Priced Above the Range 17.408*** 17.466*** 16.859***
(1.569) (1.569) (1.560)
Income Modus Operandi -8.422*** -6.480** -6.872** -6.541**
(2.246) (2.091) (2.095) (2.070)
Growth Modus Operandi 5.132** 2.926 2.525 2.214
(1.915) (1.788) (1.798) (1.777) Private Equity (PE) Ownership -7.109* -0.693 -1.043 -0.064 -0.439
(2.982) (3.146) (2.919) (2.941) (2.903) PE IPOs in Decade 0.208 0.130 0.133 0.055 0.033 (0.208) (0.204) (0.190) (0.191) (0.188) PE Tie to Underwriter 2.599 1.797 1.532 1.262 1.288 (2.133) (2.084) (1.933) (1.937) (1.906) PE Fund Size (bn) 0.005 0.002 0.003 0.002 0.004 (0.008) (0.008) (0.007) (0.007) (0.007) Issuer Marketshare 0.558 0.590 0.797 0.874 0.928* (0.497) (0.484) (0.450) (0.450) (0.451) Sector Share of GDP 0.174 0.334 0.321 0.301 0.365 (0.310) (0.304) (0.282) (0.284) (0.315) Log Number of Firms in Sector -0.153 -0.597 -0.209 -0.048 -0.450
(0.866) (0.847) (0.787) (0.792) (0.881) Log Sector Labor Productivity ($ per employees) -3.452* -4.054** -3.018* -2.787* -3.109*
(1.403) (1.373) (1.277) (1.293) (1.490)
Litigation Risk -9.914** -10.558** -7.826* -7.945** -6.625*
(3.343) (3.261) (3.036) (3.042) (3.057) Shiller Investor Confidence Index 0.231 0.311 0.292 0.275 0.405
(0.226) (0.220) (0.205) (0.291) (0.288)
Number of IPOs in Month -0.098 -0.118 -0.107 -0.246* -0.251*
(0.100) (0.097) (0.090) (0.101) (0.100)
Dividend Premium -0.245 -0.188 -0.196 -0.185 -0.226
(0.179) (0.175) (0.162) (0.237) (0.234)
NYSE Monthly Turnover -1.272 -1.441 -0.443 2.192 2.233
(5.250) (5.136) (4.766) (6.839) (6.688)
Closed-End Fund Discount 0.086 0.270 0.273 0.253 0.166
(0.354) (0.347) (0.322) (0.400) (0.394) Equity Share of Total Issuance -12.286 -10.737 -5.834 -12.575 -10.766
(14.718) (14.374) (13.344) (14.593) (14.463)
Management Ownership 2.979 3.410 0.894 0.745 0.309
(2.534) (2.744) (2.556) (2.559) (2.538)
Secondary Portion -3.557 -3.775 -2.444 -2.485 -3.628
(2.688) (2.633) (2.446) (2.451) (2.415) Average Underwriter Status 1.888** 1.426* 1.213* 1.176* 0.880
(0.646) (0.644) (0.597) (0.599) (0.599) Dow Jones Industry Sector Index 0.286** 0.277** 0.125 0.150 0.130
Table 5 (Continued)
(1) (2) (3) (4) a (5) b
Market Volatility (VIX) 0.574** 0.438* 0.400* 0.727* 0.790** (0.222) (0.217) (0.202) (0.294) (0.289) Log S&P500 Index 20.201 24.217* 19.748* 58.891** 63.691***
(10.685) (10.434) (9.689) (17.920) (17.611) S&P500 One-day Return -0.483 -0.702 -0.580 -0.541 -0.445
(0.691) (0.674) (0.626) (0.631) (0.624) S&P500 One-month Return 0.620** 0.539** 0.507** 0.433* 0.464* (0.208) (0.203) (0.188) (0.194) (0.192) Log Offering Size (mm) 2.104* 3.488*** 1.410 1.593 1.905* (0.930) (0.959) (0.909) (0.916) (0.925)
Log Age (years) -1.872* -0.837 -0.305 -0.271 -0.289
(0.920) (0.911) (0.847) (0.850) (0.846) Debt to Cap (inverse for negative) -0.047 -0.013 -0.002 -0.000 -0.000
(0.090) (0.088) (0.081) (0.081) (0.081) Revenues (standardized) -3.252*** -3.511*** -2.903*** -2.881** -2.926***
(0.968) (0.944) (0.878) (0.881) (0.883) Operating Cashflow (standardized) 0.300 0.262 0.550 0.406 0.383
(0.707) (0.689) (0.640) (0.642) (0.637) Positive Earnings 3.123* 3.797* 2.147 1.807 1.904 (1.528) (1.493) (1.393) (1.398) (1.406) Employment Growth 0.136 0.145 0.071 0.073 0.086 (0.086) (0.083) (0.078) (0.087) (0.085) Recession Month 3.191 3.404 2.819 8.313* 7.670* (3.144) (3.065) (2.844) (3.416) (3.367) Industrial Production Growth 0.029* 0.028* 0.022 0.020 0.021
(0.014) (0.013) (0.012) (0.013) (0.013) Real Consumption Growth 0.007 0.010 0.009 0.006 0.006
(0.014) (0.013) (0.012) (0.013) (0.013) Intercept -136.712 -165.830* -141.196 -435.46** -473.47***
(85.564) (83.628) (77.622) (138.258) (136.230) RANDOM EFFECTS
SIC Log Variance 4.741
Residual Log Variance 15.627
Observations 791 791 791 791 791 R2 0.138 0.184 0.298 0.312 Adjusted R2 0.101 0.146 0.265 0.271 Log-likelihood -3395.83 -3374.42 -3314.62 -3306.56 -3274.12 AIC 6861.659 6822.829 6705.245 6707.111 6644.233 BIC 7025.224 6995.741 6882.830 6926.756 6868.551
Note: Unadjusted standard errors in parentheses; heteroskedasticity-consistent standard errors do not change any of the inferences. Missing data: list-wise deletion applied to 11 cases missing underwriter status, offer price, or first- day close price information (total sample 802). (PE = Private Equity).
a Models 4: Fixed effect models controlling for offering year with dummy variables. Offering years’ fixed effects not shown.
b
Model 5: Mixed effect model simultaneously clustering issuers by four-digit SIC codes (estimation fit using REML) and controlling for offering year fixed effect with dummy variables. Offering years’ fixed effects not shown.
Robustness Analyses
Clustering issuers by SIC code is important as issuers from different industry sectors may experience systematically different return outcomes. Fixed effect coefficient estimates may change substantially due to covariation with issuer industry sector random effects, such as industry specialization in the case of private equity and venture capital investors. For example, perhaps the Growth modus operandi exerts its effect primarily because venture capital firms disproportionately invest in technology companies, and technology companies as a group experience higher offer-price revisions? Furthermore, the effects of the modus operandi may vary by industry sector, with no fixed effect remaining after controlling for these random effects. While we have clustered issuers by four-digit SIC codes (in models 6 and 7 of Tables 3 and 4), we have not allowed the modus operandi coefficients to vary by cluster. Also, more aggregated sector clusters would simplify the analysis for potential bias due to industry specialization by investment firms. I use the ten main industry sectors for the Dow Jones sector indices to cluster issuers, conducting four additional pricing above the range hierarchical mixed-effect logit models (see Table 6). Model 1 clusters issuers by Dow Jones industry sector, model 2 further allows the Income coefficient to vary by sector, model 3 allows Growth to vary, and model 4 allows both modus operandi coefficients to vary. All models also control for offering year fixed effects. The mixed effect logit models with random slopes for the modus operandi variables support H1 and strongly confirm H2, with the fixed effect coefficient estimates for Income of -0.77 to -0.80 (p<0.05 for models 1 to 3; p<0.07 for model 4) and Growth of 0.93 to 1.00 (p<0.05 or p<0.01).
Table 6: Pricing Above the Range – Alternative Sector Clustering (10 year period 2001-2010)
(1) (2) (3) (4)
FIXED EFFECTS Professional Modus Operandi
Income -0.802* -0.769* -0.777* -0.768
(0.368) (0.384) (0.371) (0.417)
Growth 0.930** 0.937** 0.965* 0.995*
(0.310) (0.311) (0.378) (0.459) Power and Agency
Private Equity (PE) Ownership -0.046 -0.058 -0.075 -0.126 (0.517) (0.518) (0.520) (0.522) PE IPOs in Decade -0.005 -0.007 -0.002 -0.001 (0.033) (0.033) (0.033) (0.033) PE Tie to Underwriter 0.272 0.283 0.255 0.252 (0.347) (0.347) (0.350) (0.352) PE Fund Size ($ bn) -0.001 -0.001 -0.001 -0.001 (0.001) (0.001) (0.001) (0.001) Issuer Marketshare -0.227 -0.205 -0.207 -0.174 (0.210) (0.204) (0.205) (0.188) Sector Share of GDP 0.007 -0.008 0.005 -0.014 (0.056) (0.056) (0.055) (0.055) Log Number of Firms in Sector -0.223 -0.193 -0.211 -0.165
(0.147) (0.144) (0.143) (0.141) Log Sector Labor Productivity ($) -0.450* -0.405 -0.444* -0.387
(0.226) (0.224) (0.222) (0.223) Cost Substitution
Litigation Risk -1.699 -1.673 -1.662 -1.728
(1.071) (1.074) (1.075) (1.077) Non-Bayesian Investor
Shiller Investor Confidence Index 0.072 0.072 0.075 0.074 (0.051) (0.051) (0.051) (0.051) Number of IPOs in Month -0.001 -0.002 -0.001 -0.001
(0.018) (0.018) (0.018) (0.018)
Dividend Premium -0.039 -0.037 -0.042 -0.041
(0.042) (0.042) (0.042) (0.042) NYSE Monthly Turnover -1.387 -1.427 -1.407 -1.529
(1.315) (1.315) (1.321) (1.329) Closed-End Fund Discount 0.004 0.000 0.005 0.001
(0.072) (0.072) (0.072) (0.072) Equity Share of Total Issuance -5.117 -5.156 -4.944 -5.097
(3.803) (3.810) (3.785) (3.807) Non-SEU Issuer Management Ownership 1.038* 1.066* 1.027* 1.084* (0.420) (0.422) (0.420) (0.425) Secondary Portion -0.675 -0.662 -0.674 -0.643 (0.424) (0.425) (0.425) (0.425) Information Asymmetry
Average Underwriter Status 0.107 0.107 0.107 0.107 (0.114) (0.114) (0.114) (0.114) Market Conditions
Dow Jones Industry Sector Return 0.075*** 0.076*** 0.074*** 0.078*** (0.018) (0.018) (0.018) (0.018) Market Volatility (VIX) 0.050 0.054 0.049 0.054
Table 6 (Continued)
(1) (2) (3) (4)
Log S&P500 Index 4.210 4.338 4.166 4.320
(3.222) (3.236) (3.227) (3.258) S&P500 One-day Return -0.065 -0.071 -0.062 -0.066
(0.125) (0.125) (0.126) (0.127) S&P500 One-month Return -0.001 -0.002 -0.001 -0.003
(0.036) (0.036) (0.036) (0.036) Fundamentals & Macroeconomic
Log Offering Size (mm) 0.803*** 0.808*** 0.808*** 0.852*** (0.162) (0.163) (0.162) (0.162)
Log Age (years) -0.196 -0.197 -0.185 -0.182
(0.145) (0.145) (0.145) (0.145) Debt to Cap (inverse for negative) -0.016 -0.016 -0.021 -0.023
(0.053) (0.052) (0.055) (0.057) Revenues (standardized) -0.179 -0.199 -0.184 -0.230
(0.195) (0.196) (0.193) (0.192) Operating Cashflow (standardized) -0.104 -0.108 -0.101 -0.112
(0.103) (0.103) (0.103) (0.104) Positive Earnings 0.667** 0.666** 0.620* 0.641* (0.249) (0.249) (0.249) (0.249) Employment Growth 0.029 0.030 0.031 0.031 (0.016) (0.016) (0.016) (0.016) Recession Month 0.689 0.703 0.650 0.735 (0.687) (0.690) (0.688) (0.695) Industrial Production Growth 0.003 0.003 0.003 0.003
(0.003) (0.003) (0.003) (0.003) Real Consumption Growth -0.000 -0.001 -0.000 -0.001
(0.003) (0.003) (0.003) (0.003)
Intercept -35.297 -37.129 -35.430 -37.853
(24.668) (24.785) (24.714) (24.963) RANDOM EFFECTS
Sector Standard Deviation 0.393 0.463 0.327 0.545
Income Standard Deviation 0.312 0.547
Growth Standard Deviation 0.536 0.909
Observations 800 800 800 800 McFadden Pseudo- R2 0.195 0.210 0.210 0.227 Nagelkerke Pseudo- R2 0.282 0.302 0.302 0.323 Log-likelihood -326.548 -320.411 -320.390 -313.676 AIC 745.097 736.822 736.779 729.353 BIC 960.589 961.683 961.640 968.268
Note: Standard errors in parentheses adjusted for clustering on sectors. Mixed effect logit models controlling for offering year fixed effects and clustering issuers by ten Dow Jones Industry Sectors: Basic Materials, Consumer Goods, Consumer Services, Industrials, Oil and Gas, Financials, Healthcare, Technology, Telecommunications, and Utilities. Estimation fit using adaptive Gauss-Hermite approximation with 7 integration points. Offering years’ fixed effects not shown. Missing data: list-wise deletion applied to 2 cases missing underwriter status or pricing above the range information (total sample 802). McFadden and Nagelkerke pseudo-R2 calculated on log likelihood of -405.8
for null model. (PE = Private Equity).
I also conduct the same analyses with offer-price revisions (see Table 7). The hierarchical mixed-effect OLS models with random slopes for the modus operandi variables strongly confirm hypotheses H1 and H2, with the fixed effect coefficient estimates for Income (-3.8 to -3.9; p<0.05) and Growth (3.5 to 4.0; p<0.05) retaining significance throughout. Hence, the impact of modus operandi on first-stage outcomes is not due to sector-level variation in issuers.
Table 7: Offer-price Revision – Alternative Sector Clustering (10 year period 2001-2010)
(1) (2) (3) (4)
FIXED EFFECTS Professional Modus Operandi
Income -3.802* -3.902* -3.827* -3.823*
(1.589) (1.632) (1.583) (1.583)
Growth 3.486* 3.568** 3.976* 3.976*
(1.367) (1.371) (1.559) (1.566) Power and Agency
Private Equity (PE) Ownership -0.544 -0.649 -0.689 -0.685 (2.241) (2.238) (2.237) (2.237) PE IPOs in Decade -0.043 -0.048 -0.034 -0.033 (0.146) (0.146) (0.146) (0.146) PE Tie to Underwriter 0.494 0.582 0.466 0.457 (1.482) (1.479) (1.480) (1.481) PE Fund Size ($ bn) 0.001 0.002 0.001 0.001 (0.005) (0.005) (0.005) (0.005) Issuer Marketshare -0.240 -0.229 -0.220 -0.220 (0.345) (0.346) (0.343) (0.343) Sector Share of GDP 0.045 0.024 0.083 0.086 (0.232) (0.229) (0.227) (0.227) Log Number of Firms in Sector -0.217 -0.174 -0.283 -0.288
(0.645) (0.640) (0.632) (0.632) Log Sector Labor Productivity ($) -2.698** -2.585* -2.605* -2.613*
(1.041) (1.042) (1.026) (1.025) Cost Substitution
Litigation Risk -3.032 -2.944 -2.792 -2.801 (2.328) (2.334) (2.328) (2.328) Non-Bayesian Investor
Shiller Investor Confidence Index 0.091 0.087 0.099 0.099 (0.223) (0.223) (0.222) (0.222) Number of IPOs in Month -0.111 -0.114 -0.118 -0.118
(0.077) (0.077) (0.077) (0.077) Dividend Premium -0.172 -0.166 -0.172 -0.173
(0.180) (0.179) (0.179) (0.179) NYSE Monthly Turnover -0.918 -0.900 -0.697 -0.697
(5.222) (5.217) (5.221) (5.222) Closed-End Fund Discount -0.090 -0.093 -0.084 -0.084
Table 7 (Continued)
(1) (2) (3) (4)
Equity Share of Total Issuance -18.022 -17.636 -18.111 -18.141 (11.117) (11.112) (11.095) (11.095) Non-SEU Issuer Management Ownership 4.467* 4.513* 4.460* 4.456* (1.944) (1.945) (1.938) (1.938) Secondary Portion -3.005 -3.004 -3.048 -3.049 (1.870) (1.868) (1.867) (1.867) Information Asymmetry
Average Underwriter Status 0.099 0.082 0.107 0.108 (0.460) (0.460) (0.459) (0.459) Market Conditions
Dow Jones Industry Sector Return 0.268*** 0.273*** 0.262*** 0.261*** (0.071) (0.071) (0.071) (0.071) Market Volatility (VIX) 0.197 0.207 0.199 0.198
(0.224) (0.224) (0.224) (0.224) Log S&P500 Index 29.570* 30.065* 29.984* 29.929*
(13.693) (13.684) (13.673) (13.673) S&P500 One-day Return -1.170* -1.177* -1.200* -1.200*
(0.500) (0.500) (0.500) (0.500) S&P500 One-month Return 0.079 0.083 0.087 0.087
(0.150) (0.150) (0.150) (0.150) Fundamentals
Log Offering Size (mm) 5.527*** 5.546*** 5.522*** 5.522*** (0.707) (0.707) (0.700) (0.700) Log Age (years) -1.805** -1.793** -1.758** -1.758**
(0.657) (0.655) (0.653) (0.653) Debt to Cap (inverse for negative) 0.041 0.034 0.033 0.033
(0.062) (0.062) (0.062) (0.062) Revenues (standardized) -1.690* -1.736** -1.679* -1.676*
(0.672) (0.673) (0.671) (0.671) Operating Cashflow (standardized) -0.766 -0.783 -0.735 -0.732
(0.489) (0.489) (0.488) (0.488) Positive Earnings 5.072*** 5.039*** 4.949*** 4.950*** (1.091) (1.093) (1.094) (1.094) Macroeconomic Employment Growth 0.115 0.120 0.129* 0.129* (0.066) (0.066) (0.066) (0.066) Recession Month 0.787 0.789 0.754 0.752 (2.617) (2.615) (2.614) (2.615) Industrial Production Growth 0.009 0.009 0.008 0.008
(0.010) (0.010) (0.010) (0.010) Real Consumption Growth -0.005 -0.005 -0.004 -0.004
(0.010) (0.010) (0.010) (0.010) Intercept -207.620* -212.776* -211.982* -211.492*
Table 7 (Continued)
(1) (2) (3) (4)
RANDOM EFFECTS
Sector Standard Deviation 2.312 2.648 1.541 1.485
Income Standard Deviation 1.096 0.101
Growth Standard Deviation 1.815 1.856
Residual Standard Deviation 12.393 12.385 12.376 12.377
Observations 795 795 795 795
Log-likelihood -3099.187 -3098.800 -3097.635 -3097.633
AIC 6292.374 6295.599 6293.270 6299.265
BIC 6512.256 6524.838 6522.509 6542.539
Note: Standard errors in parentheses adjusted for clustering on sectors. Mixed effect OLS models controlling for offering year fixed effects and clustering issuers by ten Dow Jones Industry Sectors: Basic Materials, Consumer Goods, Consumer Services, Industrials, Oil and Gas, Financials, Healthcare, Technology, Telecommunications, and Utilities. Estimation fit using REML. Offering years’ fixed effects not shown. Missing data: list-wise deletion applied to 7 cases missing underwriter status or pricing above the range information (total sample 802). (PE = Private Equity).
*p < 0.05, **p < 0.01, ***p < 0.001 (two-tailed tests)
We can also utilize SEC filing rules to mitigate another source of potential bias: unforeseen circumstances arising during the roadshow. Underwriters must re-file with the SEC if the offer is priced more than 20 percent outside the price range; issuers and underwriters seek to avoid re- filing as it significantly delays the IPO process. As such, issuers pricing more than 20 percent outside the price range are doing so due to unforeseen circumstances. While not outliers in the statistical sense, these extreme situations could nevertheless unduly influence our coefficient estimates. I conduct two-limit Tobit regressions restricting the offer-price return to +/-30 percent to test the robustness of our inferences to these situations and the exogenous shocks they may represent (see Table 8). For the majority of price ranges in the data, +/-30 percent offer-price returns approximates pricing more than 20 percent outside the price range. Again, the analyses strongly confirm hypotheses H1 and H2. Of particular note, the proxy for prospect theory (secondary portion) becomes significant once we limit the response variable.
Table 8: Offer-Price Revision – Tobit Models (10 year period 2001-2010)
(1) a (2) b FIXED EFFECTS
Professional Modus Operandi
Income -4.286** -4.126**
(1.424) (1.423)
Growth 2.892* 3.013*
(1.223) (1.227) Power and Agency
Private Equity (PE) Ownership -1.499 -0.903 (2.009) (2.004) PE IPOs in Decade -0.0111 -0.0399 (0.131) (0.131) PE Tie to Underwriter 1.084 0.831 (1.323) (1.325) PE Fund Size ($ bn) 0.00223 0.00105 (0.00477) (0.00476) Issuer Marketshare -0.160 -0.176 (0.311) (0.309) Sector Share of GDP 0.118 0.0313 (0.217) (0.208) Log Number of Firms in Sector -0.482 -0.314
(0.605) (0.577) Log Sector Labor Productivity ($) -2.537* -2.451**
(1.021) (0.932) Cost Substitution
Litigation Risk -2.114 -2.390
(2.112) (2.114) Non-Bayesian Investor
Shiller Investor Confidence Index 0.196 0.208 (0.200) (0.200) Number of IPOs in Month -0.0930 -0.0850
(0.0691) (0.0692)
Dividend Premium -0.131 -0.176
(0.162) (0.161) NYSE Monthly Turnover -1.831 -1.675
(4.655) (4.680) Closed-End Fund Discount -0.0283 -0.0813
(0.274) (0.274) Equity Share of Total Issuance -18.06 -18.39
(10.02) (9.944) Non-SEU Issuer Management Ownership 2.596 2.685 (1.749) (1.744) Secondary Portion -4.214* -4.175* (1.677) (1.675) Information Asymmetry
Average Underwriter Status 0.0921 0.110 (0.415) (0.412) Market Conditions
Dow Jones Industry Sector Return 0.257*** 0.268*** (0.0634) (0.0639) Market Volatility (VIX) 0.114 0.172
Table 8 (Continued)
(1) a (2) b
Log S&P500 Index 22.88 23.53
(12.24) (12.27) S&P500 One-day Return -1.047* -0.943*
(0.450) (0.447) S&P500 One-month Return 0.0898 0.0632 (0.135) (0.134) Fundamentals
Log Offering Size (mm) 4.946*** 4.777*** (0.631) (0.636)
Log Age (years) -1.630** -1.594**
(0.585) (0.588) Debt to Cap (inverse for negative) 0.0210 0.0430 (0.0561) (0.0553) Revenues (standardized) -1.529* -1.567** (0.610) (0.601) Operating Cashflow (standardized) -0.655 -0.637
(0.441) (0.437) Positive Earnings 5.082*** 5.138*** (0.979) (0.984) Macroeconomic Employment Growth 0.0980 0.0962 (0.0587) (0.0589) Recession Month 2.767 2.732 (2.350) (2.349) Industrial Production Growth 0.0135 0.0121
(0.00897) (0.00893) Real Consumption Growth -0.00566 -0.00443
(0.00883) (0.00874)
Intercept -166.2 -175.2
(94.61) (94.69) RANDOM EFFECTS
Sector Standard Deviation (a) (b) 3.126 2.027 Residual Standard Deviation 10.83 11.07
Observations 795 795
Log-likelihood -2977.5 -2976.6
AIC 6049.1 6047.1
BIC 6268.9 6267.0
Note: Standard errors in parentheses adjusted for clustering on sectors. Two-limit Tobit models with upper and lower bounds of +/-30 percent controlling for offering year fixed effect. Issuers must refile with the SEC if they price the offering more than 20 percent outside the price range; with the bounds of the price range primarily between $10 to $20, this approximates an offer-price return of +/-30 percent. Offering years’ fixed effects not shown. Missing data: list-wise deletion applied to 7 cases missing underwriter status or pricing above the range information (total sample 802). (PE = Private Equity).
a Model 1: Two-limit Tobit model controlling for offering year fixed effects and clustering issuers by four-digit SIC codes.
b
Model 2: Two-limit Tobit model controlling for offering year fixed effects and clustering issuers by ten Dow Jones Industry Sectors: Basic Materials, Consumer Goods, Consumer Services, Industrials, Oil and Gas, Financials, Healthcare, Technology, Telecommunications, and Utilities.
In sum, the findings are decidedly mixed for the alternative strategic and behavioral
explanations for all six sets of regressions, but in particular for our primary response variable of interest: first-stage outcomes (parameterized as either pricing above the range or offer-price revisions). Table 9 summarizes the key findings for the alternative hypotheses tested in the final models for first-stage outcomes (models 6 and 7 in Tables 3 and 4, and all models from Tables 6 through 8). Secondary selling weakly supports non-SEU issuer hypotheses based on prospect theory. Agency, litigation-risk, and information asymmetry hypotheses all find no support in the data. While the data does not support hypotheses related to the power of the private equity firm, it does indicate that issuers from industry sectors with high labor productivity exhibit lower underpricing, supporting a strategic action field perspective. The findings corroborate generic rational risk-aversion hypotheses for market conditions and issuer fundamentals. Finally, the data strongly supports the modus operandi hypotheses.
Table 9: Predictions of Alternative Hypotheses for First-Stage Outcomes
Theory Coefficient Prediction Coefficient Estimates
Professional Modus Operandi Income < 0 Growth > 0
Income < 0 Growth > 0 Power and Agency PE Indicatorsa < 0
Marketshare < 0 Industry Size < 0 Labor Productivity < 0 PE Indicatorsa = 0 Marketshare = 0 Industry Size = 0 Labor Productivity < 0 Substitution Costs Litigaton Risk > 0 Litigation Risk ≤ 0 Non-Bayesian Investor Shiller Index > 0
Baker-Wurgler Index > 0 Sentiment Proxiesb > 0
Shiller Index = 0
Baker-Wurgler Index = 0 Sentiment Proxiesb = 0 Non-SEU Issuer Management Ownership < 0
Secondary Portion < 0
Management Ownership ≥ 0 Secondary Portion ≤ 0 Information Asymmetry Underwriter Status < 0 Underwriter Status = 0 Risk-aversion Dow Jones Sector Returns > 0
Market Volatility > 0 S&P 500 Index > 0 S&P 500 Returns > 0
Dow Jones Sector Returns > 0 Market Volatility = 0
S&P 500 Index ≥ 0 S&P 500 Returns ≤ 0 Fundamentals Log Offering Size < 0
Log Age < 0 Debt to Cap > 0 Revenues < 0
Operating Cashflow < 0 Positive Earnings < 0
Log Offering Size > 0 Log Age ≤ 0
Debt to Cap = 0 Revenues < 0
Operating Cashflow = 0 Positive Earnings > 0
Note: Coefficient estimates refer to estimates across final models for first-stage outcomes controlling for offering year fixed effects and issuer industry sector random effects (models 6 and 7 in Tables 2 and 3, and all models from Tables 5 to 7). If coefficient estimate is never significant in any final model, then coefficient is regarded as equal to zero. Inequality signs refer to coefficient estimates that are statistically significantly in most final models. Finally, ≥ and ≤ refer to coefficient estimates that are consistently greater than or less than zero, respectively, but whose statistical significance is oftentimes above the p<0.05 level.
a Private Equity (PE) indicators are: PE ownership (dichotomous), PE IPOs undertaken in decade, PE ties to underwriters, PE fund size (US$ billions).
b Other proxies for investor sentiment are: number of IPOs, dividend premium, NYSE turnover, closed-end fund discount, and equity share. Dividend premium and closed-end fund discount should vary inversely with investor sentiment.
DISCUSSION
Calculative rationality viewed in isolation from professional culture cannot explain these findings. Shareholders always gain from lower first-day returns. Obviously, Growth investors have as much to gain from lower first and second-stage returns as do Income investors: the extra proceeds could be invested in growing the issuer’s business, generating stronger future growth and present values. We could question whether strategic constraints account for the pricing differences between Income and Growth investors. However, the main strategic differences between Income and Growth investors noted by market participants—secondary selling and financial leverage (debt levels)—do not diminish the extremely significant differences between the two groups. Differences in investment horizon could also impact strategic interests. For instance, if Growth investors hold onto investments longer post-IPO than Income investors, they may care less about short-term returns. However, the opposite is actually true, with venture capital firms selling shares post-IPO faster than any other institutional shareholding group (Field and Hanka 2001). Furthermore, Growth investors actually do not gain from higher first-day returns with regard to longer-term share price growth. Econometric studies repeatedly show that first-day returns are negatively correlated with longer-term returns (Ritter 1991; Krigman, Shaw and Womack 1999; Ritter and Welch 2002): acquiescing to the IPO discount is detrimental both in the immediate and longer-term for issuers and their backers. Also, lock-up expiration
provisions cannot account for differences between the two sets of investors. The regulations encouraging the industry practice of lock-up provisions for large pre-IPO institutional investors apply equally to both Income and Growth investors, with lock-up provisions becoming
and Hanka 2004).15 Hence, differences in selling shortly after the IPO cannot explain the divergence in outcomes as lock-up provisions prevent both sets of investors from doing so. Power and agency theories tested in the models share a close affinity with strategic explanations for variation in IPO pricing, but also fail to explain the dramatic differences between Income and Growth investors, whether ties to underwriters or negotiating power due to the size of funds under management. We have also controlled for offering year fixed effects, so differences in the timing of IPOs by year of offering and all annualized proxies cannot account for the divergence. Finally, we have controlled for the issuer’s industry classification, whether fine-grained at the four-digit SIC level or aggregated at the Dow Jones industry sector level, ruling out strategic differences due to choice of industry specialization by private equity and venture capital firms.
I argue that professional modus operandi qualifies financial pricing. Because of enduring professional cultures, rational actors may act in ways harmful to their self-interest in situations where their original routines no longer apply. As we have demonstrated, the enduring effect of modus operandi on IPO prices cannot be attributed to differences in holding period, sales restrictions, industry specialization, year of offering, negotiating power, issuer debt levels, secondary selling, or broader market and macroeconomic conditions. I suggest that there is an irreducible cultural component to price determination alongside the irreducible irrational component highlighted by behavioral finance (Shiller 2003). Professional routines shape how actors pursue their interests, with these constitutive effects being mutually generative with economic rationality (DiMaggio 1994). Calculative actors take into account perceived responses
15 Rule 144 applies equally to all 5 percent blockholders and investors holding shares issued outside a registered
offering (e.g., pre-IPO shares). Such shares cannot be sold until a one-year holding period has elapsed (which may already be satisfied at the time of the IPO). There are numerous regulations and reporting requirements for the sale of such shares even after the company goes public. Hence all private equity and venture capital investments pre-IPO are subject to Rule 144. Underwriters do not distinguish between private equity and venture capital firms in trying to secure a lock-up provision for the IPO.
and constraints in the environment to develop a course of action. Here, investment firms attempt to generate investment gains in the best way possible given their perception of the problem gained through the colligation process. Both Income and Growth investors are trying to
maximize value from the IPO, but they act differently based on their inferences. These inference methods may themselves originate in the rational strategies of these investment firms, but have become an enduring influence independent of those original strategies. Behavioral theories rely on individualistic cognitive biases and do not adequately explain first-stage outcomes. Pricing above the range dwarves all other effect sizes in determining first-day returns. The social influence of first-stage outcomes on first-day returns is itself a sociological rather than an atomized actor (whether rational or non-rational) outcome. Market struggles entail purposive actors influenced by their perceptions of how other actors and objects will react, with such perceptions being socially shared within professional groups. Importantly, professional culture does not override strategic concerns, but instead intertwines with calculative rationality. IPO returns are the joint outcome of socio-cultural (social valuation and professional modus
operandi), strategic (social skill), behavioral (prospect theory), and rational risk-aversion (market conditions and issuer fundamentals) processes.
CONCLUSION
The pricing of corporate securities is a professional task, with multiple professional groups