Presentation on theme: "NTU – Taipei, March 26, 2013 1. 2 MORE THAN CONNECTEDNES- HETEROGENEITY OF CEO SOCIAL NETWORKS AND FIRM VALUE Iftekhar Hasan (with Bill Francis."— Presentation transcript:
NTU – Taipei, March 26,
2 MORE THAN CONNECTEDNES- HETEROGENEITY OF CEO SOCIAL NETWORKS AND FIRM VALUE Iftekhar Hasan (with Bill Francis and Yiwei Fang) Fordham University and Bank of Finland 2
3 MORE THAN CONNECTEDNES- HETEROGENEITY OF CEO SOCIAL NETWORKS AND FIRM VALUE Iftekhar Hasan (with Bill Francis and Yiwei Fang) Fordham University and Bank of Finland 3
4 Focus of the paper Social Network Heterogeneity of Top Management and Its Potential Impact on Firm Performance?
Social Networks Social network is one of the most striking phenomena of modern society. LinkedIn, a social networking website for people in professional occupations, has more than 175 million users in around 200 countries and territories. Facebook, for another example, has more than 1 billion active users all over the world as of September, Its IPO in May 2012 was valued at $104 billion, the largest valuation to date for a newly listed public company. 5
6 Trends of Globalization 6
7 Heterogeneity of Corporate Hierarchy Figure 1: Increase of foreign-born managers 7
CEO Social Networks CEOs have extensive social networks, e.g., they meet people from alumni event, workplace, conferences, and country clubs. It is considered as an important social capital, where one can draw resources from others, obtain business opportunities, and learn new market information. 8
9 Economics of Diversity Issues of ethnicity, culture and human diversity have been at the forefront of behavioral and social science research. Heterogeneous social connections create more opportunities (Granovetter,1973 AJS; Burt, 1992). The role of social capital in the creation of human capital (Coleman, 1988 AJS; Uzzi, 1996 ASR). 9
10 Economics of Diversity Psychologists emphasize that differences in individual attributes is important for knowledge creation, leadership creativity, and firm innovation (Barron and Harrington, 1981 ARP; Amabile, 1988; Bassett‐Jones, 2005 CIM) Most groundbreaking ideas are found at the intersections of diverse fields, industries, disciplines, and cultures (Frans Johansson, 2006, The Medici Effect). 10
11 Economics of Diversity Moderate population diversity promotes economic development, but too much diversity inhibits communication and cooperation (Lazear, 1999 JPE; Alesina et al., 2000 AER; Ashraf and Galor, 2011 AER) 11
Impacts on the Financial World 12
13 What do we do? We look at the impact of CEO social network heterogeneity on firm performance. Social networks Whom you know from school, work, club, charity, army, government, and other non-profit associations. Heterogeneity of social ties Demographic (e.g. gender, ethnicity) Intellectual (e.g. degree, major, school ) Professional (e.g. industry, managerial expertise) International experience (e.g. connections to foreign companies)
14 Research Questions Does heterogeneity of CEO social network add value to corporations ? If so, through which channels ? 14
Key Findings We find that CEO social network heterogeneity has a positive impact on firms’ Tobin's Q. Greater CEO social network heterogeneity also leads to (i) more innovation, (ii) more foreign sale growth, (iii) higher investment efficiency, (iv) better M&A performance, and (v) lower cost of financing.
16 Related Academic Work on Demographic Diversity Gender and Ethnic Diversity Matter Males and females differ in risk appetite and work attitude (Barber and Odean, 2001, QJE; Hillman, Shropshire, and Cannella, 2007, AMJ). Different ethnic groups differ in beliefs and cognitive functioning, which could provide a broader view and more alternative solutions to the questions (Peffer and Salancik, 1978; Carter et al., 2003). 16
17 Related Academic Work on Diversity Intellectual and Professional Diversity Matter Different educational and professional backgrounds provide a diverse range of expertise, which enhances problem solving capability (Rodan and Galunic, 2004 SMJ; Bassett-Jones, 2005 CIM). Greater diversity in the board room can bring informational richness to the discussion and improve firm performance (Adams and Ferreira, 2009 JFE; Anderson et al., 2011 FM). 17
18 Our Contribution Our study attempts to bring a new perspective to this debate by focusing on the value implication of the diversity of social networks. Importantly, it looks at the heterogeneity within the diversity of social networks. 18
19 Our Hypotheses 19
Data and Sample Primary data BoardEx (Management Diagnostics Limited) CEOs’ social ties with schoolmates, colleagues, and other connections through club memberships, charities, army, government, etc. Biographical information of CEOs and their connections, including gender, nationality, education, and working experience Other data Compustat; NBER Patent; SDC; CRSP; ExecuComp Our final sample consists of 2216 firms’ 3100 CEOs over
25 Limitation of the data The social network diversity measures are still limited to professionals, rather than everyone in a society. We can only trace the exposure but not the real depth or intensity of the relationship. 25
Construction of social networks 26 Step 2: Check their education background. School ties are built if two people went to the same school within 3 years of each other Step 2: Check their education background. School ties are built if two people went to the same school within 3 years of each other Step 3: Check their work experience Work ties are built if two people used to worked at the same company same year Step 3: Check their work experience Work ties are built if two people used to worked at the same company same year Step 1: Our network consists of senior mangers, executives and board of directors of all US companies identified in BoardEx Step 4: Check their social activities (e.g. club, charity, army, government)? Other ties are built if two people did social activities at the same organizations Step 4: Check their social activities (e.g. club, charity, army, government)? Other ties are built if two people did social activities at the same organizations Step 5: In the end, we only look at CEOs and calculate their social network heterogeneity based on who they know.
27 An example of Coca-Cola CEO (Muhtar Kent) 27 NameGenderNationalityDegreeMajorOccupationCompanyCompany CountryIndustry Michel Naquet-RadigMFrenchMBA BoardDirectorefes breweries international nvNetherlands Hon. John KornblumMAmericanCEOlazard & co gmbhUnknown Hon. Mrs AlexisFAmericanBSBoardDirectornew venture incUnited States Tom PritzkerMAmericanJDLawCEOhyatt hotels corpUnited States Richard WolfordMAmericanCEOdel monte foods coUnited States2000 Andrew LiverisMAmericanBSCEOdow chemical coUnited States2821 Tim ShriverMAmericanPhDBoardDirectorspecial olympics internationalUnited States Senator Bill Frist SrMAmericanMDBoardDirectorcressey & company lpUnited States Sir David LoganMAmericanBoardDirectorefes breweries international nvNetherlands TuncayÖzilhanMTurkishMBA BoardDirectorefes sinai yatirim hldgs asTurkey2086 Metin TokpinarMDutchMSBoardDirectorefes sinai yatirim hldgs asTurkey2086 Ibrahim YaziciMTurkishMBA BoardDirectorefes sinai yatirim hldgs asTurkey2086 Hursit ZorluMDutchBSBoardDirectorefes sinai yatirim hldgs asTurkey2086 DemirArmanMTurkishMSFinanceBoardDirectorefes breweries international nvNetherlands Doctor Ali TigrelMDutchPhDBoardDirectorefes breweries international nvNetherlands Christos-Alexis KomMGreekBoardDirectorshelman saGreece Doctor Nakedi PhosaMPhDBoardDirectorbraemore resources plcUnited Kingdom1000 Doctor Helene GayleFAmericanMDCEOcare usaUnited States David BuceyMAmericanJDLawBoardDirectoramedisys incUnited States8082
The measure of CEO social network heterogeneity
30 Increasing trend of CEO social network heterogeneity (S&P1500) 30
Table 1: summary statistics of CEO social network heterogeneity (HHI measures) Panel A: By Industry 1-digit SIC SIC1HHI-demographicHHI-intellectualHHI-professionHHI-internationalHHI-overall Panel B: By High tech (High tech=1 if SIC2=48, SIC2=73, SIC3==283) High TechHHI-demographicHHI-intellectualHHI-professionHHI-internationalHHI-overall Panel C: By R&D (R&D=1 if R&D expenditure>0) R&DHHI-demographicHHI-intellectualHHI-professionHHI-internationalHHI-overall Panel D: By foreign business (Multinational=1 if foreign revenue>0) MultinationalHHI-demographicHHI-intellectualHHI-professionHHI-internationalHHI-overall CEOs of high tech-, R&D, and multinational firms tend to have lower HHI (higher heterogeneity).
Other variables Size of the network Various firm characteristics – Size, leverage, capx, cash flow, R&D, Tobin’s Q, innovation, foreign sale growth Board diversity – Female ratio and minority ratio CEO characteristics – Age, gender, ethnicity, education background, and past work experience CEO turnover – CAR, outside hire/inside hire, past experience M&A characteristics – CAR, BHAR, Run-up, diversified M&A, tender offer, payment method, relative size, target type, acquirer financial variables
Sample description—Firm characteristics 33 Firm and board characteristics VariableNMeanS.DMinMedianMax Q Size (million) Leverage Capex Total capital expenditure ratio Acquisition expenditure ratio Cashflow R&D intensity HiTecPharma Multinational Innovation (patent) Foreign sales growth Board female ratio Board minority ratio
Sample description—CEO characteristics 34 CEO characteristics VariableNMeanS.DMinMedianMax Age Female CEO Minority CEO MBA PhD Ivy school graduate Work mobility Oversea experience CEO social network measures Het-demographic Het-intellectual Het-profession Het-international Het_overall Centrality
Endogeneity concerns The relationships between CEO SNH and firm value can be spurious due to the possibility that – (1) better performing firms can provide CEOs opportunities to meet more people and different people (reverse causality) – (2) certain firm characteristics can simultaneously affect CEOs’ choice of social network and firm value (simultaneity bias). We employ several methods to address this issue.
CEO SNH and Firm Value: Simultaneous Equations Approach Intuition – CEO SNH could be capturing the effect of different firm types and previous performance that are correlated with firm value. – We run two stage least square equations model to correct the spurious relationships, using CEO’s personal background and experience variables as exogenous variables. Model specification: 36 CEO SNH i,t =α 0 + α (Q i,t-1, Firm char. i,t-1 ) + (CEO char. ) i, t + i,t (1-1) Q i,t = β 0 + β (Predicted SNH) i,t + (Firm char.) i, t + i,t (1-2) CEO SNH i,t =α 0 + α (Q i,t-1, Firm char. i,t-1 ) + (CEO char. ) i, t + i,t (1-1) Q i,t = β 0 + β (Predicted SNH) i,t + (Firm char.) i, t + i,t (1-2)
Baseline model: 1 st stage- Determinants of CEO network heterogeneity 37 (1)(2)(3)(4)(5) VARIABLESHet - demographicHet - intellectualHet - professionHet - internationalHet - overall Firm characteristics (t-1) Log(assets)0.010***0.040***0.026***0.004***0.021*** Leverage ***-0.046*-0.015*-0.034*** Tobin's Q **0.008**0.005***0.006*** Capextoasset * Cashflow **0.005 RDtoasset0.049*0.213***0.323*** *** High tech0.018** ***0.011*0.042*** Multinational *** * Board female ratio0.054***0.058*-0.122**0.039**0.010 Board minority ratio0.109*** ** CEO characteristics Centrality ***80.988*** *** Female0.042***0.084*** ***0.045*** Minority ***0.052***0.071*** Log (age)-0.001*** *** *** MBA0.027***0.057***0.053***0.081***0.054*** PhD0.020***0.138***0.036***0.098***0.073*** Ivy school graduate0.084***-0.067*** ***0.010** Work mobility 0.002***0.015***0.027***0.002***0.011*** Oversea experience 0.055*** **0.014* Constant1.102***0.988***1.376***1.360***1.207*** Year dummiesYes Industry dummiesYes Observations8,430 Adjusted R-squared
Baseline model: 2 nd Stage: CEO network heterogeneity and firm value 38 (1)(2)(3)(4)(5)(6)(7)(8) VARIABLESTobin's Q Network char Het_demographic1.697*** Het_intellectual0.725*** Het_profession1.155*** Het_international2.430*** Het_overall2.180***2.432***2.216***2.386*** Centrality ***99.195**51.628** ***26.529** Interaction with CEO char Het_overall*MaleCEO2.698*** Het_overall*AmericanCEO2.122*** Het_overall*NonIvySchoolCEO1.986*** Firm characteristics (t-1) Log(assets)-0.094***-0.104***-0.103***-0.100***-0.120***-0.126***-0.120***-0.116*** Leverage-0.639***-0.650***-0.637***-0.627***-0.618***-0.595***-0.618*** Capextoasset0.802***0.897***0.598*0.881***0.761**0.702**0.794***0.786*** Cashflow2.104***2.122***2.097***2.003***2.102***2.110***2.100***2.115*** RDtoasset5.292***5.226***4.998***5.303***5.004***4.966***5.010***5.058*** High tech0.401***0.406***0.262***0.365***0.312***0.300***0.309***0.288*** Multinational-0.088***-0.079***-0.129***-0.089***-0.103***-0.108***-0.104***-0.109*** Board female ratio0.508***0.586***0.776***0.458***0.567***0.646***0.568***0.561*** Board minority ratio **0.194** MaleCEO-3.958*** AmericanCEO-3.295*** NonIvySchoolCEO-2.912*** Constant0.644*1.730***0.924***-0.919** Year dummiesYes IndustrydummiesYes Observations8,430 Adjusted R-squared
Exploring economic significance 39 CEO SNHCoefficients % increase in Tobin's Q Dollar amount (mil) increase in market value as a result of 10% increase in CEO SHN Calculation Het-demographic1.679***9.21%278.5 [10% * Mean (Het-demographic) * β (Het- demographic) / Mean (Q)]*Assets = (0.1*1.119*1.679/2.041)* =278.5 Het-intellectual0.725***5.11%154.6 [10% * Mean (Het-intellectual) * β (Het- intellectual) / Mean (Q)]*Assets = (0.1*1.439*0.725/2.041)* =154.6 Het-profession1.155***9.95%301.2 [10% * Mean (Het-profession) * β (Het- professional) / Mean (Q)]*Assets = (0.1*1.760*1.155/2.041)* =301.2 Het-international2.430***17.99%554.0 [10% * Mean (Het-international) * β (Het- international) / Mean (Q)]*Assets = (0.1*1.511*2.430/2.042)* =554.0 Het-overall2.180***15.56%407.6 [10% * Mean (Het-overall) * β (Het-overall) / Mean (Q)]*Assets = (0.1*1.457*2.180/2.042)* =407.6 Centrality **3.98%12.0 [10% * Mean (Centrality) * β (Centrality) / Mean (Q)]*Assets = (0.1*0.0004* /2.041)* =12
CEO SNH and firm value: IV approach Our identification strategy – Exogenous changes in CEO social networks which are unrelated to firm performance – IV: number of death and retirement of directors (or senior managers) to whom the testing CEO is connected (Fracassi and Tate, 2011) Model specification 40 CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) I, t + i,t Q i,t = β 0 + β (Predicted SNH ) i,t + (Firm char.) i, t-1 + i,t CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) I, t + i,t Q i,t = β 0 + β (Predicted SNH ) i,t + (Firm char.) i, t-1 + i,t
Results on IV Regressions 41 (1)(2)(3)(4)(5) VARIABLES Het - demographic TobinQ Het - intellectual TobinQ Het - profession TobinQ Het - international TobinQ Het - overall TobinQ IV: death and retirement-0.021***-0.086***-0.166***-0.019***-0.070*** Het_demographic7.334** Het_intellectual1.213* Het_profession0.690* Het_international7.983** Het_overall1.740* Firm characteristics (t-1) Log(assets) **0.035***-0.095***0.016***-0.054*** ***0.013***-0.069*** Leverage *** *** *** *** *** Capextoasset * *** *** * ** Cashflow ***-0.070**1.964*** *** *** *** RDtoasset ***0.219***4.779***0.185**5.164*** ***0.187***5.388*** High tech0.023**0.396*** ***0.134***0.479***0.027***0.345**0.050***0.432*** Multinational ** *** *** ** *** Networksize0.064***-0.414*0.136*** *** ***-0.307*0.082*** Board female ratio ** *** Board minority ratio0.096*** * ***0.124 Constant0.992*** ***1.334**0.989***1.475***1.376*** ***0.223 Year dummiesYes Industry dummiesYes Observations14,429 16,191 15,404 14,389 12,157 Adjusted R-squared
Event study on CEO turnover The CEO turnover event provides a good setting to see market’s immediate reaction to the change of CEO Controlling for everything else, we relate announcement CAR of new CEO appointment to the change of SN between the new CEO and previous CEO. After matching with ExecuComp turnover data, we identify 114 turnover events with complete information on SN for both new CEO and old CEO. 42
Table 8: Investor Response to CEO Appointment Announcement Panel A. Matching on firm characteristics and CEO characteristics Variable Name Group 1: New CEO has more heterogeneous network than the old CEO Group 2: New CEO has less heterogeneous network than the old CEODiff.T-stat Firm characteristics Size Leverage Capex Cashflow R&D intensity HiTecPharma Multinational CEO characteristics Experience as CEO Outside hire Panel B: Comparison of announcement CAR CAR (-1,1)0.016* **1.742 CAR (-2,2)0.018* **2.086 CAR (-5,5)0.032** * Panel A shows that there are no significant differences among firm characteristics and new CEO characteristics Panel B shows that when the new CEO has more heterogeneous networks than the old CEO, market reactions are significantly higher.
44 Exploring potential channels Channel 1: Innovation Individual differences and intellectual diversity promote creativity Channel 2: New Revenue Generation Exposure to different cultural and international experiences allow managers to better understand the global market. Channel 3: Better Investment (Investment Efficiency and M&A) Knowledge and information obtained from people working in different professions and industries Channel 4: Lower Cost of Financing Bank value relationships and CEO can draw different resources from his or her network. 44
45 Testing Channel 1: Innovation Model specification: Rationale: if CEO SNH impacts firm value through innovation, we expect to find (1) Heterogeneity increases innovation (2) Innovation increases firm value (3) Controlling for innovation, the impact of heterogeneity on firm value is weakened. Procedure: We examine individual heterogeneity indices in addition to the aggregated heterogeneity index. 45 CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) i, t + i,t (3-1) Innovation i,t =α 0 +α (Predicted SNH) i,t + (Firm char) i, t + i,t (3-2) Q i,t = β 0 + β (Innovation) i,t-1 + α(Predicted SNH) i,t + (Firm char.) i t + i,t (3-3) CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) i, t + i,t (3-1) Innovation i,t =α 0 +α (Predicted SNH) i,t + (Firm char) i, t + i,t (3-2) Q i,t = β 0 + β (Innovation) i,t-1 + α(Predicted SNH) i,t + (Firm char.) i t + i,t (3-3)
Table 9: Regression results relating innovation channel MODEL Simultaneous equations on overall heterogeneity, innovation, and Tobin's Q Simultaneous equations on intellectual heterogeneity, innovation, and Tobin's Q Simultaneous equations on professional heterogeneity, innovation, and Tobin's Q (1)(2)(3)(4)(5)(6) Log (patent)Tobin's QLog (patent)Tobin's QLog (patent)Tobin's Q Het-overall-hat5.183***1.636 Het-intellectual-hat4.414***1.430 Het-profession-hat1.229***0.667 Log (patent) t ** 0.028*** Firm characteristics (t-1) Log(assets)0.434***-0.125***0.354***-0.153***0.502***-0.111*** Leverage-0.569***-0.529***-0.647***-0.556***-0.698***-0.578*** Capextoasset *** ***-0.998*2.677*** Cashflow0.477***1.491***0.719***1.581***0.521***1.527*** RDtoasset3.012***3.816***3.104***3.856***3.614***3.967*** High tech-0.632***0.512***-0.547***0.535***-0.463***0.546*** Networksize-0.248*** *** Board female ratio *** *** *** Board minority ratio0.490*** *** *** Constant-6.896*** ***1.588** Year dummiesYes Industry dummies (SIC1)Yes Observations6,220 Adjusted R-squared Column(1) shows that overall CEO SHN significantly increases innovation. Column(2) shows that after controlling for patent channel, the effect of heterogeneity becomes weakened. Same effects are found for intellectual heterogeneity and professional heterogeneity. So, the innovation channel is more pronounced for these two types of heterogeneity.
47 Testing Channel 2: Foreign revenue generation Model specification: If CEO SNH impacts firm value through foreign sale, we expect to find (1) CEO SNH increases foreign sale growth (2) Foreign sale growth increases firm value (3) Controlling for foreign sale growth, the impact of heterogeneity on firm value is weakened. We repeat the analysis for individual heterogeneity indices in addition to the aggregated heterogeneity index. 47 CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) i, t + i,t (4-1) Foreign sale growth i,t =α 0 + α (Predicted SNH) i,t + (Firm char) i, t + i,t (4-2) Q i,t =β 0 + β(Foreign sale growth) i,t-1 +α(Predicted SNH) i,t + (Firm char.) it, + i,t (4-3) CEO SNH i,t =α 0 + α (Deceased or retired network ties) i,t + (Firm char) i, t + i,t (4-1) Foreign sale growth i,t =α 0 + α (Predicted SNH) i,t + (Firm char) i, t + i,t (4-2) Q i,t =β 0 + β(Foreign sale growth) i,t-1 +α(Predicted SNH) i,t + (Firm char.) it, + i,t (4-3)
Table 10: Regression results relating foreign sale channel 48 MODEL Simultaneous equations on overall heterogeneity, foreign sale growth, and Tobin's Q Simultaneous equations on demographic heterogeneity, foreign sale growth, and Tobin's Q Simultaneous equations on international heterogeneity, foreign sale growth, and Tobin's Q (1)(2)(3)(4)(5)(6) VARIABLES Foreign sale growth Tobin's QForeign sale growthTobin's QForeign sale growthTobin's Q Het-overall-hat2.061*0.974 Het-demographic-hat6.769*3.200 Het-international-hat4.091*1.934 Foreign sale growth (t-1)0.080*** Firm characteristics (t-1) Log(assets)-0.062*** *** *** Leverage *** *** *** Capextoasset *** *** *** Cashflow-0.896***5.446***-1.272***5.268***-1.165***5.319*** RDtoasset-1.224*5.044***-1.421*4.951***-0.759*5.263*** High tech ***0.168*0.568*** *** Networksize-0.130* * * Board female ratio-0.401**0.550***-0.646***0.434**-0.617***0.447** Board minority ratio ** ** Constant Year dummiesYes Industry dummies (SIC1)Yes Observations5,091 Adjusted R-squared Column(1) shows that overall CEO SHN significantly increases foreign sale growth. Column(2) shows that after controlling for foreign sale channel, the effect of heterogeneity becomes weakened. Same effects are found for demographic heterogeneity and international heterogeneity. So, the channel is more pronounced for these two types of heterogeneity.
Testing Channel 3: Investment Efficiency Model specification – Estimating the investment equation (Fazarri et al., 1988) – Estimating the interaction between investment-Q sensitivity and CEO SNH Rationale: – Tobin (1969) shows that marginal q is a predictor of investment. This means that in equation (5) should be positive. Durnex et al. (2004 JF) argue that higher marginal q suggests higher investment efficiency. – Chen et al. (2007RFS) estimate investment-Q sensitivity and show that information content of a stock increases investment efficiency. – Following similar approach as Chen et al. (2007), we argue that if CEO SNH enhances investment efficiency, the interaction term β 3 should be positive. I i,t /TA i,t-1 = β 0 + β 1 Q i,t-1 + β 2 (CF i,t /TA i,t-1 )+ (Firm fixed effects) i, t + i,t (5) I i,t /TA i,t-1 = β 0 + β 1 Q i,t-1 + β 2 (CF i,t /TA i,t-1 )+ β 3 Q i,t-1 *CEO SNH (predicted) + β 4 CEO SNH (predicted) + β 5 (Firm fixed effects) + i,t (6)
Table 11: CEO SNH and Investment Efficiency (1)(2)(3)(4) VARIABLES total capital expenditure ratio acquisition expenditure ratio Het-overall-hat1.088`2.231 (1.410)(1.534) Het-overall-hat * Tobin's Q (t-1)0.299***1.043*** (2.739)(3.529) Tobin's Q (t-1)0.021***0.022***0.026***0.042*** (11.368)(8.769)(3.693)(4.731) Inverse logasset (t-1)1.884***1.863***5.490**2.481 (6.733)(4.022)(2.083)(0.861) Leverage (t-1)-0.166***-0.188***-0.385***-0.438*** ( )(-8.640)(-9.178)(-7.967) Cash flow (t-1) ***0.174* (0.873)(0.031)(3.529)(1.928) Constant0.137***0.131***0.064**0.089*** (15.976)(9.575)(2.243)(2.848) Firm fixed effectYes Observations9,7887,8206,3865,265 Number of firms2,3752,0991,8141,612 Adjusted R-squared Column (1) is the investment equation. The positive coefficient of Q on investment is consistent with Tobin (1969). Column (2) finds that the interaction term is positive and significant, which means that higher CEO SNH increases investment efficiency. In Column (3) and (4) we repeat the same analysis for acquisition expenditure because it is often inefficient. We find that the positive effect of CEO SNH on investment efficiency is strongly significant for acquisition investment.
Table 12: Regressions on CEO SNH and M&A Performance 51 (1)(2)(3)(4) VARIABLESCAR[-2,2] BHAR- 3year Het-overall-hat0.298**0.266**4.461*5.205* Het-overall-hat* Diversifying M&A0.009**1.963** Deal characteristics All stock payment-0.020**-0.019** Mix cash and stock payment Private target Public target-0.030***-0.028***-0.215**-0.238** Tender offer * Relative size ** ** Diversifing M&A-0.018***-3.021** Acquirer characteristics Log (assets)-0.013**-0.012** Leverage0.063***0.062***0.506**0.586** Market to book * Run-up Sale growth ROA0.104***0.092** Board female ratio **-0.788* Board minority ratio Constant-0.393**-0.353**-5.720*-6.299* Year dummiesYes Industry dummiesYes Observations3,7573,781 3,757 Adjusted R-squared Column (1) and (2) examine market reaction around 5 days of M&A announcement for acquirers. We find strong evidence that CEO SNH is positively associated with M&A announcement returns, especially for diversified M&As. Column (3) and (4) examine long-run post- merger performance of acquirers, measured by buy and hold abnormal returns over 3-year window. We find that CEO SNH is also positively associated with M&A performance in the long run, especially for diversified M&As.
52 Testing the channel of cost of financing Match Syndicate loan data with our sample. We investigate the cost of financing by borrowing firms as reported in these syndicate loan data. We focus on the variability of loan rate, collateral used and covenants attached to these loans. We find that CEOs with diverse social connections receive cheaper loans, need lower collateral, and experience lower intensity of covenants. 52
53 Table 13: CEO network diversity and bank loan contracts (1)(2)(3) VARIABLESLog (spread)collateralcovenant intensity Het_overall-0.234*-1.580***-0.307* (-1.675)(-3.213)(-1.646) Total assets (log)-0.166***-0.491***-0.123*** ( )(-8.506)(-5.444) Market to book-0.177*** *** ( )(-1.219)(-3.116) Book leverage0.894***2.694***0.458*** (9.763)(7.927)(3.791) Profitability-1.280***-8.325*** (-4.550)(-6.428)(-0.040) Altman_Z-0.036**-0.115**0.044 (-2.494)(-2.077)(0.307) Rating-0.022*** (-4.614)(0.668)(0.617) Loansize (log)-0.070*** *** (-4.525)(-1.549)(4.786) Maturity (log)0.106***0.623***0.010 (4.945)(8.337)(0.350) Collateral0.599***1.133*** (20.068)(5.077) Syndication *** (-0.908)(-0.137)( ) Constant5.732***3.682***0.321 (22.315)(3.670)(0.718) Year and industry dummyYes Observations2,3951,6532,395 R-squared Pseudo R-squared Key findings: CEOs with heterogeneous social networks are able to obtain bank loans with lower interest rates, less likelihood of having collaterals, and less stringent covenants.
Conclusions We find that CEOs with diverse social connections (e.g. demographic, intellectual, profession, foreign exposure) create higher value to firms. We also identify the channels and find that the diverse social network adds value through innovation, new revenue generation, better investment decisions and lower cost of funding. These results overall are consistent with the notion that greater heterogeneity allows for transfer of different knowledge, expertise, and problem-solving skills between connected people and companies, which is value-added to the firm. 54
Implications Current-day CEOs could benefit from a broader set of knowledge to response to the innovations in the new products and increased competitive business pressure in the market. Our findings suggest that a diverse social network provide a CEO with exposures to different information and resources, which ultimately improves managerial performance. Given the changing face of workforce and the increasing competition from international markets, corporate management needs to think about how diversity of social networks can be value-added for the company. 55