Skin Or Skim? Inside Investment and Hedge Fund Performance

Skin or Skim? Inside Investment and Hedge Fund Performance

Arpit Gupta (NYU Stern) Kunal Sachdeva (Rice Jones)

April 15, 2019

Q Group Spring Seminar A Tale of Two Funds – Renaissance Technologies

1 A Tale of Two Funds – BlueCrest

2 1 SD ↑ in inside investment across funds in same ﬁrm → 1.26% in annual α

Within Firm Family, Funds Outperform With More Insider Capital

3 Within Firm Family, Funds Outperform With More Insider Capital

1 SD ↑ in inside investment across funds in same ﬁrm → 1.26% in annual α

3 Approach This In Two Ways: 1. Model to Capture Effects of Investment Capacity • Berk + Green type model; insider capital, multiple funds

2. Model Predictions to be Tested Empirically • Test directly using Form ADV and commercial HF databases

Why Does Skin in the Game Matter?

• Mechanisms Driving the Results • Decreasing returns to scale technologies generate incentives to more optimally manage capacity constraints when internal funds invested • Results in improved returns for some outsiders, at the cost of investment participation

4 2. Model Predictions to be Tested Empirically • Test directly using Form ADV and commercial HF databases

Why Does Skin in the Game Matter?

Approach This In Two Ways: 1. Model to Capture Effects of Investment Capacity • Berk + Green type model; insider capital, multiple funds

4 Why Does Skin in the Game Matter?

Approach This In Two Ways: 1. Model to Capture Effects of Investment Capacity • Berk + Green type model; insider capital, multiple funds

2. Model Predictions to be Tested Empirically • Test directly using Form ADV and commercial HF databases 4 1. Model Framework for understanding the relationship between inside investment and fund performance

4 Decreasing Return to Scale for an Investment Strategy

• Capital by insiders (I) and outsiders (O), [no borrowing]

T I O qn ≡ qn + qn

• An active manager specializes in N strategies:

T Rn,t = αn − Cn qn,t

• Cost function: T an T 2 Cn q , = q , n t 2 n t • Alpha, ‘α’, has standard interpretation Scale cost ‘a’ captures the scalability of the strategy

5 Insiders Tradeoff Beneﬁt from Fees vs. Cost of Outside Capital

Insider’s value add: Return on own capital + management fees:

VI = qI α − C qT + qOf | {z } |{z} Insider Return on Capital Mgmt fee

Outsider’s value add: Return on invested capital - management fees (taken as given):

VO = qO α − C qT − f | {z } Net Return

Objective: maximize insider value add subject to the participation of the outsiders and scarce insider capital 6 Insider Funds Tend to Be Smaller

200

150

100 Capital

0 0.25 0.50 0.75 1.00 Percent Insider Component Insider Capital Outsider Capital Total Capital

7 Components of Capital Source Allocating Internal Capital to Multiple Strategies

Consider two strategies where α1 > α2, but also a1 > a2:

T R1 = α1 − C1 q1 T R2 = α2 − C2 q2

This will result in different sized funds. The intuition is that insiders will ﬁrst contribute to the high α fund while:

I I dV1 dV2 I ≥ I dq1 dq2

8 Two Strategies, Intuition

0.100

0.075

0.050

0.025 Derivative of Value Add for Strategy Add for of Value Derivative

0.000 0 100 200 300 Insider Capital Strategy High Alpha, High Scale Cost Low Alpha, Low Scale Cost

Decision of Where Insiders Allocate Capital 9

Capital Tradeoff Value Decomposition Single Strategy • Where does it come from? • Skill vs. Scale: Tradeoff fees against return on own capital • Incentive alignment: GPs internalize dilutive impact of new capital raising on returns of existing investors • At the cost of participation by rationed new investors

• Key friction: limited commitment. Insiders cannot credibly commit to future strategy allocation and fund raising

Berk + Green + inside capital = Predictions on Performance

• Predictions from model: 1. Scale/size costs vary → inside capital not evenly allocated 2. High skin funds are smaller 3. High skin funds should outperform, ex-fees

10 • Key friction: limited commitment. Insiders cannot credibly commit to future strategy allocation and fund raising

Berk + Green + inside capital = Predictions on Performance

• Predictions from model: 1. Scale/size costs vary → inside capital not evenly allocated 2. High skin funds are smaller 3. High skin funds should outperform, ex-fees

• Where does it come from? • Skill vs. Scale: Tradeoff fees against return on own capital • Incentive alignment: GPs internalize dilutive impact of new capital raising on returns of existing investors • At the cost of participation by rationed new investors

10 Berk + Green + inside capital = Predictions on Performance

• Predictions from model: 1. Scale/size costs vary → inside capital not evenly allocated 2. High skin funds are smaller 3. High skin funds should outperform, ex-fees

• Key friction: limited commitment. Insiders cannot credibly commit to future strategy allocation and fund raising

10 2. Data Form ADV enables novel analysis of inside investment in hedge funds

10 Novel Linkage of Regulatory Data and Hedge Fund Returns

• Form ADV • Required disclosure form for investment advisors (> $100m), 2001 — Present • Dodd-Frank — Hedge Funds required to disclose, report internal investments • Survival-bias free, comprehensive

• Commercial Hedge Fund Return Databases • eVestment, HFR, BarclayHedge, CISDM, Eureka Hedge • Linkage based on SEC Identiﬁer or name, hand-checked

Merge Rates

11 Sample ADVs – RenTech Firm Level

12 Sample ADVs – Medallion LP, High Skin

13 Sample ADVs – Medallion LP, High Skin

14 Sample ADVs – Medallion LP, High Skin

Statistic Mean SD Sponsor of GP 0.741 0.438 Other Investment Advisor 0.501 0.500 Commodity Pool 0.401 0.490 Broker/Dealer 0.160 0.367 Insurance 0.065 0.246 Sponsor of LP 0.046 0.210 Bank or Thrift 0.045 0.207 Trust 0.042 0.201 Pension 0.027 0.161 Accountant 0.025 0.156 Real Estate 0.024 0.153 Lawyer 0.019 0.138 Municipal Advisor 0.013 0.113 Futures Merchant 0.009 0.094 Swap Dealer 0.007 0.081 Swap Participant 0.001 0.026 Share Supervised Persons 74% Share Ofﬁce 59%

15 Sample ADVs – RIEF LLC, Low Skin

16 Sample ADVs – RIEF LLC, Low Skin

17 Sample ADVs – RIEF LLC, Low Skin

18 Summary Statistics: ADV Dataset, Firm Level

Names Total Median Mean Std.Dev Custodial AUM ($m) 8, 525, 754.0 775.5 6, 458.9 28, 332.9 Regulatory AUM ($m) 18, 084, 715 1, 166.7 13, 700.5 72, 114.3 Discretionary AUM ($m) 17, 518, 589 1, 030.8 13, 271.7 71, 040.1 Non-Discretionary AUM ($m) 566, 126 0 428.9 2, 585.1 Number of Employees 139, 264 13 57.2 199.0 − Support Staff 81, 033 5 33.3 132.9 − Advisors 58, 231 7 23.9 75.6 Firms with Only One Fund 682 Number of Firms 2, 433

19 Summary Statistics: ADV Dataset, Fund Level

Names Total Median Mean Std.Dev Gross Asset Value ($m) 6, 177, 174.0 127.8 632.7 3, 060.7 Gross Assets, Inside Investment ($m) 772, 663 3.8 79.1 553.2 Gross Assets, Fund of Funds ($m) 1, 160, 354.0 0 118.9 873 Gross Assets, Non-US Investors ($m) 2, 492, 344.0 4.7 255.3 1, 698.6 Number of Owners 19 66.8 544.3 Minimum Investment ($m) 1 7.5 70.3 Inside Investment (%) 3 16.7 28.6 Fund of Funds (%) 0 15.9 29.5 Non-US Investors (%) 4 30.7 39.0 Number of Hedge Funds 9, 763 Number of Fund of Funds 2, 322 20 Revealing RenTech’s Dark Matter

1 Number of Funds

0 0 25 50 75 100 Insider or Related (%)

AQR 2-sigma Appaloosa Blue Crest 21 Data Enable Novel Analysis of Insider Investment

Related Parties Gross Investment

0.075

0.050 Density

0.025

0.000 0 25 50 75 100 % of Assets Invested by Self or Related 22 Personal Stakes Particularly Important in Hedge Funds

1.00

0.75

0.50 CDF

0.25

0.00 0% 25% 50% 75% 100% Inside Ownership Hedge Fund Private Equity Securitized Asset Fund Other Real Estate Fund Venture Capital 23 Managerial Compensation from Personal Stakes ≈ Either Source of Fee Income

100%

75% 1. Return on privately invested capital 50% 2. Performance Fee 3. Management Fee

Cumulative Compensation Cumulative 25% Monthly Number

0% 2012 2013 2014 2015 2016 Date 24 3. Results Inside Investment → Performance, and Mechanisms

24 Reminder – Predictions From The Model

1. High skin funds should outperform, ex-fees

2. Model Mechanism: Limit fund scale to maximize returns at expense of outsider participation

3. Firms strategically allocate their insider capital

25 1. Ownership-Performance Relationship

FH Excess Returns FFC Excess Returns Inside Investment (Percent) 0.0024∗∗∗ 0.0048∗∗∗ 0.0024∗∗∗ 0.0048∗∗∗ (0.0009) (0.0015) (0.0009) (0.0013)

Year FE No Yes No Yes Firm FE No Yes No Yes Fund Controls No Yes No Yes Log(Fund Size) Yes Yes Yes Yes Observations 41,097 41,097 41,097 41,097 R2 0.0003 0.0368 0.0009 0.0404

Rit − Rft = β1(RMt − Rft) + β2SMBt + β3HMLt + β4MOMt + εit

Shown in table: second stage panel Beta Exposures Beta Exposures by Inside Investment

26 1. Ownership-Performance Relationship

FH Excess Returns FFC Excess Returns All Controls All Controls

Inside Investment (Percent) 0.0024∗∗∗ 0.0048∗∗∗ 0.0024∗∗∗ 0.0048∗∗∗ (0.0009) (0.0015) (0.0009) (0.0013)

Year FE No Yes No Yes Firm FE No Yes No Yes Fund Controls No Yes No Yes Log(Fund Size) Yes Yes Yes Yes Observations 41,097 41,097 41,097 41,097 R2 0.0003 0.0368 0.0009 0.0404

Additional percent of insider investment adds 0.48 bps of α, monthly; or a 1 SD shift within ﬁrm adds 1.24% yearly α. Gross Inside Investment Fama MacBeth Lagged Returns

27 2. Groucho Marx Theory of Investment Don’t Want to Invest in a Fund that will have you as a LP

Open for Investors Excess Returns (FH) Excess Returns (FFC) (1) (2) (3) (4) (5) (6) Inside Investment (%) −0.0013∗∗∗ −0.0021∗∗∗ (0.0003) (0.0003)

Open for Investors −0.2291∗∗ −0.2186∗∗∗ −0.4463∗∗∗ −0.3141∗∗∗ (0.0971) (0.0746) (0.0660) (0.0706)

Fixed Effects No Yes No Yes No Yes

Funds closed to new investment outperform by 2-4% yearly Inside capital displaces outside investment; capital rationing on external margin

28 2. Insider Funds also Accept Less Inﬂows on the Intensive Margin

0 Flow (% of AUM) Flow

−50

−10 0 10 Lagged Excess Return Ownership of Fund Insider Outsider

29 Insider: Above median inside investment 2. Successful Insider Funds do not Drive Flows to Own or Other Funds in Family

Flow Percent Flow Percent >0 Complementary Complementary Complementary Complementary Flow Percent Flow Percent >0 Flow Percent Flow Percent >0 ∗∗∗ ∗∗∗ ∗ αt−1 0.1880 0.7927 −0.1365 −0.3876 0.0121 −0.3650 (0.0509) (0.2714) (0.1238) (0.2340) (0.1206) (0.2230)

∗∗∗ ∗∗∗ ∗ ∗∗ αt−2 0.1780 0.7715 −0.2110 −0.7776 −0.1350 −0.6218 (0.0639) (0.2541) (0.1216) (0.3324) (0.1275) (0.4062)

∗∗∗ ∗∗∗ ∗ ∗∗ ∗ αt−3 0.2122 0.7774 −0.1875 −0.7340 −0.1188 −0.5208 (0.0685) (0.2081) (0.0967) (0.2971) (0.0810) (0.3019)

αt−1×Insider −0.0523 −0.1669 (0.0713) (0.2577)

∗∗ αt−2×Insider −0.0960 −0.4574 (0.0900) (0.1960)

αt−3×Insider 0.0056 −0.1894 (0.0792) (0.3437)

High Ownership No No No No Yes Yes

30 2. High Inside Investment Funds are Smaller

AUM from Merged Dataset ($m) Gross Value from ADV ($m) Skin (Percent) −3.82∗∗∗ −7.86∗∗∗ −6.34∗∗∗ −10.14∗∗∗ (0.24) (1.20) (0.89) (1.12)

Year FE No Yes No Yes Firm FE No Yes No Yes Fund Controls No Yes No Yes Observations 2,633 2,633 35,960 35,960 R2 0.01 0.88 0.002 0.57 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

Interpretation: One additional percent of inside investment associated with a $7-10m smaller fund

31 2. Effects driven by specialist, arb, and equity funds

0.012

0.008

0.004 Estimate

0.000

CTA FOF equity

event driven fixed income relative value multi−strategy Strategy

e Coefﬁcient on main regression (Ri,t = βOwnershipi,t−1) by type of fund

32 2. Within equity: by long-short, emerging

0.010

0.005 Estimate

0.000

−0.005

long sector other

long−short Strategy

e Coefﬁcient on main regression (Ri,t = βOwnershipi,t−1)by type of fund among equity funds 33 3. Results Reﬂect Active Discretion in Insider Capital Allocation

All Controls All Controls Adjusted Inside Investment (Percent) 0.0039∗∗∗ 0.0085∗∗∗ 0.0028∗∗∗ 0.0069∗∗∗ (0.0013) (0.0017) (0.0009) (0.0014)

Log(Fund Size) No Yes No Yes Fixed Effects No Yes No Yes Observations 41,097 41,097 41,097 41,097 R2 0.0008 0.0377 0.0011 0.0410

Adjusted: assumes fees are mechanically reinvested; removed from inside investment Young Firms

34 3. Agency Conﬂict "Skimming" Event Study: Returns Follow Skin

DiD Regression

20 Cumulative Return (%) Cumulative

−20 0 20 Event Time (Month) Original Fund High Skin Low Skin

35 Tracks excess return of original fund after new fund is created with outsider or insider money Additional Tests

A. Where in Distribution? B. Insider Information C. Nefarious Actions D. Firm-level Ownership E. Equal-weighted Link F. Fung-Hsieh risk factor Link G. Other risk factors Link H. Merge Bias Link I. Relation to Fees Link J. GP Restriction on Related Parties Link K. Including 0, 100% Skin Funds Link

Conclusion 36 A. Ownership-Performance Relationship, Effects Strongest for High-Skin Funds 0.010 0.005 0.000 −0.005

0.2 0.4 0.6 0.8

Quantile Regression of Inside Investment on Excess Returns

37 B. Insider Information Changes in inside investment don’t matter, only levels

Ri,t−1→t = βInsiderInﬂowi,t−1 + γOutsiderInﬂowi,t−1 + εit

Excess Return Lagged Insider Flow (%) 0.0003 −0.0000 0.0001 (0.0003) (0.0003) (0.0003)

∗∗ ∗∗ Lagged Outsider Flow (%) 0.0002 0.0002 −0.0001 (0.0001) (0.0001) (0.0001)

Year FE No Yes Yes Firm FE No No Yes Observations 936 936 936 R2 0.0035 0.2465 0.6035

Interpretation: a 1% increase in Insider Flow associated with a 1bps annual increase in excess return 38 C. Nefarious Actions Fraud weakly linked to to other characteristics

Corporate Governance of Hedge Funds

Dep. Var: Excess Return Inside Investment

Ever Civil Judgement −0.08 −2.01 (0.05) (1.27)

Ever Criminal Judgement −0.23∗∗ −0.28 (0.10) (2.52)

Observations 63,978 63,978 5,062 5,062 R2 0.0000 0.0001 0.0005 0.0000 ∗∗p<0.05; ∗∗∗p<0.01

39 D. Firm-level Ownership Ownership of Partnership

Direct Equity Ownership of AQR

40 D. Firm-level Ownership Shell Companies

Direct Equity Ownership of AQR

41 D. Firm-level Ownership Really Cliff Just Owns This

Indirect Equity Ownership of AQR

42 D. Firm-level Ownership Placing Bounds on Firm-level Ownership

0.3

0.2

Density 0.1

0.0 0 50 100 150 200 Firm Level Equity (%) Max Equity Method Mid Equity Method Min Equity Method

Distribution of min and max estimates of Firm Ownership 43

Expect min estimate to be < 100 and max estimate ≥ 100 D. Firm-level Ownership Dispersion of Equity Ownership (HHI)

0.20

0.15

Density 0.10

0.05

0.00 0.00 0.25 0.50 0.75 1.00 HHI

44 HHI Index of Equity Ownership D. Equity Ownership Dispersion and Performance

Monthly Excess Return (FF) Skin (Percent) 0.0025∗∗∗ 0.0021∗∗∗ 0.0025∗∗∗ (0.0004) (0.0004) (0.0004)

# of Equity Holders −0.0165∗∗∗ −0.0170∗∗∗ (0.0028) (0.0032)

HHI of Firm Equity 0.0840∗∗ −0.0142 (0.0355) (0.0399)

Year Yes Yes Yes Log(Size) Yes Yes Yes Observations 63,978 63,978 63,978 R2 0.0142 0.0132 0.0143 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

45 Our Contribution Relates to Several Literatures

• Inside Investment and Mutual Fund Performance: Khorana et al. (2007), Evans (2008), Chen et al. (2008), Cremers et al. (2009) Hedge funds: Qiu et al. (2016), Brown et al. (2008)

• Assessing Managerial α: Kosowski et al. (2006), Fama and French (2010), Kacperczyk, Nieuwerburgh, and Veldkamp (2014), Berk and van Binsbergen (2015), Khorana, et al. (2007), Evans (2008), Chen, Goldstein and Jiang (2008), Koijen (2014)

• Fund Families: Massa (2003), Berk et al. (2017)

• Financial Compensation and Incentives: Das et al. (2002), Ibert et al. (2017), Ma et al. (2016) Hedge funds: Agarwal et al. (2009), Burasachi et al. (2014) Model: Berk and Green (2004), Berk and van Binsbergen (2015) Inequality: Kaplan and Rauh (2013), Philippon and Reshef (2012), Alvaredo et al. (2013) • Ownership and Firm Performance: Berle and Means (1932), Jensen and Meckling (1976), Fama and Jensen (1983), Holmstrom (1985), Randall, 46 Shleifer and Vishny (1988) Where are the Investors’ Yachts?

• We’re ignoring a major component of hedge fund compensation: insider returns • Managers tradeoff high capacity-management fee funds with low capacity-inside money funds • Predictions confirmed by merging comprehensive dataset • We should also consider allocation rights as a form of compensation • Suggests why hedge fund manager profits so persistent despite seeming competition and low performance • This relationship is driven by profit maximization of the asset manager • Insiders benefit from alignment of incentives: but only if they get to participate

• Public ADV Data: https://www.skinorskim.org/

47 Thank You!

47 Where are the Investors’ Yachts?

Relationship is driven by proﬁt maximization of the asset manager

• Limited Commitment • We should also consider allocation rights as a form of compensation

• Major Component of Hedge Fund Compensation • Beyond fees, insider returns are important • Access to specialized investment technology

48 Managerial Compensation from Personal Stakes ≈ Either Source of Fee Income

1,500

1,000 1. Return on privately invested capital 500 2. Performance Fee 3. Management Fee Monthly Compensation 0 Back

−500

2012 2013 2014 2015 2016 Date 49 3. Results Reﬂect Active Discretion in Insider Capital Allocation

All Controls All Controls Adjusted Inside Investment (Percent) 0.0034∗∗ 0.0085∗∗∗ 0.0026∗∗ 0.0078∗∗∗ (0.0015) (0.0015) (0.0012) (0.0017)

Log(Fund Size) No Yes No Yes Fixed Effects No Yes No Yes Observations 25,938 25,938 25,938 25,938 R2 0.0005 0.0415 0.0007 0.0424 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

Adjusted: assumes fees are mechanically reinvested; removes from inside Back investment 50 Merge Rates

10 Merge Rate (%)

[ 0, 5)[ 5, 10) [ 10, 15)[ 15, 20)[ 20, 25)[ 25, 30)[ 30, 35)[ 35, 40)[ 40, 45)[ 45, 50)[ 50, 55)[ 55, 60)[ 60, 65)[ 65, 70)[ 70, 75)[ 75, 80)[ 80, 85)[ 85, 90)[ 90, 95)[ 95,100] Back Insider Ownership (%)

51 Merge Rates

6 Merge Rate (%)

HFR CISDM EUREKA BARCLAYS EVESTMENT Source

52 1. Ownership-Performance Relationship Back

FH Excess Returns FFC Excess Returns (1) (2) (3) (4) Inside Investment (Gross) 0.0397∗∗ 0.0710∗∗ 0.0297∗∗ 0.0856∗∗∗ (0.0184) (0.0284) (0.0150) (0.0235)

Year FE No Yes No Yes Firm FE No Yes No Yes Fund Controls No Yes No Yes Log(Fund Size) Yes Yes Yes Yes Observations 41,097 41,097 41,097 41,097 R2 0.0002 0.0367 0.0008 0.0404

53 1. Ownership-Performance Relationship Back

FH Excess Returns FFC Excess Returns (1) (2) (3) (4) Inside Investment (Percent) 0.0024∗∗∗ 0.0020∗∗ 0.0020∗∗ 0.0021∗∗ (0.0008) (0.0008) (0.0008) (0.0009)

Year FE No No No No Firm FE No No No No Fund Controls No Yes No Yes Log(Fund Size) Yes Yes Yes Yes Observations 41,097 41,097 41,097 41,097 R2 0.1662 0.2034 0.0469 0.0690 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01

54 1. Ownership-Performance Relationship Back

FH Excess Returns FFC Excess Returns ∗∗ ∗∗∗ ∗∗ ∗∗∗ Inside Investment (Percent) 0.0025 0.0055 0.0021 0.0045 (0.0010) (0.0014) (0.0009) (0.0013)

∗∗∗ ∗∗∗ Returni,t−1 0.0066 −0.0249 0.1027 0.0837 (0.0384) (0.0363) (0.0206) (0.0218)

Returni,t−2 −0.0014 −0.0273 0.0334 0.0143 (0.0331) (0.0315) (0.0238) (0.0219)

Returni,t−3 0.0341 0.0117 0.0138 −0.0081 (0.0369) (0.0370) (0.0202) (0.0199)

Log(Fund Size) No Yes No Yes Fixed Effects No Yes No Yes Observations 37,958 37,958 37,958 37,958 R2 0.0020 0.0425 0.0212 0.0535

55 Decreasing Return to Scale for an Investment Strategy

• An active manager specializes in N strategies:

T Rn,t+1 = αn − Cn qn,t

• Cost function: T an T 2 Cn q , = q , n t 2 n t

• Alpha, ‘α’, has standard interpretation Scale cost ‘a’ captures the scalability of the strategy T I O • Capital by insiders (I) and outsiders (O): qn ≡ qn + qn (no borrowing)

56 Insiders Tradeoff Beneﬁt from Fees vs. Cost of Outside Capital

Insider’s value add, Return on own capital plus management fees:

VI = qI α − C qT + qOf | {z } |{z} Insider Return on Capital Mgmt fee

Outsider’s value add, proﬁt from invested capital minus management fees (taken as given):

VO = qO α − C qT − f | {z } Net Return

The insider’s objective is to maximize insider value add subject to the participation of the outsiders and scarce insider capital 57 One Strategy, Unconstrained Insider Capital

First consider a capital unconstrained insider. Their is no beneﬁt to collecting fees. The insider’s problem reduces to:

I I T O arg max Vt+1 = qt α − C qt + fq qI,qO

Can set qT = qI, qO = 0, optimal level of capital for an insider that is unconstrained:

r 2α q¯I∗ = t 3a

Back

58 One Strategy, Constrained Insider Capital

I I∗ Now consider a capital constrained insider, qt ∈ 0, q¯t :

I I T O arg max Vt+1 = qt α − C qt + fq qI,qO

qI is constrained, and the only choice variable is qO. The outsiders capital that maximizes value add is:

f qO = − qI t aqI t

While fees are such that must have a non-negative value add:

O T O 0 ≤ q α − C q − fq 59 One Strategy, Constrained Insider Capital (Cont.)

Fees are set to ensure non-negative value-add:

s ∗ I2 I2 2α f = −a q + a q 1 + a (qI)2

Total optimal investment reduces to:

r 2α qT∗ = −qI + (qI)2 + a

• Key Intuition: • Total fund size is decreasing in insider capital • Total funds size is increasing in α, decreasing scale cost a

Robustness 60 Blue Crest

4 Number of Funds

0 0 25 50 75 100 61 Insider or Related (%) Back AQR

20 Number of Funds

0 0 25 50 75 100 62 Insider or Related (%) Back Two Sigma

4.0

3.5

3.0

2.5

2.0

1.5 Number of Funds

1.0

0.5

0.0 0 25 50 75 100 63 Insider or Related (%) Back Appaloosa

1 Number of Funds

0 0 25 50 75 100 64 Insider or Related (%) Back Data Enable Novel Analysis of Insider Investment

0.25

0.20

0.15 Density 0.10

0.05

0.00 15 20 Log AUM Invested by Self or Related

65 Back Effects Hold Under More Restrictive Measures of Related Parties

FH Excess Returns FFC Excess Returns All Controls All Controls (1) (2) (3) (4) Inside Investment (Percent) 0.0035∗∗∗ 0.0055∗∗∗ 0.0030∗∗∗ 0.0056∗∗∗ (0.0011) (0.0017) (0.0010) (0.0014)

Log(Fund Size) No Yes No Yes Fixed Effects No Yes No Yes Observations 29,691 29,691 29,691 29,691 R2 0.0006 0.0387 0.0015 0.0492

Back 66 Effects not Driven by Alternative Risks

1.00 ● ● ●

● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0.75 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

● 0.50

0.25 Getmansky et al. (2004) Measure

0.00 0 25 50 75 100 Inside Investment (%)

67 Back Effects not Driven by Alternative Risks

1.00

0.75

0.50

●

● ● 0.25 ● ● ● ● ● ● ● ● Maximum Drawdown of Fund Drawdown Maximum ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●

0.00 0 25 50 75 100 Inside Investment (%)

68 Back Also Holds for Fung-Hsieh Factor Adjustment

Risk Adjusted Excess Returns (FH) All Controls All Controls Skin (percent) 0.0060∗∗∗ 0.0110∗∗∗ (0.0022) (0.0038)

Skin (log of gross) 0.0902∗∗∗ 0.1824∗∗∗

(0.0219) (0.0539) Robustness

Year FE No Yes No Yes Firm FE No Yes No Yes Log(Fund Size) No Yes Yes Yes Fund Controls No Yes No Yes Observations 63,978 63,978 63,978 63,978 R2 0.0022 0.1022 0.0033 0.1040

69 Equal Weighted

Risk Adjusted Excess Returns (FF) All Controls All Controls Skin (percent) 0.0018∗∗ 0.0031∗∗∗ (0.0008) (0.0008)

Skin (log of gross) 0.0238 0.0499∗∗∗ (0.0146) (0.0127)

Year FE No Yes No Yes Firm FE No Yes No Yes Log(Fund Size) No Yes Yes Yes Fund Controls No Yes No Yes Observations 63,978 63,978 63,978 63,978 R2 0.0003 0.0938 0.0002 0.0937

Robustness

70 Equal Weighted

Risk Adjusted Excess Returns (FH) All Controls All Controls Skin (percent) 0.0017∗∗ 0.0035∗∗∗ (0.0007) (0.0008)

Skin (log of gross) 0.0244 0.0726∗∗∗ (0.0149) (0.0143)

Year FE No Yes No Yes Firm FE No Yes No Yes Log(Fund Size) No Yes Yes Yes Fund Controls No Yes No Yes Observations 63,978 63,978 63,978 63,978 R2 0.0003 0.1048 0.0003 0.1049

Robustness

71 Merge Bias is Constant Except at 0% and 100% of Ownership

Robustness

2.0

1.5

1.0 Merge Bias

0.5

0 25 50 75 100 Percent Ownership

Merge Bias by Ownership Percentage 72 Genuinely Survival-Bias Free HF Series

Total Merged Sample Total Gross Asset Value 6, 013 236 Number of Employees 2, 524, 566 26, 091 Non-Advisory Employees 1, 699, 510 15, 669 Advisory Employees 825, 056 10, 422

73 Skin Level R2 Sharpe Ratio Residual SE (Intercept) Mkt.RF SMB HML MOM 1 0.34 0.06 1.63 0.11 0.33 0.1 0.08 0.14 (0.39) (3.91) (0.86) (0.48) (1.42)

2 0.61 0.08 0.88 0.08 0.33 -0.02 -0.08 0.02 (0.53) (7.21) (-0.31) (-0.87) (0.38)

3 0.64 0.12 0.86 0.11 0.32 0.1 0.01 0.03 (0.78) (7.17) (1.67) (0.08) (0.53)

4 0.54 0.34 0.84 0.32 0.24 0.07 -0.1 -0.02 (2.24) (5.6) (1.15) (-1.18) (-0.42)

5 0.62 0.27 0.86 0.25 0.31 0.06 -0.03 -0.03 (1.77) (6.91) (0.94) (-0.31) (-0.54)

Back 74 Skin Level R2 Sharpe Ratio Residual SE (Intercept) PTFSBD PTFSFX PTFSCOM SNPMRF SCMLC BD10RET BAAMTSY 1 0.3 0.21 1.74 0.4 -0.02 0.01 -0.04 0.2 -0.06 0.01 -0.04 (1.39) (-0.62) (0.68) (-2.23) (2.06) (-1.75) (0.49) (-1.51)

2 0.26 0.38 1.27 0.52 -0.02 0.01 -0.03 0.01 -0.01 0.01 -0.03 (2.49) (-1.16) (1) (-2.34) (0.08) (-0.32) (0.74) (-1.63)

3 0.3 0.42 1.25 0.57 -0.01 0.02 -0.02 -0.03 0 0.01 -0.04 (2.73) (-0.66) (1.18) (-1.76) (-0.37) (-0.15) (1.26) (-2.56)

4 0.25 0.51 1.12 0.63 -0.01 0 -0.02 -0.01 0 0 -0.04 (3.37) (-0.63) (0.33) (-1.66) (-0.1) (0.14) (0.07) (-2.36)

5 0.36 0.49 1.17 0.62 -0.01 0.01 -0.02 0.02 0 0.01 -0.05 (3.19) (-0.35) (0.49) (-2.12) (0.28) (-0.02) (0.96) (-2.88)

Back

75 Relationship Between Ownership and Returns Persists Under Factor Adjustment

Regression of Percent Inside Investment against α

Dep Var: Excess Return Excess Return Excess Return Excess Return Within Firm Firm Level 1F1F No FE (1) (2) (3) (4) Skin (percent) 0.0021∗∗ 0.0020∗∗∗ 0.0075∗∗∗ 0.0015∗∗∗ (0.0009) (0.0005) (0.0013) (0.0005) Year FE Yes Yes Yes No Firm FE Yes No No No Observations 33,056 33,056 4,127 33,056

Rit − Rft = β1(RMt − Rft) + β2SMBt + β3HMLt + β4MOMt + εit

76 Relationship Between Ownership and Returns Persists Under Factor Adjustment

Regression of Percent Inside Investment against α

Dep Var: Excess Return Excess Return Excess Return Excess Return Within Firm Firm Level 1F1F No FE (1) (2) (3) (4)

Skin (percent) 0.0021∗∗ 0.0020∗∗∗ 0.0075∗∗∗ 0.0015∗∗∗ (0.0009) (0.0005) (0.0013) (0.0005) Year FE Yes Yes Yes No Firm FE Yes No No No Observations 33,056 33,056 4,127 33,056

Interpretation: each additional percent of insider investment adds .2 bps of α, monthly; or from 0 → 100 skin adds 2.5% yearly α.

77 Fund Beta Exposures — MKT

0.08

0.06

0.04

0.02

0.00

−1 0 1 MKT.RF

Back

78 Fund Beta Exposures — HML

0.075

0.050

0.025

0.000 −1 0 1

Back

79 Fund Beta Exposures — SMB

0.10

0.05

0.00 −1 0 1 SMB

Back

80 Fund Beta Exposures — MOM

0.15

0.10

0.05

0.00 −1 0 1 MOM

Back

81 Fund Beta Exposures — R2

0.04

0.03

0.02

0.01

0.00 0.00 0.25 0.50 0.75 1.00

Back

82 Fund Beta Exposures — SNPMRF

0.09

0.06

0.03

0.00 −1.0 −0.5 0.0 0.5 1.0

Back

83 Fund Beta Exposures — SCMLC

0.06

0.04

0.02

0.00 −1.0 −0.5 0.0 0.5 1.0

Back

84 Fund Beta Exposures — DGS10PCH

0.15

0.10

0.05

0.00 −1.0 −0.5 0.0 0.5 1.0

Back

85 Fund Beta Exposures — CSFACTOR

0.075

0.050

0.025

0.000 −1.0 −0.5 0.0 0.5 1.0

Back

86 Fund Beta Exposures — PTFSBD

0.3

0.2

0.1

0.0 −1.0 −0.5 0.0 0.5 1.0

Back

87 Fund Beta Exposures — PTFSCOM

0.4

0.3

0.2

0.1

0.0 −1.0 −0.5 0.0 0.5 1.0

Back

88 Fund Beta Exposures — PTFSFX

0.4

0.3

0.2

0.1

0.0 −1.0 −0.5 0.0 0.5 1.0

Back

89 Fund Beta Exposures — R2

0.06

0.04

0.02

0.00 0.00 0.25 0.50 0.75 1.00

Back

90 Fund Beta Exposures by Inside Investment

0.5 ● ●

● ● ● ● ● 0.4 ● ●

●

0.3

● Mkt.RF ● HML 0.2 ● ● SMB ● MOM Estimate ● ● ● ● ●

0.1 ● ●

● ● ● ● ● ● ● ● 0.0 ● ● ● ● ● ● ● ● ● ● ● ● −0.1 ● 1 2 3 4 5 6 7 8 9 10 Inside Investment Decile

Back

91 Fund Beta Exposures by Inside Investment

0.05 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● 0.00 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● PTFSBD ● ● ● ● PTFSFX ● ● ● ● ● ● ● ● PTFSCOM −0.05 ● SNPMRF ● SCMLC ● ● Estimate DGS10_PCH ● CSFACTOR ● ●

● ● ● ● ●

● ● −0.10 ● ●

● ● ● ● ● ● ●

● −0.15 1 2 3 4 5 6 7 8 9 10 Inside Investment Decile

Back

92 Corporate Governance Obligations of Hedge Funds

As noted in Nowak (2009) and quoted in Morley (2014), the manager:

is required to devote to the [fund] only that amount of time and attention that the [manager] in its sole discretion deems reasonably necessary to achieve the [fund’s] objectives.

Back

93 7.5

5.0

Value Add Value 2.5

0.0 0.00 0.25 0.50 0.75 1.00 Insider Fraction Component vI vO vT

Tradeoff Between Insider and Outside Capital

200

150

100 Capital

0 0.00 0.25 0.50 0.75 1.00 Insider Fraction Capital Provider qI qO qT

94 Tradeoff Between Insider and Outside Capital

200

150

100 Capital

0 0.00 0.25 0.50 0.75 1.00 Insider Fraction Capital Provider qI qO qT

7.5

5.0

Value Add Value 2.5

0.0 0.00 0.25 0.50 0.75 1.00 94 Insider Fraction Component vI vO vT Decomposition of Value to Insider

7.5

5.0

Value Add Value 2.5

0.0 0 50 100 Insider Capital Component fees investments vI

Value to Insider as Components

Back 95 Related Parties are Typically Vehicles for Ownership by GPs

Statistic Mean SD Sponsor of GP0 .7410 .438 Other Investment Advisor 0.501 0.500 Commodity Pool 0.401 0.490 Broker/Dealer 0.160 0.367 Insurance 0.065 0.246 Sponsor of LP 0.046 0.210 Bank or Thrift 0.045 0.207 Trust 0.042 0.201 Pension 0.027 0.161 Accountant 0.025 0.156 Real Estate 0.024 0.153 Lawyer 0.019 0.138 Municipal Advisor 0.013 0.113 Futures Merchant 0.009 0.094 Swap Dealer 0.007 0.081 Swap Participant 0.001 0.026

Share Supervised Persons 74% 96 Share Ofﬁce 59% Note: Measures whether related party is present; does not add to 100%

Back Agency Conﬂict Event Study Suggestive of "Skimming"

DiD Monthly Return ∗∗ Post 0.263 (0.123)

High −0.214 (0.134)

∗∗∗ Post x High 0.706 (0.170)

∗∗∗ Constant 0.389 (0.096)

Observations 3,885 R2 0.020 ∗∗∗ F Statistic 27.034 (df = 3; 3881) ∗ ∗∗ ∗∗∗ Note: p<0.1; p<0.05; p<0.01

Back 97 Relation to Fees

Back

98 Liquidity and Inside Investment

Back

1.00 ● ● ●

● 0.50

0.25 Getmansky et al. (2004) Measure

0.00 0 25 50 75 100 Inside Investment (%) 99 Maximum Drawdown and Inside Investment

Back

1.00

0.75

0.50

●

0.00 0 25 50 75 100 Inside Investment (%) 100 Monthly Compensation

Back

1,500

1,000

500

Monthly Compensation 0

−500

2012 2013 2014 2015 2016 Date

managerial fees, performance fees, privately invested capital 101 Cumulative Compensation

Back

100%

75%

50%

Cumulative Compensation Cumulative 25%

0% 2012 2013 2014 2015 2016 Date

managerial fees, performance fees, privately invested capital 102 Top 10 Hedge Fund Manger Paychecks, 2016

Back

Rank Name Fund 2016 Paycheck

1 James Simons Renaissance Technologies $1.6 billion 2 Ray Dalio Bridgewater Associates $1.4 billion 3 John Overdeck Two Sigma $750 million 4 David Siegel Two Sigma $750 million 5 David Tepper Appaloosa Management $700 million 6 Kenneth Grifﬁn Citadel $600 million 7 Paul Singer Elliot Management Corp. $590 million 8 Michael Hintze CQS $450 million 9 David Shaw D.E. Shaw Group $415 million 10 Israel Englander Millennium Management $410 million

Total: $6.9 billion

103 0 and 100 Skin Funds

Robustness

FH Excess Returns FFC Excess Returns (1) (2) (3) (4) Skin (Percent) 0.0017∗∗ 0.0035∗∗∗ 0.0026∗∗∗ 0.0044∗∗∗ (0.0007) (0.0012) (0.0007) (0.0011)

Log(Fund Size) Yes Yes Yes Yes Fixed Effects No Yes No Yes Observations 47,589 47,589 47,589 47,589 R2 0.0002 0.0348 0.0010 0.0393 Adjusted R2 0.0001 0.0188 0.0010 0.0234

104 Skin Doesn’t Predict Fees

Back

Management Fee Performance Fee Management Fee Performance Fee (1) (2) (3) (4) Skin (Percent) −0.0030∗ 0.0040 −0.0014 0.0056 (0.0016) (0.0153) (0.0014) (0.0128)

Log(Fund Size) No No Yes Yes Year FE No No Yes Yes Inception Year FE No No Yes Yes Strategy FE No No Yes Yes Observations 5,925 5,848 5,925 5,848 R2 0.0137 0.0002 0.3216 0.5405 Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01105