The Quality and Service of Investment Banks’ Service: Evidence from the PIPE Market
Na Dai, University of New MexicoHoje Jo, Santa Clara University
John Schatzberg, University of New Mexico
What is a PIPE?
Private Investment in Public Equity PIPE securities are generally issued pursuant
to Section 4(2) of the Securities Act or Regulation D under the Securities Act, which provide an exemption from registration for a non-public offering by an issuer.
Size of the PIPE Market
Year PIPE ($M) SEO($M)
1995 1870 46327
1996 9068 57294
1997 12800 63488
1998 13489 54731
1999 25609 75546
2000 59442 90746
2001 79559 63189
2002 38170 55744
2003 92866 58297
2004 56741 75243
2005 52899 70112
2006 87960 83217
2007 98477 64570
0
20000
40000
60000
80000
100000
120000
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Year
Amou
nt R
aise
d ($
M)
PIPE ($M)
SEO($M)
Major Players in the PIPE Market
Issuers Placement Agents
Investors
Hedge Funds
VC Funds
PE Funds
Mutual Funds
Broker/Dealers
Banks/Insurance Companies
Pension Funds
Small
Young
Sometimes distressed
High information asymmetry
Traditional investment bank
Specialized PIPE placement agents
Research Questions
We investigate the market structure and the pricing by placement agents in private investments in public equities (PIPEs). How do firms and placement agents choose each other? Do placement agents help lower transaction costs and
information costs? Do placement agents with strong reputations provide
higher-quality service? Does the higher quality service enable placement agents
to charge higher fees?
Motivation
Theoretical models about investment banks (underwriters)’ general behavior in the IPO and SEO market
Chemmanur and Fulghieri (1994) In equilibrium, reputable investment banks underwrite less risky issues,
obtain higher prices for the issuers, and receive higher compensation. Issuer quality and the pricing of investment banks’ services solve the
matching problem. Fernando, Gatchev, and Spindt (2005)
Model the matching between issuers and underwriters as a two-sided process
The underwriting spread is the result of a bargaining process and does not require a specific matching of issuers and underwriters
Key differences between these two models Does the fee structure determine the matching process or is
it negotiated after the matching is concluded? Do more reputable underwriters charge higher fees?
Motivation (cont.)
The existing empirical evidence is mixed. Fang (2005): More reputable underwriters provide
higher quality services and command a fee premium in the corporate bond market.
Chen and Ritter (2000): In the US market, more than 90% of IPOs raising $20-80 million have spreads of exactly seven percent.
Krigman, Shaw, and Womack (2001): Their issuer survey data reveal that fee structure received the lowest ranking among all decision criteria when selecting a lead underwriter in the IPO market.
Motivation (cont.)
The PIPE market is an emerging market where firms raise capital when traditional financing approaches become difficult. There is very little study or evidence about investment banks’ behavior in this market.
The PIPE market is an attractive venue for testing whether financial intermediaries help lower the transaction and
information costs of the financing process. The PIPE market possesses particularly high levels of information
asymmetry given that most of the issuers are small, young, and with high growth potential.
Hypotheses
H1: Higher-quality PIPE firms are associated with more reputable placement agents.
H2: More reputable placement agents provide higher-quality services, but not necessarily charge higher fees.
Methodology
The endogeneity issueControl for the endogeneity using
Instrumental Variable Framework The Lee (1978) Two-Stage Approach Simultaneous Equation Analysis
Sample and Data
PIPE sample is provided by Sagient Research Sample period: from 1996 to 2005 Final sample: 1,148 common stock PIPE
transactions where CRSP and Compustat data one year prior to the PIPE are available.
A total of 215 placement agents with varying levels of participation in the PIPE market
Measures of Agent Reputation
C&M ranking Total market share Discounts/CAR Market share in previous three-year period
The percentage of total gross proceeds of all PIPE deals led by the placement agent over the last three years
The top 15 placement agents in every three-year period as reputable placement agents.
√
Measures of Firm Quality
Analyst coverage: the maximum number of analysts following the issuer during the 12 months prior to the issuance as reported by I\B\E\S
Delist Dummy: equal to one if the issuer is delisted within 24 months subsequent to the issuance and zero otherwise
Volatility: the standard deviation of the daily returns in the last 12 months
EBITDA/Assets Book to market (BM) ratio Long term debt/Assets
Measures of Placement Agents’ Service Quality
Discounts (closing price the day before the PIPE deal
closes-offer price)/offer price Additional benefits
Analyst following Turnover Spread Volatility
The Matching of Placement Agents and Issuers---- Likelihood of Having Agent
Coefficient p-valueIntercept -1.116** 0.011Firm QualityLn (Analyst) -0.123* 0.075EBITDA/Assets 0.031 0.644B/M -0.052 0.232Financial Leverage -0.488** 0.021Delist Dummy -0.068 0.548Ln (Volatility) 0.159 0.277Other VariablesFirm Size -0.217*** 0.000Ln (Age) 0.083 0.156Ln (Proceeds) 0.655*** 0.000Warrants 0.557*** 0.000Industry Dummies YesYear Dummies YesN 1,148Pseudo R-square (%) 15.86
The Matching of Placement Agents and Issuers---- Likelihood of Having Reputable Agent
Coefficient p-valueIntercept -1.239* 0.055Firm QualityLn (Analyst) 0.192** 0.042EBITDA/Assets 0.226** 0.050B/M 0.022 0.792Financial Leverage -0.007 0.984Delist Dummy -0.130 0.456Ln (Volatility) -0.050 0.793Other VariablesFirm Size -0.095 0.276Ln (Age) -0.110 0.180Ln (Proceeds) 0.357*** 0.000Warrants -0.056 0.681IPO/Previous SEO Underwriter 0.178 0.656Previous PIPE Agent 0.420*** 0.004Industry Dummies YesYear Dummies YesN 707Pseudo R-square (%) 12.02
Do Placement Agents Help Reduce Discounts?
IV ApproachThe Lee Model
With Agent Without Agent
Coefficient p-value Coefficient p-value Coefficient p-valueIntercept 15.821** 0.026 -41.379 0.146 0.809 0.956Ln (Analyst) -4.342*** 0.004 -4.143*** 0.003 -10.108** 0.016EBITDA/Assets -0.260 0.821 0.335 0.916 1.472 0.662B/M -1.608* 0.060 -2.344** 0.031 -1.291 0.332Financial Leverage -7.804 0.111 -9.628 0.193 -18.157** 0.038Delist Dummy 0.115 0.959 1.502 0.642 -4.912 0.273Ln (Volatility) 9.303*** 0.002 11.398*** 0.007 12.437* 0.077Ln (Proceeds) 9.630*** 0.001 5.945 0.116 28.265*** 0.004Warrants 11.813*** 0.001 6.005 0.316 36.595*** 0.001Ln (Cash) -2.085** 0.025 0.183 0.901 -4.649** 0.025Insiders -0.331 0.924 -3.881 0.625 5.807 0.477Block Investor -0.519 0.785 0.224 0.913 6.489* 0.071Hedge Funds 4.106** 0.027 1.421 0.489 4.340 0.292With Agent -61.225*** 0.000Inverse Mills Ratio 44.977* 0.065 85.074*** 0.001Industry Dummies Yes Yes YesYear Dummies Yes Yes YesN 1148 707 441Adjusted R-square (%) 6.20 14.80 15.46
Are More Reputable Placement Agents More Capable of Reducing Discounts?
IV ApproachThe Lee Model
More Reputable Agents Less Reputable AgentCoefficient p-value Coefficient p-value Coefficient p-value
Intercept 34.457*** 0.000 -5.254 0.838 23.122 0.128Ln (Analyst) 0.615 0.708 -0.800 0.771 1.926 0.630EBITDA/Assets 0.198 0.864 -2.552 0.591 0.663 0.795B/M -1.079 0.344 0.519 0.763 -1.284 0.226Financial Leverage 1.205 0.827 -9.467 0.259 2.571 0.697Delist Dummy 0.503 0.848 -1.024 0.778 0.286 0.944Ln (Volatility) 5.832* 0.056 1.777 0.633 6.638 0.241Ln (Proceeds) -2.019 0.242 -0.750 0.772 -1.673 0.572Warrants -7.737*** 0.000 -0.495 0.185 -8.838** 0.010Ln (Cash) 1.162 0.260 -0.183 0.894 1.611 0.508Insiders -3.988 0.411 -11.840 0.162 -2.581 0.791Block Investor -2.073 0.297 -2.944 0.285 -1.857 0.349Hedge Fund 1.346 0.500 0.878 0.749 0.531 0.827Reputable Agent -42.421*** 0.003Inverse Mills Ratio 14.156 0.179 38.742** 0.050Industry Dummies Yes Yes YesYear Dummies Yes Yes YesN 707 154 553Adjusted R-square (%) 9.44 31.48 11.27
Analyst Coverage and Stock Liquidity after PIPEs
Before After p-value
Full sample Mean Median Mean Median Mean Median
Analyst Coverage 2.11 1.00 2.60 2.00 0.000*** 0.000***
Turnover 0.18 0.11 0.19 0.12 0.395 0.063*
Spread 2.60 1.90 1.96 1.27 0.000*** 0.000***
Volatility (%) 6.04 5.59 5.33 4.83 0.000*** 0.000***
N 1148 1086
Analyst Coverage and Stock Liquidity after PIPEs: With Agents vs. Without Agents
Firms with Agents Firms without Agents p-value
Mean Median Mean Median Mean Median
Δ Analyst Coverage 0.73 0.00 0.27 0.00 0.001*** 0.000***
Δ Turnover 0.01 0.01 0.00 -0.00 0.785 0.058*
Δ Spread -0.78 -0.50 -0.51 -0.35 0.063* 0.009***
Δ Volatility (%) -0.92 -0.71 -0.53 -0.47 0.008*** 0.023**
N 671 415
Analyst Coverage and Stock Liquidity after PIPEs: With Agents vs. Without Agents
Firms with more reputable agents
Firms with less reputable agents p-value
Mean Median Mean Median Mean Median
Δ Analyst Coverage 0.90 1.00 0.69 0.00 0.361 0.112
Δ Turnover 0.03 0.00 0.00 0.01 0.390 0.800
Δ Spread -0.84 -0.56 -0.56 -0.34 0.038** 0.068*
Δ Volatility (%) -0.96 -0.74 -0.90 -0.71 0.437 0.406
N 147 524
The Distribution of Placement Agents’ Fees
0
20
40
60
80
100
120
140
160
180
(0, 1] (1, 2] (2, 3] (3, 4] (4, 5] (5, 6] (6, 7] (7, 8] (8, 9] (9, 10] (10, 11] (11, 12] >12
Agent Fees (%)
Nu
mb
er
of
De
als
Does More Reputable Placement Agent Charge a Fee Premium?
IV ApproachThe Lee Model
More Reputable Agents Less Reputable AgentCoefficient p-value Coefficient p-value Coefficient p-value
Intercept 5.856*** 0.000 7.209 0.108 7.474 0.000Ln (Analyst) -0.359** 0.034 -0.201 0.573 -0.488** 0.021EBITDA/Assets -0.032 0.793 0.371 0.721 -0.053 0.825B/M -0.086 0.475 -0.934 0.200 0.084 0.486Financial Leverage -0.145 0.799 2.209 0.396 -0.704 0.298Delist Dummy 0.041 0.881 0.250 0.748 0.165 0.640Ln (Volatility) 0.315 0.317 1.256 0.315 -0.028 0.938Warrants -0.091 0.673 0.336 0.586 -0.186 0.470Ln (Proceeds) -0.477** 0.012 -1.109** 0.050 -0.409* 0.059Future Business -0.372 0.115 -0.229 0.681 -0.457* 0.052Reputable Agent 0.320 0.855Inverse Mills Ratio -0.305 0.848 -0.535 0.674Industry Dummies Yes Yes YesYear Dummies Yes Yes YesN 707 154 553Adjusted R-square (%) 10.82 41.43 10.79
Discounts, Agent Fees, and Future Market Share
Market Share at t Market Share at t+1 Market Share at t+2
Coefficient p-value Coefficient p-value Coefficient p-value
Intercept 0.008* 0.069 0.013*** 0.002 0.013** 0.010
Market share (t-3, t-1) 0.196*** 0.000 0.061 0.155 0.120** 0.044
Mean discounts (t-3, t-1) -0.017** 0.016 -0.011 0.125 -0.008 0.341
Mean agent fee (t-3, t-1) 0.072 0.257 0.003 0.966 -0.014 0.851
N 209 145 96
Adjusted R-square (%) 10.67 0.77 2.47
Conclusions
How do firms and placement agents choose each other? There exists a positive assortative matching of placement agents and
issuing firms. Specifically, firms with more analyst coverage (less information asymmetry) and better profitability, and larger offers are associated with more reputable placement agents.
Do placement agents help lower transaction costs and information costs? Yes. Placement agents help lower PIPE discounts and improve
firms’ information environment and stock liquidity after the offering.
Do placement agents with strong reputations provide higher-quality service?
Yes. Firms associated with more reputable placement agents pay lower discounts to PIPE investors and obtain other benefits suggesting improved stock liquidity after the offering.
Conclusions (cont.)
Does the higher quality service enable placement agents to charge higher fees? No. More reputable placement agents do not appear to
charge higher fees than less reputable placement agents for their higher-quality service. However, their higher-quality service enables them to increase future market share.
Overall, our results support the model of Fernando, Gatchev, and Spindt (2005). Rather than fees per se, the quality of the issuing firm and the reputational concern of the placement agent are the key factors that drive the equilibrium in the PIPE market.