Only time will tell : the effect of uncertainty on earnings management

Only time will tell

The effect of uncertainty on earnings management

Name: Ewout Baaij

Student number: 11145811

Thesis supervisor: dr. P. Ghazizadeh Date: Augusts 20, 2018

MSc Accountancy & Control, specialization Control Amsterdam Business School

Statement of Originality

This document is written by student Ewout Baaij who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Abstract

This thesis studies the effect of firm age on earnings management. The main research question is whether there exists a clear relationship between the age of a firm, as a proxy variable for the level of shareholder uncertainty, and the amount of earnings management occurring in the firm. After reviewing the literature on this subject, we arrive at the hypothesis that a firm with a lower age, as a proxy variable for the level of uncertainty, has a higher amount of earnings

management occurring in its organization. The hypothesis tested in this thesis has been confirmed by the empirical results of the OLS regression model used. The five different variables for earnings management consist of the residuals of the total accruals from the Jones model, Modified Jones model, Jones model including lagged ROA, Modified Jones model including lagged ROA and discretionary current accruals. The existing research which supports the regression results argues that lower firm age means that it is listed for a shorter time period on the stock exchange and because of this there would be less information available about this firm. This results in a higher level of uncertainty for the shareholders, in the way that is harder for them to determine the true current state of the firm they are investing in or are planning to do so in the future. Furthermore, older firms have more restraints towards accruals management and thus less uncertainty for the shareholders. Finally, firms with greater endogenous accruals-generating potential have greater uncertainty about reported earnings.

1

Table of Contents

2.3 Firm Age and Uncertainty………...………8

2.4 Earnings Management and Uncertainty ..………9

2.5 Implications and Hypothesis ………..11

3. Methodology……….………12

3.1 Jones model ……….……….………….12

3.2 Measuring earnings management and uncertainty ..………14

4. Data and Descriptive Statistics……….15

4.1 Earnings management and firm age………15

4.2 Descriptive Statistics………..………15

5. Results………17

5.1 OLS regression results………..……….………17

5.2 Robustness checks and R-squared……….………20

6. Conclusion and Discussion……….………21

2

1. Introduction

In the past decade, there have been numerous scandals around earnings management in the world. For example, the case of Toshiba in 2015. The technology company announced in May 2015 that it might had to revise its profits for the previous three years because of a possible recently discovered accounting scandal. The CEO of the company, Hisao Tanaka, left in July of the same year and referred to the scandal as "the most damaging event for our brand in the company's 140-year history." It turned out that during the seven years prior to the discovery of the fraud, the profits had been inflated by a total of $1.2 billion.

The Research I will be conducting focuses on this concept of earnings management. Earnings management is the practice of presenting relatively positive numbers and views of a company's financial position and business activities with the help of accounting techniques. The possibility of engaging in earnings management activities arises for these companies, because of the fact that their managements are asked to make judgments for the application of most of the accounting rules and principles used in their financial statements.

Accruals are defined as the combination of revenues and expenses that impact a

company’s income statement and balance sheet. The non-discretionary accruals are considered to be the component of the total accruals that is imposed by the accounting regulator regarding the adjustments of a company’s cash flows. The discretionary accruals are considered to be the component of total accruals that managers can choose within the flexibility of accounting regulations regarding the adjustments of a company’s cash flows. These discretionary accruals thus represent an important indicator of earnings management. When a company has a higher number of total accruals as a percentage of total assets, this increases the possibility that the earnings quality of this company is relatively low.

Gunny (2010) argues that there are two different types of earnings management that can be used by a company’s management to achieve its goals: real activities manipulation (RM) and accruals management. Real activities manipulation (RM) involves adjusting the firm’s

underlying operations with the sole purpose of increasing its income. On the other hand, accruals management involves the different possibilities of accounting methods used to document these operations. Gunny (2010) also states that a firm’s management may decide to engage in accruals management rather than in real activities manipulation, because they are able to engage in accruals management at the point in time, the end of the fiscal year, when they are most certain

3

about their need for earnings management. Because the above explains why managers may prefer the accruals management option, I will focus my research on this specific type of earnings management.

In the past decade, there have been numerous scandals around earnings management in the world. The gap in the existing literature that I would like to fill is to combine the recent research on earnings management with that on firm age and shareholder uncertainty, to find out if there is evidence that firm age is an important indicator of earnings management and can help explain this phenomenon. When evaluating the literature on these topics, one could argue that higher firm age has a negative effect on the level of earnings management. This because younger firms provide more uncertainty for shareholders (Barry and Brown, 1985) and earnings

management occurs more when there is more uncertainty (Francis, Maydew, and Sparks ,1996). So, it will be very interesting to see if there is also more earnings management occurring in younger firms.

On the other hand, one could argue that firm age has a negative effect on the quality of corporate governance (Kieschnick and Moussawi, 2017 & Loderer and Waelchli, 2010) and in these poorer governed firms, we would expect to find a higher level of earnings management (Ali and Hirshleifer, 2015). This remarkable gap in the existing literature will be the starting point for my thesis and it will be interesting to see if my research can contribute to the available knowledge on the concept of earnings management. Therefore, my main research question becomes:

How does the age of a firm, as a proxy variable for the level of shareholder uncertainty, affect earnings management in the organization?

My findings will have important implications for regulators, corporate managers and investors.

I will use data for this research obtained from the Wharton WRDS database, for the period of 2000 - 2017. Firstly, firm age, our proxy variable for the level of uncertainty, is measured as the number of years that a firm is listed on the CRSP database. Continuing, for our data on our earnings management variables, I will begin with all firm-year observations from the COMPUSTAT Industrial Annual and Research files for the same time period. Following the

4

method of Kothari et. al (2005), my variables for earnings management will consist of the residuals of the total accruals (TAt) from the Jones model, Modified Jones model, Jones model

including lagged ROA, Modified Jones model including lagged ROA and discretionary current accruals. Total accruals are constructed from the change in non-cash current assets minus the change in current liabilities excluding the current portion of long-term debt minus depreciation and amortization.

The remainder of this thesis is structured as follows: the next chapter reviews the existing literature on earnings management, uncertainty and firm age and how I applied this to arrive at my research question. The third chapter continues with an explanation of the research design and methodology that has been used for answering the main question of my thesis. Subsequently, the data for the dependent and independent variables used and its descriptive statistics will be

provided. Thereafter, I display the results from the empirical analysis that has been conducted. Finally, this thesis finishes with a conclusion and a discussion of the results obtained during my research.

5

2. Literature Review

The following chapter will discuss the different concepts and main theories used to arrive at and answer the research question of this thesis. At the end of this chapter, the hypothesis used to answer the main question will be derived and explained.

2.1 Earnings management

Earnings management is the practice of presenting relatively positive numbers and views of a company's financial position and business activities with the help of accounting techniques. Earnings management provides the market with financial statements which include inflated total assets and/ or revenues and earnings, by exploiting the application of the current accounting rules in place. The possibility of engaging in earnings management activities arises for these

companies, because of the fact that their managements are asked to make judgments for the application of most of the accounting rules and principles used in their financial statements.

Accruals are defined as the combination of revenues and expenses that impact a company’s income statement and balance sheet. For example, future interest expenses and

accounts receivable are possible accruals. When a company has a higher number of total accruals as a percentage of total assets, this increases the possibility that the earnings quality of this company is relatively low. The non-discretionary accruals are considered to be the component of the total accruals that is imposed by the accounting regulator regarding the adjustments of a company’s cash flows. The discretionary accruals are considered to be the component of total accruals that managers can choose within the flexibility of accounting regulations regarding the adjustments of a company’s cash flows.

Bergstresser and Philippon (2004) provide evidence that the use of discretionary accruals to manipulate the reported earnings occurs more often at firms where CEO compensation is strongly related to the value of stock and option holdings. On top of that, the authors observe that during years of high accruals for those companies, their CEOs, among other insiders, sold a higher amount of their shares and exercised particular large amounts of options as well. In addition, Cornett and Tehranian (2009) find that more independent boards appear to constrain earnings management.

Continuing, Ali and Zhang (2015) examine changes in CEOs' incentives to manage their firms' reported earnings during their tenure. They observe that in the early years of CEOs'

6

service, earnings overstatement occurs more often than in the later years, and this phenomenon seems to happen less often at companies where more internal and external monitoring occurs. The authors interpret their findings as evidence that newly appointed CEOs try to manipulate the market’s overall perception of their skills during the beginning of their tenure, when there is more uncertainty present in the market. Moreover, the observed amount of earnings

overstatement is also higher during a CEO’s last year of service before he leaves the company. One could argue that if earnings management happens more in firms where a new CEO is just appointed and the market is more uncertain, then perhaps it would also happen more often at relatively younger firms, when considering firm age as a form of uncertainty. This would imply that there would be a possibility of obtaining interesting findings when looking at the

relationship between earnings management and a firm’s age.

Gunny (2010) states that there are two different types of earnings management that can be used by a company’s management to achieve its goals: real activities manipulation (RM) and accruals management. Real activities manipulation (RM) involves adjusting the firm’s

underlying operations with the sole purpose of increasing its income. Examples of RM are decreasing sales or R&D expenses. On the other hand, accruals management cannot be achieved by adjusting a company’s operating activities, however it does involve the different possibilities of accounting methods used to document these operations. Accruals management has to happen at the end of the fiscal quarter or year and it is uncertain for a firm’s managers which specific accounting techniques will be allowed or forbidden by the auditor at this point in time. The firm’s management decides on the operating activities while decisions on the accounting techniques depend on the judgement of the auditor.

Gunny (2010) also states that a firm’s management may decide to engage in accruals management rather than in real activities manipulation, because they are able to engage in accruals management at the point in time, the end of the fiscal year, when they are most certain about their need for earnings management. While RM decisions have to be made prior to the end of the fiscal year, when the need for earnings management is much less certain.

Because the above explains why managers may prefer the accruals management option, I will focus my research on this specific type of earnings management and will use the Jones-model and its modified version for the discretionary accruals. Discretionary accruals are the

non-7

obligatory expenses, for example the future bonuses that still must be paid out to the management of the company but are already recorded in the books.

Total Accruals (TAit) will be my variable for earnings management and will be defined as

the change in non-cash current assets minus the change in current liabilities excluding the current portion of long-term debt minus depreciation and amortization following the research of Kothari et. al (2005) and with their reference to COMPUSTAT data items, TA = (Data4 - Data1 - Data5 + Data34 - Data14)/lagged Data6.

2.2 Firm age

Regarding the topic of firm age, Fort et. al (2013) provide evidence in their paper that young firms, which usually also have a small firm size, respond differently to cyclical dynamics than older firms. The younger and smaller firms respond stronger to changes in these cyclical dynamics than the older firms with a larger firm size. The results that the authors have found when they compared the same phenomenon between relatively old firms with a large firm size and those with a small firm size are miscellaneous.

Continuing, Thornhill and Amit (2012) provide evidence that the reasons why older and younger firms fail are very different. After they had collected and analyzed data on the

bankruptcies of Canadian companies, they concluded that the reason for bankruptcy of older firms is often due to the lack of their capabilities to adapt themselves to environmental change. However, bankruptcies of younger firms where often due to shortcomings of the company’s management team, both in forms of their knowledge as well as their financial management abilities.

Furthermore, Kieschnick and Moussawi (2017) find that a company’s choices on whether they will issue debt and when they decide to do so, the total amount they will borrow from their debtors, is directly affected by their corporate governance structures and their age. They reveal a couple of striking findings in their paper. First of all, they find that even though older companies use debt more often than younger ones, the effects of firm age on the amount of debt used are the other way around. Younger firms thus tend to use a relatively high amount of debt once they decide to use debt, compared to older firms. The authors mainly seek the explanation for these findings in the relationship between a company’s age and its corporate governance structures. They interpret these results as an implication that a company’s decisions on capital structures are

8

heavily influenced by a manager’s risk preferences. This follows from their conclusion that companies tend to use less debt during their ageing when insiders are able to exercise more power.

Moreover, Loderer and Waelchli (2010) argue that the profitability of companies declines when their age increases. They provide two potential explanations for their findings. The first reason being a higher firm age could potentially mean that the company is relatively slow and that organizational changes are difficult to implement. Furthermore, this could also include a decline in R&D expenses, other investments, the firm’s growth rate and value of assets and also potentially a rise in its costs. The second explanation from the authors is that, supported by the lower level of corporate governance, larger amount of CEO pay and higher average number of board members they observe in older companies, the higher firm age could lead to an increase in rent-seeking behavior inside the company. Rent-seeking behavior is defined as a firm that is trying to create wealth and economic gain for itself at the cost of society, for example a lobby for subsidies.

Regarding the field of corporate governance, Dai et al. (2014) find that compared with insiders of poorer-governed companies, insiders of better-governed companies earn significantly smaller profits from their opportunistic insider trades. In addition, Ali and Hirshleifer (2015) argue that companies with more opportunistic insiders experience higher levels of earnings management.

Summarizing the literature used above, one could argue that firm age has a negative effect on the quality of corporate governance and in these poorer governed firms, I would expect to find a higher level of earnings management. On the other hand, one could argue that if

earnings management happens more in firm’s where a new CEO is just appointed, and the market is more uncertain, then perhaps it also would happen more at relatively younger firms, thus the other way around. Finally, I will determine firm age as the number of years that a firm is listed on CRSP.

2.3 Firm age and uncertainty

Firms with a long history are considered to have more information available to the market (Barry and Brown, 1985). This originates from the fact that older firms are more likely to be in more mature industries and firm age also captures the underlying volatility at the industry level.

9

Furthermore, there is obviously no extensive track-record of past performance available for the shareholders regarding newly listed firms on the stock exchange. This implies that the type of uncertainty present here is the one for the shareholders to determine the true current state of the firm they are investing in. So, they consider firms that are relatively recently listed as more uncertain.

In addition to this, Ling et al. (2007) argue that companies with a younger age must cope with many potential uncertainties and threats related to the existing literature on organizational lifecycle. Choi & Shepherd (2004) state that younger firms must anticipate rapidly when an investment opportunity arises. However, it usually proves to be a difficult matter to achieve this goal, because all top managers must be quickly convinced about the value creation of the

investment opportunity and due to their different judgements, it is hard to reach an agreement on the investment in time (Baron, 2006; Keh, Foo, & Lim, 2002). Baron & Markman (2005) add to the findings above that due to their issues with timely decisions on investment opportunities, younger companies are less able to benefit from the possible advantages of founder-CEO’s collectivism than older firms. Where founder-CEO’s collectivism is defined as them focusing more on the growth of their entire company and workforce than on solely their personal wealth.

Finally, after considering the different possible types of uncertainty associated with firm age, the type of uncertainty that I will use for my research is shareholder uncertainty, in the way that is harder for them to determine the true current state of the firm they are investing in or are planning to do so in the future.

2.4 Earnings management and uncertainty

Francis, Maydew, and Sparks (1996) argue that companies face a higher level of uncertainty about reported earnings when they experience greater endogenous accruals-generating potential. This originates from the fact that outsiders (shareholders) find it harder to distinguish

discretionary and nondiscretionary accruals. Here, the type of uncertainty meant is again the uncertainty for the shareholders about determining the true current state of the firm they are investing in. In addition, Becker et. al (1998) argue that companies with these characteristics would be wise to hire a Big Six auditor to signal to their shareholders that they are trying to prevent earnings management in their firm, because this will decrease the uncertainty of the shareholders and will lead to a higher share price.

10

More evidence on the relationship between uncertainty for shareholders and earnings management is found by Barton and Simko (2002). They state that the possibilities for a company to manage their accruals are usually limited. Their possible decisions on accruals management in the current year are constrained by the management’s decisions in its prior years. This implies that younger firms have a better opportunity to manage accruals and thus are a greater source of uncertainty to the shareholders. On the other hand, older firms are more

constrained regarding accruals management due to the values for the accruals they have reported in prior years and thus it is easier for the shareholders to determine the true current state of the older firm.

As mentioned before, Gunny (2010) states that accruals management has to happen at the end of the fiscal quarter or year and it is uncertain for a firm’s managers which specific

accounting techniques will be allowed or forbidden by the auditor at this point in time. The author adds to this type of uncertainty that under the current accounting rules, R&D expenditures must be charged to expense as incurred because the investment leads to much uncertainty about the future benefits related to it. This uncertainty about the R&D investments could potentially lead to a company’s manager deciding to focus more on increasing short-term income rather than long-term value creation for the firm. Because he expects to obtain the potential benefits related to the R&D investment during a future period and thus decides to stop investing in R&D during the current period and the manager thereby potentially destroys value for the company.

Degeorge et al. (1999) find that the decision of a firm’s management on whether they will engage in earnings management relies heavily on the level of uncertainty of their second-period earnings. If the company’s management has invested in an opportunity for which they have borrowed money, and which has a bonus threshold, they are more likely to manipulate their earnings to secure their bonus for the first period because large borrowings are less likely to be sacrificed for the second year’s threshold. When the bonus the firm’s management receives for crossing the threshold decreases relative to the proceeds per unit of reported earnings, the amount of observed earnings management also decreases because it becomes relatively costlier.

Finally, Doupnik (2008) investigates data obtained from many countries around the world to observe the influence of national culture on earnings management. Their results suggest that even after they control their regression model for investor protection and other legal institutional factors, the amount of earnings management occurring is significantly influenced by the cultural

11

dimensions of individualism and, most importantly, uncertainty avoidance. Han et al. (2010) also find a negative relationship between the level of uncertainty avoidance in a country and the amount of earnings discretion engaged in by a firm’s management. So, this implies that a higher level of uncertainty in a firm, results in higher amount of earnings management occurring in that firm.

2.5 Implications and Hypothesis

When reviewing the literature above, we would expect a higher amount of earnings management occurring in a firm with a lower age. This because the lower age means that it is listed for a shorter time period on the stock exchange and because of this there would be less information available about this firm. This results in a higher level of uncertainty for the shareholders, in the way that is harder for them to determine the true current state of the firm they are investing in or are planning to do so in the future. Furthermore, older firms have more restraints towards

accruals management and thus less uncertainty for the shareholders. Finally, firms with greater endogenous accruals-generating potential have greater uncertainty about reported earnings. This leads to the following hypothesis for my research:

A firm with a lower age, as a proxy variable for the level of uncertainty, has a higher amount of earnings management occurring in its organization.

In other words: the longer a firm is listed on the stock exchange, the lower the level of

12

3. Methodology

This chapter discusses the methodology used to answer the research question. The Jones model and its modified version to obtain the required values for the discretionary accruals are explained as well as how its relationship with the variable firm age will be tested. In the next chapter I will discuss the data and descriptive statistics that I have analyzed to arrive at my results and conclusions.

3.1 Jones model

After reviewing possible suitable research designs to answer my main research question, I arrived at the conclusion that the use of discretionary accruals to test for earnings management is common practice in the existing literature and thus would be a good starting point. Kothari et. al (2005) provided the best guidance for the design of my research model because they recommend the renowned Jones (1991) model as the perfect starting point for proper research on earnings management.

The total accruals of the base Jones model will be defined as the change in non-cash current assets minus the change in current liabilities excluding the current portion of long-term debt minus depreciation and amortization following the research of Kothari et. al (2005). The residuals from these total accruals (TAt) are used as the discretionary accruals. The

residuals from the total accruals from the Jones model are estimated for each firm and year as follows:

TA

= a

/ASSETS

+ a

SALES

+ a

PPE

+ e

Where

ASSETS

and PPEt is net property, plant and equipment scaled by lagged assets.

Dechow et al. (1995) share Kothari’s opinion on the good fit of the Jones model for research on earnings management and also recommend modifying the base Jones model, while still using the same parameters as the regular model and apply those to a modified sales change variable where the change in accounts receivable is subtracted from the base sales change variable. The equation for obtaining the Modified Jones total accruals becomes:

13

TA

= a

/ASSETS

+ a

[SALES

-

AR

+ a

PPE

+ e

Continuing, Teoh et al. (1998a) further develop the Jones and Modified Jones model by adding lagged return on assets, ROA, to the regression. They explain that this helps to control the model for observed growth in receivables that is not due to earnings management by the firm but can be attributed to those firms experiencing rapid growth. This suggests that companies which experience these increases in the real values for their assets to receivables ratio, usually over estimate their discretionary accruals and underestimate their non-discretionary accruals.

Return on assets = net income / total assets Net income = total revenue – total expenses

When adjusting the regression equation for the (Modified) Jones model with the addition of lagged return on assets, our model becomes:

TA

= a

/ASSETS

+ a

SALES

+ a

PPE

+ a

ROA

+ e

The Modified Jones model with the addition of lagged return on assets to obtain the total accruals and its residuals becomes:

TA

= a

/ASSETS

+ a

[SALES

-

AR

+ a

PPE

+ a

ROA

+ e

Finally, I will estimate a discretionary current accruals (DCA) measure, which excludes the depreciation accruals from the discretionary accruals, that is increasingly being used in

accounting research. Teoh et al. (1998a) find evidence that high discretionary current accruals imply long-run earnings and stock return underperformance. The equation for the accruals of the DCA model, which also includes the modified sales change variable, becomes:

14

3.2 Measuring earnings management and uncertainty

To test the hypothesis stated at the end of the previous chapter, my regression model will become:

𝑌𝑒𝑎𝑟𝑛𝑖𝑛𝑔𝑠 𝑚𝑎𝑛𝑎𝑔𝑒𝑚𝑒𝑛𝑡 = 𝛽0 + 𝛽1∗ 𝐹𝑖𝑟𝑚 𝐴𝑔𝑒 + 𝛽2∗ 𝐹𝑖𝑛𝑎𝑛𝑐𝑖𝑎𝑙 𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑠 + 𝐹𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠𝑌𝑒𝑎𝑟+ ɛ

With this OLS regression model, I will try to find a direct significant relationship between firm age and earnings management, measured by discretionary accruals in different models and forms, occurring in the firm. The equations to arrive at the different measures for the accruals include a variety of financial numbers of the firms to control for possible endogeneity. Furthermore, I will also control for year fixed effects.

Firm age, our proxy variable for the level of uncertainty, is measured as the number of years that a firm is listed on the CRSP database. As explained before, the residuals of the total accruals (TAt) from the Jones model, Modified Jones model, Jones model including lagged

ROA, Modified Jones model including lagged ROA and discretionary current accruals are used as the discretionary accruals and these will be my variables for earnings management. It will be interesting to observe and compare the different outcomes for the relationship between firm age and the five different accruals models.

The time frame used to test this model is 2000 – 2017 which is a period of 18 years. After reviewing the existing literature on this topic and our hypothesis, a negative coefficient for the variable firm age in our regression model would be expected. This because a lower value for our firm age variable means that it is listed for a shorter time period on the stock exchange and because of this there would be less information available about this firm. This results in a higher level of uncertainty for the shareholders, in the way that is harder for them to determine the true current state of the firm they are investing in or are planning to do so in the future. On top of that, older firms have more restraints towards accruals management and thus less uncertainty for the shareholders. Thus, the longer a firm is listed on the stock exchange, the lower the level of uncertainty for the shareholders and the lower the amount of earnings management expected.

15

4. Data and Descriptive Statistics

4.1 Earnings management and firm age

The data used for this research is accessible through Wharton WRDS, for the period of 2000 - 2017. Firstly, firm age, our proxy variable for the level of uncertainty, is measured as the number of years that a firm is listed on the CRSP database.

Continuing, for our data on our earnings management variable, I will begin with all firm-year observations from the COMPUSTAT Industrial Annual and Research files from 2000 till 2017. As explained before, my variables for earnings management will consist of the residuals of the total accruals (TAt) from the Jones model, Modified Jones model, Jones model including

lagged ROA, Modified Jones model including lagged ROA and discretionary current accruals. Total accruals are constructed from the change in non-cash current assets minus the change in current liabilities excluding the current portion of long-term debt minus depreciation and amortization.

Further following Kothari et. al (2005), I will exclude firm-year observations that do not have sufficient data to compute total accruals or where the absolute value of total accruals scaled by total assets is greater than one. In addition, I will winsorize extreme observations for all discretionary accrual measures by setting the values in the bottom and top one percent of observations to the values of the 1st and 99th percentiles.

4.2 Descriptive Statistics

Variable Description

Age Firm age in years

TA Total accruals

DAJM Discretionary Accruals Jones Model

DAMJM Discretionary Accruals Modified Jones Model

JMROA Lagged ROA Discretionary Accruals Jones Model

MJMROA Lagged ROA Discretionary Accruals Modified Jones Model

DCA Discretionary Current Accruals

ROA Lagged Return on assets

16

Variable Obs Mean Std.Dev 25% 75%

Age 71.652 15 13 5 20 TA 71.652 -723 4859 -286 12 DAJM 71.564 -47 1605 -85 46 DAMJM 71.563 -47 1611 -195 160 JMROA 71.564 -2.3 3669 -85 49 MJMROA 71.563 2.0 3661 -194 161 DCA 71.588 -39 1622 -99 35 ROA 71.646 .11 .50 .02 .18

Table 2. Summary Statistics

Table 1 presents and explains the dependent and independent variables used for this research and in addition Table 2 contains the corresponding summary statistics. The amount of observations for the different variables fluctuates between 71.500 and 71.700 observations. We observe that the average time that a firm, on which we have collected data, is listed on the CRPS database, is 15 years.

Furthermore, the summary statistics for the dependent variables of the discretionary accruals for the Jones model, the modified version and the current accruals all seem pretty similar, with the spread of the numbers for the modified versions seem to be slightly larger. In addition, the standard deviations of the versions of the Jones model including the lagged ROA are much higher than those for the models which do not control for this variable.

Continuing, the average number for total accruals I have calculated has turned out to be negative. This means that the change in non-cash current assets is smaller than the sum of the change in current liabilities excluding the current portion of long-term debt plus depreciation and amortization.

Finally, the average 11% for the return on assets ratio seems like a reasonable number, especially when considering the percentages for the 25th and 75th percentile.

17

5. Results

OLS

DAJM DAMJM JMROA MJMROA DCA

Age -33.57 -33.05 -52.98 -52.31 -30.12

(1.338)*** (1.339)*** (3.066)*** (3.066)*** (1.353)***

Intercept 1,445 968.7 2,902 2,390 1,509

(66.97)*** (67.02)*** (153.5)*** (153.5)*** (67.72)***

Year Fixed Effects yes yes yes yes yes

Observations 71,564 71,563 71,564 71,563 71,588

Number of Years 18 18 18 18 18

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

Table 3: Regression output, dependent variable is earnings management

5.1 OLS regression results

Table 3 contains the ordinary least squares (OLS) regression results of the different models used. All different models used to obtain the discretionary accruals are controlled for year fixed

effects, the number of years observed is 18 for every column and the amount of observations included is almost the same for all models, namely in the range between 71,564 and 71,588.

In column 1, firm age, as a proxy for the level of shareholder uncertainty, is the

independent variable and is regressed on the discretionary accruals, which are the residuals from the Jones model and used as our first dependent variable representing earnings management. Before being able to obtain the residuals from the Jones model, the total accruals of the model are calculated by using change in total assets, change in current liabilities and depreciation and amortization. In an additional step before obtaining the residuals, the lagged total assets, change in sales scaled by lagged total assets and PPE scaled by lagged total assets are used and these variables are thus also being controlled for in the regression model.

The results show a negative regression coefficient which is highly significant at the 99 percent confidence level. The coefficient has a value of -33.57, which means that a one-year increase in firm age results in a 33.57 decrease in the value for the discretionary accruals as obtained by the standard Jones model. This corresponds to the hypothesis of this research and the

18

literature reviewed on this topic and could be considered an interesting finding. This because a decrease in discretionary accruals means a decrease in the amount of earnings management. Since this occurs after an increase in the variable for firm age, our first regression results immediately imply that older firms, as a proxy for less shareholder uncertainty, experience less earnings management.

Continuing, column 2 represents a similar regression model to column 1. However, instead of the regular Jones model, the Modified Jones model is being used to obtain the

discretionary accruals as the dependent variable representing earnings management. This means that we will still be using the same parameters as the regular Jones model, however, we apply those to a modified sales change variable where there is being controlled for the change in accounts receivable and this is subtracted from the base sales change variable.

The results show a negative regression coefficient which is highly significant at the 99 percent confidence level. The coefficient has a value of -33.05, which means that a one-year increase in firm age results in a 33.05 decrease in the value for the discretionary accruals as obtained by the Modified Jones model. Even though the number of the coefficient is slightly smaller for the modified version of the Jones model compared to the standard model, this outcome still contributes greatly to the confidence in the correctness of the hypothesis of this research. Because it implies again that younger firms seem to face more difficulties regarding earnings management than younger firms.

Moreover, column 3 represents a similar regression model to column 1. However, it includes lagged return on assets (ROA) in the equation used to obtain the total accruals of the Jones model, which are again being used to provide the discretionary accruals. These are adjusted for lagged ROA, however, they will serve as the dependent variable for earnings management with the default sales change variable.

The regression results show again a negative coefficient which is highly significant at the 99 percent confidence level. The coefficient has a value of -52.98, which means that a one-year increase in firm age results in a 52.98 decrease in the value for the discretionary accruals as obtained by the standard Jones model. This number of the regression coefficient is larger than the results of the standard and modified version of the Jones model. Thus, when we control for lagged ROA, this helps to control the model for observed growth in receivables that is not due to earnings management by the firm but can be attributed to those firms experiencing rapid growth.

19

Meanwhile, controlling for ROA enhances the relationship between firm age and earnings

management in the regression model. This outcome further strengthens the predicted relationship of the hypothesis of this research. It is again implied that the older firms grow, they seem to face less difficulties regarding earnings management than younger firms. Especially when there is being controlled for firms experiencing rapid growth.

Furthermore, column 4 represents a similar regression model to column 3. However, instead of to the regular Jones model, the lagged ROA variable is this time added to the Modified Jones model to obtain the discretionary accruals as the dependent variable used for earnings management. This means that we again will be using the same parameters as the regular Jones model with lagged ROA and apply those to a modified sales change variable where there is being controlled for the change in accounts receivable and this is subtracted from the base sales change variable.

Once again, the results show a negative regression coefficient which is highly significant at the 99 percent confidence level. The coefficient has a value of -52.31, which means that a one-year increase in firm age results in a 52.31decrease in the value for the discretionary accruals as obtained by the Modified Jones model. Even though the number of the coefficient is slightly smaller for the modified version of the Jones model compared to the standard model, this result corresponds to the existing literature and the hypothesis of this research. Because once again, it implies that younger firms seem to face more difficulties regarding earnings management than younger firms.

Finally, column 5 estimates the discretionary current accruals (DCA), which excludes the depreciation accruals and the lagged ROA variable from the discretionary accruals equation. However, the modified sales change variable, where there is being controlled for the change in accounts receivable, is included. Teoh et al. (1998a) found evidence that high discretionary current accruals imply long-run earnings and stock return underperformance.

The results for the DCA regression provide a negative coefficient which is highly

significant at the 99 percent confidence level. The coefficient has a value of -30.12, which means that a one-year increase in firm age results in a 30.12 decrease in the value for the discretionary current accruals. Even though this final DCA coefficient is the lowest of the five tested

20

earnings management remains unchanged after excluding one of the accruals, depreciation, on which management decisions have the most influence, implies that our hypothesis is correct.

5.2 Robustness checks and R-squared

In addition to the regression results, I have conducted robustness tests for the OLS regression models on multicollinearity, with help of the Vif test. Vif is the abbreviation for variance

inflation factor and when the mean Vif value for an independent variable is below 10, it does not suffer from multicollinearity. Firm age has a Vif value of 2.43, so it could be considered a suitable independent variable to test for earnings management.

Finally, even though the explanatory power of the regression models is not considerably high, around two percent of the variance for the discretionary accruals is explained by the model, it is still important to realize that although it might not be one of the major reasons for earnings management, firm age still is one of many factors that needs to be taken into account when researching this topic.

21

6. Conclusion and Discussion

This thesis has studied the effect of firm age on earnings management. The main research question is whether there exists a clear relationship between the age of a firm, as a proxy variable for the level of shareholder uncertainty, and the amount of earnings management occurring in the firm.

After reviewing the literature on this subject, we arrived at the hypothesis that a firm with a lower age, as a proxy variable for the level of uncertainty, has a higher amount of earnings management occurring in its organization. In other words: the longer a firm is listed on the stock exchange, the lower the level of uncertainty for the shareholders and the lower the amount of earnings management occurring.

There has been an extensive discussion of literature that supported this theory and some conclusions were drawn regarding firm age, uncertainty and earnings management. First, Barry and Brown (1985) argued that lower age means that it is listed for a shorter time period on the stock exchange and because of this there would be less information available about this firm. This results in a higher level of uncertainty for the shareholders, in the way that is harder for them to determine the true current state of the firm they are investing in or are planning to do so in the future. Furthermore, older firms have more restraints towards accruals management and thus less uncertainty for the shareholders (Barton and Simko, 2002). Finally, Francis, Maydew, and Sparks (1996) stated that firms with greater endogenous accruals-generating potential have greater uncertainty about reported earnings.

The hypothesis tested in this thesis has been confirmed by the empirical results of the research conducted. As explained before, the five different variables for earnings management consist of the residuals of the total accruals (TAt) from the Jones model, Modified Jones model,

Jones model including lagged ROA, Modified Jones model including lagged ROA and

discretionary current accruals. I have constructed an OLS regression model to test whether there exists a significant negative relationship between an increase in firm age and the amount of earnings management occurring in the same firm, in the five different forms mentioned above.

All five different OLS regression models result in a highly significant negative regression coefficient for the variable firm age. Even though the number of the coefficient is slightly

smaller for the modified versions of the Jones model compared to the standard models, their outcomes still contribute greatly to the confidence in the correctness of the hypothesis of this

22

research. The value of the negative coefficient becomes even larger when we control for lagged ROA, this helps to control the model for observed growth in receivables that is not due to

earnings management by the firm but can be attributed to those firms experiencing rapid growth. In addition, even though the final DCA coefficient is the lowest of the five tested

regression models, the fact that the significant negative relationship between firm age and earnings management remains unchanged after excluding one of the accruals, depreciation, on which management decisions have the most influence, provides further proof that our hypothesis is correct.

Further implications of the results of the different regression models on earnings management suggest that investors, especially relatively inexperienced ones, should keep in mind that they face a higher risk of investing in a company which is engaging in earnings management when they decide to invest in firms that are relatively shortly listed on the stock exchange. In addition, corporate management should also be aware of potential earnings management occurring at their target firm when considering acquiring a relatively newly listed company. Finally, my results imply that regulators should slightly focus the development of their future regulations at monitoring earnings management at relatively newly listed firms.

The obtained regression results provide interesting suggestions for further research in this field. One could, for example, conduct research on the question if the same negative relationship between firm age and earnings management would be observed when real activities manipulation (RM) would be used instead of accruals management. The difference between the two types of earnings management originates from the fact that a firm’s management decides on the operating activities while decisions on the accounting techniques depend on the judgement of the auditor. This could provide interesting new insights in understanding the concept of earnings

management.

Finally, this research is limited in the sense that it uses firm age as the only proxy for shareholder uncertainty. Other papers have used different proxies such as disagreement amongst professional forecasters or the frequency of newspaper articles about economic uncertainty (Bloom, 2014). Furthermore, including some additional control variables from the literature to improve the value for R-squared and performing some robustness tests on the assumptions of the OLS regression model, would have further enhanced the precision and quality of this thesis.

23

However, even though the explanatory power of the regression models is not

considerably high, around two percent of the variance for the discretionary accruals is explained by the model, it is still important to realize that although it might not be one of the major reasons for earnings management, firm age still is one of many factors that needs to be taken into account when researching this topic.

24

7. Reference List

Ali, A. and Zhang, W., 2015. CEO tenure and earnings management. Journal of Accounting and

Economics, 59(1), pp.60-79.

Ali, U. & Hirshleifer, D. 2015. “Opportunism as a Firm and Managerial Trait: Predicting Insider Trading Profits and Misconduct.” Working paper. Available at:

http://papers.ssrn.com/sol3/Papers.cfm?abstract_id=2635257

Baron, R. A. 2006. Entrepreneurship: A process perspective. The psychology of

entrepreneurship.

Baron, R. A., & Markman, G. D. 2005. Toward a process view of entrepreneurship: The

changing relevance of individual-level variables across phases of new firm development. Current

topics in management, vol. 9: 45-64.

Barry, C.B., & Brown, S.J. 1985. Differential information and security market equilibrium,

Journal of Financial and Quantitative Analysis 20, 407–422.

Barton, J. and Simko, P.J., 2002. The balance sheet as an earnings management constraint. The

accounting review, 77(s-1), pp.1-27.

Becker, C.L., DeFond, M.L., Jiambalvo, J. and Subramanyam, K.R., 1998. The effect of audit quality on earnings management. Contemporary accounting research, 15(1), pp.1-24.

Bergstresser, D. and Philippon, T., 2006. CEO incentives and earnings management. Journal of

financial economics, 80(3), pp.511-529.

Bloom, N., 2014. Fluctuations in uncertainty. Journal of Economic Perspectives, 28(2), pp.153-76.

Brown, L.D. and Caylor, M.L., 2006. Corporate governance and firm valuation. Journal of

accounting and public policy, 25(4), pp.409-434.

Choi, Y. R., & Shepherd, D. A. 2004. Entrepreneurs’ decisions to exploit opportunities. Journal

of Management, 30: 377-395.

Cornett, M.M., McNutt, J.J. and Tehranian, H., 2009. Corporate governance and earnings management at large US bank holding companies. Journal of Corporate finance, 15(4), pp.412-430.

Dai et al. 2014. “Internal Corporate Governance and the Profitability of Insider Trading.”

Available at:

www.business.kaist.ac.kr%2Ffaculty%2Finmoo%2FArticles%2FDaiFuKangLee2015Jan.pdf&us g=AFQjCNEok3inCkQ4zc93jVF_3P2uHV1gvg

25

Dechow, P., Kothari, S., Watts, R., 1998, The relation between earnings and cash flows, Journal

of Accounting & Economics 25, pp.133-168.

Dechow, P., Sloan, R., Sweeney, A., 1995, Detecting earnings management, The Accounting

Review 70, pp.193-225.

Degeorge, F., Patel, J. and Zeckhauser, R., 1999. Earnings management to exceed thresholds. The Journal of Business, 72(1), pp.1-33.

Doupnik, T.S., 2008. Influence of culture on earnings management: A note. Abacus, 44(3), pp.317-340.

Fort, T.C., Haltiwanger, J., Jarmin, R.S. and Miranda, J., 2013. How firms respond to business cycles: The role of firm age and firm size. IMF Economic Review, 61(3), pp.520-559.

Francis, J., Maydew, E.L. and Sparks, H.C., 1996. Earnings management opportunities, auditor quality, and external monitoring. Working paper, University of Missouri at Columbia.

Gunny, K.A., 2010. The relation between earnings management using real activities manipulation and future performance: Evidence from meeting earnings

benchmarks. Contemporary Accounting Research, 27(3), pp.855-888.

Han, S., Kang, T., Salter, S. and Yoo, Y.K., 2010. A cross-country study on the effects of national culture on earnings management. Journal of International Business Studies, 41(1), pp.123-141.

Jones, J., 1991, Earnings management during import relief investigations, Journal of Accounting

Research 29, pp.193-228.

Keh, H. T., Foo, M. D., Lim, B. C. 2002. Opportunity evaluation under risky conditions: The cognitive processes of entrepreneurs. Entrepreneurship: Theory & Practice, 27(2): 125-148. Zhang, X., 2006. Information uncertainty and stock returns. The Journal of Finance, 61(1), pp.105-137.

Kieschnick, R. and Moussawi, R., 2018. Firm age, corporate governance, and capital structure choices. Journal of Corporate Finance, 48, pp.597-614.

Kothari, S.P., Leone, A.J. and Wasley, C.E., 2005. Performance matched discretionary accrual measures. Journal of accounting and economics, 39(1), pp.163-197.

Ling, Y., Zhao, H. and Baron, R.A., 2007. Influence of founder—CEOs' personal values on firm performance: Moderating effects of firm age and size. Journal of Management, 33(5), pp.673-696.

26

Teoh, S., Welch, I., Wong, T., 1998a, Earnings management and the long-run underperformance of seasoned equity offerings, Journal of Financial Economics 50, pp.63-100.

Teoh, S., Welch, I., Wong, T., 1998b, Earnings management and the long-run underperformance of initial public offerings, Journal of Finance 53, pp.1935-1974.

Thornhill, S. and Amit, R., 2003. Learning about failure: Bankruptcy, firm age, and the resource-based view. Organization science, 14(5), pp.497-509.