results to the (previously) 15- and 10-year sample periods ending in 2006 does not reduce the edge of active fixed-income over active equity.
With 2009 results added, median manager alphas are now substantially positive for all active fixed-income universes, whereas the results through 2008 showed median average alphas either very small or negative for core and full discretion core universes. Active fixed-income manager tracking errors increase further when 2009 experience is included. Fixed-income managers’ sharply positive alpha in 2009 contributed to higher tracking error just as did sharply negative alpha in 2008. In contrast, including 2007-09 results had little impact on active equity tracking error.
We have pointed out elsewhere that while stocks’ total return suffered a larger loss in 2008 than did fixed-income sectors, it was still the case that on a relative basis, the 2007-08 crisis hit corporate bond and mortgage markets much harder than stocks.4 This is reflected in the fact that active fixed-income performance has been affected by the 2007-09 financial crisis much more than has active equity.
The same conclusion is evident in Exhibits 3 through 6. These charts show median manager alpha over time (with manager sets varying from year to year) versus total returns in the corresponding benchmarks. For the active equity universe data in Exhibit 3, there appears to be nothing unusual about the recent behavior of manager alpha compared to previous experience. Active equity managers did poorly during the bull market of the late 1990s and very well during the bear market of 2000-02, and their fortunes also moved counter to the overall market through the last two years, with recent swings little different—milder, if anything—than before.
In contrast, all active fixed-income universes showed under- and over-performance in 2008 and 2009, respectively, that was orders of magnitude larger than anything previously experienced, even relative to previous credit crises in 1989-90, 1998-99 and 2002-03. The only apparent difference among active fixed-income universes is that for core and long duration managers, the 2009 overperformance was a good deal larger than the 2008 underperformance, while for full discretion core, they were about the same size.
Simply put, in none of this evidence is there any indication that active equity dominates—or is even merely superior to—active fixed-income. Even including the 2007–09 experience that was so traumatic for active bond managers, active fixed-income performance is generally on par with that of active equity, and upon abstracting from the recent financial crisis experience, active fixed-income performance generally looks superior to that of active equity.
Perhaps the most striking contrast between active equity and active fixed-income performance is the tendency for active stock managers to consistently underperform in bull markets and outperform in bear markets (whereas active fixed-income performance showed no correlation with overall market trends until the extremely severe crisis struck in 2008). This is why the apparent superiority of active equity in the 1999–2009 sample period is of dubious meaning, given the severe bear market for stocks in place then. (Witness the sharply negative equity premium for this period.)
The contrarian performance of active equity would seem to recommend it as a defensive tactic, a means of reducing downside risks in the event of adverse market movements. However, this is hardly what commentators appear to have in mind when they recommend a passive bonds/active stocks approach. Furthermore, even if active equity were used as a defensive tactic, there is nothing in that approach that precludes active fixed-income, unless one expects another once-in-a century financial crisis. Meanwhile, there is nothing in the general results that supports an active stocks and passive bonds approach.
An investor seeking active returns should diversify active allocations across all sectors in which there is an expectation of positive alpha. Empirical results would have to show thorough dominance of active equity information ratios over those of active fixed-income to justify the assertion that there are no expectations of positive alpha in the fixed-income universe (but there are expectations of positive alpha in stocks). The results analyzed here show nothing close to such thorough dominance. If anything, the results for periods excluding the recent financial crisis show active fixed-income dominating active equity. With 2007-09 included, the active management results are about equal between stocks and bonds.
Nowhere have we stated that positive alpha is easy to attain. There is no question that it is tough to beat an efficient or nearly efficient market. Nevertheless, active managers do show results over time that are supportive of an active management approach. Our analysis covered all the Rogerscaey Investworks database, without exclusion. The only parsing of results we applied was the sorting of managers into appropriate universes as per their self-description, so that appropriate benchmarks could be applied. Given those appropriate screens, the results failed to indicate a preference for active equity over active fixed-income over a range of relevant sample periods.
- Active Equity Passive Bonds?, May 2009.
- These sample statistics were taken with respect to compound—or logarithmic—returns, with information ratios also the ratios of these.
- See "Is It Time To Rebalance?" April 2009, as well as the 2009 Quarterly Notes of Steve Walsh and Ken Leech. The “relative” referred to in the text concerns historical risk/return relationships and performance in previous periods of financial stress.