The Impact of the credit crisis on accounting conservatism

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Chapter 1: Introduction

In this literature review the topic conservatism in accounting will be discussed. Conservative accounting exist already for many years. Evidence is found that trading partner in the early of the 15th century used conservative accounting (Penndorf, 1930). If it exits for so long then their has to be some reason why conservative accounting is still practiced. Watts (2003) provide those reasons of why conservative accounting is still applied and explained why it is an important factor of accounting.

But what is conservative accounting? There is no general accepted definition but, several researchers have come with their definition of conservative accounting. Basu (1997) for example interpreted conservative accounting as followed: ''accountants' tendency to require a higher degree of verification for recognizing good news than bad news in financial statements''. In the literature review several other definition of accounting conservatism will be provided.

Conservative accounting is an important principle of accounting therefore there exits many research of this topic. The variety of this topic is very large. For instance research that studied the explanations of conservatism, or research that studied the link between conservatism and earnings management. Therefore I have narrowed the conservatism in this literature review to a few specific topics. One of those topics is conservatism and crisis. Those studies have looked at conservatism during a crisis. The results were that companies report less conservative during a crisis. But those results are bases on the Asian crisis that occurred in 1997.

1.2 Research question

A lot have changed in the past years from 1997 until 2011, for instance the credit crisis that has hit the economy worldwide. But what was the impact for the accountancy?

Therefore this literature review has the following research question:

''What was the impact of the credit crisis on accounting conservatism?''

I expect that companies were less conservative during the crisis. This is consistent with the prior research. Although those studies were based on the Asian crisis and conducted in a different period. I expect that companies reacted the same in the credit crisis as companies did in the Asian crisis.

This literature review is relevant for practitioners, financial analyst, accounting students and other who are interested on the impact of the credit crisis on accounting conservatism. As stated in the previous paragraph there has been some research to conservative accounting in relation with a crisis. But those studies where based on the Asian crisis. In this literature conservative accounting in relation with the credit crisis will be studied. Therefore this literature review provides some new insight on the relation between conservative accounting and crisis.

1.3 Structure

The structure of the literature review will be as followed. In the chapter two a theoretical background of conservative accounting will be provided. The definitions and explanations of conservatism will be discussed. Chapter three is about the method that measure conservatism. The method and limitations will be discussed in this chapter. The fourth chapter prior research on accounting conservatism will be presented. Finally chapter five consists of a short summary of the literature review.

Chapter 2: conservative accounting

In this chapter I will discuss the literature about conservatism. At the beginning of this chapter the accounting principle conservatism will be discussed. The reasons of existing of conservatism will be explained.

2.1 definitions

The difficulty of the research of conservatism is that there is no specific definition for what conservatism is, although it is a very important and well know characteristic of the Generally Accepted Accounting Principle (Huang et al., 2008).

The FASB tries to give an explanation of what conservatism is. Statement of Financial Accounting Concepts (SFAC) No.2 (FASB, 1980) paragraph 95 states: "if two estimates of amounts to be received or paid in the future are about equally likely, conservatism dictates using the less optimistic estimate".

This means that the company should choose the lowest estimate. Therefore the company is being cautiously in measuring the assets and liabilities, or amounts to be received or paid in the future. As a result the risk that the company has overstated his assets will be much lower. Even if something happens in the future that could have an impact on the company, the risk of overstating is less.

Because of the fact that there is no exact definition of conservatism. Researcher has interpreted accounting conservatism in different ways.

Accountants have interpreted traditionally conservatism rule as: "anticipate no profit, but anticipate all losses". (Bliss, 1924). This is conservatism is its extreme form, because you don't anticipate/ recognize any profits and that you recognize all losses even if they aren't yet verifiable. The profit will only be recognized when they are completely verifiable and there is a legal claim on the profit.

(Basu,1997) has interpreted this conservatism rule in accounting as: ''the accounting tendency to require a higher degree of verification to recognize good news as gains that to recognize bad news and losses''. For this interpretation he used several accounting conservatism examples for instance: (Accounting Principle Board (APB) opinion 6, APB,1965) ''writing down of physical assets to reflect obsolescence of impairments, but not revaluing them upwards. For this and other examples he concluded that conservatism results in a greater probability of more timely accounting recognition of bad news and good news (Basu, 1997).

(Watts,2003) defines conservatism as followed. ''Conservatism is the asymmetrical verification requirements for gains and losses''. This means that there exist a difference in requirement for recognizing profits and losses. This definition allows different degrees of conservatism. It depends on the degree of verification of profit and losses. The greater the difference of verification the greater the degree of conservatism.

The definitions of (Basu,1997) and (Watts, 2003) are the most used definitions of conservatism in the recent papers and studies. Therefore those definitions of conservatism will be further used for this paper.

In the accounting literature there is a distinction made between conditional conservatism and unconditional conservatism. Unconditional conservatism also called ex ante or news dependent and conditional conservatism also referred as ex post or news independent (Beaver & Ryan, 2005). What is meant under unconditional conservatism is the following: ''the predetermined understatement of the book value of net assets, as occurs with the immediate expensing of the cost of most intangibles''(Ryan, 2006). This type of conservatism is also referred as balance sheet conservatism. With conditional conservatism is meant the difference of timely recognition of bad news and goods news also called the asymmetrical timeliness (Ryan,2006). Conditional conservatism is also referred earnings conservatism.

2.2 explanations of conservatism

As mentioned in the previous part of this paper. There is not a general definition of conservatism despite the fact that it is an important principle. Conservatism has influenced accounting for many centuries. In fact it has influenced the accounting practice for at least 500 years (Basu, 1997). There are historical records found of trading partnership in the early 15th century who practice conservatism (Penndorf, 1930). With this records it is proven that conservatism in Europe exist for many years. The fact that conservatism exist for so many years must be a positive thing, because if it had a negative impact on accounting it would have been deleted many years ago, as in line with accounting Darwinism that only the fittest will survive.

In the literature there exist several explanations for the existence conservatism in accounting. Those explanations are: contracting, litigation, taxation and accounting regulation (Basu,1997: Watts, 2003a&b). In the next section a will explain and elaborate each of the four explanations stated above.

2.2.1 Contracting

The first explanation is contracting. ''Conservatism arises because it is part of the efficient technology employed in the organization of the firm and its contracts with various parties'' (Watts, 2003a). With conservatism in accounting the moral hazard problem is limited and agency cost is reduced. Agency cost is based on the agency theory (Jensen and Meckling,1976) they define agency relationship as: ''a contract under which one or more (principles) engage another person (the agent) to perform some service on their behalf which involves delegating some decision-making authority to the agent''.

In other words this means that the agent (manager) has to work for the interest of the principle (shareholders), but the agent has his own interest. Therefore the interests of the agent en the principle are different and that can cause costs. Those costs are called agency cost. The moral hazard is that people act differently if they do not have to pay for the cost or only a small part. As a result they don't bear the cost/risk. For example a manager will invest in a highly risky project, because the money is from the shareholders and if the investment turns out to be good he gets the credits but when it turns out to be negative the shareholders have lost their money not the manager.

The problem with moral hazard and agency theory is that it comes from the information asymmetry between the agent and the principal (Jensen and Meckling; 1976). Managers have more information about the firm than shareholders, because the managers are working at the firm and therefore know what is happening. The shareholders rely on the information provided by the managers. This difference of information level is called the information asymmetry. (Lafond and Watts, 2008) found evidence that this is a reason that conservatism is applied. In chapter four I will elaborate the results some more.

Conservatism reduces the optimistically views of the managers, because they have to recognize the full losses and only partly recognize gains. As a result the managers can pay fewer dividends to the shareholders and receive less bonus rewards. The latter is also called the compensation contracts (Watts, 2003a).

Conservatism also reduces the risk for debt holders that, the management over spend the money and that the assets are overstated. In other words the debt contracting (Watts 2003a). Again there is evidence that debt contracting is linked with conservatism and that it results in benefits (Beatty et al., 2008; Zhang, 2008). A final subgroup of contracting is corporate governance. Timely signals of bad project will be a result of conservatism. This protects the shareholders because, they can investigate the bad project and take appropriate actions (Watts, 2003a).

2.2.2 Litigation cost

This explanation for conservatism looks at the risk of litigations. The risk of litigation for overstating the assets and results is much higher than for understating the assets and results (Watts, 2003a). The reason is that when the company overstated his assets and results and the market finds out that the company overstates their assets and results there will be a drop in the market value of the company. This drop of market value causes financial harm to shareholder. But when it's the other way around (understating) the company has actually a bigger market value. Therefore the shareholders are not immediately harmed. As a result of overstating results in a higher litigation risk. (Chung et al.,2008) found evidence that there is a relation between litigation and conservatism.

2.2.3 Taxation

Watts (2003a) links conservatism with taxation for the following reason. Companies will lower their reported income to lower their taxable income. Therefore they defer the payable taxes to the future. As a result they have to pay a lower tax in their current fiscal year, because they have lowered their income by full recognizing their cost and only partly recognizing their gains.

2.2.4 Accounting regulation

In this final explanation for accounting conservatism Watts (2003a) concludes that conservatism is also a result of the rules and regulation. As mentioned earlier in paragraph 2.1 the example of the FASB that the company has to choose the lowest of the two valuation methods. This results in an understatement of the assets (conservatism).

FASB is not the only regulator with rules that results in conservatism. For instance the IFRS (2004) sees conservatism as the prudence principle. With this principle you have to be cautious with estimating the assets, so that assets aren't overstated and liabilities aren't understated. As a result when you apply this principle from the IFRS you have to be conservative.

The reason that regulators prefers conservatism is that the political consequences will be lower (Watts, 2003a). When the rules leads to overstating the public will held the regulators responsible for their losses. Therefore the regulators prefer conservatism which results in fewer critics.

The contracting and litigation explanation are the most important reasons for conservatism, taxation and accounting regulation have weaker evidence (Watts, 2003a).

Chapter 3: measuring conservatism

In this chapter I will explain and elaborate some models that measures accounting conservatism. Those models uses different approaches but they are grouped into 2 main topics. Those are information perspective and the measurement perspective. In the next subparagraph a will try to explain those perspectives a bit further. After those perspectives I will briefly discus the market efficiency hypotheses followed by categorization that Watts (2003b) made of models that measures accounting conservatism. Finally the models will be discussed.

3.1 market-based accounting

Conservatism in accounting is a topic in market-based accounting research. In this research approach the focus lays on the relation between share prices and accounting numbers.

The market-based approach uses the market efficiency theory from (Fama et al., 1969). He provided three hypotheses of market efficiency. Those are: strong form of efficiency, semi strong efficiency and the weak form of efficiency. The strong form indicates that all available information including private and public information is correctly interpreted and quickly spread. The semi strong indicates that not all private information is correctly interpreted and available. The weak form indicates that only prior information about the share prices and trading volumes are reflected in the current share prices.

The market-based approach has two perspectives those where already mentioned in the introduction but, I will give an explanation of those perspectives. With the information perspective there is expected that when new information becomes available there will be a share price reaction. The new information must be useful otherwise it isn't new information that leads to reactions of investors. (Scott,2008). The other perspective is the measurement perspective. The measurement perspective according to (Scott,2008,p. 177). ''to decision usefulness is an approach to financial reporting under which accountants undertake a responsibility to incorporate current values into the financial statements proper, providing that this can be done with reasonable reliability, thereby recognizing an increased obligation to assist investors to predict firm performance and value''.

Watts doesn't look at the differences in information perspective and measurement perspectives. Instead he categorized the methods used to measure conservatism in the following groups.

Net assets measures, Earnings and accrual measures, Earnings/ stock returns relation measures (Watts,2003b).

Under net assets measure conservatism leads to understatement of net assets. They are valued below the market value. (Watts, 2003). One of those methods to measure conservatism was developed by the theory of Feltham and Ohlson (1995) the market to book ratio (Givoly and Hayn,2000)

Earnings and accruals measures. Under this method conservatism is seen as that profit aren't fully recognized but the losses are fully recognized. This results in an asymmetry in accruals because of the different treatment of profits and losses. Conservatism can be measured as the negative periodic net accruals and negative cumulative accruals accumulated over periods (Watts, 2003). Several papers have used this method to measure conservatism such as (Basu,1997; Givoly and Hayn, 2000; Ball and Shivakumar,2005)

Earnings/stock return relation measure. In this methods ''the stock market prices reflect the value changes at the time those changes occur those changes imply losses or gains in assets value- stock returns tend to be timely''(Watts,2003, p.290). There have been many variations on this method but the method that had been mostly applied in paper is the method of Basu (1997).

In the following subparagraph I will discuss some of the methods that have been mentioned more and provide some others methods.

3.2 measurement method

In this paragraph the methods that are used to measure conservatism in accounting will be discussed in every subparagraph a different method is then further elaborated. Finally some limitations of the methods will be provided.

3.2.1 Basu Model

As already mentioned in paragraph 2.1 Basu (1997) interpret accounting conservatism as the result that bad earnings are more quickly reflected then good news. Therefore there is a systematic difference in the timeliness and persistence in earnings. According to (Basu,1997) stock prices reflect information received from sources other than current earnings. This means that all publicly available information is reflected in the share prices. And due to conservatism there is an asymmetrical recognition of goods news and bad news. Therefore Basu predict that the earnings are more sensitive for bad news than for good news. To measure the conservatism he uses stock prices as a proxy for good and bad news (Basu,1997). The stock returns can be both positive and negative. When it is negative it is an indication for conservatism, because losses are recognized earlier then profits. Therefore you will get a negative stock return with losses/bad news.

Basu(1997) uses the following formula: Xit / Pit 1 = αo + α1DRit + β0Rit + β1Rit*DRit

Xit / Pit 1 means: the earnings per share for firm i in the fiscal year t/ price per share at the beginning of the fiscal year. The D is the dummy variable in the regression. The dummy variable is 1 or 0. This depends on the R. If R is positive then is will be 0 but when R is negative dummy variable is 1. R stands for the return. The return is the dividend and the difference in share price. Rit stand for the firm i in the period starting 9 months before the end of the fiscal year t and ending 3 months after the fiscal year end. Basu have made this distinction of 9 and 3 months because the annual report is expected between those two periods.

Basu expect that earnings are more sensitive for bad news then for goods news. β1 measures this difference in sensitivity. Therefore the more conservative a firm is the bigger β1 will be. To further indicate the relative sensitivity he uses the following ratio (β1+ β0)/ β0.

In his paper Basu made several predictions and models. I have only discussed the first prediction and model because, the other predictions and models are related to the first one. And in other papers like (Givoly and Hayn,2000; khan and Watts,2009) the first regression model is only used or criticized. And the reverse earnings-returns regression from Basu (1997) is the most frequently used measure for timely loss recognition ( Dechow et al.,2010). limitation of the Basu model

Although the method of Basu (1997) with some variations on his methodology is used in many other studies(Watts, 2003b), there exist some critic on his method. In this section a will provide some criticism on the Basu method.

The first limitation of the method is that it depends on the stock price movements for bad news and good news (Givoly and Hayn,2000). The limitation of this is the fact that bad news is only reflected by negative stock returns. In other words Basu assumes that every negative stock returns is bad news and therefore a signal for conservatism.

But the negative stock returns could also be caused by different things such as a crisis, or another trend. Ball and Shivakumar (2005) provided some more limitations. The model can only identify the existence of gains and losses not whether if they are timely or not. And second is that the model can't distinguish between the profits and losses components in to earnings from random errors.

(Huang et al.,2008) says that the Basu model implicated that the firm specific characteristics (size, leverage etc.)are homogenous, but those characteristics are heterogeneous. And finally (Givoly et al.,2007) and (Dietrich et al., 2007) provides some more criticism. Givoly et al. (2007) found other factors like time periods, countries and reporting regimes influence the measurement method of Basu. (Dietrich et al.,2007) conclude that the method of Basu results in biases and therefore those biases measures conservatism.

3.2.2 Givoly and Hayn model

Givoly and Hayn (2000) use four methods to measure conservatism. one of those methods is the method from Basu discussed in the previous subparagraph. The other three methods are level of accumulated negative non-operating accruals, measures based on the time-series properties of earnings and cash flows and market-to-book ratio. I will elaborate each method in the next subparagraphs. Non-operating accrual method

Accrual accounting is the opposite of cash accounting. With accrual accounting you match the cost and gains with actual income and expenses. Cash flows are unaffected by accruals (Givoly and Hayn, 2000). Therefore the accruals could be seen as the difference of what the company actually earns (income) and the cash flows. As a result of this you can use the accruals as a measure for conservatism which implicates that more negative accruals leads to more conservatism and the other way around. Givoly and Hayn (2000) use the following formula: non-operating accruals= total accruals (before depreciation) - operating accruals

Total accruals (before depreciation)are: (net income+ depreciation)- cash flow from operation. Givoly and Hayn (2000) uses total accruals before depreciation because, the accruals left are those of bad debt provision and loss provision.

Operating accruals are: Δ account receivable + Δ inventories + Δ prepaid expenses - Δ account payable - Δ taxes payable.

One limitation of this measure of conservatism is, that this measure is consistent with 'big bath'. And big bath is a result of earnings management and not of accounting conservatism (Zhang, 2008). Big bath theory is based on (Healy,1985), meaning that new managers will take extra losses and blame there predecessor. The result is that the managers have a bigger change of reaching their target profit the following year because of the extra right off/ losses they have already recognized. measures based on the time-series properties of earnings and cash flows

This method consist of two measurement methods. I will discuss those two methods separately. The first method is the skewness the second is variability.

Conservatism leads to full recognition of losses/cost and partially recognition with a delay of good news. Because of this asymmetry in recognizing losses and gains there exist a negative skewed distribution of earnings (Givoly and Hayn, 2000). This means that earnings don't have a normal distribution. With a normal distribution the skewness is 0. With a negative skewness the left tail of the distribution is longer. And with a positive skewness it is the other way around. A negative skewness will indicate conservatism.

Givoly and Hayn (2000) uses the following formula to measure conservatism: [E(x-µ)3]/σ3

The x stands for return on assets (ROA) or cash flow from operation (CFO)/assets. The µ is estimated by the mean of the x distribution. The σ is estimated as the standard deviation of the distribution of x.

One limitation of this measure of conservatism is, that this measure is again consistent with 'big bath' (Zhang, 2008).

The second method is the variability method. Conservatism is measurement by the standard deviation of net income to total assets (ROA). A high standard deviation is linked to conservatism (Givoly and Hayn,2000). Therefore a higher standard deviation leads to a higher accounting conservatism. The market-to-book ratio

The final measurement method of accounting conservatism provided by Givoly and Hayn (2000) is the market-to-book ratio. They use the theoretical framework of Feltham and Ohlson (1995). With this method market-to-book ratio is used as a proxy for accounting conservatism. Through accounting conservatism assets are undervalued because losses are fully incorporated and gains partially or delayed. This results in an undervaluation of the book value against the market value. This undervaluation is accounting conservatism. When the market-to-book ratio becomes higher there is expected that the conservatism will also be higher.

One limitation of this method is that the market-to-book ratio could reflect market growth expectation instead of conservatism (Givoly and Hayn, 2000). The increase of the ratio could indicate that the company simply has grown instead of increasing conservatism.

3.2.3 Khan and Watts model

The final measurement method I will discuss is the measurement of Khan and Watts (2009). There method is based on the regression method of Basu (1997), but they have modified the method so that they can measure firm year level conservatism. As one of the limitations of the Basu model was that it doesn't look at firm specific characteristics (Huang et al.,2008) .

Khan and Watts (2009) uses the following measurement of conservatism. they use the standard regression formula of Basu (1997).

Xi = β1 + β2DRi + β3Ri + β4Ri*DiRi + ei . e indicates the residual error.

G_score= β3 = µ1 + µ1Sizei + µ1 M/Bi + µ1 Levi

C_score= β4 = λ1 + λ2Sizei + λ3 M/Bi + λ4 Levi

The C_score measures the firm-year conservatism. the G-score measures the timeliness of good news. The total bad news timeliness is measured by the sum of the G_score and the C_score. The empirical estimators µ and λ are constant across the firms. However they vary over time because they are estimated from the annual cross-sectional regressions. i= between 1-4. Size stand for the natural log of market value of equity. Lev stands for the leverage. Leverage is the long term debt and short term debt deflated by the market value of equity. The M/B is the market-to-book ratio (Khan and Watts,2009). They have used those firm specific characteristics because, those characteristics are a proxy for accounting conservatism. And more importantly those characteristics vary with conservatism ( Khan and Watts, 2009). Therefore they are important firm specific characteristics that have an influence on the level of accounting conservatism in a firm. There exist many more firm specific characteristics but, Khan and Watts have limited those variables for the reason of parsimony. Meaning that with fewer variables they can use a bigger and complete sample size.

The C_score and the G_score are substituted into the regression formula. The final formula is then:

Xi = β1 + β2DRi +Ri (µ1 + µ1Sizei + µ1 M/Bi + µ1 Levi) + DiRi (λ1 + λ2Sizei + λ3 M/Bi + λ4 Levi)+ (δ1 Sizei +δ2 M/Bi +δ3 Levi +δ4 DiSizei+δ5Di M/Bi + δ6 DiLevi)+ ei

In the final formula there are some extra terms included, because this formula includes interaction terms between returns and firm characteristics. Therefore the firm characteristics have to be separately controlled (Khan and Watts, 2009).

Chapter 4: overview of prior research on conservatism

4.1 Introduction

As already mentioned in the paper conservatism is an important principle for accounting, even though a clear definition of conservatism is missing. The fact that conservatism is important and that the definitions are provided by researchers, has resulted in allot of research in accounting conservatism. The research topics in accounting conservatism are very broad. For example some research is done to provide some explanation for conservatism. Others look at conservatism in relation with earnings management or the difference of conservatism between different countries.

In this chapter I will discuss some papers that have done some research in the context of accounting conservatism. Some papers are already mentioned before, but in this section the papers will be more elaborated. And the results of those studies will be further discussed. The papers will be categorized in specific research subtopic of accounting conservatism. The categorizes are: the progress of conservatism, criticism on method that measures conservatism, explanations of conservatism and finally conservatism and crisis. Every subparagraph will discuss a category in more detail. In the appendix there will be a table that provides an overview of the literature.

4.2 the progress of conservatism

The first paper in this category is the paper from Basu (1997). He used the reverse regression method to measure conservatism. He found that earnings where two to six times as sensitive for negative returns then for positive returns, therefore he concluded that accounting conservatism exists. The results where robust for several other methods. Basu also concluded that positive earnings are more persistent then negative earnings. This difference implies that positive earnings have a higher earnings response coefficient meaning that the unexpected abnormal earnings are bigger for positive earnings. Finally he concluded that accounting conservatism, in the United Stated of America (USA), have increased in time. For the last three decades( (the period 1963-1990) conservatism has increased, a possible reason could be the increased legal auditor's legal liability exposure (Basu, 1997).

Givoly and Hayn (2000) had several methods to measure conservatism as stated in chapter three. They used several methods because, of the fact that an excepted definition of conservatism is missing. Givoly and Hayn found a trend that over time conservatism in accounting have increased. That is consistent with the finding of Basu (1997). However Givoly and Hayn used several method of conservatism and they investigated a longer period 49 years (1950-1998) for firms in the USA. Givoly and Hayn think that the increasing conservatism is a result of the accountant tendency to omit the intangibles from the accounting books of the firm. And intangibles assets has become more prominent for firms therefore the increasing conservatism could be explained by this factor.

Both papers are consistent with each other finding that accounting conservatism has grown over time for firm in the United States of America. But Basu (1997) thinks that the increase could be caused by the litigation risk and Givoly and Hayn (2000) think that not recognizing intangible assets could be the reason of increasing conservatism in accounting.

4.3 criticism on method that measures conservatism

Dietrich et al., (2007) examined the relationship between earnings on stock returns and accounting conservatism. Many papers such as Basu (1997) found that bad news is more timelier incorporated then good news, therefore asymmetrical timeliness indicates accounting conservatism. But Dietrich et al. (2007) found empirical evidence that the method timely loss recognition is biased. And those biases are interpreted as evidence that accounting conservatism exists. Those biases are inherent related on the TLR method and therefore irresolvable (Dietrich et al (2007). They use a economic model that indicates the relation among the economic income, accounting earnings, non-earnings information and stock returns. With this model accounting conservatism can be includes or excluded. To prove that accounting conservatism is biased they used two test sample of the year 1963-1990 (the same as Basu 1997) and 1991-2001. Both test samples were tested with actual data and simulated data, as a result the could prove the biases of timely loss recognition. Therefore the results of this regression cannot be interpreted as accounting conservatism. Finally Dietrich et al (2007) suggested that researcher who wants to measure conservatism should used different methods such as market-to-book ratio or the negative non-operating accrual method.

The next paper that has also some critics on the Basu (1997) timely loss recognition (TLR) is that of Givoly et al. (2007). They have tested the power and the reliability of the TLR. To test this they have used the regression formula of Basu (paragraph 3.2.1) ,for the test period of 1950-2001 for firms in the USA, and have checked the sensitivity of this formula against certain unrelated conservatism characteristics. Those characteristics where: aggregation effect, the nature of the economic events effect and the disclosure policy effect (Givoly et al.,2007). TLR uses stock returns as the measure of conservatism, but those stock returns are aggregated data. Therefore the TLR is measuring data that is gathered together to form a total quantity.

Givoly et al (2007) found evidence that the TLR is influenced by those characteristics. Thus the TLR method leads to abnormal results. Their final conclusion was that the TLR measures only a small fraction of conservatism, for that reason researcher should use more measurement method to capture conservatism. Otherwise when a researcher relies only on a single measurement method the result will lead to incorrect conclusion (Givoly et al., 2007).

The final paper of this category is the paper of Huang et al. (2008). They investigated the heterogeneity of the firm specific characteristics in the TLR. The used a modified regression model of Basu (1997) which includes some firm specific characteristics to measure the level of conservatism of firm in de USA in the period 1976-2005. Huang et al (2008) conclude that the level of conservatism is much lower as indicated in prior research and that the level of conservatism did not increase over time. This is inconsistent with the finding of Basu (1997) and Givoly and Hayn (2000). The reason could be that the sample period is different and more importantly Huang et al.(2008) take into account that firm specific characteristics aren't homogenous. As a result of that the characteristics will change and are therefore not the same for every firm.

All three papers have comments on the timely loss recognition method. They are consistent that this method has several shortcomings. However the papers indicate different shortcomings. Varying from biases that are inherent on the method, to characteristics that influence the method. And finally not controlling for heterogeneity of characteristics.

4.4 explanations of conservatism

The first paper in this category will be Beatty et al. (2008).. Beatty et al. (2008) used several methods to measure conservatism form the period 1994-2004 and link them to agency cost. Agency cost are measured by contract specific and firm specific variables. They found evidence that conservative reporting is used with conservative debt covenants, by firms and lender, to resolve their agency conflicts. Finally Beatty et al. (2008) concluded that contractual modifications alone do not fulfil lenders demands for conservatism. Meaning that lenders are in favour of conservatism and that they prefer conservatism in debt covenants and in financial reporting.

Zhang (2008) investigate the benefits of conservatism in the debt contracting process. He measures conservatism with several different methods for the period 1994-2003 and loan information for the debt contracting process. Zhang (2008) found evidence that conservatism leads to lower interest rates. This suggest that both lenders and borrowers benefit of conservatism. Otherwise lenders would not provide a lower interest rate. This finding are consistent with (Beatty et al., 2008), that debt contracting in conservatism is useful. However the conclusion made by Zhang (2008) are based on a restricted sample. Therefore his conclusion could be biased through the small sample size.

Chung and Wynn (2008) conducted research on the link between conservatism and litigation. They have measured conservatism through the TLR and the litigation risk was measured trough the sum of liability insurance coverage and the cash for indemnification for Canadian firm for the years 1998-2004. Indemnifications are provision taken by the company to compensate their manager for legal liability costs. As a results when managers acted honestly they don't have to pay for them self for the liability costs. Chung and Wynn (2008) used Canadian firms because for those firm their insurance data is publicly available. They found that firms with a high legal liability coverage are less conservative. Meaning that the risk for legal liability for managers becomes lower when they have a high legal liability coverage and therefore the earnings became less conservative. The conclusion is that when a company has a high legal liability risk, the company will be more conservative (Chung and Wynn, 2008).

Lafond and Watts (2008) studied the impact of information asymmetry on conservatism. They used the TLR of Basu (1997) to measure conservatism and information asymmetry was measured trough the PIN score and the bid-ask spread for USA firm in the period 1983-2001. The PIN score is a measure of abnormal order flow, which indicates that private information is used to sell or buy stocks (Lafond and Watts, 2008). Lafond and Watts (2008) found evidence that asymmetry is positively correlated to conservatism. This means that when asymmetry of information is high, than conservatism in accounting will also be high.

Khan and Watts (2009) have developed a firm year measure of conservatism. Khan and Watts have used the C_score method (paragraph 3.2.3) for a sample of USA firm for the period 1962-2005. They found that with the C-score they can predict the flow of conservatism for a period of three years ahead. They also found evidence that companies with a high information asymmetry have higher level of accounting conservatism. Khan and Watts also found that when a company has a high chance of litigation the company is more conservative. These results are consistent with prior research of that of Chung and Wynn (2008) and Lafond and Watts (2008).

All papers discussed in this category provide evidence that conservatism exists for a reason. Although the papers provide different explanations. They all concluded that conservatism in accounting have some benefits and therefore conservatism is still used in accounting

4.5 conservatism and crisis

Gul et al. [1] have done research to the link between conservatism and audit fees in a period of a crisis. They have used the TLR of Basu (1997) to measure conservatism and used an audit fee regression model to capture the audit fee. Gul et al. looked at listed Hong Kong firms for the period 1990-1997. Then they divided the sample into two period (1994-1995) and (1996-1997). The first period are the non-downturn years and the latter are the downturn years which indicates the crisis year (Gul et al.). Gul et al. found evidence that conservatism is lower during the crisis then the period before the crisis. And they found that the audit fees increases when conservatism decreases.

The paper of Herrmann et al. (2008) studied the impact of the Asian crisis on the conservatism of the auditor. To measure conservatism they used a modified Basu (1997) model which includes dummy variables for big 4 audit firms. Herman et al. (2008) looked at listed companies in Thailand for the period 1997-2003. They found that companies who are audited by a big 4 audit firm are more conservative then companies who are audited by non big 4 audit firms. Herrmann et al. (2008) found evidence that companies, audited by a big 4 audit firm or not, are less conservative during a crisis and report more conservative after the crisis. The evidence that firms are less conservative during crisis is consistent with the findings of Gul et al..

Eng et al. (2010) also found evidence that are consistent with the finding of Herrmann et al. (2008) and Gul et al.. According to Eng et al. (2010) companies are less conservative during a crisis and they are more conservative after the crisis. Only Eng et al. (2010) found evidence that companies are even more conservative after the crisis then before the crisis. Their conclusion where made on a sample of listed firms in Hong Kong, Malaysia, Singapore and Thailand for the period 1995-2004. To measure conservatism they uses the Basu (1997) model and non operating accrual method from Givoly and Hayn (2000).

The paper from Wu (2010) doesn't investigated if companies are less conservative during a crisis, but instead he studied the consequences of conservatism during the financial crisis. He found evidence that companies who are more conservative suffer less during the financial crisis then companies who are less conservative. Therefore conservative accounting is an efficient mechanism to control the agency costs (Wu, 2010). Therefore conservatism benefits the shareholders, because the agency problem is mitigated with conservatism. This finding is consistent with the explanation of conservatism (paragraph 4.4) that conservative accounting has some benefits.

Wu (2010) used the C_score method ( Khan and Watts), the non operating accruals method and the market to book ratio (Givoly and Hayn, 2000) to measure conservatism. He looked at listed firms in the USA for the period October 2007- December 2008. He used this specific period because there is no official date that indicates that the credit crisis has begun. Therefore he started in October 2007 and ended in 2008, because the data of 2009 was not yet available. As a result of the small sample size the finding of Wu (2010) could be biased and therefore the wrong conclusion where made.

The first three papers provided evidence that companies are less conservative during a crisis. Meaning that companies provide more good news and wait with their bad news during a crisis. And they found evidence that companies become more conservative after the crisis. Finally Wu (2010) provided evidence that more conservative firm suffer less during a crisis then less conservative firms. He also contributes some evidence that conservatism brings benefits to shareholders.

Chapter 5: Conclusion

At the beginning of the literature review accounting conservatism was discussed. Several definition were provided. Although a general accepted definition of conservatism is missing, conservatism can be seen as the higher verification requirement for good news/ gains then the verification requirement for bad news/ losses.

Conservatism is a important factor in accounting. People already used conservative accounting in the early 15th century. Therefore there has to be a reason why conservative accounting exists. Those reasons are: contracting, litigation, taxation and accounting regulation.

There are several method to measure conservatism in accounting. The Basu (1997) model is the most popular one in researches. But you can also measure conservatism with market-to-book ratio, non operating accruals (Givoly and Hayn, 2000) or C_score (Khan and Watts, 2010). All the methods have some limitations. But more importantly is the fact that the methods only measure a single part of accounting conservatism. Therefore you cannot measure conservatism in totality, when you use only one single method to measure conservatism (Givoly et al., 2007).

There has been a lot of research on the topic conservatism. The prior research of accounting conservatism indicates that conservatism has grown over time. Also they showed that there are limitations on the timely loss recognition method. Further the prior research resulted in evidence that conservative accounting has benefits. Finally prior research indicated that companies are less conservative during a crisis and more conservative after a crisis.

Research Design

As stated in the previous chapter. Conservatism is an important principle in accounting. It exist for many years. Conservatism is not only important for the company itself but, shareholder and debt holders also benefit from conservatism. Prior research found evidence that companies are less conservative during a crisis. However those research was based on the Asian crisis. And therefore had only an impact on Asian countries. Nowadays the recent credit crisis had an worldwide impact. More importantly it had an impact on the western developed countries.

Therefore I would like to investigate in this paper the following research question: What was the impact of the credit crisis on accounting conservatism?

I would expect that companies were less conservative during the crisis, what is consistent with prior research. The decreases in conservatism would be less than the decrease of conservatism in prior research. Because the accounting regulation has changed a lot during those years. And nowadays there are more in favour of fair value accounting. Fair value accounting you cannot be conservative because you have to use the market value.

The method I intent to use is the C_score (Khan and Watts, 2009), because they use firm specific characteristics in the regression model. The original model from Basu (1997) had the limitation that it assumes that firm characteristics are homogenous. While empirical evidence suggest that those characteristics are heterogeneous.

Because of the fact that one measurement method cannot measure conservatism in totality (Givoly et al., 2007). I want to use a second method to measure conservatism. This method will be the market-to-book ratio or the non operating accruals method. It depends on the amount of data that is available. For the first method less data is needed then for the latter one.

For the sample I had in mind to look at the listed companies in the Netherlands. I would like to use the companies that are listed on the AEX, AMX and the ASCX. Those listed companies are working in many different industry sectors. Therefore you will get an small sample of Dutch firms, which are spread across a lot of different industries.