The question of private equity performance and its added value is becoming an important topic for plans, as traditional portfolios are increasingly augmented with alternative assets. For private markets, the conventional methodology for performance reporting is based on since-inception internal rate of return (IRR) and net multiples. Furthermore, private equity professionals often argue that investors need to wait until most of the capital is distributed to judge performance. However, waiting eight or ten years before passing any judgement on private programs is often not practicable. Given the increasing importance of private asset for plan sponsors, performance and value added analysis has to be more timely and relevant. Board members and investment managers have to make regular decisions about the allocation to private asset classes, sub-asset classes and specific managers, while, at the same time, plans need to regularly communicate results to boards and other stakeholders.
Plan sponsors need to ask the following questions:
- Is the private program adding value to the plan?
- Is private portfolio construction and allocation to particular sub-strategies or geographies added value?
- And, which funds within each sub-asset class are the outperformers and which are the underperformers?
Simple peer ranking may, at best, answer one question: were above average managers selected? I say ‘at best’ as most private equity professionals appreciate, by now, the concerns and limitations inherent in the commercial peer universes, concerns such as survivorship bias, lack of transparency on data integrity, delays in reporting and often small sample sizes. Furthermore, simply categorizing the vintage year differently from the database (which is not very difficult as there is no accepted definition) will create significant distortions to peer ranking. Concern over small sample size is especially acute when trying to benchmark non-US assets, where sample sizes are growing but are still too small to draw any significant conclusions on peer performance.
Comparing a fund’s since-inception performance against a sample peer universe will always have an important role in measuring a fund’s performance against its peer group. However, supplanting a peer analysis with a public market equivalent (PME) analysis significantly improves the performance analysis, enhancing it by offering a value-added measurement. There are two commonly accepted methodologies: Long & Nickels’ index comparison method[1] and Rouvinez’s PME+ methodology[2]. Both methodologies show similar results for most funds. Long & Nickels’ method, however, produces some unusual results that could not be explained by the performance of the public markets’ index during any part of the investment period. At the same time, the pro-rata distribution assumption created by the residual resulting from the ‘distribution scaling factor’ in Rouvinez’s methodology seemed to smooth out results. Nonetheless, the benefits of Rouvinez’s methodology far outweighed the limitations. (Click on image to enlarge).
The main appeal when comparing a private equity fund’s performance to a public index is that it enables plans to answer the question of whether the program, any sub-strategy, vintage or fund added value to a mixed portfolio of stocks and bonds, while taking into account the risk parity of the asset classes. Put another way: even if a portfolio contains above-average funds, do these funds add value to the plan? To answer the above question, one must compare performance of a private equity fund to an investable and transparent index. Clearly, the peer universes do not meet these qualifications.
What one realizes when benchmarking performance using a PME analysis is that some asset classes or vintages, even if they end up being well above peer average performing strategies, may fail to add value when compared to the returns of risk-equivalent public index. The failure on the part of certain periods or vintages to add value may arise from a number of reasons, such as the excess of private capital in an asset class compared to the opportunity set, high cost of leverage, lack of technology innovation, lack of good exit opportunities, etc. The high fee structure of private funds creates a high hurdle for outperformance when compared to cheaper forms of equity investing such as passive indexing. This is one of the reasons why, in the absence of a strong view on vintage year allocation, equally weighing the vintage year diversification makes sense for private equity investors. It will ensure that if plans are skilled at picking private equity funds, that skill will ‘pay off’ over time.
One asset class that falls glaringly short is venture capital. Recent venture funds that are ranked as top quartile will likely fail to add value when compared to the NASDAQ 100 or a small cap technology sub-index (this is without even taking into account the cost of liquidity or incremental early stage risk). In certain vintages, investors would need to be in top decile funds to guarantee outperformance against equivalent public indices. Furthermore, some energy funds may perform poorly against the limited peer universe; however, they will nonetheless add value compared to an energy sub-index.
The public markets equivalent analysis also helps to compensate for the small peer sample sizes of regional funds, sector focussed funds and other sub-asset classes. Comparing these funds to the broader universe may create the impression of significant over or under performance. However, no public investments manager would contemplate comparing a growth manager to a value index, oil and gas manager to the S&P 500 index or an emerging markets manager to an MSCI-World Index.
Beyond the J-curve
One of the arguments against using a PME analysis is the outsized weight of the unrealized value of a private equity fund in the middle of the fund’s life and the lack of certainty surrounding its value. The unrealized value is certainly conditional on subjective valuation methodologies, especially when a discounted cash flow methodology is used. Mitigating this concern is the ever-increasing rigor of fair market valuations, which gives the PME analysis more credibility. Furthermore, one always has the ability to reduce the unrealized value of a fund by a discount. Another challenge with PME analysis is benchmarking manager’s performance during the early years when funds exhibit negative performance due to their heavier relative fee drag compared to a public index, which is not subject to a J-curve. Comparing a fund’s performance to a public index in a fund’s early years is not easy using any benchmarking tool. It is certainly more difficult to get around the J-curve in benchmarking using PME, and as a result, one should not put much weight on it.
Another issue that requires some consideration is whether a PME analysis correctly takes into account a private equity managers’ skill in timing acquisitions and dispositions. For example, in a case where two private equity managers were formed in the same vintage period, the ‘unskilled’ manager did not anticipate the downturn and invested capital early during an over-priced period, while the ‘skilled’ manager anticipated the downturn and invested the capital during a lower priced period. The peer analysis, which is based on funds’ vintage year, will show the ‘skilled’ fund as outperforming, while PME analysis may not. In fact a PME analysis may even show that the ‘skilled’ manager underperformed the ‘unskilled’ manager when that relative performance is measured against to public markets, given that ‘skilled’ manager earned management fees during the early years when no capital was deployed and, furthermore, incentive fees would likely be higher. Those fees may be large drag on relative performance. The concern that PME analysis does not reflect market timing is dismissed when one considers that the opportunity cost of a timing bet is not absolute performance, but rather the performance relative to the risk equivalent assets: passive public markets strategies. Put another way, when one funds a capital call for a private fund, the capital is typically drawn by selling a public holdings, which is precisely what the PME analysis captures. In fact, this highlights one of the main weaknesses of peer analysis: grouping funds by vintage year glosses major differences in the timing of capital deployment and return of capital. Furthermore, it also highlights the limited weight one should put on the skill of fund managers to time acquisitions and dispositions. These mainly impact the absolute returns and fund managers` carry rather than alpha.
Despite the strong support for PME analysis, private equity universes will have an important role in performance analysis. Peer rankings are still helpful in answering questions such as whether funds are above or below average when compared to their peers, and whether certain investment professionals are skilled at picking the better funds given the opportunity set. In future years, as sample sizes increase and classifications improve, the value of peer ranking will also improve, especially as it relates to various niche private equity strategies.
Benjamin Abramov is Head, Canadian private markets research at Aon Hewitt.
Endnotes
[1] Long, Austin M. and Nickles, Craig J., A Private Investment Benchmark, AIMR Conference on Venture Capital Investing, 1996
[2] Rouvinez, Christophe, Private Equity Benchmarking with PME+, Private Equity International, August 2003