Practical analysis for investment professionals
23 July 2014

Evaluating Private Equity Performance: PME vs. Direct Alpha

In a recent paper titled “Benchmarking Private Equity: The Direct Alpha Method,” authors Oleg Gredil, Barry E. Griffiths, and Rüdiger Stucke propose a new technique called “direct alpha” to overcome key issues with some typically used private equity benchmarking methodologies.

The standard approach to evaluating an investment’s performance is to compare it with the performance of a related reference index. Generally, an investment’s return is thought to comprise two components: an alpha and a beta. Beta represents the return obtainable from passive benchmark investing, whereas alpha is the return attributed to nonmarket sources.

Alpha is considered to represent the manager’s skill and is used as a key input for making asset allocation and investment decisions.

The statistical method to calculate alpha is to use standard regression techniques from the modern portfolio theory toolkit. However, these techniques work well (at least from a statistical perspective) only for those instruments that are publicly traded and are highly liquid. This is a major problem for private equity (PE) investments as they are not only “private” and illiquid but also exhibit serious smoothing issues because of subjective appraisals and valuation lags.

In recent years, however, a number of heuristic (not rigorous mathematically, more of an empirical approach) methodologies have been employed to estimate the alpha of PE investments. These methodologies, commonly known as public market equivalent (PME), are heuristic in nature as they try to deconstruct alpha indirectly by comparing it with the return of a related public market benchmark.

Basically, these methods try to evaluate the value of a private equity investment by assessing its opportunity cost vis-à-vis to investing in other available vehicles.

PME Methodologies

The common basis of PME methodologies is to calculate an alternate internal rate of return (IRR) by applying the investment cash flows of the private equity investment to a reference benchmark. Alpha is then heuristically determined as the difference of the actual private equity IRR and the alternate PME IRR. Commonly used PME approaches are:

  • ICM/PME,
  • PME+, and
  • mPME.

The Long–Nickels index comparison method (ICM/PME), proposed in 1996, is considered to be the first of the PME approaches. In this approach, every capital contribution and distribution of the private equity investment is matched by an equal and timely investment and sale of the reference benchmark, respectively. This results in an identical set of contributions and distributions but a different residual net asset value (NAV). The resulting PME IRR provides a basis for appraisal against the investment’s actual IRR.

Although the ICM/PME is easy to understand, one key issue is that it does not liquidate as the private equity investment does. A strong outperformance or underperformance results in the reference portfolio carrying a large long or short position in later years. Also, as the investment approaches maturity, large swings in the benchmark may have no effect on the unrealized investments, while it has a big effect on the residual value of the reference portfolio.

The PME+ method, introduced by Christophe Rouvinez and Capital Dynamics, tries to rectify the previous issues by applying a fixed scaling factor to distributions to arrive at the same residual value. One major drawback of this method is the effect of the fixed scaling factor on the calculated IRR, as IRR is very sensitive to early distributions. A down-scaling or up-scaling of distributions due to outperformance or underperformance has an inflating effect on the calculated IRR. Also, PME+ is not an investable methodology and thus cannot be replicated.

The mPME method, developed by Cambridge Associates, is similar to the PME+ technique but uses a time-varying scaling factor based on interim private equity NAVs instead of a fixed one. This, however, does not remedy the effect of scaling on the calculated IRR as noted previously. In fact, rescaling distributions based on interim private equity NAVs introduces additional biases in case of any pricing issues with the interim NAV calculations.

The Direct Alpha Method

The main attraction of the direct alpha method is that it actually formalizes the calculation of the exact alpha to a chosen reference benchmark, avoiding most of the issues of the heuristic methodologies discussed above.

In this method, all private equity cash flows are compounded by the returns of the reference benchmark to the same single point in time, which when combined with the final NAV, forms a series of future values of net cash flows. By doing this, the impact of any changes in the reference benchmark on the actual private equity cash flows is effectively neutralized. Thus, the resulting net cash flows are not affected by any changes in the reference index but reflect only the sole private equity returns relative to the index returns.

The direct alpha method is closely related to another PME measure introduced by Steven N. Kaplan and Antoinette Schoar, referred to as KS-PME, which seeks to measure the wealth multiple effect of investing in the private equity investment versus the reference benchmark. KS-PME above (below) a level of one indicates that the private equity investment generated higher (lower) returns relative to the reference benchmark. Direct alpha can be thought of as annualizing the KS-PME and is zero whenever the KS-PME is one.

Empirical analysis of the different methods shows that direct alpha compares favorably versus the other heuristic methods, which have several cases that result in significant errors.


ICM/PME Approach (Numerical Examples from the Study)

ICM/PME Approach


PME+ Approach

PME+ Approach


mPME Approach

mPME Approach


The Direct Alpha Approach

The Direct Alpha Approach


Conclusion: PME or Direct Alpha? A Question of Perspective

After this discussion, is it straightforward enough to conclude that direct alpha is better than and preferable to PME? Based on my experience, it seems that direct alpha is more robust than PME analysis. IRR computation using PME is sometimes very sensitive to even relatively minor changes in cash flows or index values. Also, direct alpha is the closest thing we have in terms of applying modern portfolio theory to private equity.

Finally, one very important question to consider is, Does the reference benchmark accurately represent the opportunity cost of the fund investment? This is to ensure that any analysis is grounded in reality and does not end up just becoming a futile quantitative exercise.

If you enjoyed this post, you should subscribe to The Enterprising Investor

If you liked this post, don’t forget to subscribe to the Enterprising Investor.


All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.

Photo credit: ©iStockphoto.com/retrorocket

About the Author(s)
Prasad Ramani, CFA

Prasad Ramani, CFA, is the founder and CEO of Syntoniq, a behavioral tech company that seeks to transform the financial services practice by productizing cutting-edge behavioral finance research into easily usable tech applications. Ramani launched Syntoniq in 2017 to address inconsistencies in traditional financial service models following 18-plus years of experience in financial services, behavioral finance, and quantitative modeling. Ramani holds an MS in quantitative and computational finance (QCF) from the Georgia Institute of Technology. He is also a regular guest speaker at the London Business School where he teaches behavioral Finance and decision science.

4 thoughts on “Evaluating Private Equity Performance: PME vs. Direct Alpha”

  1. Ross Kasarda says:

    Prasad,
    Nice write up. The problem with PME and this new variant – they all assume that the beta of the private equity fund to the reference index is 1. This is most likely not the case. Venture has been shown by many researchers to to have a beta closer to 2, for example. Private equity also can include many other factor loadings, such as small cap and illiquidity. Without estimating these betas, this isn’t really a measure of alpha, instead it only measures excess return.

    There is another paper by Ang, Chen, Goetzmann, and Phalippou from 2014 available on SSRN titled “Estimating Private Equity Returns form Limited Partner Cash Flows” that provides a more robust way to estimate the alpha and betas simultaneously, thus removing the beta = 1 assumption of PME. The betas they estimate include the market, size, value, and illiquidity. I would recommend your readers take a look.

    All the best,
    Ross

    1. Oleg Gredil says:

      Dear Ross,

      Thank you for your comment. Note however that KS-PME and, thus, Direct Alpha do not assume beta of one. We discuss this in “Appendix C: Robustness” of the paper that Prasad referenced. For further details, consider Sørensen and Jagannathan (2013) and Korteweg and Nagel (2013) [both available through ssrn].

      Note also that with methods like in Ang et al., one cannot obtain fund-specific estimates of alpha or beta. It will have to be an average across a large group of funds.

      Still, the estimation error on Direct Alpha / PME might be smaller if one scales SPX returns by beta estimates from Ang et al. or elsewhere (e.g. industry-matched public equity). So the bests of two worlds can be combined in this case.

      Regards,
      -Oleg

  2. Thanks for the information! my brother is currently in school learning about how private equity performance works. It’s really interesting how an investment’s return relies on the return that can be collected by a company from passive benchmark investing and from other nonmarket sources. I thought that the information about different methods for estimating the alpha of PE investments was really interesting. I didn’t know that the public market equivalent can be used to compare the alpha with the return of a related public market benchmark in order to indirectly deconstruct the alpha.

  3. sumit says:

    Hi,

    thanks for sharing this informative article. I am confused in Mpme method how did you calculated the Dm PME and NAV MPME. need more information on this.

Leave a Reply

Your email address will not be published. Required fields are marked *



By continuing to use the site, you agree to the use of cookies. more information

The cookie settings on this website are set to "allow cookies" to give you the best browsing experience possible. If you continue to use this website without changing your cookie settings or you click "Accept" below then you are consenting to this.

Close