The consequence of discretionary accounting: the CEO succession and the return on the corporate net income
Name: Johnny Xian Student number: 10015280 Study: Economics and Business
Supervisor: Dr. Chih Mao Hsieh Date: 15-07-2014
Content Abstract
1. Introduction
2. Related literature and past researches 3. Data description and setup model 4. Results
5. Discussion
6. Problems due to limitation 6.1 Limitations
6.2 Omitted variables Literature
1 Introduction
1.1 Value of CEO to a firm
New CEO announcements or firing CEO are appealing publication in financial papers. Who will be the next successor, but more obvious what are the reasons behind the replacement. Over the last decades
researches are done in the field of the relationship between CEO’s and the firm performances. The CEO nowadays has become more than just a CEO by title and in terms of value. They fill even more than tasks of directing the company. They can be compared to the company, because of their responsibility and
performances. Further, when a new CEO is announced and the news has reached the public community, the concerning stock prices of the company reacts on this. That is because the company is followed by large number of analyst who makes new reports and analysis, based on prosperity of the new CEO. Analysts pay even more attention to the deviation of the expected performance rather than long term performance (K.A. Farrel, 2003). This shows slightly the value of what a CEO means to firm and to the public.
1.2 Succession on firm performance
The global tenure of CEO is 7.6 years while in 1995 it was 9.5 years (Charan, 2005). A tendency to more flexible succession has become a part of the corporate life. The replacement of bad performing CEO’s triggered this study to investigate how succeeding CEO’s are performing in the new firms. As stated in the first paragraph, the CEO is valuable to the firm, so a succession leads (in) directly changes in the firm performance. The decisions of CEO can be divided into non-discretionary and discretionary. The parts of discretionary decisions have influences on the accounting numbers. So discretionary decision will have influences on the corporate net income. It all depends on how the corporate policy is structured according to the new CEO. In the end succession it self, out of control of the CEO, has an impact on the firm
performances.
1.3 Utilization of discretionary decisions for strategic purposes
In Murphy’s et al. research (1993) three different kind categories succession regard to CEO are made. Three categories of CEO turnover lead to all different discretionary decision for strategic purpose. The first one is an approaching retirement. CEOs make accounting or investment decisions to improve earning in their final year, so their earning based compensation increases. Second, outgoing CEOs make accounting decision in an attempt to cover up the firm’s health. This happens in cases where the firms are performing poorly. Final, succeeding CEOs take a bath, which means rather suffer a big loss or declining certain not profitable
operations and put that loss in a year accounting which lowers the assets. In the following accounting periods when there’s a small improvement, they can report a slightly increased income as “improvement” of their effort. At cost of the changing year, new CEOs stimulate future earnings by writing off unwanted operations and unprofitable divisions. The focus is on the latter group of incoming CEO’s.
1.4 Brief summary
The thesis examines performances of firms and CEO succession that uses the discretionary accounting will be compared to firms without CEO succession. The data consists of CEO succession and accounting numbers on firm levels for a serial subsequence years, variables will conducted from previous researches to adjust for omitted variable bias.
.2. Theory development
A firm faces many adjustments and the financial statements are presented each year. Accounting in financial statements has to follow up the guideline that is explicit written in the Generally Accepted Accounting Principles, leaving options open to land-specific rules of accounting reporting. In particular, one can structurally abusing the accounting discretion for their own benefit or using the options to maximize the shareholders’ value of the firm (Watts & Zimmerman, 1978). Bowen et al. (2008) build accounting discretion in an index consisting of significant accrual usage and smoothing of corporate earnings. In a complete contingent contract, which in order led to optimal governance choices, there would be no relation of governance structure and accounting discretion. The investigation of Becker et al. (1998) in discretionary accounting did find significant evidence, under incomplete contracts, towards poor governance quality. The poor governance structure encourages managers to manipulate accounting numbers and apply discretionary accounting. Especially discretion due to poor governance has some relation with the operation cash flow in the future. After the contract is signed, hidden characteristics of the manager due to moral hazard is the point where the manager can apply discretion accounting in the short run to enriches himself. Using optimal contracting CEO, the level of discretion accounting should already be included in the optimal contract (Core & Larcker, 1999). Shortages in contracts seem to be significant factor in which discretion for own benefit is likely to occur. The exploitation of discretion accounting is making use of this opportunity which contributes to the CEO’s benefit at the expense of the shareholder wealth (Bowen et al, 2008). The influence of
discretionary decision has impact onto the firm performances and in the research of DeAngelo et al. of accrual accounting, they measures the change of operating cash flow. Operating cash flow can be defined as the operating income minus change in working capital. While stock prices and accounting earning are a good measurement (Weisbach, 1988), the discretionary decisions involving R&D, advertising and capital
expenses as well. It will have an effect on the accounting earning or accrual accounting (Murphy & Zimmerman, 1993). For example, Weisbach found out that using accounting earnings gives a strong relationship for performance and turnover of a CEO while Smith and Warner demonstrated that accounting data are less noisy than stock measurement. Hence the use of accounting earnings in this research is the same. Discretion decisions following from the incomplete contract have extrinsic as intrinsic motivation as well.
One of the researches that have been done in the field of CEO turnover is the paper of Clayton, Hartzell and Rosenberg (2000). They make several hypotheses around CEO turnover and the effect on the stock prices. The results of their finding are an increased annual stock return volatility of 3% in the
subsequent 4 years after CEO succession. If the succession of CEO was forced, then an additional 20% are added to the volatility of the stock return. However, the research did not show how the stock prices related to CEO turnover affect the performance. There is no clue whether the turnover of the CEO leads to a better or worse performance. The importance of effect after a change in executive will still have an effect 2 years after date.
In the research of Murphy, instead of focusing on one financial variable affecting the CEO
succession, multiple variables are tested. Some of these financial variables are discretionary based. To be more specific, the thesis focuses on the following discretionary variables: R&D, advertising and capital expenditures. To clarify, discretion means in the following context that mentioned variables are influenced by decisions made by the CEO, expressed in accounting numbers. Discretion is not a mandatory action, but a manually or driven by incentives action which affects other parties. It refers to the ability of deciding certain action in its interest. In contrast, sales, asset and stock price performance are less discretion to the CEO (Murphy & Zimmerman, 1993). Research by Moore has showed that there was a significant change of income reducing accounting when there was change in management. The test in his research with 165 observations has showed that there were 36 cases confirmed as a discretionary accounting decision involvement. 23 out of 36 had the net effect of decreasing income and 1 had as result increased income (Moore, 1973). Because of discretionary decisions, there will be large significant declines in operating performances which will be followed up by large improvement in performances in the following periods (Denis & Denis, 1995).
Several reasons around CEO succession are mentioned. E.g. retirement, end of term/contract and the outstanding bad performance of the company. As mentioned in section 1.3, there are 3 different categories of departures of CEO. Potential discretionary accounting of retirement are most likely to happen where the firm do have a healthy finance and therefore the CEO can improve his remuneration in final year by making steady investment in advance. The other 2 categories of strategic discretionary accounting are likely to happen when the firm is facing poor economic performances (Murphy & Zimmerman, 1993). Supporting the statement of discretionary decisions for own purpose/benefits of the CEO is found in the
research of Finkelstein et al. (1998). It shows some proof onto alignment. They have found out that when discretion is aligned with the pay of CEO, the firm performance is higher. In strategic view this is efficient, but contradicting with what other researches have concluded. If CEO does have the power to regulate certain accounting data, affecting the wage, then an opposite effect will be expected. A high rate of succession of CEO does not necessarily leads to bad performance, but a bad performing firm comes together with a succession. This points out that bad performing firms or poor governance structure results in a CEO succession, but more important in particular is the poor governance structure encourages discretion accounting. In the same research, the role of discretion may explain differences of interest. Discretion could totally be outside the control of the CEO, but also totally in control by the CEO and even half control. For this reason, the discretion of CEO has a significant impact on the organization. The level of discretion determines the impact by the CEO on the firm and it is the marginal contribution, because in the end the compensation is deduced from his contribution to the firm (Finkelstein & Hambrick, 1988). Since the compensation is depending on the level of discretion, it implied also a risk factor. Risk is greater because the action that is done or the decision that is made by the CEO has unpredictable outcome which cause an increase in uncertainty and complexity.
There is a relationship between the compensation scheme of CEO and CEO discretion, because a CEO of high discretion firm earns more than CEO of low discretion firms. Higher earnings compensate the risk involvement of the CEO for the decision he made. Finkelstein & Boyd (1998) showed a strong support for a positive relation between the managerial discretion and the compensation scheme. Also, the size of the firm is a factor which determines a part of the discretionary accounting. This proof is given by Chakrabarti & Halperin (1990). They investigated that the level of R&D expenditures is dependent on the firm size. The net income and firm size is a factor which determines the level of R&D expenditures beside the managerial discretion in large firms, while R&D expenditures of small firms are related to the net sales.
In this research it needs to be clear how strong the relationship is between the R&D expense and the concerning net income and firm size. If the relationship is strong between net income and firm size to the R&D expense, then a part of the R&D expense is explained by those two factors.
Hypothesis 1: R&D expense has a strong relation to the net income and firm size.
According to these researches the expectation of a CEO succession will be followed up by a change in the corporate net income, especially if there is incentive for the CEO to use discretionary accounting. The level of discretionary accounting is depending on the compensation scheme of the CEO, so the greater the level of discretionary accounting (allocated to the CEO), the lower the corporate net income will be and because, a successions of the CEO has also a negative impact on the corporate net income on the long term. A negative relationship will be expected between discretionary accounting and corporate net income
Hypothesis 2: A negative relationship will be expected between discretionary accounting and corporate net income. .
3. Model setup and data description 3.1 Econometric model
The following model explains the effect of discretionary accounting together with CEO replacement. Net incomei,j = α0 + α1*CEODummyi,j + α2*DiscrDummy + α3*CEODummy*DiscrDummy +
α4*R&Dexp*CEODummy + α5*CAPexp*DiscrDummy + α6*ADVexp*DiscrDummy + α7*Firmsize + εi,j The CEOdummy is dummy variable for the CEO succession. The data consists of accounting data, specific on firm level. With these accounting data each fiscal year, it is possible to measure the modifications effect of discretionary accounting. The corporate net income functions as the performance measurement of the companies. Of great importance is the independent variable CEOdummy, because this variable measures the effect of the CEO replacement on the return of the firm in the following 3 years. It takes the value of one when a CEO is replaced in a particular year and zero when none is replaced. Explanation of i=company and j=year. The other variables include advertising expense, capital expense and R&D expense as the
discretionary accounting. Using firm size as the control variable to avoid omitted variable bias. To test the effect of CEO turnover and discretionary decision, it involves a cross-sectional data research spread over years. For this reason, the regression is a panel regression. αi is the intercept/constant for each company. There is need to generate a dummy for CEO turnover.
3.2 Sample data collection
This study emphasis the effect of the CEO succession. For this reason no distinction are made between forced, voluntary and retirement of the CEO, while most of the previous done researches do make a distinction. For the convenient, CEO that are placed for a certain fixed period after a sudden event, are not included in the data, because these interim CEO’s are not assigned as permanent and therefore the time range is too narrow to influence the corporate accounting on the long term. The CEO’s succession process by the board plays no effect in this research. There will be no difference between internal or external attracted CEO in the company. Corporate income statements data, in particular accounting data, can be acquired through Wharton Research Data Services (WRDS). WRDS is an internet based dataset available
for global academic researches. It contains data from different economics fields. It provides data about the turnover of CEOs and specific accounting data from companies in the United States. However the CEO turnover range is limited in this research to 1997 until 2012, but the discretionary and accounting data can be collected from before 1997 and after 2012. The data used for thesis contains of the following numbers: The sample contains a total of 1566 different sizes firm in U.S. listed in Compustat Annual Updates obtained from the Northern U.S. pool mixed firms. The Northern U.S. pool mixed firms stands for firms coming from all different sectors and are in the Northern U.S. the biggest corporate firms. In total, the amount CEO from the sample is 2451. From these CEO, if the company has a CEO succession and only if the successions are between 1997 and 2012, and then it is defined as the effect of CEO succession of that specific firm.
Following from the previous stated statement if and only if, there were 1802 times turnovers in 15 years used sample range. The time range is from 1997 until 2012, conducting a serial observation per firm over the years.
3.3 Variables
Distinguishing CEO title
Data of CEO turnover is obtained from the WRDS Risk Metrics (ISS) Director. The data comes from 2012. The dataset gives a range between 2007 and 2012, but for distinguishing who the CEO is for the particular company, the recent data from 2012 gives more information about job titles in a company. Since Risk Metrics does not contain a variable that properly assign the CEO of the company, for the reason that they does not make a distinction between e.g. a chief officer for administration and CEO, there is need to distinct who the CEO is for that year and when he acquired that title or succeeds the previous CEO. To identify the CEO, the company name variable and primary company name variable have to be the same in order to retrieve the CEO title. The primary company name variable describes which company is the primary company where the employee works for, besides other activities at other companies. Two separate tables are created, first when company name variable is the same as primary company name variable, then the CEO title is appointed to that unique employee in the year that he succeeded the previous CEO and the second table reflects the year where the CEO began his services, since the observations are from 1997-2012, only succession in or after 1997 matters to the CEOdummy. But, because this manner of distinction is open for errors, a second check for errors is done manually. The errors happen in cases where company name and primary company name differs in small occasions of a symbol or not corresponding correctly exactly. These are excluded from the research. Another small notification due to manually checking has arisen. Some of the companies show error that is retrieved from WRDS Risk Metrics (ISS) Director. A tracker which is not corresponding with the company is omitted from this dataset.
CEO discretionary
Bowen et al. (2008) used the variable SMOOTH to describe the action the CEO takes to smooth the earnings. This earnings smoothing is a component of what is considered as discretionary accounting, because using excessive accruals to smooth the earnings than in comparison what other firms do, leads to discretion accounting. Discretion itself is an inner aspect, which is not observable. To determine discretionary accounting, only estimators of actions where discretion is involved and accounting data that can be verified as discretionary decision are plausible, for example what Bowen (2008) used in his research. When there is more variability in the operating cash flow than the earnings, then there is accrual to smooth the earnings of the company. The estimator is dependent of the standard error of the operating cash flow and earnings. Divide the standard error of the operating cash flow by the standard error of earnings, when the ratio exceeds 1, then it is assumed that the accounting accruals are used to smooth the earnings. For the econometric model, the earnings and operating cash flow are collected and divided yearly to compute the CEODiscretionary dummy variable.
Capital expense and advertisement expense
The two variables are built into the model, because are influenced by direct discretion decisions of the CEO. However, different than R&D expense, these two variables are mainly dependent on whether the CEO uses discretion accounting or not. The capital expenses represent the portion of funds used for acquiring property, plant and equipment, but exclude the amount of funds through acquisitions. The funds coming from
acquisitions are excluded, because the shift in the capital expense is not in line with other companies which do not have this acquisition.
R&D Expense
The R&D expense is also a part of discretion accounting where the CEO can apply his decisions on, but if there is relation to the firm size and net income, and then a part will be explained by those factors. This expense is different in regarding to the other two expenses that is built into model, because Financial Accounting Standards Board has decided that R&D expense can only be deducted in the year that the expense has incurred. So, for the CEO exploitation he will apply in the last year of his duty to boosts his final remuneration. The variable is therefore only dependent whether these was a CEO succession and therefore, eliminates the partial effect of net income and firm size. This data will be collected annually.
Firm size
The firm size variable is displayed as the numbers of employees in the firm working as a control variable. It represents the human capital of the firm and with these numbers a distinction can be made between small and big firms, because the level of net income depends on it. A large firm has a different scale than a small firm in proportion to the net income. However, there is more possibility to determine firm size, but all of them are strong correlated making no differences which one to choose for. The firm size is a dummy variable for a large firm with more than 1000 workers and a small firm less than 1000 workers. Another criterion of large firms is when an R&D expense is more than 1 million dollars, so the firm size dummy variable is the cross-product of workers-dummy and R&D expense dummy.
3.3 Time invariant assumption
The observations are from multiple, different companies and over multiple years or different time. The panel data analysis is used to test the relationship between CEO succession and the effect on corporate net income. Because of observations over years, some of the variables do not change over time, since these are correlated within the company itself and therefore time-invariant in the model. One assumption needs to be made for this: Unobservable variables that affects the regression is not affected by time, otherwise it is not possible to observe the effect. Making this assumption it is now possible to observe the effect of succeeding CEO and discretionary decision within a company. The essential part of the research is controlling
discretionary decisions, so that is a variable which differs across companies. In fact, it accounts the CEO turnover as the individual heterogeneity.
3.4 Retrieving accounting data
The data is retrieved from WRDS Compustat Annual Updates. The sample dataset has to correspond with the acquired data of the CEO, which are the variables that describe the operating results. In order to the choice of what Murphy & Zimmerman consider as discretionary decision of the CEO consist of the following variables: the research and development expenses, advertising expenses and capital expenses. The variables give the implication of reducing income while increasing expenses. The time period after CEO succeeding is 3 years that will be tested. As mentioned above in section 2 the effect of an executive change is 2 years, in this research a longer period of 3 years will be tested.
In this section, the descriptive statistics and the measurements of the regression are presented. The results of the discussed variables in the previous sections will be displayed in this section.
Table 1 Relation between R&D & net income and firm size
From the sample regression the correlation matrix shows the relation of the variables. This independent matrix is built to investigate the relationship between R&D expenses and the firm size and net income as stated in hypothesis 1. The firm size takes value of 1 when the company has more 1000 workers, which qualifies as a large firm. In table 1 the correlation between firm size and R&D expenses are low, but positive related to each other. This is not supported by the hypothesis that is made. While net income and R&D expense shows a relatively high and positive relation. Net income and R&D expenses have correlation of 44%.
Table 2 CEO discretionary accounting and successions
The regression functions in table 2 consider the variables which could explain the hypothesis 2. The
joint hypothesis. The sample is Normal distributed. The dependent variable is Net Income, which represents the corporate net income. The interpretation is explained in the following section. According to the
regression, the constant has the value of 347.855 (25.38) in dollars. The CEODummy shows a minus of 251.835 (65.53) significant at 5%, so the CEO succession has negative influence on the corporate net income in comparison with other firms who do not experience a CEO succession. DiscrDummy shows negative 263.333 (27.13) and can be interpreted as when the CEO use earning smoothing as discretion, the corporate net income will lower at significant rate. Both CEO succession and CEO discretionary accounting leads to a lower corporate net income, but CEODummy*DiscrDummy shows a positive value to the net income, so if there is a CEO succession and the CEO use discretionary accounting, it boost the corporate net income by 203.172 (93.67). What was stated in hypothesis 2 is not in line with the findings here. The previous two variables of CEO succession and CEO discretionary accounting both should have negative influence on the corporate net income, which they do. But together, it has the opposite effect on the
corporate net income, however, this coefficient is only significant at 10%>. The R&Dexp*CEODummy shows a positive significant effect, which means that if there was CEO succession in relation with the R&D
expenses, the net income was higher. This significant coefficient is in line what was expected, because if the CEO leaves the firm, he exercises discretionary accounting by cutting R&D expenses and having a higher net income in his final service year to retrieve a higher remuneration bonus. Both CAPexp*DiscrDummy and ADVexp*DiscDummy shows a positive relation to the net income. This is not what was expected, because if there was a discretionary accounting, the level of net income would be lower, but both of the expenses have significant positive relation to the corporate net income. The control variable firm size has a positive relation to the net income, a large firm will have on average 48.1582 (26.91) higher net income than a small firm. However, this coefficient is only significant at 10%.
5. Discussion
In this part the results from the last sections will be compared to the two hypothesis stated in the section 2 and the literature found in the same section. If the hypothesis is not supported by the results found, section 6 will discuss the eventually explanations.
The first hypothesis is to research the reciprocal relationship between on one side the R&D expense and on the other side the firm size and net corporate income. First, the relationship i.e. correlation between firm size and R&D expense is 4.46%. Considering, the predictive power i.e. the degree of correlation relation between R&D expense and firm size is weak related and therefore, firm size does only have a low influence on the level of R&D expenses. Firm size has therefore not a dependency on the level of the R&D expenses of the firm. Second, the hypothesis 1 also investigates the level of R&D expense by the net income. The variable R&D expenses and net income results from section 4 are strong related. Important finding of this
research is that there is a reciprocal relationship between R&D expense and net income. R&D expense is strong correlated to net income, because from the result, it shows that net income determines for 44% the level of R&D expense of the company. The results supports hypothesis 1, because the level of R&D expense is depending of the level of net income and the main reason is corresponding from the literature by
Chakrabarti & Halperin (1990). The support for this hypothesis provides information of the level of discretion by the CEO. The explanation is a higher correlation between R&D expense and net income, the lower the level of discretion will be for the CEO. Since 44% of the R&D expense is determined by the net income of net income, discretion influence onto R&D expense will therefore be lower. Discretion by the CEO to influence the R&D expense for its own stake i.e. the level of influence on R&D expense is less likely to be determined by the CEO than the net income. For this reason, the R&D expense is not fully allocated to the discretion of CEO. The results of table 2 shows CEO succession and the discretionary R&D expense both together boost the net income by 0.9678 (0.0812). In another words, CEO leaving the company that partly can influence the R&D expense, will cut the R&D expense in his final year to boost the net income in order to grant a higher remuneration for its service.
The second hypothesis predicts a negative relationship between discretionary accounting and corporate net income. The results from section 4 show a support for this hypothesis. The descriptive statistics in table 2 shows significant results for DiscrDummy, suggesting that hypothesis 2 is not rejected, because discretionary accounting did contribute to a lower net income according to the research. The idea behind this is leaving too much accrual such that the CEO can use these accruals to smooth the corporate net income. The same result is concluded from the research of DeAngelo et al. (1994). Their finding is that discretionary accounting as the accruals have influence on the firm performances. This model shows only a negative relation between the firm performances as in net income and accruals as in discretionary
accounting (earning smoothing). However, if this discretionary accounting goes together with the CEO succession, a positive relationship has been shown, significant at 10%>. This coefficient can either be significant at 10% or insignificant. In case where this coefficient is significant, it will be not in line from what was concluded from the previous researches. Discretionary accounting together with CEO succession should lead to a lower corporate net income, while in this research the model shows a positive coefficient. Only one research that corresponds with this result is from Murphy & Zimmerman (1993). According to them, there is no CEO discretion in strongly performing firms of retiring CEO. In this research the assumption was made that all successions are treated equally, probably there was distinction between the performance level of the firms and the distinction between regular succession and forced succession.
From the research that is done by Murphy & Zimmerman (1993) confirmed higher capital expenses and advertising expenses due to discretionary accounting leads to a lower firm performance. The regression in model 2 shows noticeable remarks in the regression. Discretion in capital expenses and advertising leads to a significant positive related to the corporate net income. None of studies done before has found this
results. What might be possible is that there is some difference between this study and other study. Causes might be the firm performance measurement or there is no direct link of net income and these two expenses, while other studies do suggest this. The study to discretion of CEO (a psychological phenomenal) and accounting (economics numbers) is a research in different fields. The direct link suggests that the study by Murphy & Zimmerman (1993) uses growth rate of the expenses and in this study the expense are absolute numbers. The difference might occur because the data used in this study are the biggest corporate firms, but the scale of expenses varies from 1 million to max 100million. This remark will be discussed in section 6.3, in the section of problem, limitation and recommendations.
6. Problems due to limitations and recommendations 6.1 Limitations
An improvement to this research is using quarterly accountings numbers rather than yearly from Compustat. The main reason for yearly accountings number in this research is the limitation of Risk Metrics, which provides only the year of succeeding CEO instead of the exact date. Together with this limitation, there was no access to the reason for succession of the CEO. As it is shown in the section 5, the reason behind the succession and from the research of Murphy & Zimmerman (1994) turns out that there is difference between the ways the CEO is dismissed.
6.2 omitted variables
The regression suffers from bias of omitted variables. In general, these are independent variables which are included in the regression, while having correlation with the error term. Beside the capital, R&D and
advertising expenses, the discretionary decision are dependent on more factors. Intrinsic motivation or the contentment of the CEO within the company plays a factor that contributes to the level of discretion he exercise. The role of stress, work, conflict and financial all play a part in the job and life for the CEO, which could all be a variable on what decision the CEO would make (Daily & Near, 2000).
6.3 Firm performance
Researches of this topic done by others have different manners to measure the firm performances. Firm performance can be either measured by stock price, accounting data and many more. To improve this study to discretionary accounting, in the future, including multiple performance measurement might help to understand whether in this study, corporate net income, is a good measurement for discretion of the CEO.
Also, the scale must first be translated into growth rates or percentages. In this study, it might be better to use percentages depending of the firm performance (corporate net income) instead of absolute number for expenses. Control variable for firm size was not enough, because the scale of the expense varies too wide and this has influences on the weight of each company in the data.
Literature:
Bowen, R.M., Raigopal, S. & Venkatachalam, M. (2008) Accounting discretion, corporate governance, and firm performance, Contemporary accounting research, 25(2), 351-405
Becker, C., DeFond, M., Jiambalvo, J. & Subramanyam, K.R. (1998) The effect of audit quality on earnings management. Contemporary Accounting Research, 15 (1), 1–24.
Chakrabarti, A. & Halperin, M. (1990) Technical performance and firm size: Analysis of patents and publications of U.S. firms, Small business economics, 2(3), 183-190
Charan, R. (2005). Ending the CEO succession crisis. Harvard Business Review, 83, 72-81.
Core, J., Holthausen, R. & Larcker, D.F. (1999) Corporate governance, chief executive officer compensation, and firm performance. Journal of Financial Economics, 51 (3), 371–406.
Catherine, D., & Near, J. (2000) CEO satisfaction and firm performance in family firms: divergence between theory and practice. Social indicators research, 51(2), 125-170
DeAngelo, H., DeAngelo, L., & Skinner D.J. (1994) Accounting choice in troubled companies. Journal of accounting and economics, 17, 113-143
Denis, D.J., & Denis, D.K. (1995). Performance changes following top management dismissals. The journal of finance, 50, 1029-1057.
Eaton, J. & Rosen, H.S. (1983) Agency, delayed compensation, and the structure of executive remuneration. Journal of Finance, 38, 1489-1505.
Farrell, K.A., & Whidbee, D.A. (2003). Impact of firm performance expectations on CEO turnover and replacement decisions. Journal of accounting and economics, 36, 165-196.
Finkelstein, S. & Boyd, B.K. (1998) How much does the CEO matter? The role of managerial discretion in the setting of CEO compensation, Academy of Management Journal, 41, 179-199
Moore, M.L., (1973). Management Changes and Discretionary Accounting Decisions. Journal of Accounting Research, 11, 100-107
Murphy, K.J. & Zimmerman, J.L. (1993). Financial performance surrounding CEO turnover. Journal of accounting and economics, 16, 273-315.
Rosenberg, J.V., Clayton, M.J., & Hartzell, J.C. (2000). The impact of CEO turnover on equity volatility. Federal Reserve Bank of New York Research Paper Series - Staff Report, 166
Smith, C.W., & J. B. Warner, (1979) On financial contracting: An analysis of bond covenants. Journal of Financial Economics, 7, 117–161.
Watts, R., and J. Zimmerman. 1978. Towards a positive theory of the determination of accounting standards.
The Accounting Review, 53 (1), 112–134.
