study had been conducted during a severe downturn in the market, the results alone could be much more negative and potentially detrimental to the investor. 51 The founder-CEO index has 65 positive periods and 43 negative periods during the December 2006–December 2015 examination period. Consequently, the upward bias of positive periods over negative periods, coupled with an exag- gerated benefit (with strong periods being relatively stronger than the negative periods), combine for a net benefit to this investment strategy. 52 As Exhibit 11 shows, the IT and Health Care sectors perform appreciably better than the peer benchmark. How- ever, in the case of IT, the founder-CEO index has a sector weight that is almost 10% below the benchmark weight (19.54% versus 29.72% for VUG) and a low sector weight for Health Care (9.78% for founder-CEO index versus 12.54% sector weight for VUG). 53 Given the strong growth orientation of the founder- CEO index and heavy sector concentration, we thought it might be possible (likely) that a growth or sector factor might negate much of the selection effect of the founder-CEO index. As the factor analysis demonstrates, the founder-CEO selection factor is, by far, the most dominant factor among the 20 + factors we examine. 54 The Fama–French three-factor model (see Fama and French [1993]) is well represented in academic literature and provides a solid basis to assess key factors in portfolio returns. 55 The Carhart model shows an improvement from the Fama–French three-factor model (improving from 13.54% to 15.71%), owing to the inclusion of the momentum factor. However, much of the 60.71% of the return is still without explanation. 56 This analysis is likely the first documentation to rec- ognize the finding in factor form, when compared side-by- side to other well-known factors. In time, if corroborated by other academic scholars, it may well be recognized as common knowledge. 57 In total, we examined 30 different factors in an attempt to help evaluate the total active return. We utilized the capa- bilities of Bloomberg analytic tool to see if the founder-CEO factor might be encompassed within another factor. As we will show, the founder-CEO variable during the time of our study has been extremely helpful in explaining excess returns. 58 All of the factors other than the selection effect founder CEO sum to approximately 1.47% (total active return of 60.79% equals the selection effect of 59.32% + 1.47% for all other factors). 59 The challenge should not be taken lightly. Although it is easy to suggest a longer time period, the complexity of cor- rectly assembling a thorough sample of 20 + or 30 + years would be an enormous undertaking. To avoid a self-selection bias, each publicly traded firm needs to be checked (individually)
for the proper founder and CEO. Because no current database currently holds these data from inception, it becomes a very manual process. Many firms disappear due to takeovers, bank- ruptcies, mergers, delisting, and so on. Moreover, even current records show inconsistencies among the top four or five useful databases (Bloomberg, FactSet, Capital IQ, SEC company dis- closures, company websites). Preparing the nine-year database of founder CEOs likely required 1,000–1,500 + hours. Going further back in time 20 + years would take well over 3,000 hours, and many missing data points are likely to exist. 60 We note that there may be many reasons why a founder CEO may depart, including retirement, death, acqui- sition, bankruptcy, and merger, among others. To minimize bias, we include only the founder-CEO companies for which we have a complete dataset for periods both before and after the CEO departs. 61 The return differential is striking and statistically significant past the 1% level. To incorporate differing com- parison periods, we apply market-adjusted returns (excess returns) for each year both before and after CEO depar- ture. Given the strong contrast in the return differential, this research issue will gain more attention. Our future research report will show the numbers of mergers, acquisitions, bank- ruptcies, deaths, and retirements associated with the founder CEO departure and the outcomes of each. 62 As of October, 2017, the founder-CEO index com- prised 15.4% of the VUG and 6.6% of the S&P 500 Index. These weights are determined by examining a side-by-side comparison on Bloomberg for both indexes and compiling the respective weights of each. 63 Active share refers to the weight of a security within a portfolio relative to the weight of the security in the compa- rable benchmark. Securities that have high active share will have a disproportionate impact on the performance of a port- folio relative to the benchmark. The portfolio will benefit from high active share when its key holdings perform well (above the average security in the benchmark); conversely, when secu- rities with high active share perform poorly, the portfolio will have a greater likelihood of underperforming its benchmark. 64 As we discussed earlier, the founder-CEO index has a disproportionate weight in three key sectors: IT, Consumer Discretionary, and Health Care. As of October 2017, the founder-CEO index had 66% in those three sectors. How- ever, VUG, owing to its growth orientation, is also heavily weighted in the same three sectors. As of October 2017, the VUG also has approximately 66% in IT, Consumer Discre- tionary, and Health Care; the S&P 500 Growth has 68%; and the Russell 1000 Growth Index has approximately 69%. Consequently, the distinction is not as significant as one might initially anticipate. In contrast, the S&P 500 Index (not growth) only has 49% in IT, Consumer Discretionary, and Health Care. Therefore, a tighter benchmark for a smart
L EADERSHIP M ATTERS : C RAFTING A S MART B ETA P ORTFOLIO WITH A F OUNDER -CEO T WIST
W INTER 2017
Made with FlippingBook - Online magazine maker