With The Right Benchmark, Anyone Can Outperform
By Ted Barac :
A rising tide can lift all ships while the converse can also be true. As such, relative performance and benchmarking is an important tool for judging the performance of any investment manager. The problem with benchmarking, however, is that it also has limitations and can often be misleading. I believe a good example of this relates to the benchmarking of the University of Texas' Investment Management Company ("UTIMCO"), $17 billion fund (the "Permanent University Fund" or the "PUF").
UTIMCO invests heavily in hedge funds and private equity investments - paying high fees for outside managers to invest their capital. Given the very high fees of hedge funds and the poor performance of the hedge fund industry over recent years, you would imagine that it's been a tough several years for the relative performance of UTIMCO. And it has been.
When comparing the PUF's prior eight-year performance to a traditional 60%/40% (stock/bond) multi-asset benchmark, the fund has cumulatively underperformed by 24% (2.2% on an annualized basis). Interestingly, the PUF underperformed this benchmark in bear (2008), bull (2012, 2013, and 2014), and broadly flat market years (2011 and 2015). In fact, the fund only outperformed in two of the prior eight years (2009 and 2010; both bull market years).
While the PUF's recent performance doesn't appear to be impressive, UTIMCO's performance reports tell a very different story. UTIMCO reports that it has consistently "added value" versus the PUF's benchmark in EVERY one of the last eight calendar years. How can this be and what is the cause of the apparent disconnect in how it perceives its relative performance?
Well, it appears that UTIMCO basically ignores the impact of its asset-class allocation decisions (which appear to have been the basis for their relatively poor performance over the past eight years) and measures their performance based on how investments do within their respective asset-class categories. In other words, the tactical decision to direct a high percentage of assets towards expensive hedge funds (and all other asset-class allocation decisions) are basically ignored for the purpose of determining value-added.
This seems odd as UTIMCO defines "value added" as the "result of the active management decisions made by UTIMCO staff and external managers". Shouldn't the allocation decision to go heavily into hedge funds and private equity, as targeted through its benchmark, be considered a decision made by UTIMCO's staff and external managers?
To me, it seems that ignoring the performance impact of asset-class allocations only makes sense for specific and static fund strategies where asset allocation decisions aren't really part of the investment process. Examples of this would include a focused large-cap U.S. equity fund or an investment-grade corporate bond fund.
For a manager like UTIMCO, however, it seems that its detailed and non-traditional asset allocation decisions (which are periodically reviewed and adjusted) are meant to optimize performance. As such, it doesn't make much sense to ignore the performance impact of these allocation decisions, which appear to have had a substantial (negative) impact on the returns of the PUF over the past eight years.
UTIMCO's benchmarking seems analogous to a fund manager tactically deciding to go heavily into energy stocks, right before an energy crash, and then claiming outperformance after their picks lose only 20% while the broader energy index loses 25%. To be clear, that kind of sector outperformance data is informative and useful, but it shouldn't be the only comparison to be considered. The fund manager's tactical decision to go heavily into energy stocks should also be evaluated and considered, as should UTIMCO's decision to go so heavily into hedge funds and other alternative investments.
Undoubtedly, UTIMCO's high hedge fund exposure is meant to cushion losses in down periods and reduce the fund's correlation with the stock market. Notwithstanding the fact that there are many lower-fee alternatives for obtaining such diversification (most obviously, a healthy allocation to high-grade bonds and/or cash), this is a fair point to take into account. Of additional consideration is the fact that the market has been going up for seven of the past eight years and the hedge fund industry claims to add most value in down years.
With all that in mind, let's look in more detail at the PUF's performance during the last stock market crash in 2008. For the calendar year ending December 31, 2008, the PUF lost 27% while the 60/40 benchmark substantially outperformed and lost only 22%. Interestingly, the only two (of the past eight) years in which the PUF actually did outperform the 60/40 benchmark were both bull market years (2009 and 2010).
While I don't have the data available to look back further to other bear market periods, this example does help support my belief that the PUF's significant long-term underperformance versus a traditional stock/bond benchmark mix is difficult to justify. This is particularly true when considering that the 60/40 mix is quite conservative and comprised of lower volatility (large capitalization domestic) equities and high-quality U.S. bonds. Furthermore, it should be noted that both asset classes in the 60/40 benchmark are very well diversified, and the latter typically has a negative correlation with the stock market.
Ignoring asset allocation considerations, is the PUF really consistently "adding value" versus its benchmark?
Ignoring the poorly performing asset allocation decisions, beating the PUF's benchmark for eight consecutive years still has to be considered impressive, right? Well, not necessarily.
The PUF's benchmark target for fiscal year 2016 show a 29% allocated to hedge "fund-of-funds" and another 31% to other fund-of-funds (which appear to be related to other "private investments"). Prior year benchmarks also show a heavy weighting towards fund-of-funds and this could explain a lot.
Fund-of-funds incur not only the high fees associated with the underlying funds (for hedge funds, often 2% of capital and 20% of profits), but they also incur an additional layer of fees at the top. These are the additional fees paid to the fund-of-funds, themselves, for managing/selecting the underlying hedge funds. With such an expensive benchmark component - and with such poor hedge fund performance over recent years - one can see why "adding value" versus the PUF benchmark was much easier than it would have been versus a more traditional (broad market equity/bond) benchmark.
In fact, it seems that UTIMCO's "added value" over time is almost guaranteed - at least in the fund-of-fund sub-sectors which now comprise 60% of the PUF's benchmark. This is partly because the PUF's hedge fund exposure is so diversified. As of fiscal year-ending August 31, 2015, the PUF reported owning 69 different hedge funds of varying strategies worth over $5.2bn. With such broad diversification, it seems likely that the underlying performance of its hedge funds should closely resemble that of the overall hedge-fund industry over time.
Given similar underlying fund performance, and with the very top layer of fees/expenses being lower for UTIMCO (as a non-profit) versus other fund-of-funds, outperformance in this major asset class seems to be very likely over time. This may help explain why it claims to have "added value" versus the PUF benchmark in every one of the past eight years despite substantially underperforming a simple and conservative 60/40 allocation strategy over that same time frame.
But, why hasn't anyone noticed?
None of this underperformance seems to have received much attention despite the billions of dollars involved. Why hasn't anyone seemed to notice the substantial underperformance of the PUF versus the broad equity/bond markets and more traditional benchmarks?
I believe that one of the reasons is because its investment allocations and benchmarking methodologies are so detailed and complex, that it's easier to simply take UTIMCO's "value added" reports at face value. Furthermore, while there is considerable reporting and general public knowledge about the calendar-year performance for the broad stock and bond indexes, comparisons against PUF's fiscal year-ending August 31st results are not going to be nearly as obvious.
To be clear, I am not judging UTIMCO's performance in selecting individual hedge funds to be either good or bad. In fact, I attribute its underperformance to the headwinds resulting from the high fees that it is charged (and which characterizes the overall hedge fund industry) and not to its individual hedge fund selections.
While there are individual exceptions, I believe that the high fees of the hedge fund industry are value destructive (except for the hedge funds themselves) and can also misalign investor/manager interests. I also believe that few hedge funds justify their high fees, and when one diversifies across many different funds, alpha generation is very unlikely and underperformance over time should be expected (fellow contributor SL Advisors wrote an interesting article on this phenomenon). This is a similar argument to that behind Warren Buffett's (BRK.B) famous bet, from eight years ago, when he argued that a broad market index fund would beat a hedge fund-of-fund over the long-term (so far, Buffet is way ahead on that bet).
I am also not saying that it isn't possible for UTIMCO's hedge fund and other investments to ultimately deliver long-term value and outperformance versus the broad markets and more traditional benchmarks (although I don't expect this to be the case, because of their high-fee headwinds). What I do take issue with, however, is claiming "value added" in these years when the PUF is clearly not doing so and for reasons that seem to have more to do with the structure of the benchmark rather than management's investment decisions. This is particularly true given that taxpayer-funded bonuses for UTIMCO employees are tied to these questionable "value-added" determinations.
While UTIMCO's PUF is just one example of potentially flawed benchmarking, the problem is obviously more widespread and can happen amongst institutional and retail investors alike. In fact, it's a problem that's even more prevalent amongst retail investors (who don't have the regulatory and investor scrutiny of institutional funds).
For example, retail investors often measure the performance of their brokerage accounts (including dividends) versus a broad market index like the S&P 500 (which typically doesn't include dividends). Unless these investors underperform the market by more than the dividend yield of the index (which is over 2% these days), they are able to claim outperformance - whether or not they actually did so when using an apples-to-apples comparison.
As appears to be the case with UTIMCO, the benchmark of these retail investors is structured such that they have a constant tailwind that is supporting their relative performance. The lesson here is to always be careful as to how someone benchmarks themselves. While performance comparisons can be subjective, and there's no perfect benchmark for any fund, some are certainly better than others.
See also Oil Price Bottom: Analysts Are Looking In The Wrong Place on seekingalpha.com