1. David Bernard San Francisco, 26 July 2005 Thomson Financial Venture Economics The Art and Science of Benchmarking in Today’s Fundraising Climate T H O M S O N F I N A N C I A L 12 th Annual Private Equity CFO Conference

2. T H O M S O N F I N A N C I A L 26 July 20052 Overview What are we measuring and why is it so difficult? What/how do we benchmark? What are the actual results for the industry?

3. T H O M S O N F I N A N C I A L 26 July 20053 Current Environment More transparency for the asset class vs. more disclosure (FOIA)? Increased transparency for fund raising, reporting, asset allocation and fund due diligence Confusion: What is the return being reported? How was it derived? How can you put it in context? Valuation guidelines (www.privateequityvaluation.com) being adopted internationallywww.privateequityvaluation.com

4. T H O M S O N F I N A N C I A L Part 1 What are we measuring and why is it so complicated?

5. T H O M S O N F I N A N C I A L 26 July 20055 Is a return of 200% good enough? A return of 200%? –200% total return: having invested $1m, we get $2m back –200% percentage change: we get $3m back (let’s assume this) Over what time period? –Over two years -- great at 73% per year (1.73^2=3) –Over ten? --- hmmm!! At 11.6% per year (1.116^10=3) 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?

6. T H O M S O N F I N A N C I A L 26 July 20056 Why an IRR? Why the difference with most stock indices? You can’t just look at the value at two points in time, i.e. today and some point in the past, with no transactions or cashflows in between – it would assume that you buy and hold You don’t invest the money all at once, and you also take money out over a period of time 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

7. T H O M S O N F I N A N C I A L Part 2 What/how do we benchmark?

8. T H O M S O N F I N A N C I A L 26 July 20058 Why a benchmark? Return is mathematical algorithm – it is an absolute measure Performance is a relative measure – can only be determined by comparing return to something else – for example past returns, benchmarks, etc. So you need a benchmark

9. T H O M S O N F I N A N C I A L 26 July 20059 Why a benchmark? the Naïve Investor example Investor has choice of 2 investments. Other things being equal, with no additional information, optimal allocation for naïve manager is 50-50 So any decision you make different than this should be better performance – so benchmark is performance of 50-50 allocation. You are benchmarking the decision of the allocation That’s why public indices is used so often in stock market benchmarks – it’s the naïve manager decision Any investment decision you make different than allocation to, say, S&P500 should be better if you are worth the fees you are being paid

10. T H O M S O N F I N A N C I A L 26 July 200510 Whose decision are you benchmarking? Several decisions to benchmark for the LP investor –The allocation to private equity –The allocation between private equity sub asset classes –The timing decision of when to invest –The decision of one manager over the other –(The investment decision of the fund)

11. T H O M S O N F I N A N C I A L 26 July 200511 Principal benchmarks Cumulative IRR Cumulative Realization Multiples Time Weighted Return Investment Horizon Return Public Market Comparables – Index method

12. T H O M S O N F I N A N C I A L 26 July 200512 Which benchmark to use – The LP Investor Benchmark the decisions under the control of the manager –If time is under control of the manager – timing of investments should be rewarded / penalized – so use IRR – either cumulative or investment horizon –If time is not under control of the manager – use Time Weighted Return that takes time out of the calculation of the benchmarks Several decisions to benchmark –The allocation to private equity – time weighted return –The allocation between private equity sub asset classes – time weighted return –The temporal (timing) decision of when to invest – investment horizon or cumulative IRR –Decision of one fund over the other – cumulative portfolio IRR* –The investment decisions of the fund – IRR/realization by vintage year * AIMR, GIPS, standard practice

13. T H O M S O N F I N A N C I A L 26 July 200513 Some definitions (1/2) Limited Partners Private Equity Firm (General Partners) Fund I Fund II Company 1 Company 2 Cash take-down Cash/stock distribution Takedown : actual money paid into partnerships, a.k.a. capital calls, paid in capital Distributions : cash or stock returned to LP investors NAV (net asset value*), a.k.a. residual value: ending value of the fund for the period being measured – net of carry Vintage Year: year fund started investing Pooled Return: portfolio return by pooling cashflows of several funds Management fees Carry * as calculated and reported by the GPs

14. T H O M S O N F I N A N C I A L 26 July 200514 Some definitions (2/2) Best fund Worst fund IRRs in decreasing order Maximum IRR Upper Quartile Median Lower Quartile Minimum IRR Top Quarter 2 nd Quarter 3 rd Quarter 4 th Quarter …

15. T H O M S O N F I N A N C I A L 26 July 200515 Fund Returns Calculations Cash / stock returns to investors = ‘Distribution’ Invested capital = ‘Paid-In’ time Principal metric is IRR since inception calculated net to limited partner. Beginning point is fixed, endpoint is variable The IRR is calculated as an annualised effective compounded rate of return using daily cash flows and annual/quarterly valuations. The IRR is the return (discount rate) that will equalise the present value of all invested capital with the present value of all returns, or where the net present value of all cash flows (positive and negative) is zero: where CFi is the cash flow for period i (negative for takedowns, positive for distributions) r -i

16. T H O M S O N F I N A N C I A L 26 July 200516 Typical Fund Cashflow - Simple example of IRR calculation YearTakedownsDistributionsNAV 1992(5,201.8)5,201.8 1993(12,749.5)17,300.2 1994(15,299.4)32,246.0 1995(5,099.8)7,988.049,128.1 1996(5,099.8)73,777.1 1997(7,649.7)30,770.566,416.4 199816,740.938,853.7 199911,484.725,046.8 -5.201.8 1 + IRR + -12,749.5 (1 + IRR) 2 + -15.299.4 (1 + IRR) 3 = 0 -5,099.8 + 7,988.0 + 49,128.1 + Column A Row 1(5,201.8) Row 2(12,749.5) Row 3(15,299.4) Row 452,016.3 THE RAW DATATHE CALCULATION IN MS EXCEL THE FORMULA =irr(A1:A4,0) =28.9%

17. T H O M S O N F I N A N C I A L 26 July 200517 Cashflows for Cumulative Returns CF series to 1993 CF 1994CF 1995CF 1996CF 1997CF 1998CF 1999 1992(5,201.8) 19934,550.7(12,749.5) 199416,946.6(15,299.4) 199552,016.32,888.2 199668,677.3(5,099.8) 199789,537.223,120.8 199855,594.616,740.9 199936,531.5 IRR-12.5%-4.4%28.9%32.5%29.4%20.7%17.9% Series of actual annual cash flows with NAV added as a positive cash flow in last year 17,300.2 -12,749.5

18. T H O M S O N F I N A N C I A L 26 July 200518 Realization Multiples DPI = Distributions / Paid In Ratio, a.k.a. realized multiple RVPI = Residual Value / Paid In Ratio, a.k.a. unrealized multiple TVPI = DPI + RVPI YearTakedownsDistributionsNAV Cumul IRR DPIRVPITVPI 1992(5,201.8)5,201.800.001.00 1993(12,749.5)17,300.2-12.5%0.000.96 1994(15,299.4)32,246.0-4.4%0.000.97 1995(5,099.8)7,988.049,128.128.9%0.211.281.49 1996(5,099.8)73,777.132.5%0.181.701.88 1997(7,649.7)30,770.566,416.429.4%0.761.302.06 199816,740.938,853.720.7%1.090.761.85 199911,484.725,046.817.9%1.310.491.80

19. T H O M S O N F I N A N C I A L 26 July 200519 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 IRR for each period and then compound together Shortfalls –Creates aberrations: 100 + 20% = 120 120 - 20% = 96 –Returns heavily dependent on valuations. Wrong valuations affect future returns –Assumes money can come and go freely at the beginning and end of each period

20. T H O M S O N F I N A N C I A L 26 July 200520 Cashflows for Time Weighted Returns Year 1Year 2Year 3Year 4Year 5Year 6Year 7 1992(5,201.8) 19934,550.7(17,300.2) 199416,946.6(32,246.0) 199552,016.3(49,128.1) 199668,677.3(73,777.1) 199789,537.2(66,416.4) 199855,594.6(38,853.7) 199936,531.5 TWR-12.5%-2.0%61.3%39.8%21.4%-16.3%-6.0% 32,246.0 -15,299.4

21. T H O M S O N F I N A N C I A L 26 July 200521 Investment Horizon Return Calculates backwards – what is the return over the last year, 3 years, etc. Came about because some funds are quick out of the gate but LPs want to know – what have they done for me lately Indicates what impact overall market is having most recently Beginning point is variable and endpoint is fixed IRR is calculated for each “investment horizon” IRR is calculated net to limited partner Composites are calculated on a “pooled” basis as if from one investment

22. T H O M S O N F I N A N C I A L 26 July 200522 Cashflows for Horizon Returns 1-year2-year3-year4-year5-year6-year7-year 1992(5,201.8) 1993(17,300.2)(12,749.5) 1994(32,246.0)(15,299.4) 1995(49,128.1)2,888.2 1996(73,777.1)(5,099.8) 1997(66,416.4)23,120.8 1998(38,853.7)16,740.9 199936,531.5 -6.0%-12.2%1.6%11.9%22.5%19.2%17.9% Series of actual cash flows during the period, with NAV at the end added as a positive cash flow in last year, and NAV at the beginning added as a negative cash flow at beginning 11,484.7 +25,046.8

23. T H O M S O N F I N A N C I A L 26 July 200523 So what do you have to ask to benchmark appropriately? Net IRR annualised since inception, net of fees and carried interest compounded at least quarterly, preferably daily The vintage year (be sure it agrees with Thomson Venture Economics’ vintage year – the year of the first capital call whether for investment or just management fees) Optional – total distributions, total paid in capital, ending NAV

24. T H O M S O N F I N A N C I A L 26 July 200524 Thomson Venture Economics Private Equity Performance Database Maintained by Venture Economics since 1988, online since 1991 1771 US funds formed 1969-2004 956 European funds 1979-2004 230 other international funds formed 1980-2004 74 funds of funds Available online in VentureXpert, where you can define your own performance sample (by country, vintage, size, focus, etc.)

25. T H O M S O N F I N A N C I A L 26 July 200525

26. T H O M S O N F I N A N C I A L 26 July 200526 Sources ~50% from GPs upon request from LPs who contract our benchmarking services ~50% from GPs who need data for their own benchmarking and fund raising needs Since we get data from LPs in addition to GPs there is not a consistent or significant self reporting bias We calculate IRR ourselves (we do not use self-reported IRRs) based on the underlying cashflows, and we verify against general partner financial reports to LPs We treat confidentiality very carefully – all data reported is strictly anonymous

27. T H O M S O N F I N A N C I A L Part 3 What are the actual results for the industry?

28. T H O M S O N F I N A N C I A L 26 July 200528 US Private Equity Cumulative IRRs by Vintage Year as of 31-Dec-04 Source: Thomson Venture Economics (VentureXpert database) / NVCA J-curve effect

29. T H O M S O N F I N A N C I A L 26 July 200529 US Private Equity Realization Multiples (DPI/RVPI) by Vintage Year as of 31-Dec-2004 Source: Thomson Venture Economics (VentureXpert database) / NVCA

30. T H O M S O N F I N A N C I A L 26 July 200530 US Private Equity Performance Benchmarks US Limited Partnerships Formed 1969-2004 Returns Since Inception Net to Investors as of 31-Dec-2004 StageNumberPooled IRRAverage DPIStd Deviation Manager Risk Coefficient Seed Capital6510.2%1.0635.1%3.4 Early Stage47720.8%1.2965.2%3.1 Balanced43213.6%1.0226.6%2.0 Later Stage18114.0%1.0827.2%1.9 All Venture Funds115415.9%1.1247.2%3.0 Buy-outs 0-$250m*17624.6%1.3031.9%1.3 Buy-outs $250-$500m10617.0%1.0924.4%1.4 Buy-outs $500-$1bn8514.7%0.9821.6%1.5 Buy-outs $1bn+1069.8%0.6622.3%2.3 All Buy-Out Funds47313.1%0.8126.9%2.1 Mezzanine649.4%0.8315.1%1.6 All Private Equity Funds177114.1%0.8940.8%2.9 Source: Thomson Venture Economics (VentureXpert database) / NVCA * fund size

31. T H O M S O N F I N A N C I A L 26 July 200531 European vs. US Private Equity Cumulative IRR Since Inception by Calendar Year Source: Thomson Venture Economics (VentureXpert database) / NVCA / EVCA Note: returns calculated with cash flows in US dollars for US funds, and with cash flows in Euros for European funds

32. T H O M S O N F I N A N C I A L 26 July 200532 US Private Equity Funds Formed 1969-2004 Net Investment Horizon Return as of 31-Dec-2004 1-year3-year5-year10-year20-year Early/Seed Stage29.5%-7.5%-1.4%44.9%19.8% Balanced14.5%0.2%0.3%18.0%13.3% Later Stage10.1%-0.3%-4.4%16.0%13.9% All Venture Funds17.7%-2.7%-1.2%26.1%15.7% Buy-Out Funds18.5%8.2%3.1%8.9%13.2% All Private Equity18.0%4.5%2.0%13.0%14.0% Source: Thomson Venture Economics (VentureXpert database) / NVCA

33. T H O M S O N F I N A N C I A L 26 July 200533 Want to know more? VentureXpert, the most complete private equity database globally (www.venturexpert.com)www.venturexpert.com –Profiles and directories (LPs, firms & funds, portfolio companies) –Analytics (includes investments, divestments, fund raising, fund performance) david.bernard@thomson.com, +44 (0)20 7336 1930 david.bernard@thomson.com For your data contributions and surveys: sandra.ribeiro@thomson.com, 212-806-3149sandra.ribeiro@thomson.com For sales enquiries: bill.moore@thomson.com, 646-822-3479bill.moore@thomson.com www.thomsonfinancial.com