FREE ELECTRONIC LIBRARY - Dissertations, online materials

Pages:     | 1 |   ...   | 17 | 18 || 20 | 21 |   ...   | 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 19 ] --

–  –  –

4.4 Variable definitions I need to define CEO turnover and board turbulence. In table 4.1 three measures of CEO turnover emerge. First of all, I follow Yermack (2004) in considering all departures as relevant. Second, I supplement with a proxy of forced dismissals constructed as the product of CEO and chairman turnover. As is evident from table 4.1, CEO and chairman turnovers are simultaneous in more than 50 per cent of the CEO turnover cases. I suppose that these cases of coincidence signify dismissals. Third, a measure of forced departures for CEOs is obtained from previous work on CEO turnover in Norwegian companies (Bøhren et al., 2002).

I identify four reasons for not concentrating on forced CEO departures alone. First, the timing of CEO turnover and board turbulence may reveal board control type, for whatever reason the departing CEO left the company. Thus, for this reason alone, all CEO changes belong to the data set. Second, public information on dismissals is likely to emerge only in a minority of cases, probably where the conflict is most acute, since both the firm and the departing CEO want to defend their reputational capital.

Gilson (1990) shows that directors in defaulting firms have trouble obtaining new directorships. This induces them to keep conflicts between CEO and board, or among board members, out of the public eye. Using only this subsample of turnovers is, consequently, likely to result in a seriously biased sample. To top it all, some studies judge a departure as forced if firm performance in advance has been weak. Then, there is no surprise in finding weak performance to be associated with CEO or director departure. Third, the often-used Parrino (1997) criteria3 for selecting forced departures include the explicit dismissals, but then assign a turnover to the forced category if the departing CEO is less than 60 years old. In practice, this means that nearly all departures of CEOs less than 60 years of age are forced. This comes close to the all CEO turnover measure. And even though the public announcement may be that the CEO or the director has found new employment or wishes to leave, the protection of reputational 3 Parrino classifies turnovers as either forced or voluntary, ascertained from the Wall Street Journal (WSJ). The turnover is forced in two main cases: (1) If WSJ reported that the CEO was fired, forced from the position, or departed due to unspecified policy differences;

(2) If the departing CEO is under the age of 60 and the WSJ announcement of the succession (2a) does not report the reason for the departure as involving death, poor health, or the acceptance of another position; or (2b) reports that the CEO is retiring, but does not announce the retirement at least six months before the succession. Searches in other parts of the business and trade press are undertaken for the second group in order to reclassify, if necessary.


capital makes all such announcements untrustworthy. Franks and Mayer (2001) move even one step further, when they define board turnover as the number of directors leaving the board during the year, for reasons other than death or retirement.

Fourth, CEO turnovers in my sample show that departing and arriving CEOs have about the same age distribution, see figure 4.1.

–  –  –

The measure is continuous and always positive and is new to the literature. A value of zero will signify no replacements and no board size change. Table 4.1 shows that the average turbulence is 1.43. Looking ahead, table 4.2 reports that the percentage of new directors on a board is 26.1 on the average Norwegian board. Thus, the turnover of both CEO and directors appears to be high in my Norwegian data. I return to this point in section 4.6.

This paper employs two measures of firm performance; one is the stock return, a market measure, and one accounting measure, the return on assets (ROA). Hermalin and Weisbach (1998) argue that the board will have a greater faith in the accounting measure than in a market measure. The Bøhren and Strøm (2007) measure of independence, that is, tenure difference between the shareholder elected directors and the CEO, is employed.


4.5 Data and methods 4.5.1 Data My sample comprises all non-financial firms listed on the Oslo Stock Exchange (OSE) at year-end at least twice over the period 1989 to 20024. To reduce censoring bias in the tenure measures, I start collecting director data in 1986. The ownership structure data covers every equity holding by every investor in every sample firm. The public securities register provided the ownership data, accounting and share price data is from the OSE’s data provider, and board data was collected manually from Kierulf’s Håndbok and a public electronic register from 1995.

4.5.2 Methods

Since the CEO turnover is a discrete event, taking two values, 1 for a new CEO and 0 for the continuation of the incumbent, and the board turbulence measure is continuous, different estimation methods are employed.

For the CEO turnover the probit model is the estimation vehicle, and I use the general methods of moments (GMM) for the board turbulence. Both relationships are estimated with panel data methods.

I first set out the GMM method. Since the measure of board turbulence is continuous, I may then use fixed effects panel data estimation (Woolridge, 2002). In a panel of firms, the specific firm’s heterogeneity will cause residuals to be dependent. Therefore data pooling will lead to biased estimates. Fixed effects estimation has the advantage of removing the dependency in residuals, since each firm’s overall average on a particular variable is subtracted from the given year’s observation. The transformed observations will be independent. With firm heterogeneity thus removed, the fixed effects method makes it unnecessary to include a host of control variables in regressions. In ordinary least squares estimation on pooled data the objective for the inclusion of control variables is, of course, to remove firm heterogeneity. In the regressions that follow, I retain only the control variables firm size and risk.

4 The OSE had an aggregate market capitalization of 68 bill. USD equivalents by yearend 2002, ranking the OSE sixteenth among the twenty–two European stock exchanges for which comparable data is available. During my sample period from 1989 to 2002, the number of firms listed increased from 129 to 203, market capitalization grew by 8% per annum, and market liquidity, measured as transaction value over market value, increased from 52% in 1989 to 72% in 2002 (sources: www.ose.no and www.fibv.com).


The fixed effects estimations are implemented with GMM, whose great advantage over other methods is that few assumptions are needed. For instance, OLS needs assumptions of homoskedasticity and absence of serial independence. The consistency of the GMM estimator follows only from the fact that it satisfies certain moment conditions (Davidson and MacKinnon, 1993, page 585). When these are satisfied, instruments are given naturally. In this paper, the instruments are constructed from the explanatory variables. Using Davidson and MacKinnon (1993, page 584), the expected value of the dependent variable yt given the information set Ωt is written E (yt |Ωt ) = Xt β t = 1,..., n (4.2) where Xt is a vector of explanatory variables, and β the vector of k corresponding coefficients. Since (4.2) provides the conditional moment condition E (ut |Ωt ) = 0, it follows that for any vector W with elements wt ; t = 1,..., n such that wt ∈ Ωt, the unconditional moments E (wt (yt − Xt β)) =

0. Now, the regressors Xt belong to the information set Ωt, and there are precisely k of them. The practical implication is that the k regressors dene k unconditional moment conditions. I have gone one step further and added transformations of the explanatory variables as instruments (see below). This means that I have more than the necessary k instruments for identification.

The transformations are as follows. First, I use the raw (Amemiya and MaCurdy, 1986), the time-demeaned, and the squared time-demeaned explanatory variables. Furthermore, I include the average and standard deviation of firm-demeaned explanatory variables (Breusch et al., 1989). Finally, the panel data structure allows the inclusion of contemporary as well as lagged instruments. Since explanatory variables are lagged, I use lagged instruments.

Next, I want to estimate the probability that the CEO is new. For this purpose, the probit method, applied to panel data, is used. Early studies (Hermalin and Weisbach, 1988 and Kaplan and Minton, 1994) employ the logit model on pooled data. Panel data methods for the study of CEO and director turnover have only recently come into use, one example being Fich and Shivdasani (2006).

With panel data, the unobserved effects probit model is in general P (CEO Turnoverit ) = 1|(Explanatory variables)it, ci (4.3) = Φ ((Explanatory variables)it β + ci ) t = 1,..., T where P is probability, ci is unobserved firm i heterogeneity, and β is the

4.6. DESCRIPTIVE EVIDENCE 121 vector of coefficients to the explanatory variables. The Φ symbol indicates the standard normal distribution. Two common assumptions are added to (4.3). The first is that the CEO turnover variable is independently distributed across time t, conditional on the explanatory variables and unobserved firm heterogeneity. The second assumption is that unobserved firm heterogeneity is normally distributed with zero mean and a fixed standard deviation ci ∼ Φ 0, σc. With these assumptions in hand, I estimate the relationship with maximum likelihood methods.

For panel data the probit models meet with the incidental parameters problem (Woolridge, 2002, page 484) when assuming a fixed effects model and performing within transformations, leading to inconsistent estimates.

For this reason, the fixed effects model is dropped, and instead I add 14 year and 19 industry dummies to control for firm heterogeneity as much as possible. The dummy coefficients are not reported. Since unobserved firm heterogeneity cannot be removed, the coefficient values will be biased. However, the coefficients will show the correct direction of impact of each explanatory variable. Furthermore, since the bias is known, the slope coefficients are observable at the average of the distribution, that is, when ci = 0 given the assumptions for (4.3). These so-called average parI report the most tial effects (APE) βc are estimated as βc = β/ 1 + σc important in the text.

4.6 Descriptive evidence

The objective in this section is to give some stylized facts about CEO turnover and board turbulence. First, I report some overall statistics, and then turn to the simultaneity of CEO, chairman and director changes around the CEO turnover event.

Table 4.1 in section 4.

3.2 shows the main characteristics of the various definitions of turnover. It is plain that overall CEO turnover is about 20 per cent. This is higher than in American data, where Fich and Shivdasani (2006) report a departure rate of 11.28 per cent. The contemporaneous departure of CEO and chairman takes place at an average rate of 11 per cent.

Identified forced CEO dismissals happen at an average rate of approximately 4 per cent. In Fich and Shivdasani (2006) the forced CEO dismissals are 18 per cent of all dismissals, that is, comparable to the proportion in our data. Thus, my proxy for forced dismissals implies a high simultaneous CEO and chairman turnover. As will become evident, this is repeated for board turnover as well.


Do CEO, chairman and director changes occur with different intensities over the period? Table 4.2 shows changes in each year of my sample.

–  –  –

Clearly, the changes happen with different intensities over the period.

At the lowest, 10.3 per cent of CEOs are new (1995), while the highest is in 2001 at 28.3 per cent of CEOs. The result for the CEO confirms the Huson et al. (2001) finding that CEO firings have been trending upwards during the sample period, although the influx of new directors is at its lowest in the middle. The variation in changes affects the CEO, the chairman and the directors usually in the same year.

In the same table, firm performance, expressed as stock return and return on assets (ROA), is set out for each year. The expectation is that turnover is high in years of low firm performance, but low in years of high firm performance, as former studies show. Yet, it is hard to detect any pattern. These summary statistics show no such definite relationship.

The simultaneity of CEO and chairman departures is shown in figure 4.2.

Figure 4.2

It turns out that most of the chairman changes occur simultaneously with the change of new CEO. Furthermore, notice that changes are fairly symmetric around the year of CEO turnover, with slightly more chairman changes taking place before the CEO turnover.

Figure 4.3 shows the director changes around the CEO turnover event.

Figure 4.3

The same-year spike shows up for director turnover as it does for the chairman. New directors arrive either because they substitute old, or because the board size is enlarged. Again changes are distributed around the CEO turnover year, but this time is clearly higher before CEO turnover than after. The values at the tails of the distribution are less reliable, since fewer observations appear here.

The conclusion to this simple overview is that chairman and director turnover tend to occur simultaneously with CEO turnover, and when they do not, they occur before rather than after the new CEO is in place. This contradicts the CEO control hypothesis, and supports the joint control hypothesis. Thus, the board and the CEO seem to form a team.


4.7 Econometric evidence

Is the simultaneity in CEO and board departures in figures 4.2 and 4.3 also present in a multivariate setting? This section tests for this and for the effects of other variables. Section 4.7.1 reports the results for CEO turnover.

Pages:     | 1 |   ...   | 17 | 18 || 20 | 21 |   ...   | 22 |

Similar works:

«INTERNATIONAL PROJECT FINANCE Law and Practice This page intentionally left blank INTERNATIONAL PROJECT FINANCE Law and Practice Edited by John Dewar 1 1 Great Clarendon Street, Oxford ox dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City...»

«A report by Cambridge Economic Policy Associates (CEPA) for the Department for Transport The extent to which airlines’ interests are aligned with those of passengers Note This work carried out by CEPA for the Department for Transport was commissioned in March 2010 and this report was completed in June 2010. Therefore, any policy references contained herein are those of the previous Government and do not necessarily reflect current Government policy.THE EXTENT TO WHICH AIRLINES’ INTERESTS...»

«February 6, 2011 CURRICULUM VITAE STEPHEN J. CONROY Associate Professor of Economics School of Business Administration University of San Diego, 5998 Alcala Park San Diego, CA 92110-2492 Tel: 619.260.7883 / Fax: 619.260.4891 sconroy@sandiego.edu http://www.sandiego.edu/~sconroy EDUCATION Ph.D. University of Southern California, (Economics) Los Angeles, CA; December 1998. (Dissertation Committee Chair: Richard Easterlin) M.A. University of Southern California, (Economics) Los Angeles, CA; May...»

«The Complete Guide to Public Speaking Other Books by Jeff Davidson The One-Minute Procrastinator (Adams Media) The Complete Idiot’s Guide to Managing Your Time (Pearson Education) 101 Internet Marketing Secrets (Entrepreneur Press) The Complete Idiot’s Guide to Managing Change (Pearson Education) The Complete Idiot’s Guide to Reinventing Yourself (Pearson Education) The Ten Minute Guide to Managing Stress (Pearson Education) The Ten Minute Guide to Project Management (Pearson Education)...»

«International Finance Corporation Climate risk case study: Terminal Maritimo Muelles El Bosque June 2010 Draft report for Vladimir Stenek Climate Change Unit International Finance Corporation 2121 Pennsylvania Ave., NW Mailstop F 3K-300 Washington, DC 20433 USA Our reference IFC002_IDA/MEB Project manager Jean-Christophe Amado Approved by Dr Richenda Connell Acclimatise Hexgreave Hall Farnsfield Newark Nottinghamshire NG22 8LS T: +44 (0) 1865 292010 E: jc.amado@acclimatise.uk.com,...»

«UNIVERSITY OF SZEGED FACULTY OF ECONOMICS PHD SCHOOL EC IN ONOMICS ECONOMIC PSYCHOLOGY RESEAR CH WORKSHOP Szabolcs Prónay Adherence and construction of self in youth consumption – Analysis on the relationship between consumer loyalty and self-image PhD THESIS Thesis supervisor: Dr. Erzsébet Hetesi Szeged, 11 April 2011 I. Topic relevance and objectives of dissertation The issue of brand loyalty has been one of the most current business scientific areas in the past few decades. It burst into...»

«ECONOMIC IMPACTS OF THE PORT OF LOS ANGELES August 20, 2007 Prepared for: Port of Los Angeles Martin Associates 941 Wheatland Avenue Suite 203 Lancaster, PA 17603 www.martinassoc.net TABLE OF CONTENTS I. OVERVIEW OF THE ANALYSIS 1. IMPACT DEFINITIONS 2. METHODOLOGY 3. ECONOMIC IMPACT MODELS 3.1 Marine Cargo Sensitivity 3.2 Cruise Activity 3.3 Port of Los Angeles Marina Tenants 3.4 Fish Processing Activity 3.5 Non-Maritime Real Estate Sensitivity 4. SUMMARY OF RESULTS 4.1. Total Los Angeles Area...»

«SOCIO-ECONOMIC ENVIRONMENT FOR THE ESTABLISHMENT AND OPERATION OF REGIONAL FOOD SUPPLY SYSTEMS PhD DISSERTATION THESES Péter Boros Supervisor: Dr. Zoltán Lakner associate professor, CSc Budapest 1. WORK HISTORY, OBJECTIVES The development of Hungarian food economy has been hallmarked by the faith in the omnipotence of efficiency and productivity for a long time. This involved the development of significant, occasionally unparalleled levels of concentrated capacities even in Europe from the...»

«The Sustainability of Global Financial Imbalances Clemens J.M. Kool* Utrecht University School of Economics November 17, 2009 Abstract In this paper, I provide evidence on the development of gross and net asset positions for a group of 23 developed countries over the period 1960-2008. The data show that both gross asset and liability positions and net foreign asset positions have increased tremendously for most developed countries, especially since the early 1990s. Apparently, the global...»

«2014 Human Rights Institute “Rethinking Education Reform: A Human Rights Perspective” Speaker Biographies Sylvain Aubry, Global Initiative for Economic, Social and Cultural Rights Sylvain Aubry is currently working as a researcher on the right to education for the Global Initiative for Economic, Social and Cultural rights, with which he’s coordinating a research and advocacy project on privatisation in education and human rights in 8 countries. He is also doing freelance work on economic,...»

«Comparative Connections A Triannual E-Journal on East Asian Bilateral Relations Regional Overview: A Frustrating Start to 2015 for the US Ralph A. Cossa, Pacific Forum CSIS Brad Glosserman, Pacific Forum CSIS President Obama initiated his long-awaited (and long overdue) quest for “fast track” or Trade Promotion Authority (TPA) from the US Congress, understanding that final negotiations and eventual passage (or not) of his Asian “rebalance” economic centerpiece, the Trans-Pacific...»

«FIRST MINING FINANCE CORP. MANAGEMENT’S DISCUSSION AND ANALYSIS FOR THE THREE AND NINE MONTHS ENDED SEPTEMBER 30, 2015 GENERAL This Management’s Discussion and Analysis (“MD&A”) should be read in conjunction with the unaudited condensed interim consolidated financial statements of First Mining Finance Corp. (formerly Albion Petroleum Ltd. (“Albion”)) (the “Company” or “First Mining”) for the three and nine months ended September 30, 2015, and the audited consolidated...»

<<  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.