Lecture 2: Performance Measurement in a Nutshell Jesse Reyes

What are we measuring and why is it so difficult Overview What are we measuring and why is it so difficult What/how do we benchmark Industry performance

Date Return 12/31/11 12/31/12 10% 12/31/13 20% 12/31/14 -30% 12/31/15 QUIZ TIME Date Return 12/31/11 12/31/12 10% 12/31/13 20% 12/31/14 -30% 12/31/15 50% WHAT IS THE AVG ANNUAL RETURN?

Date Return 12/31/11 12/31/12 -30% 12/31/13 -50% 12/31/14 20% 12/31/15 QUIZ TIME Date Return 12/31/11 12/31/12 -30% 12/31/13 -50% 12/31/14 20% 12/31/15 10% WHAT IS THE AVG ANNUAL RETURN?

QUIZ TIME Date Index 12/31/11 1000 12/31/12 1100 12/31/13 1320 12/31/14 924 12/31/15 1386 WHAT IS THE ANNUAL RETURN From 12/31/11 to 12/31/15?

QUIZ TIME Date Index 12/31/11 1000 12/31/12 700 12/31/13 1050 12/31/14 1260 12/31/15 1386 WHAT IS THE ANNUAL RETURN From 12/31/11 to 12/31/15?

“Our/Some Fund Has a Return of 300 Percent” Is that good? Depends? Typical Question “Our/Some Fund Has a Return of 300 Percent” Is that good? Depends?

Time Weighted/IRR/Realization/Horizon? Depends on What? Over What Time period Over Two Years -- great at 173% per year Over Ten? --- hmmm!! At 11.61% Is it return on the investments the fund made or is it the return to the investors in the fund. Time Weighted/IRR/Realization/Horizon?

Example 1 Invest $1.5 million Worth $5 million What is the return?

5/1.5=333% Total Return or percentage change of 233% What is the Return? Invest $1.5 million Worth $5 million What is the return? 5/1.5=333% Total Return or percentage change of 233% Note – we haven’t defined a time period

Worth $5 million five years later What is the return? Example 2 Invest $1.5 million Worth $5 million five years later What is the return?

Example 2 What is the Return? Invest $1.5 million Worth $5 million five years later What is the return? 5/1.5 = 333% Total Return Lets annualize with formula (total return)^(1/5) -1 = 27.2% This is the annual rate of return (compound annual growth rate, Geometric average return, Time weighted return, etc)

Worth $5 million at the end of year five Example 3? Invest .5 in three years .5 at the beginning of year 1 .5 at the beginning of year 2 .5 at the beginning of year 3 Worth $5 million at the end of year five

Why The difference? First two examples are just examining two points in time. i.e .today and some point in the past. No transactions or cashflows occurred in the in between. The third example, complicated the calculations because we didn’t invest the $1.5 all at once. Notice that stock indices use the first approach-you only calculate returns at two points in time – assumes that you buy and hold. The third has shortened the average investment holding period so the time value of money says that the return is higher other things being equal.

Why The difference? Gets even more complicated if we measure returns for an investment where we also take money out over a period of time as well as put money in over a period of time.

Why The difference? Same with your savings account – if you put money in at one point and take it out at another point, you can calculate the return for that period of time which will be the APR, but if you put money in, take money out over a period time, the actual return over the life of your investment getting is not the APR but something different that depends on the average time you held the money. The APR assumes you put money in a lump and hold it

Why The difference? You have the same complications if you measure the return to an investment manager/mutual fund in public stocks --money goes in at different points in time, and is taken out at different points in time by the investor. – more on this later. Don’t have this complication with a simple stock index – you are just measuring values at two points in time – no transactions in the middle, so you can use time weighted/total return calculations

So what do we do? So with investments either in private equity or any investment manager, if you have cashflows in and out of an investment simple percentage change/total return calculations can no longer be done to get the true Return on investment. So we turn to IRR, a form of ROI that takes the time value of money into account as it accounts for the timing of the transactions in the investment.

Example 3 What is the Return? Invest .5 in three years .5 at the beginning of year 1 .5 at the beginning of year 2 .5 at the beginning of year 3 Worth $5 million five years later What is the return? 5/1.5 = 333% Total Return IRR = 46.97% annual IRR Compare to return on example 2 of 27.2% annualized

Distributions – cash or stock returned to LP investors. Definitions Takedown – actual money paid into partnerships, aka capital calls, paid in capital Distributions – cash or stock returned to LP investors. NAV – aka residual value, ending value of the fund for the period being measured – net of fees and carry

Definitions continued Vintage Year -- year fund started investing IRR -- Net to investor after fees and carry DPI -- cumulative distribution/cumulative Paid in capital – cash in – cash out (CICO) RVPI -- net asset value / cumulative paid in capital TVPI -- DPI + RVPI PICC-- cumulative paid in capital to committed capital (dry powder ratio)

Cumulative Realization Multiples (Cash-in/Cash-out) Principal metrics Cumulative IRR Cumulative Realization Multiples (Cash-in/Cash-out) Time Weighted Return Public Market Comparables – Index method

Realized Investment Multiples Other valid measures Realized Investment Multiples Cash on Cash (CICO) (cash in cash out) Ignore NAV – measure ratio of cumulative distributions since inception to cumulative paid in capital since inception (DPI) – measure of realized return Unrealized Investment multiple How much of the return is unrealized and still locked up in portfolio company investments – ratio of ending NAV (residual value) to cumulative paid in capital (RVPI) Total Investment Multiple – the addition of the the two (TVPI ) total value to paid in capital ratio

Intro to Fund accounting for performance (kindergarten version) Done from LP perspective Three inputs: Cash in, cash out, terminal net asset value Contributions (also known as capital calls or paid-in capital or takedowns) increase net asset value. In cashflow analysis these are treated as negative numbers (cash in) Distributions (either in cash or in stock) decrease net asset value. In cashflow analysis, these are treated as positive numbers (cash out) Net Asset Value-(net of fees and carried interest) is treated as a positive number Note: Management fees are either deducted from net asset value or are billed to the limited partner as another contribution. In either case they decrease net asset value. But timing of fee recognition is huge issue. Net asset value ultimately is determined by the underlying value of the assets/companies being held so it may increase/decrease depending on value of those investments.

Principal measure of Fund Returns Principal metric is IRR since inception Beginning Point is fixed, endpoint is variable Principal Composite is “Vintage Year” (designating at the year of the first capital call) IRR is calculated net to limited partner

Time Weighted Returns Time weighted return calculates a return for each period -- quarterly, annually. Beginning point is variable, endpoint is variable Calculate using net asset value at beginning and end of period and cashflows between periods. Calculate return for each period and then compound together

Time Weighted Returns - Shortfalls Creates aberrations: 100 + 20% = 120 120 - 20% = 96 Periodic valuations heavily dependent on valuations. Wrong valuations affect future returns. Assumes money can come and go freely at the beginning and end of each period

Kindergarten Financial Math FV=Future value, PV=Present Value, r rate of return, n=number of periods

Internal Rate of Return, IRR=internal rate of return, n=period ,of cashflow, N=total periods, CF= cashflow

Time weighted Return – Periodic returns Periodic Returns can be calculated Before modern math based tools Dietz mid-point weighted Modified Dietz Now IRR can be calculated readily but Dietze is still prevalent Returns are calculated for each sub period then they are linked and geometically averaged True time weighed calcuates a return every time there is a transaction Practical time weighted return calculates return on a periodic interval (daily/monthly/yearly) Private equity can practically only calculate returns quarterily as that’s as frequent as funds are valued

Before the IRR and even now Dietz Midpoint and Modified Dietz BMV=Beginning Market Value, EMV=Ending Marketin Value, CF=interim cashflow, W=percent of time cashflow was in portfolio

Why can’t we all be friends? Is the IRR all that Bad? Why can’t we all be friends?

Myths every good MBA knows about IRR Myth #1: IRR is inferior to NPV Myth #2:IRR can give you different results than NPV Myth #3: IRR can give you Indeterminate multiple solutions with cashflow sign changes Myth #4: IRR is flawed because of single re-investment rate assumption Myth #5: IRR is inferior to other techniques such as modified IRR.

Yes Virginia, There are Private Equity Benchmarks

What are you comparing to? Returns Performance

Returns Performance Benchmarks

Everything I learned about benchmarking I learned in Boy Scouts

So what is a benchmark? A benchmark is a point whose position is known to a high degree of accuracy and is normally marked in some way

Unambiguous – Securities clearly delineated Benchmark Criteria Unambiguous – Securities clearly delineated Investable – Option of passive or active mgt Measurable – Calculable on a frequent basis Appropriate – Consistent with managers style Opinion Reflective – Mgr has opinion of benchmark securities Specified in advance – Constructed before use Source: Benchmark Portfolios and the Manager/Plan Sponsor Relationship,”Jeffery Bailey, Thomas Richards and David Tierney

Sources of PE Benchmarks PE Fund/Portfolio PE Industry Public Markets Source; Alignment Capital

Sources of PE Benchmarks PE Fund/Portfolio PE Industry A Source; Alignment Capital

Sources of PE Benchmarks PE Fund/Portfolio B Public Markets Source; Alignment Capital

Sources of PE Benchmarks PE Industry C Public Markets Source; Alignment Capital

Sources of PE Benchmarks PE Fund/Portfolio PE Industry D Public Markets Source; Alignment Capital

Sources of PE Benchmarks PE Fund/Portfolio PE Industry D B C Public Markets Source; Alignment Capital

Question: How do we benchmark? Choosing Benchmarks Question: How do we benchmark? Answer: Depends on the Decision you are trying to evaluate?

Decisions, Decisions A benchmark should reward/penalize the decisions made under the control of the decision maker.

A naïve manager would invest 50% in A and 50% in B Naïve Manager Example I have a choice to invest in either investment A or investment B in any amount A naïve manager would invest 50% in A and 50% in B Any decision that I make should be compared to the above decision and should be superior if I’m to be rewarded for my decision making.

Decision/Benchmark Example I invest in vintage years 1999, 2006, 2008 Do I compare my returns to all funds in just those three years? Do I compare my returns to all funds between 1999 and 2008? Do I compare my returns to all funds between 1999 and present? These scenarios benchmark different decisions primarily based on opportunity cost

LP has control over which managers to pick Who has control? LP has control over which managers to pick LP has control over the amount / when to commit GP has control over capital calls, exits, distributions IRR rewards/penalizes timing TWR is timing agnostic

Decisions an LP controls The fund(s) you choose? – Manager Selection When you committed? - Timing The amount you invested? Pacing/Timing Secondary Sales

Do PE Benchmarks Fit Criteria A (PE-IND) B (PE-PUB) C (IND-PUB D (PE-IND-PUB) Unambiguous N Y Investable Measureable Appropriate Opinion Reflective Specified in Advance

Comparing public / private equity IRRs and TWR are incommensurable Reinvestment rate assumption IRR – timing matters, TWR – timing doesn’t Public Equity/Private equity have different valuation schema Public equity has a contemporaneous price Private equity is an appraisal asset with significantly lagged values So how does and investor compare public/private equity Enter the Public Market Equivalent

Public Market Equivalents (PME) Rather than calculate TWR for public equity, create a synthetic IRR for public equities First formulation independently developed by: Coller / Reyes (investor/researcher) in 1993 (Public Market Equivalent or PME) Long / Nickles(alignment capital) in 1996 (called ICM “index comparison method) Bannock/BVCA (industry association) in 1994 (called comparators) Idea is to invest the cashflows from private equity into an index and then create a synthetic cashflow from that investment and calculate and IRR which can be compared to the original IRR When it works, it works great, but does create some mathematical dominance stability issues under certain circumstances

Public Market Equivalent History Original PME developed independently by Reyes/Coller Long/Nickles, Bannock/BVCA/EVCA (1992-1996) Long and Nickels created an Adjusted ICM to deal with negative NAV (1998) Roubinez and Kubr (investment managers) patented a PME method called PME+ which tries to remediate the stability issue (2003) Kaplan and Schoar (academic researchers) created another method which isn’t a return but is a multiple (2005) Recently, Cambridge Associates created mPME to further address stability issues (2013) New “Direct Alpha” method developed by researchers at Landmark Partners tries to address some of the issues of the original PME while providing a direct comparison that doesn’t need an intermediate step (2014) Global Endowment Foundation independently created same formula as Direct Alpha (2014) Bison has created a version of PME as well. (2015)

Original PME/ICM/Comparators Developed by Reyes/Coller then Long/Nickles and then Bannock/EVCA A) Invest all capital calls into an index – compound as a positive value forward from date of capital call to the end date being measured B) Divest all distributions from the index – compound as negative value forward from date of distribution to end date being measured C) sum A and B to get a synthetic NAV D) create cashflow series using original fund cashflows with their natural signs Capital calls negative and distributions positive and use the synthetic NAV in C as the terminal cashflow E) calculate IRR on the cashflow in D F) This is the return to the public market if you invested cashflows in the public market – can be compared to fund IRR

Trouble with Original PME Under certain conditions: Incalculable Negative NAV when PE outperforms substantially Long/Nickles tried to address Capital dynamics created version ot PME which they patented which scales the cashflows to address the problem. Cambridge Associates mPME and Direct Alpha try to address the same

Scales the cashflows in a variety of methods Capital Dynamics PME+ Patented method developed by Capital Dynamics see patent number US7698196 Method and System for Benchmarking and Modeling Private Equity and Applications of Same Scales the cashflows in a variety of methods Takes concept of long/nickles by investing/divesting but compounds capital calls/distributions separately and then creates a ratio from those two series that is applied to the original cashflows so that the NAV never becomes negative – this is one application of their patent. Then takes the scaled cashflows and calculates an IRR that can be compared to original fund IRR

Kaplan & Schoar PME Formulated by Steven Kaplan (Univ chicago) and Antoinette Schoar (MIT) as a byproduct of a paper on private equity returns persistence A) Invest capital calls into an index by compounding them as positive values from the date of the capital call until the end date being measured B) Divest distributions into an index by compounding them as positive values from the date of the distribution until the end date being measured C)Sum up A and B separately D) add B to the original fund NAV and divide by A E) creates a scaleless ratio – greater than 1, PE dominates, less than 1, PE underperforms public market. F) can’t be compared to original fund but could be said to be the total return alpha to the fund over the life of the investment

Direct Alpha Created by Landmark Partners and academics working with Landmark Partners Uses concepts from original PME but the result is the direct alpha difference between the private equity fund and the public markets A) Invest all capital calls into an index – compound as a positive value forward from date of capital call to the end date being measured. Date this synthetic cashflow result on the date of the capital call B) Divest all distributions from the index – compound as negative value forward from date of distribution to end date being measured. Date this synthetic cahflow result on the date of the distribution D) create cashflow series using synthetic cashflows created in A and B and combine with original fund NAV. distributions positive and use the synthetic NAV in C as the terminal cashflow E) calculate IRR on the cashflow in D F) This is the difference between the fund cashflows and the public markets. The comparison to the original fund is done automatically.

mPME Created by Cambridge Associates Like PME+, scales cashflows so you never distribute more than you have A) calculate a distribution weight for each distribution based on current NAV B) compute a theoretical NAV using the the distribution weight D) create a distribution weighted cashflow by using the distribution weight and the public market index E) create a cashflow using the original capital calls and the weighted distributions F) calculate IRR on the cashflow in E G) Unlike the other methods, needs quarterly NAVs and calculates the weighted cashflows for each period rather than over the life of the fund. Cambridge Associates and Capital Dynamics were involved in a patent infringement lawsuite of the use of scaling distributions.

Bison PME Created by Bison – a PE benchmarks provider Makes the assumption that size of cashflows are as important as timing so tries to incorporate a version of the TVPI. Scales cashflows using the equivalent of the Kaplan & Schoar ratio. The original version is more complicated, but ultimately decomposes to this simplifying method A) calculate a K&S for the fund B) Scale the distributions by the K&S – ultimately this is a version of capital dynamics pME+ as lambda in PME+ is 1/K&S D) create a cashflow stream using original capital calls and scaled distributions E) calculate IRR on the cashflow in D

GEM Implied Performance Premium Created by Global Endowment management Thought experiment: what premium would I have to add to a public market return to equalize the FV of contributions to the FV of distributions and NAV. Equivalent to Direct Alpha and developed independently about the same time except. Direct Alpha assumes geometric difference. IPP assumes arithmetic difference. Direct Alpha has a closed form solution, IPP requires iterative estimation like an IRR

Later versions use scaling to minimize math problems Problem with PME Early version have mathematical stability problems under certain common conditions Later versions use scaling to minimize math problems But all suffer from: Comparison of public markets to private markets not the other way around Hard to put into investor presentations upstream to management as you can’t eaily compare to other asset classes Mimics equivalent decision making in public markets but that may be an artificial decision making framework. Ignores commitment and dry powder problem Benchmarks GP decisions not LP decisions

Benchmarks for LPs may be different than Benchmarks for GPs

Benchmarking LP Portfolios LPs make the following decisions How much to commit When to commit Who to commit to Cashflows afterwards are out of LPs control so benchmarks should reflect that lack of decision control.

Heresy: Maybe IRRs shouldn’t be used for LP portfolios

Heresy: Maybe we get rid of vintage year

Heresy: Maybe we benchmark returns on committed capital

How should we benchmark LPS? Maybe we should benchmark commitment opportunity cost Maybe calculate return on commitment in addition to return on paid-in Benchmark from date of commitment rather than date of cashflows? Use wealth creation on commitment compared to wealth creation of industry and public markets Shudder -- Maybe use time weighted returns on LP Portfolios

Things our Current Benchmarks Framework Ignore At Our Peril

Manager Selection Risk Wealth Creation Attribution Things We Ignore Risk Risk Risk Temporal Allocation Manager Selection Risk Wealth Creation Attribution Excess Returns Analysis Tracking Error