WWW.DISSERTATION.XLIBX.INFO
FREE ELECTRONIC LIBRARY - Dissertations, online materials
 
<< HOME
CONTACTS



Pages:     | 1 |   ...   | 12 | 13 || 15 | 16 |   ...   | 22 |

«Three essays on corporate boards R. Øystein Strøm A dissertation submitted to BI Norwegian School of Management for the degree of Dr.Oecon SERIES ...»

-- [ Page 14 ] --

The board index and the free cash flow hypotheses come out more in line with expectations in the co-determined firms too. Now a significant result for the board index towards average wage appears. Leverage turns out to be negative and significant towards average wage, while positive in

CHAPTER 3. CO-DETERMINATION

the overall sample. In shareholder determined firms, significant results are fewer and of different sign. Leverage is positively correlated with average wage, in contrast to the co-determined firms.

In both sub-samples the board index and leverage are negatively related. Thus, the substitution result from the overall sample is confirmed in sub-samples. We also see that the past firm performance endogeneity hypothesis gains less support in the sub-samples than in the overall. In fact, only the negative leverage result in the co-determined sub-sample is significant.

Another difference exists for firm size. Firm size is negative and significant in the firm performance equation in shareholder determined firms, while positive in co-determined. Also, in the leverage equation the signs are reversed, and significant in shareholder determined firms only. The latter confirms “stylized facts” about the positive relationship between firm size and leverage (Harris and Raviv, 1991).

An interpretation of the difference in sub-samples is that in shareholder determined firms the board composition is closer to the optimal, and therefore, exogenous characteristics such as firm size play a larger role. The large differences between samples confirm the Buchanan and Tullock (1962) theory.

Are the results arrived at so far driven by a firm size effect? Table 3.6 shows regressions for all firms with more than 200 employees in the upper part, while the lower part is limited to the largest co-determined firms.

Table 3.6

The 200+ employee sample shows results very similar to those in the entire sample in table 3.4 in the upper part, and for the co-determined in table 3.5 in the lower. Thus, the former results are not due to some firm size effect. In fact, even among firms where co-determination is compulsory, the main co-determination hypothesis is confirmed.

Looking back, the co-determination and governance hypotheses are confirmed. Tests in sub-samples do not overturn these conclusions, on the contrary, they add to their strength. For instance, while the employee director is negative for leverage in the overall sample, it is positive in the co-determined sub-sample, as the hypothesis predicts. Thus, having representatives of one stakeholder group, the employees, in addition to shareholders on the board does not improve firm performance, as a stakeholder (Freeman and Reed, 1983; Blair, 1995) or a new economy position (Zingales, 2000; Becht et al., 2003) implies. Instead, the results point to conflict

3.7. ROBUSTNESS CHECKS 81 of interests among the stakeholders. Furthermore, evidence of substitution between the board index and leverage is present in all regressions.

I also find evidence of endogeneity (or reverse causation) from past firm performance, but with opposite signs to those predicted in Hermalin and Weisbach (1998). However, the indirect effects of employee directors and past firm performance upon firm performance through the board index and leverage are small compared to the direct effects from the board index and leverage. Endogeneity counts, but has low economic significance.

The negative relation between employee directors and firm performance is in agreement with FitzRoy and Kraft (1993); Schmid and Seger (1998); Gorton and Schmid (2000); Falaye et al. (2006) and Bøhren and Strøm (2007), but at odds with Fauver and Fuerst (2006). None of these studies contain simultaneous equations models, and only the Bøhren and Strøm (2007) paper investigates the endogeneity of board mechanisms. I will return to the Fauver and Fuerst (2006) and Bøhren and Strøm (2007) articles in the following robustness section.

3.7 Robustness checks

I perform robustness checks on the definitions of the board index, firm performance, and leverage. In addition, I check for the absence of serial dependence of the firm performance, that is, whether lagged firm performance is zero. Finally, I check the Fauver and Fuerst (2006) results in two sub-samples of information industries and other industries. With simultaneous equations, changes in one place are likely to propagate throughout the system. Thus, different coefficient values and significance from the original formulation are quite likely to appear. Fortunately, the results largely confirm those in section 3.6.

Do the co-determination results survive when the individual board mechanisms are used in place of the board index? Table 3.7 shows simultaneous regressions results when all four board characteristics making up the board index enter the regressions individually.

Table 3.7

Former results for co-determination largely apply. The employee director variable is negative to Tobin’s Q, and positive to average wage and leverage. For the board characteristics, only the relation to board size is significant. On the other hand, the hypotheses on governance variables are upheld for all board characteristics but the gender variable. It turns

CHAPTER 3. CO-DETERMINATION





out to be non-significant in the Tobin’s Q relation. The other variables are as expected, and their coefficients are close to those Bøhren and Strøm (2007) find in partial GMM estimations. They also discuss endogeneity.

Even though the estimations are not directly comparable, none of the significant results in table 3.7 conflict with the endogeneity results in Bøhren and Strøm (2007). The second endogeneity effect from lagged firm performance is significant in the leverage, but not in any of the board variables.

However, the signs on the individual board variables conform to the positive sign of the board index in earlier tables.

Besides these main points, table 3.7 contains many new details, which it is beyond this paper to explore. For instance, the substitution effect between the board index and leverage in former tables now turns out to concern network, while leverage is a complement to board size and gender. Thus, overall the results are well in line with former findings, except for the lagged firm performance relationship to governance variables.

In the next table 3.8 I have modified the board index to include all board variables used in Bøhren and Strøm (2007) as specified in footnote 13 to check if the board index is sensitive to the selection of board characteristics.

Table 3.8

The overall Wald tests are strong and the significance of the coefficients are almost similar to what earlier full sample results in table 3.4 show. We note that the impact of the employee director variable is less in the new board index, and is now significant in its positive relationship to average wage. Thus, the co-determination hypothesis is supported with this new board index, although with lower coefficient values. The endogeneity effect of a lagged firm performance loses significance in the board index relation. The same happens when individual board characteristics replace the board index, and also disappear in the shareholder determined subsample. Thus, a preliminary conclusion is that the reverse causation in the board index relation seems to be sensitive to the specification of the index and in sub-samples.

The conclusion from the discussion of the two last tables is that the results are upheld, in particular, the co-determination hypothesis is conrmed.

Now I turn to variations on firm performance, using the stock return and ROA instead of Tobin’s Q. The stock return and ROA may be seen as two extremes in performance measurement, the one only market based,

3.7. ROBUSTNESS CHECKS 83 the other only accounting based. Bhagat and Jefferis (2002) argue in favour of accounting measures, noting that market measures may contain an anticipation bias, since accounting numbers may be manipulated during a given year. Since our data span fourteen years, this accounting manipulation should be a minor concern. These two measures of firm performance should together provide an adequate framework for robustness tests.

The results for the full sample are given in table 3.9. Since the results in sub-samples largely parallel those found for the full sample, the subsample results are not reported.

Table 3.9

The results in table 3.9 largely replicate those already found for Tobin’s Q in table 3.4. The co-determination and the governance hypotheses show the same confirmations. As before, leverage is negative in the firm performance equation. Again, the board index and leverage are substitutes. Endogeneity (or reverse causation) is evident in both firm performance specifications, although at different variables. For the stock return the lagged stock return is significant in firm performance and leverage, as before. One would expect this to happen with accounting numbers due to earnings management or conservative accounting practices (Watts, 2003), which would induce serial correlation. However, lagged performance is significant for only the board index for the accounting measure ROA. Overall, table 3.9 supports earlier findings.

The upshot is that alternative performance measures do not upset conclusions reached with Tobin’s Q. Therefore, further robustness tests may well proceed with Tobin’s Q as the dependent variable.

Next, table 3.10 shows results when the dividend payout rate replaces the leverage, and Tobin’s Q is the firm performance in the upper part, while in the lower part the lagged firm performance is removed. Dividend payout rate is gauged as the annual dividend as a fraction of the earnings before interest, taxes, depreciation, and accruals (EBITDA). During the period of study, share buybacks were illegal in Norway.

Table 3.10

The striking results are first that the dividend payout rate is nowhere significant as an independent variable, and second, as a dependent variable no variable in the system is related in a significant way. In fact the Wald test cannot reject the hypothesis that all coefficients in the dividend

CHAPTER 3. CO-DETERMINATION

payout rate equation are zero. An exclusion test (not reported) for the dividend payout rate cannot confirm that the variable coefficient is different from zero. Thus, the dividend payout rate is an inferior substitute for the leverage. Second, the results for the other variables are not affected, even though changes in one part of a simultaneous system may bring about new values in other parts. Therefore, the results in table 3.10 increase the confidence in the original model.

The lower part of table 3.10 shows results when the lagged firm performance is left out. The reason for the removal is that lagged firm performance induces bias (Hsiao, 2003, p. 71-2), since the errors are no longer independent of the regressors. The smaller the bias, the larger is the number of periods in the panel and the closer to zero is the auto-correlation coefficient on lagged firm performance. Furthermore, if the explanatory variables apart from the lagged firm performance have very persistent elements, the bias will not disappear. This persistence can be a concern in governance studies. For instance, the firm’s board size is likely to be fairly stable. To test for the seriousness of this bias, I include static system regressions, that is, with no lagged performance.

Comparing the results from the no lagged firm performance regression to the original estimates in table 3.4 we see that practically all signs are maintained, and also that coefficient values are quite similar. The codetermination hypothesis is confirmed. For average wage on firm performance, the variable is significant in the static specification but not in the dynamic. But overall the results from the dynamic estimations are upheld.

Apparently, the low auto-correlation coefficient, the rather long time period and small persistence in the explanatory variables warrants the use of the dynamic specification in table 3.4.

I also run a regression (not reported) with all explanatory variables lagged one period for the entire sample. This regression shows far fewer significant results, and although the signs are the same as before, this specification is far inferior to the main regression in table 3.4. Again, this points to a contemporaneity in governance mechanisms.

Finally, I run a test for the Fauver and Fuerst (2006) information hypothesis in sub-samples. The authors assume information significance to trade, transportation, and manufacturing industries. Using the same GICS industry classification as in table 3.3, I allocate Capital goods, Transport, Consumer articles, Retailing, Food and staples retailing, Health care equipment and supplies, and Telecommunications to the information intensive industries, while the rest is in other industries. Co-determined

3.8. CONCLUSION 85

firms are distributed in the two sub-samples almost as in the total population, with 61.1 per cent without employee directors in the Other industries category against 57.4 in the full sample. A test for the Fauver and Fuerst (2006) information hypothesis is that the employee director variable is positive in the information intensive industries. Table 3.11 shows results.

Table 3.11

The main interest is in the employee director, that is, the co-determination hypothesis. Both sub-samples show a negative and significant coefficient on the employee director variable. The Chow test shows that the two subsamples are different, but the main Fauver and Fuerst (2006) hypothesis is not supported.

Overall, the results for the robustness test do not invalidate the results found in table 3.4.

3.8 Conclusion



Pages:     | 1 |   ...   | 12 | 13 || 15 | 16 |   ...   | 22 |


Similar works:

«COFFEE ME EDIT 5/10/2012 10:14 AM THE POLITICAL ECONOMY OF DODD-FRANK: WHY FINANCIAL REFORM TENDS TO BE FRUSTRATED AND SYSTEMIC RISK PERPETUATED John C. Coffee, Jr.† INTRODUCTION I. THE REGULATORY SINE CURVE AND STATUTORY CORRECTION. 110 II. SOX REVISITED: THE DOWNSIZING OF REFORM A. Section 404 and Internal Control Reports B. Executive Loans and Section 402 C. Attorneys as Whistle Blowers and Section 307 D. An Evaluation III. THE DODD-FRANK ACT: PREMISES AND POLICY OPTIONS. 127 A. Moral...»

«City Council City Council Chambers  524 West Stephenson Street  Freeport, IL 61032 COMMITTEE OF THE WHOLE REPORT REGULAR MEETING Monday, October 13, 2015 1. CALL TO ORDER The regular meeting of the Freeport, Illinois, Committee of the Whole was called to order by Mayor James Gitz with a quorum being present at 6:00 p.m. on Monday, October 13, 2015, in council chambers. 2. ROLL CALL Present on roll call: Mayor James Gitz and council members Tom Klemm, Peter McClanathan, Art Ross, Jodi...»

«1 NEW BERY MEDAL ACCEPTANCE NUMBER THE STARS, 1990 For years, I have carried in my wallet, stuck in there among the frequent flyer cards and the MasterCard carbons and the cash register receipts from the supermarket, two little wrinkled slips of paper which are fortunes from Chinese fortune cookies. Like everyone else, I've gotten countless fortunes from Chinese fortune cookies, and, like everyone else, I tend to throw them away. Let's face it: fortune cookie fortunes' are generally boring....»

«THE DECKS RAN RED! The grimmest of all wartime shipboard duty was experienced by the valiant medical corpsmen who manned the PCE (R) rescue ships the Naval equivalent of seagoing ambulances One of the least-known but most significant warship variants of WWII was the PCE(R) the Navy’s equivalent of a seagoing ambulance. Only 13 were commissioned and these saw extensive duty only in the later campaigns of the Pacific. Though few in number the heroic role of the PCE(R)s in Pacific invasions was...»

«VOLUME 21 NUMBER 2 THE CYPRUS REVIEW A Journal of Social, Economic and Political Issues The Cyprus Review, a Journal of Social, Economic and Political Issues, P.O. Box 24005 1700 Nicosia, Cyprus. Telephone: 22-353702 ext 301, 22-841500 E-mail: cy_review@unic.ac.cy Telefax: 22-353682, 22-357481, www.unic.ac.cy To access site: Research UNic Publications Subscription Office: The Cyprus Review University of Nicosia 46 Makedonitissas Avenue 1700 Nicosia, Cyprus Copyright: © 2009 University of...»

«Ilham Haouas Department of Economics and Finance College of Business Administration Abu Dhabi University P.O. Box 59911, Abu Dhabi, UAE Phone: +971-2-5015658, Fax: +971-2-5860184 E-mail: ilham.haouas@adu.ac.ae Education Ph D, Paris I, Sorbonne University, 2003. Major: Macroeconomics and Labour market Dissertation Title: The Impact of Trade Liberalization in Tunisian Labour Market MA, University of Science and Technology of Lille I, 1998. Major: Applied Economics Dissertation Title: Impact of...»

«Curriculum Vitae September 2011 Name Christopher Michael ADAM Addresses Offices: Home: Dean’s Unit, Room 626 10 Paul Street Australian School of Business Bondi Junction University of New South Wales New South Wales Sydney, New South Wales AUSTRALIA 2022 AUSTRALIA 2052 Tel.: (+61 2) 9389 6288 Tel.: (+61 2) 9385 9695 E-mail: c_m_adam@hotmail.com Fax: (+61 2) 9385 5933 E-mail: c.adam@unsw.edu.au Personal Data Date of birth: 31 August 1951 Place of birth: Suva, FIJI Tertiary Education Ph.D.,...»

«CHAPTER 1 THE FEDERAL HOME LOAN BANK SYSTEM Introduction Many Federal credit programs are targeted toward supplying credit to rural markets, which tend to be comprised of smaller borrowers, smaller businesses, and smaller banks. Wholesale funding has traditionally been made by government-sponsored enterprises1 (GSEs), such as the Farm Credit System, Farmer Mac, and the Federal Home Loan Bank (FHLB) System. Each GSE has a defined and separate mission, though some inter-GSE competition exists...»

«Issues in Informing Science and Information Technology Volume 7, 2010 The Adoption of Single Sign-On and Multifactor Authentication in Organisations – A Critical Evaluation Using TOE Framework Marise-Marie Michael Lane D’Costa-Alphonso School of Information Systems, University of Southern Faculty of Business, Queensland, Melbourne, University of Southern Victoria, Australia Queensland, Toowoomba, Queensland, Australia Marise.Alphonso@gmail.com Michael.Lane@usq.edu.au Abstract The...»

«Innovation and Knowledge Diffusion in the Global Economy A thesis presented by Jasjit Singh to The Department of Business Economics in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the subject of Business Economics Harvard University Cambridge, Massachusetts April 2004 © 2004 – Jasjit Singh All rights reserved. Innovation and Knowledge Diffusion in the Global Economy Thesis Chair: Professor Tarun Khanna Author: Jasjit Singh Abstract The first part of this...»

«sanity, humanity and science real-world economics review Formerly the post-autistic economics review ISSN 1755-9472 Issue no. 52, 10 March 2010 back issues at www.paecon.net Subscribers: 11, 312 from over 150 countries In this issue: Pragmatism versus economics ideology: China versus Russia 2 David Ellerman Racism and Economics Free enterprise and the economics of slavery 28 Marvin Brown Why some countries are poor and some rich 40 a non-Eurocentric view Deniz Kellecioglu The GFC Declaring...»

«The Structure of Rural Financial Markets in Mexico1 Susan M. Richter University of California, Davis srichterprimal.ucadvis.edu Steve Boucher University of California, Davis boucherprimal.ucadvis.edu Chris Woodruff University of California, San Diego cwoodrufpopmail.ucsd.edu Selected Paper prepared for presentation at the American Agricultural Economics Association Annual Meeting, Long Beach, California, July 23-26, 2006. Copyright 2006 by Susan M. Richter, Steve Boucher, and Chris Woodruff....»





 
<<  HOME   |    CONTACTS
2016 www.dissertation.xlibx.info - Dissertations, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.