The relationship between CFO expertise and firm performance

The relationship between CFO expertise and firm performance Master Thesis Bsc. Edonne C.Z.L. Girigori November 2013 ANR: 459910 Department of Finance Tilburg University Supervisor: Oliver G. Spalt

The relationship between CFO expertise and firm performance Abstract This paper studies the association between chief financial officers expertise and firm performance. In particular, focus is put on three key characteristics. Namely, Worked before in the company, M.B.A. degree and Experience elsewhere. I find only support for two of them in a novel, hand-collected dataset of US CEOs and CFOs firms from 2001 until 2011. Firstly, I find that CFOs that have worked in the company before becoming CFO, on average, significantly positively affect firm performance. In addition, CFOs that have experience from other industries than the current one, on average, negatively affect firm performance. Weak evidence is found suggesting that CFOs that have an M.B.A. degree, on average, negatively affect firm performance. The results are robust to alternative measures of performance, outliers and numerous controls. 2

I. Introduction In this paper I examine the relationship between chief financial officer s (CFO) expertise and firm profitability. A look will be taken at how certain CFO characteristics relate to firm profitability and alternative firm performance measures. There have been numerous studies that analyzed executives and/or managers characteristics and how they affect firm performance (e.g., Bertrand and Schoar, 2007; Gottesman and Morey, 2006; Barker and Muller, 2002). However, they never exclusively focused on CFO characteristics and how that relates to firm performance. I contribute to the existing literature by exclusively focusing on CFO characteristics and studying what kind of effect they have on firm performance. In past decades the role of the CFO within companies has changed. A recent study done by Ernst and Young 1 on CFOs reported that CFOs are becoming more and more strategic. They have started taking an active role in developing and defining the overall strategy for their company instead of only overseeing financial aspects. In addition, CFOs are also taking on operational responsibilities. Areas such as IT, property and logistics are now all part of the CFO tasks (Ernst and Young, 2010). Clearly chief financial officers are starting to take a progressive role in contributing to firm performance. In a study done by Frank and Goyal (2007), for example, they found that the CFO matters more than the chief executive officer (CEO) when it comes to determining corporate leverage. This all suggests that the CFO may be an important contributor within a company when it comes to firm performance. This all leads me to pose the following question; how do CFOs characteristics affect corporate performance? In line with a previous study (Aier et al., 2005) done on financial expertise of CFOs, I use proxies that are consistent with both the Blue Ribbon Committee s definition of financial expertise 2 and guidance concerning the acquisition of financial expertise mandated in the Sarbanes-Oxley Act of 2002 3. Specific characteristics that I study are whether the CFO has a 1 The Ernst and Young report is based on a survey, conducted by the Economist Intelligence Unit, of 669 senior finance professionals in Europe, the Middle East, India and Africa, and a program of in-depth interviews with leading CFOs and finance directors from these regions allowing them to explore the distinctive qualities of this broad community of professionals. 2 Financial expertise has been defined by The National Association of Securities Dealers (NASD) and the New York Stock Exchange (NYSE) as past employment in finance or accounting, requisite professional certification in accounting, or any other comparable experience or background which results in the individual s financial sophistication, including being or having been a CEO or other senior officer with financial oversight capabilities (Blue Ribbon Committee 1999, 25). 3 Under Section 407 of the Sarbanes-Oxley Act of 2002, an audit committee member may acquire financial expertise through experience as principal financial officer, principal accounting officer, controller, public aaccaccountant or auditor, by supervising functions, by monitoring the auditing or evaluation of financial statements, or through other relevant experience (Aier et al., 2005). 3

master s degree in business administration (M.B.A.) or not, whether or not the CFO has prior experience in other industries, and whether or not the CFO has worked at their current company before obtaining the CFO title. Data on CFO characteristics has been collected manually for 395 US companies over the period 2001 until 2011. Firm performance can be measured by many factors. Many researchers have used Tobin s Q, stock returns or firm profitability to measure firm performance. In this study the main dependent variable will be firm profitability. Firm profitability (return on asset) gives an idea as to how efficient management is at using its assets to generate earnings. It serves as an acceptable measure to evaluate how the CFO is able to affect firm profitability. In addition, I too will also take a look at these alternative firm performance measures to see if the same relationship holds between them and the CFO characteristics. Over the years there have been a lot of papers written on firm performance. A study done by Guest (2009), for example, showed that board size is strongly negatively correlated with firm profitability, Tobin s Q and share returns. Another study done by Adams et al. (2005) found that stock returns are more variable for firms run by powerful CEOs. Their findings suggest that the interaction between executive characteristics and organizational variables has important consequences for firm performance. Lastly, in the study of Demsetz and Villalonga (2001) it was established that there is no significant relation between ownership structure and firm performance. One important reason why it is interesting to investigate the relationship between CFO expertise and firm performance is that not much is known on how a CFO affects firm performance. Since the CFO is responsible for the overall financial situation of a company, he plays a major role in how a firm performs. Many researchers have predominantly focused on CEOs as they have final responsibility in a corporation and can significantly influence the performance of a firm from their position (Adams et al., 2005). My research will shed more light on how the aforementioned CFO characteristics influence firm performance. In a sample of 395 S&P 500 largest firms with observations over the period 2001 until 2011, I find a positive relationship between CFOs that have worked at the company before becoming CFOs and firm performance. In addition, I also find that CFOs that have worked in other industries than the current one and CFOs that have an M.B.A. degree tend to negatively affect firm performance. The associations hold for several firm performance measures. Lastly, fixed effects models have been used to control for unobserved variables. 4

This paper will be structured as follows. Section II presents a brief literature review of the previous research done on firm performance and CFO characteristics. Section III presents the data description. In this section the main dependent and independent variables will be further explained. Section IV explains the methodology used in this study. Section V presents the empirical results and I conclude in Section VI with limitations and recommendations at the end. II. Literature review In recent years researchers have studied the behavior of the chief executive officer (CEOs) and chief financial officers (CFOs). The studies focused on what kind of effect their behavior has on corporate performance. It has become an important aspect of behavioral finance as one would like to know why different managers make different choices. Corporate managers all have their own unique set of capabilities and characteristics that drive them to make certain decisions. How these capabilities and characteristics influence corporate decisions and performance have been studied more thoroughly over the last decade (Bertrand and Schoar, 2003), (Malmendier and Tate, 2005). Although it has gained in importance, there is still little literature on the influence of CFOs on corporate performance. Previous researches have mainly focused on CEOs and not so much on CFOs. However, lately more emphasis has been put on the behavior of both the CEO and CFO. Jiang et al. (2008) investigated who had the most influence on earnings management. They separately and jointly examined the association between CFO and CEO equity incentives and earnings management. They found that CFO equity incentives play an independent role in firms earnings management activities, even after controlling for CEO equity incentives. Furthermore, a study done by Chava and Purnanandam (2010) also shows that CFOs significantly influence their firms financial policies. Chava and Purnanandam looked at the effect of managerial risk-taking incentives and how it affects corporate financial policies. What they have found is that corporate risk-taking is significantly related to the risk-taking incentives of the firms key managers. The chief financial officer matters significantly when it comes to financing decisions of the company, more so than the CEO (Frank and Goyal, 2007). In the study of Frank et al. (2007) they look at the effect of top managers on corporate financing decisions. They zoom in on chief executive officer and chief financial officer personal characteristics and how that relates to corporate leverage choices. They find that CEO personal characteristics are not closely 5

connected to corporate leverage choices. Notably, they find that the CFO seems to play at least as important a role as the CEO in determining corporate leverage. This further proves that the CFO also has an important role within the company and can be an important factor when it comes to influencing corporate policies and even corporate performance. As mentioned before the current literature focuses predominantly on CEOs and how their decisions and behavior affect corporate performance. Nevertheless, recently researchers have started to pay more attention to the roles of CFOs within a company and what the implications are for corporate policies and performance (Ernst and Young, 2010). My study will add to this existing literature. Moreover, what makes this study unique is that the data on the characteristics is hand-collected and thus provides me with a unique database. This could lead to interesting results not seen before in other studies. When it comes to literature on firm performance there is plenty. A lot of research has been done on how firm performance is associated to corporate governance and other factors. In a study done by Guest (2009) he finds that board size is negatively correlated with firm performance. He looks at different measures of firm performance in his study to confirm that the same relationship holds across different measures. I will also follow a part of his methodology in this paper. Additionally, Demsetz and Villalonga (2001) also use firm performance in their study. They researched the relationship between ownership structure and firm performance. They also employed different measures of firm performance, with Tobin s Q as the main one. What they found was that there is no statistically significant relation between ownership structure and firm performance. Furthermore, a study done on firm performance determinants by Hansen and Wernerfelt (1989) provides insight on how economical, organizational and industrial factors affect firm performance. They find that organizational variables explain about twice as much variance in profit ratios as economic factors. In conclusion, a recent study done by Gottesman and Morey (2006) provided some interesting insight on how CEO education relates to firm performance. In this study they find no evidence that CEOs from more prestigious schools perform better than firms with CEOs from less prestigious schools. A second finding was that CEOs with M.B.A. degrees or law degrees do not perform better than CEOs without graduate degrees. As a matter of fact, they negatively affect firm performance. As measures for firm performance they used simple excess returns, 4-index alpha, Tobin s Q, return on equity and return on asset. This is quite 6

interesting for my research as I too will look at the education of CFOs and how that relates to firm performance. III. Data description The sample consists of chief financial officers (CFOs) of S&P500 firms listed in Standard and Poor s Execucomp database. The sample period is from 2001 to 2011. Financial firms and utilities have been excluded from the sample 4. Furthermore, firms with observations less than five years are excluded from the sample. In addition, I have merged the Execucomp data with the financial data from Compustat. For this research most of the personal data on CFOs is hand-collected, hence a unique database. The data on CFO characteristics is predominantly retrieved from Lexis-Nexis, Securities and Exchange Commission (SEC), annual reports, 10-K filings and 10-Q filings. This collection of personal data was done in a group by fellow students and me. The final sample has observations on 395 firms. The key variable definitions can be found in Appendix A. The variables are winsorized at the 1% level in both tails to replace outliers and the most extremely misreported data. The key dependent variable of this study is firm profitability measured by return on asset. The main independent variables of interest are the dummy variables worked in the company before, master s degree in business administration and experience elsewhere. Firm profitability For this study I use return on asset as a profitability measure. Firm profitability gives an idea as to how efficient management is at using its assets to generate earnings. It is the ratio of income before extraordinary items and discontinued operations divided by total assets. The ratio gives investors an idea of how effectively the company is converting the money it has to invest into net income. This measure of performance is backwards looking. Accordingly, it measures what has been accomplished already. Although the measure can be subject to accounting practices such as using different methods of determining the value, it is not the only performance measure that can be affected by such practices (Demsetz and Villalonga, 2001). In this study I will also use other alternative performance measures such as return on equity, Tobin s Q and industry-adjusted firm profitability to determine whether the same 4 Financial companies and utility companies are excluded from the sample due to the special regulatory environment in which they operate. Regulation masks efficiency differences across firms (Vafeas and Theodoruo, 1998). 7

relationship holds between the CFO characteristics and the alternative performance measures. More importantly, I decided not to look at stock returns as a measure of firm performance since they are not, to my beliefs, a reliable measure for firm performance in this research setting. It is believed that the true value of management is reflected in the stock price. Although this is true over the long run, it does not always hold in the short run. The market can behave strangely in the short term. When strong stock performance is perceived one cannot automatically assume that management has done a good job 5. Furthermore, stock returns are also driven by investors subjectivity. They could be misleading when trying to interpret them. On the other hand, firm profitability provides a more solid and clean measure of firm performance. Table A. Descriptive statistics This table reports descriptive statistics for the selected financial variables and CFO characteristics. Variables Observations Mean Standard deviation Firm profitability 0.059 0.090 Industry-adjusted firm profitability 0.006 0.083 Return on equity 0.136 0.420 Tobin s Q 1.83 1.271 Leverage 0.340 0.247 Liquidity 2.028 1.276 Growth 0.077 0.133 Size 8.685 1.223 Dummy experience elsewhere 0.708 0.455 Dummy worked in the company 0.698 0.459 before Dummy M.B.A. degree 0.50 0.50 Minimum -0.425-0.408-2.028 0.376 0 0.536 0.004 3.911 0 0 0 Maximum 0.247 0.208 2.192 7.363 1.177 8.172 0.890 12.71 1 1 1 5 One important event that proved this was the late financial crises and the dot.com bubble burst. Moreover, if a firm is performing badly, yet the stock is fairly priced, it does not matter if you buy or sell the stock. If a firm is performing well, yet the stock is fairly priced, again it does not matter if you buy or sell the stock. In either case, it is very hard to determine from the level of stock returns whether good management had anything to do with it. 8

Worked at the company before This is a dummy variable that indicates whether or not the CFO has worked at the firm before becoming the chief financial officer. Working at the company before obtaining the title of CFO, provides the CFO with an opportunity to better understand how the firm operates (Aier et al., 2005). I conjecture that the CFO can then make decisions that improve the performance of the firm based on knowledge acquired during previous years working at the company. In figure 1 the proportions of CFOs who have worked before in the company are displayed per year. Here it is clear that the proportions do not vary over time a lot. Figure 1. Proportions per year of CFOs who have worked at the company before 0.2.4.6.8 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 9

Master s degree in Business Administration M.B.A. degree is a dummy variable that indicates whether or not the CFO has a master s degree in business administration. According to the Institute of Management Accountants, an M.B.A. degree allows an accountant to build a better understanding of his/her company, and thus allows him/her to play a much better broader, more strategic role in the management and operation of his/her businesses (Aier et al., 2005). Looking at figure 2 one can see that the proportions of CFOs with an M.B.A. degree stay relatively stable over the 11 year period. If a CFO has an M.B.A. degree does not necessarily guarantee that it will translate into better corporate performance (Gottesman and Morey, 2002). However, it does provide an opportunity to better corporate performance. Figure 2. Proportions per year of CFOs who have an M.B.A. degree 0.1.2.3.4.5 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 10

Experience elsewhere Experience elsewhere is a dummy variable that takes the value 1 if the CFO has worked for another industry than the current one, 0 otherwise. I conjecture that CFOs that have worked in other industries bring that experience to bear when taking decisions regarding financial firm performance (Aier et al., 2005). The proportions of CFOs with experience elsewhere are displayed in figure 3. It shows a slight increase over the 11 year period. However, what causes these rises and falls in the proportions of CFOs are beyond the scope of this research. Figure 3. Proportions per year of CFOs who have worked elsewhere 0.2.4.6.8 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Control variables In this paper I use a set of financial variables to control for the dependent variable. Following Hansen and Wernerfelt (1989), Guest (2009), Gottesman and Morey (2006) and Yermack (1996), I use size, growth, leverage and liquidity as financial control variables. 11

IV. Methodology My econometric approach follows that of Guest (2009) and Yermack (1996). The specific econometric regression that I run is as follows: (1) Firstly, I estimate equation (1) with only year dummies to address for year fixed effects. Secondly, I estimate equation (1) with only industry dummies to address for industry fixed effects. Thirdly, I estimate equation (1) with both year and industry dummies to address simultaneously for fixed effects. Also, a closer look will be taken as to how interactions amongst the dummy variables relate to firm performance. Lastly, I run a firm fixed effects model to account for unobservable firm specific characteristics. In order to control for clustering of standard errors within firms, t-statistics are based on robust standard errors in which observations are clustered at the firm level. Again, to address the concern that profitability and the CFO characteristics are jointly determined by unobservable firm or industry specific variables, I use fixed-effects models. This is a common method for controlling for omitted variables in a panel data set (Yermack, 1996). In addition, I will look at alternative performance measures. The purpose of these regressions is to see whether the same relationship holds between the CFO characteristics and alternative performance measures. The same econometric regression will be applied but with different firm performance measures as dependent variables. Although, I already employ industry dummies, I will also include an industry-adjusted return on asset variable to control for further bias in the standard errors. Furthermore, I will include past profitability in the regressions to control for further possible omitted variable bias in the standard errors. All of the regression results are reported in tables B, C, D, E, F and G in Appendix B. Main hypotheses: 1. CFOs that have worked before in the company positively affect firm performance. 2. CFOs that have an M.B.A. degree positively affect firm performance. 3. CFOs that have worked elsewhere negatively affect firm performance. 12

V. Empirical results Main results Table B reports regression results for return on asset with the CFO characteristics and financial control variables including year and industry dummies. The first column reports results with only year dummies. The second column report results with only industry dummies. Thirdly, the third column reports results with both year and industry dummies. In the last column, results with firm fixed effects are presented. In tables C, D and E the same conditions apply. First of all, the signs of the CFO characteristics are to be looked at. In all columns, the same relationship holds between the CFO characteristics firm profitability. CFOs that have worked in the company before positively affect firm profitability, while CFOs that have an M.B.A. degree and CFOs that have worked elsewhere negatively affect firm profitability. This shows that an M.B.A. degree does not necessarily ensure that the CFO will perform better. This finding is consistent with a study done by Gottesman and Morey (2006). Furthermore, looking at the adjusted R 2 from the first column in Table C this model only explains 8.65% of the variation in firm profitability. The adjusted R 2 increases when controlling for industry fixed effects in column 2 and year and industry effects simultaneously in column 3. Out of the three CFO characteristics Worked before in the company and Experience elsewhere are the only statistically significant characteristic in the first column. They are both significant at the 5% and 10% level. Also, Worked before in the company proves to be economically significant. The variable has a rather large relative effect on firm profitability. For instance, CFOs that have worked before in the company increase firm profitability by 0.018, holding all the other factors fixed. The mean firm profitability is 0.059. This means an increase of 30.5% relative to the mean firm profitability. Looking at the other two characteristics, Experience elsewhere is also economically significant. CFOs that have worked in other industries before decrease profitability by -0.01, holding all other factors fixed. This means a decrease of 16.9% relative to the mean firm profitability. M.B.A. degree is neither statistically nor economically significant. 13

In columns 2 and 3 almost the same magnitude of economic significance is seen for the CFO expertise dummies. In column 2 variables Worked before in the company and Experience elsewhere are both statistically and economically significant. This provides evidence suggesting that the CFO can most likely significantly affect firm performance. For the financial control variables, liquidity and size positively affect return on asset while leverage and growth negatively affect return on asset. In column 4 regression results with firm fixed effects are presented. The adjusted R-squares are much higher than in the previous columns, indicating that the firm fixed effects contribute substantially to explaining the variation in firm profitability. In this setting, when controlling for unobserved firm factors, only one characteristic, Experience elsewhere, turned out to be statistically significant at both 5% and 10% level. In conclusion, all the dummy variables are interacted with each other and included in the regressions to see how they relate to firm profitability. Only two interaction terms were found to be statistically significant. First, CFOs that have worked in the company before and are in possession of an M.B.A. degree are found to have a positive effect on firm profitability. It means that CFOs that are in the possession of an M.B.A. degree and have worked at the company before positively affect firm profitability compared to CFOs that do not have any of these characteristics. Moreover, holding all factors fixed, firm profitability increases by 0.02. This is an increase of 32.2% relative to the mean firm profitability. Second, CFOs that have experience in other industries and have an M.B.A. degree tend to positively affect firm profitability. However, this result is only marginally statistically significant. The former two results show that having an M.B.A. degree alone does not guarantee that the CFO will positively influence firm performance. However, in the case that the CFO possesses both of the aforementioned characteristics a positive association is found. 14

Alternative performance measures Return on equity In Table D regression results are presented for the return on equity (ROE). This regression has the same set on control variables and CFO characteristics. The first column reports results with only year dummies. The second column report results with only industry dummies. Lastly, the third column reports results with both year and industry dummies. In column four coefficients are presented from the firm fixed effects regression. In general, the same relationship holds between the CFO characteristics and return on equity as it did with return on asset. Judging from the adjusted R-squares, the independent variables do no explain much of the variation in return on equity. Here too, CFOs that have worked in the company before positively affect firm performance, where CFOs that have worked elsewhere and CFOs that have an M.B.A. degree do not. The only difference is the magnitude of the economic effect. In column 3, when controlling for industry and year effects simultaneously, CFOs that have worked in the company before increase the return on equity by 0.059. The mean return on equity is 0.136. This translates into a percentage increase of 43.4 relative to mean ROE. This is a rather large relative effect on the return on equity. Looking at the financial control variables, leverage and growth negatively affect return on equity. However, size and liquidity both flip signs when controlling for industry and firm effects. When controlling for industry fixed effects, liquidity negatively affect return on equity and flip sign when controlling for firm fixed effects. The former holds in reverse for size. Tobin s Q The second alternative firm performance measure I include is Tobin s Q. It is a widely used measure for firm performance. The results are presented in Table E. Contrary to the other previous results with year fixed effects; all of the CFO characteristics do have a positive relationship with the Tobin s Q. Nevertheless, none of them are statistically significant. In fact, they are insignificant 6. Columns 2, 3 and 4 do report signs consistent with the findings for firm profitability and return on equity, when controlling for industry and firm unobserved effects. Judging from the R-squares, this model seems to be more fitted than the previous ones. The 6 The coefficients for M.B.A. degree, Worked before in the company and Experience elsewhere had p-values of 0.263, 0.663 and 0.642, respectively. 15

large difference between the R-squares in the first three columns and the last column indicates that again firm unobserved effects contribute a lot to explaining the variation in the dependent variable. When it comes to the economic effect of the coefficients, then we do find significant results. For example, in column 2 we see that CFOs who have experience elsewhere decrease the Tobin s Q by -0.176. The average firm in the sample has a Tobin s Q of 1.83. This will result in a 9.6% decrease relative to the mean Tobin s Q. Coefficient estimates for the financial control variables show that leverage and size have a negative relationship with the Tobin s Q. The opposite holds for liquidity and growth. Industry-adjusted firm profitability One concern is that the results are driven by industry factors. Although I use industry dummies and to control for unobserved industry factors, I also employ an industry adjusted firm profitability variable. It is calculated as the difference between return on asset and average industry return on asset for all firms within the same industry indicated by their standard industry classification codes (SIC). The results in Table F are very similar to those in Table B. The coefficients of the CFO characteristics are practically the same and have the same signs. CFO characteristics, past firm performance and current performance In order to control for further bias in the standard errors, possibly due to omitted variables, I include past profitability. Past profitability can certainly help explain current profitability, hence the reason to include it as an explanatory variable. Return on asset is lagged by one year. The purpose of this regression is to see whether the same conditions hold when controlling for past performance. Results are reported in Table G. In all columns past profitability is statistically significant and positively related to current profitability as expected. Also, the adjusted R-squares are much higher compared to those in Table B and D, indicating that past performance is contributing to explaining the variation in the dependent variable. More importantly, the same correlation holds between the CFO characteristics and current profitability. Furthermore, the magnitudes of the coefficients are more or less the same as in Table B. 16

VI. Conclusion This study provides empirical evidence on the association between CFO characteristics and firm performance based on a sample of 395 firms over the years 2001 until 2011. Specifically, I focus on three key characteristics of CFOs. I find that CFOs that have worked in the company before becoming chief financial officer, on average, consistently positively affect firm performance. On the other hand, CFOs that have worked in other industries and CFOs that have an M.B.A. degree, on average, negatively affect firm performance. CFOs that have worked in the company before have been found to significantly positively affect firm performance compared to CFOs who do not posses any of the three key characteristics in this research. It seems that working in the firm before provides (potential) CFOs the opportunity to get to know the firm better and eventually it will lead to positive influence on firm performance. On the other hand, CFOs that have worked elsewhere significantly negatively affect firm performance. The fact that CFOs have worked in other industries than the current one, does not necessarily mean that the experience they take with them will affect firm performance in a positive manner. Clearly, CFOs that did work in the same industry and fully understand it perform better than those who have not. Furthermore, CFOs that have an M.B.A. degree tend to negatively affect firm performance. The results are marginally significant in some cases, but in most of the cases no significant results were found. However, the negative sign does coincide with findings of Gottesman and Morey (2007). They found that the education background of CEOs, in particular whether they had an M.B.A. degree or not, did not positively affect firm performance. Apparently, in a certain extent this also holds for chief financial officers. All in all, the results do indicate that CFOs have a significant impact on firm performance in some cases. Current literature has shown that CFOs are taking a more and more prominent role within a company. It is crucial to understand how their actions affect firm performance as their actions do have a profound impact on firm performance. Given the job title, perhaps that should not be too surprising. Further attention on CFO characteristics might prove interesting. 17

Implications This study sheds light on how certain CFO characteristics affect firm performance. It has shown that chief financial officers do certainly have a significant impact on firm performance. Since CFOs have been taking a more prominent role within companies, it is quite important to study and understand how they affect corporate performance. In this study it has been established that CFOs that have worked in the company before significantly positively affect firm performance. On the other hand, CFOs that have worked in other industries than the current one and CFOs that do not have an M.B.A. degree negatively affect firm performance. This could have practical implications in the sense that when selecting CFOs companies may want to take certain characteristics that the CFO has into account and how they will contribute to the performance of the firm. Limitations and recommendations First of all, the data on CFO characteristics in this research was hand-collected by a group of five selected students. Evidently, it is this that also imposes the first limitation. The data on CFOs was very limited compared to data on CEOs. This drastically reduced the number of observations for CFOs in my research. In addition, this posed a restriction on the construction of the CFO characteristics. Secondly, in this study there were only three CFO characteristics used in the analysis. For future reference, it could be useful to go in more detail on choosing additional proxies for characteristics of CFOs. Lastly, an important recommendation for next time is to include CEO characteristics as well and see how CFOs and CEOs jointly affect firm performance. The job of a CFO is a very complex one. This study has only looked at the tip of the iceberg. 18

References Adams, R.B., Almeida, H. and Ferreira, D. (2005), Powerful CEOs and their impact on corporate performance. Oxford Journals, Vol. 18, No. 4, pp. 1403-1432 Aier, J., Comprix, J., Gunlock, M. and Lee, D. (2005), The financial expertise of CFOs and accounting restatements. Accounting Horizons, Vol. 19, No. 3, pp. 123-135 Barua, A., Davidson, L., Rama, D. and Thiruvadi, S. (2010), CFO gender and accruals quality. Accounting Horizons, Vol. 24, No. 1, pp. 25-39 Barker, V.L., and Muller, G.C. (2002), CEO characteristics and firm R&D spending, Management Science, Vol. 48, No. 6, pp. 782-801 Bertrand, M. and Schoar, A. (2003), Managing with style: the effect of managers on firm policies. The Quarterly Journal of Economics, Vol. 118, No. 4, pp. 1169-1208 Chava, S. and Purnanandam, A. (2010), CEOs vs CFOs: Incentives and corporate policies. Journal of Financial Economics, Vol. 97, pp. 263-278 Chung, K.H. and Pruitt, S.W. (1994), A simple approximation of Tobin s q, Financial Management, Vol. 23, pp. 70-74 Conyon, M.J. and Peck, S.I. (1998), Board size and corporate performance: Evidence from European countries, The European Journal of Finance, Vol. 4, pp. 291-304 Demsetz, H. and Villalonga, B. (2001), Ownership structure and corporate performance. Journal of Corporate Finance, Vol. 7, pp. 209-233 Ernst and Young. (2010). The DNA of the CFO: a study of what makes a chief financial officer. London, United Kingdom: Clifford, L., Wood, P. and Dowding, E. Guest, P.M. (2009), The impact of board size on firm performance: evidence from the UK. The European Journal of Finance, Vol. 15, No. 4, pp. 385-404 Frank, M.Z. and Goyal, V.K. (2007), Corporate leverage: how much do managers really matter? SSRN working paper, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=971082 Gottesman, A. and Morey, M. (2006), Does a better education make for better managers? An empirical examination of CEO education quality and firm performance. SSRN working paper, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=564443 Hansen, G. and Wernerfelt, B. (1989), Determinants of firm performance: the relative importance of economic and organizational factors. Strategic Management Journal, Vol. 10, No. 5, pp. 399-411 Jiang, J.X., Petroni, K.R. and Wang, I.Y. (2010), CFOs and CEOs: Who have the most influence on earnings management? Journal of Financial Economics, Vol. 96, pp. 513-526 Malmendier, U. and Tate, G. (2005), Does overconfidence affect corporate investment? CEO overconfidence measures revisited. European Financial Management, Vol. 11, No. 5, pp. 649-659 19

Perfect, S.B. and Wiles, K.W., (1994), Alternative construction of Tobin s q An empirical comparison, Journal of Empirical Finance, Vol. 1, pp. 313-341 Powell, T. (1996), How much does industry matter? An alternative empirical test. Strategic Management Journal, Vol. 17, No. 4, pp. 323-334 Vafeas, N. and Theodorou, E. (1998), The relationship between board structure and firm performance in the UK. British Accounting Review, Vol. 30, pp. 383-407 White, H. (1980), A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity, Econometrica, Vol. 48, pp. 817-838 Yermack, D. (1996), Higher market valuation of companies with a small board of directors. Journal of Financial Economics, Vol. 40, pp. 185-211 20

Appendix A. Variables definitions The specific variables used in this research are defined as follows: Firm performance variables Firm profitability (ROA) is defined as the ratio on net income before extraordinary items and discontinued operations to the book value of assets. Return on equity (ROE) is defined as net income divided by shareholders equity. Tobin s Q is defined as the ratio of market value of assets divided by the book value of assets. Industry-adjusted firm profitability is the mean return on asset by industry subtracted from the return on asset year by year separately. Financial control variables Leverage is defined as total long-term debt plus debt in current liabilities divided by longterm debt plus debt in current liabilities plus book value of common equity. Size is defined as the natural log of book value of assets. Growth is defined as capital expenditures divided by total sales. Liquidity ratio is defined as current assets divided by current liabilities. Chief financial officer characteristics M.B.A. degree is a dummy variable equal to 1 if the CFO has an M.B.A. degree, 0 otherwise. Worked in the company before is a dummy variable equal to 1 if the CFO has worked at the current company before (becoming the CFO), 0 otherwise. Experience elsewhere is a dummy variable equal to 1 if the CFO has prior experience in other industries, 0 otherwise. 21

Appendix B. All tables with results Table B. Regression results The impact of CFO characteristics on firm profitability Dependent variable: Return on asset (1) (2) (3) (4) Leverage Liquidity Growth Size M.B.A. degree Worked in the company before Experience elsewhere Constant -0.071*** (4.05) 0.0003 (0.07) -0.043** (3.02) 0.005 (1.62) -0.004 (0.80) 0.018** (2.95) -0.01* (1.93) 0.039 (1.22) -0.088*** (4.82) 0.009** (2.31) -0.049 (1.39) 0.013*** (4.88) -0.008* (1.76) 0.016** (2.92) -0.015** (2.82) -0.037 (1.34) -0.081*** (4.39) 0.009** (2.43) -0.052 (1.45) 0.011*** (3.81) -0.008* (1.79) 0.016** (3.02) -0.018** (3.35) -0.038 (1.34) -0.091*** (5.28) 0.006 (1.58) -0.14*** (4.52) -0.004** (0.57) -0.001 (0.20) 0.005 (0.72) -0.013** (2.48) 0.108* (1.65) Year dummies Industry dummies Firm dummies Adj. R-squared No. observations 0.0865 0.2509 0.2775 0.4550 This table reports the estimates of fixed effects regressions for firm profitability. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on Hubert-White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 22

Table C. Regression results The impact of interacted CFO characteristics on firm profitability Dependent variable: Return on asset (1) (2) (3) (4) Leverage -0.071*** (4.06) -0.09*** (4.93) -0.082*** (4.50) -0.092*** (5.35) Liquidity.0002 (0.06) 0.009** (2.28) 0.009** (2.41) 0.005 (1.54) Growth -0.044** (3.12) -0.055 (1.55) -0.057 (1.61) -0.145*** (4.64) Size 0.005 (1.64) 0.013*** (4.99) 0.011*** (3.90) -0.005 (0.61) M.B.A. degree 0.0003 (0.02) -0.01 (0.94) -0.009 (0.91) 0.013 (1.00) Worked in the company before 0.002 (0.19) -0.006 (0.56) -0.006 (0.54) -0.005 (0.37) Experience elsewhere -0.013 (1.04) -0.018 (1.51) -0.022* (1.73) -0.005 (0.38) Worked in the company before * M.B.A. degree 0.009 (0.74) 0.02** (2.08) 0.02** (2.10) 0.007 (0.61) Worked in the company before * Experience elsewhere 0.013 (1.10) 0.016 (1.32) 0.016 (1.35) 0.005 (0.39) M.B.A. degree * Experience elsewhere -0.014 (1.50) -0.016* (1.70) -0.017* (1.79) -0.02 (1.56) Constant 0.046-0.025-0.026 0.11* (1.30) (0.87) (0.88) (1.68) Year dummies Industry dummies Firm dummies Adj. R-squared 0.0882 0.2540 0.2807 0.4565 No. observations This table reports the estimates of fixed effects regressions for firm profitability. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 23

Table D. Regression results The impact of CFO characteristics on return on equity Dependent variable: Return on equity (1) (2) (3) (4) Leverage Liquidity Growth Size M.B.A. degree Worked in the company before Experience elsewhere Constant -0.05 (0.60) -0.025** (2.66) -0.185*** (4.23) 0.002 (0.28) -0.019 (1.06) 0.045** (2.24) -0.027 (1.41) 0.196** (2.46) -0.14 (1.46) -0.004 (0.43) -0.181 (1.54) 0.013 (1.55) -0.027 (1.54) 0.059** (3.04) -0.045** (2.41) 0.095 (1.20) -0.127 (1.31) -0.005 (0.53) -0.189 (1.62) 0.002 (0.20) -0.028 (1.63) 0.059** (3.11) -0.059** (3.20) 0.137 (1.61) -0.298** (5.94) 0.002 (0.17) -0.453** (3.29) -0.121** (6.34) -0.025 (1.19) 0.025 (1.07) -0.037 (1.49) 1.225 (7.34) Year dummies Industry dummies Firm dummies Adj. R-squared No. observations 0.0202 0.0946 0.1046 0.1665 This table reports the estimates of fixed effects regressions for the return on equity. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 24

Table E. Regression results The impact of CFO characteristics on Tobin s Q Dependent variable: Tobin s Q (1) (2) (3) (4) Leverage Liquidity Growth Size M.B.A. degree Worked in the company before Experience elsewhere Constant -0.833** (3.44) 0.255*** (4.64) -0.329** (2.02) -0.149** (3.23) 0.087 (1.12) 0.039 (0.44) 0.039 (0.47) 2.815*** (6.03) -0.697** (3.43) 0.22* (4.35) 0.561** (2.06) -0.132** (3.35) -0.092 (1.38) 0.096 (1.15) -0.176** (2.30) 2.827*** (6.89) -0.572** (2.79) 0.242*** (4.86) 0.531* (1.89) -0.084** (2.13) -0.082 (1.24) 0.108 (1.32) -0.121 (1.58) 2.649*** (6.62) -0.613*** (4.13) 0.052 (1.59) 0.860** (2.96) -0.405*** (4.83) -0.094 (1.30) 0.138* (1.80) -0.058 (0.85) 5.646*** (7.96) Year dummies Industry dummies Firm dummies Adj. R-squared No. observations 0.2184 0.4825 0.5191 0.7335 This table reports the estimates of fixed effects regressions for Tobin s Q. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on Hubert-White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 25

Table F. Regression results The impact of CFO characteristics on industry-adjusted firm profitability Dependent variable: Industryadjusted firm profitability (1) (2) (3) (4) Leverage Liquidity Growth Size M.B.A. degree Worked in the company before Experience elsewhere Constant -0.05*** (4.24) 0.007** (2.22) -0.026** (2.06) 0.006** (2.67) -0.004 (0.95) 0.011** (2.28) -0.007 (1.65) -0.046 (1.93) -0.086*** (4.89) 0.009** (2.37) -0.052 (1.49) 0.013*** (4.73) -0.008* (1.83) 0.016** (2.96) -0.015** (2.81) -0.088** (3.19) -0.078*** (4.42) 0.009** (2.51) -0.055 (1.55) 0.011*** (3.69) -0.008* (1.86) 0.016** (3.07) -0.018** (3.33) -0.090** (3.15) -0.09*** (5.28) 0.006 (1.57) -0.144*** (4.55) -0.005 (0.67) -0.001 (0.22) 0.005 (0.78) -0.013** (2.46) 0.060 (0.91) Year dummies Industry dummies Firm dummies Adj. R-squared No. observations 0.0894 0.1065 0.1383 0.3510 This table reports the estimates of fixed effects regressions for industry-adjusted firm profitability. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on Hubert-White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 26

Table G. Regression results CFO characteristics, past performance and current performance Dependent variable: Return on asset (1) (2) (3) (4) Leverage Liquidity Growth Size Lagged roa t-1 M.B.A. degree Worked in the company before Experience elsewhere Constant Year dummies Industry dummies Firm dummies Adj. R-squared No. observations -0.058*** (4.67) 0.002 (0.94) -0.034** (3.39) 0.002 (0.84) 0.322*** (8.20) -0.002 (0.65) 0.009** (2.32) -0.009** (2.57) 0.05** (2.71) 0.3348 3256-0.081*** (5.01) 0.007** (2.72) -0.036 (1.14) 0.008** (3.41) 0.245*** (7.23) -0.006* (1.81) 0.011** (2.77) -0.015*** (3.59) 0.0003 (0.02) 0.3841 3256-0.075*** (4.61) 0.007** (2.93) -0.038 (1.15) 0.007** (3.05) 0.240*** (7.12) -0.006* (1.74) 0.012** (2.83) -0.016*** (3.80) -0.014 (0.62) 0.3988 3256-0.101*** (6.24) 0.005* (1.77) -0.104** (2.85) -0.001 (0.09) 0.124*** (4.31) -0.001 (0.16) 0.003 (0.51) -0.011** (2.16) 0.076 (1.26) 0.4936 3256 This table reports the estimates of fixed effects regressions for past and current firm profitability. Columns (1), (2), (3) and (4) present ordinary least squares estimates. Absolute values of t-statistics are reported in parentheses. For all columns t-statistics are based on Hubert-White (1980) robust standard errors in which observations are clustered at the firm level. ***, ** and * indicates significance level of 1, 5 and 10 percent respectively. 27