One-sample
Abstract
This chapter studies the impact of financial reporting quality on firms’ market performance in a sample of LATAM corporations. We infer that, especially in contexts of high information asymmetry, investors are not able to effectively discern the quality of the information they are provided with and can therefore be misled in their investment decisions by managerial opportunism. Our theoretical framework is built upon a combined agency theory and cognitive approach. We thereby seek to provide a valuable method to better understand how investors could be making suboptimal choices as a consequence of managers’ opportunistic behaviour. Empirically, we use the Generalized Method of Moments (GMM) model, hypothesizing that a positive relationship should be observed between the opportunistic manipulation of earnings (that is, the misuse of accounting accruals) and the firm’s market performance (that is, the consequential behaviour of investors). Through this ‘pioneering’ methodology, applied to the relatively under-researched LATAM region, we find that: (1) Financial data are identifiably and consistently manipulated through discretionary accruals in these countries. (2) As manipulation increases, markets do tend to appear more attractive to investors. (3) The elasticity of the market reaction to this manipulation is higher in what we term ‘opaque’ countries.
Keywords
- corporate governance
- earnings quality
- market performance
- LATAM
- agency theory
- social cognitive approach
1. Introduction
Familiar Quotations, 10th ed. 1909.
Recent corporate scandals across the globe have drawn attention to the field of corporate governance. Users of financial information such as investors, governments and regulators are increasingly concerned about how earnings numbers are derived [1]. This is due in large part to the countless examples of managers who have used their discretionary decision-making power to misreport their firms’ profits. Petrobras in Brazil, which overpriced contracts for private benefits, or Disco in Argentina, which was found to have inappropriately recorded the financial results from several joint ventures, are just two examples of high profile firms that have inflated their earnings, to the detriment of investors and in direct contradiction to the provisions of governments and regulators.
This chapter studies the impact of financial reporting quality on firms’ market performance in a sample of LATAM corporations, using these data to examine the perception processes of investors as a mediating variable between reporting quality and market performance. Specifically, we address whether the perceived performance of an organization is in reality based upon actual organizational performance, or is instead more a function of the results overtly exhibited in the organization’s financial reporting structures, which may have been discretionally manipulated. We propose that, especially in contexts of high asymmetry of information, investors are not able to effectively discern the quality of information provided to them for decision-making purposes and can therefore be easily ‘fooled’ by managerial opportunism.
Empirically, we base this upon data collected in six Latin American countries by applying the Generalized Method of Moments (GMM) model [2], thereby hypothesizing that a positive relationship will be observed between the opportunistic manipulation of earnings and the firm’s market performance. We then examine these results using a lens that combines agency theory with a social cognitive approach, to analyse the manipulated perception process that occurs as a result of that relationship.
There are a number of principal contributions from this chapter. We begin by viewing the process of manipulation with a holistic approach that integrates both a cognitive and agency perspective and allows us to better understand the relationship between earnings management, financial reporting quality and market performance as a whole, thereby providing a more comprehensive vision of the entire process. This contribution is even more valuable because we have situated our study in the under-researched context of Latin America. We also believe that we are the first to apply the GMM methodology in this context, empirically showing how financial manipulation occurs and then impacts upon investor decisions, thus influencing the organization’s market valuation. By doing so we have created an algorithm and adapted an overall model that can be more generally used to rank the quality of any earnings reports, thereby contributing to a more transparent market information system. Finally, we hope that our research will go on to inform and serve decision-makers who analyse firms’ financial statements, as well as act as a catalyst to governments, institutions and policy-makers in deriving policy and promoting market efficiency.
Our most important findings can be summarized to be the following. Our results show: (1) Financial data are identifiably and consistently manipulated through discretionary accruals in these countries. (2) As manipulation increases, markets do tend to appear more attractive to investors. (3) The elasticity of the market reaction to this manipulation is higher in what we term ‘opaque’ countries.
The steps we take in this chapter begin with our theoretical framework, where we review the relevant literature, illustrate this with a comprehensive model of the overall process and then state our hypothesis. As a second step, we then proceed with the empirical analysis through the operationalization of our baseline model and the construction of our variables. In our third step, we present and discuss our results. We do this by displaying and analysing both the univariate and multivariate analysis and by segmenting the sample into clusters based on the country-level governance system, calling them ‘opaque’ (lower level of governance) or ‘transparent’ (higher level of governance) countries. Our conclusions and final remarks are presented at the end of the chapter, where we also address policy recommendations.
2. Theoretical framework
The extent to which financial statements reflect actual operating fundamentals is of growing concern throughout the world, especially in emerging markets where managerial controls and practices can vary substantially from those in the USA or Europe. Some more economically developed countries have passed legislation to ensure better corporate governance and have adopted codes of good conduct in order to reduce the asymmetries of information between shareholders and the firm and to increase the rational component of the decision-making process around choosing one’s investments [3, 4]. At the same time, a large difference in the quality of financial reporting across countries has been extensively documented [5]. This has led, according to the behavioural finance approach, to the conclusion that the perception of market participants is likely to be biased as a consequence of the lack of transparency in pricing and the poor quality of financial information [6]. Such losses in the quality of financial information have been modelled through earnings management [7].
Earnings management can be defined as the adjustment of a firm’s reported economic performance by insiders, done either to mislead some stakeholders or to influence contractual outcomes [8, 9]. Earnings management is considered to be the most informative and trustworthy to investors if it is supported by what is perceived to be a good system of governance. However, the act of managing earnings does not necessarily reflect the true performance of the company, a situation that may contribute to shareholders and investors making inaccurate judgements about the company [9]. To examine this, we first turn to agency theory, well used in the financial arena, which holds that managerial behaviour can be opportunistic and fuelled by self-interest. Most importantly for our purposes, it accounts for the existence of asymmetries of information between managers and shareholders. Executives accept agent status because they perceive an opportunity to maximize their own utility [10]. Consequently, agency theory holds that managers may take advantage of the information they have and their latitude in making accounting and reporting decisions to overstate financial information. They generally do this by acting in what they perceive to be in their own interest [11]. Reducing agency costs by imposing internal mechanisms of control should therefore encourage managers to behave in the best interest of shareholders instead of in their own interests. However, because controls are imperfect, we would expect some degree of opportunism to remain [10]. Since managers are widely paid based on firm’s performance, it is plausible to expect that active earnings manipulation will occur in order to enhance managerial compensation packages [12]. This approach is highly focussed on bounded rational decision making around incentives, information and self-interest. Thus, it is a viewpoint that suggests that it may be necessary to limit managers’ discretion with respect to accounting, since investors, as a consequence of asymmetrically distributed market information, cannot unravel the valuation effect of reported earnings in a timely manner under current reporting standards.
We suggest, however, that in addition to agency theory, a more cognitive viewpoint can also be used, to guide and further understand managerial behaviour. Social cognitive theory advocates that behaviour, cognitive and other personal factors and the external environment are the three main factors that drive the decision-making process [13]. These three factors are known to be asymmetrical, similar to the asymmetry of information in agency theory, in that they do not influence each other simultaneously, instantly or with equal strength, but they do influence each other multidirectionally. As a result, both investors and managers can be understood to be making decisions based upon a combination of factors that include a triad of perceptual, environmental and behavioural elements, all converging to ultimately produce an investment decision.
Regarding cognition, two of the most relevant elements related to decision making are managerial biases and heuristics. The most common biases that managers revert to using include representativeness, availability and anchoring-and-adjustment [14] although there are now many other biases and heuristics that have been studied at length in a financial context [15]. The use of heuristics is considered a necessary way for humans to cope with our more limited capacity to process information [16]. More specifically to our study, many researchers have identified the biases and heuristics used in making financial decisions as highly relevant to understand the human and cognitive elements of the processes involved [6, 17]. Thus, according to the social cognitive approach, the market may wrongly perceive the actual firm performance disclosed in financial reports, as a consequence of biases and heuristics held perceptually and socially, in addition to behavioural and environmental elements. Thus, when managers overstate a firm’s earnings, due to their bounded rationality and to information asymmetries, they can be easily misled to overprice the firm’s shares.
As described in Figure 1, by suggesting a complementary relation between agency and social cognition theories, we have produced a model that further explains how this process occurs and shows how the process is reinforced by the lack of sound corporate governance systems in the institutional context of Latin American countries, as is our case.
Financial markets in the region are still in a stage of early development, which allows managers to make use of accounting discretion to manipulate financial information. In immature financial markets, where there are large imbalances of information and opacity, and where the markets are not integrated, investors may not be able to discriminate between companies that provide high or low quality information [18, 19]. Therefore, in the midst of inefficient financial markets in Latin America, managers have more room to manipulate financial statements. Leuz et al. [20] present evidence that the level of outside investor protection endogenously determines the quality of financial information reported to outsiders, showing how legal protection influences the agency conflicts between investors and controlling shareholders. In Latin America investor protection is weak, and this therefore gives insiders more incentives to manipulate earnings. In conclusion, in the institutional context of Latin America, investors suffer more acutely the consequences of earnings manipulation by managers when compared to more developed financial systems, and therefore they may not be able to make optimal investment decisions.
Therefore, based upon the previous arguments regarding agency theory, cognitive theory and the institutional setting in Latin America, we hypothesize that:
3. Baseline model and empirical analysis
3.1. Sample
The sample we use corresponds to 896 representative large non-financial firms from Argentina, Brazil, Chile, Colombia, Mexico and Peru. Kaufmann et al. [21].1 The sample corresponds to unbalanced panel data with a total of 9647 firm-year observations over the period from 1997 to 2013.
We use the Generalized Method of Moments (GMM) model to deal with the characteristic econometric problems of unobservable heterogeneity of individual firms and endogeneity [2, 22]. Several statistical contrasts are used as diagnostic tests for our panel data structures (e.g. the Hansen test for the validity of instruments, the second-order serial correlation test AR(2), the Wald test of joint significance of parameters, and the variance inflation factor (VIF) as a formal multicollinearity test). Additionally, non-linear restriction tests are used for those interacted (multiplicative) variables.
3.2. Variable construction
Key to this study is the definition and analysis of our proxy independent variable of earnings manipulation, which corresponds to our measure of managerial discretion and quality of financial reporting. Two alternative estimations of earnings management are used based on absolute discretionary accruals. Since total accruals are known, the discretionary accruals must be estimated. Based on Dechow et al. [23], the total accrual (
where
Thus, once the total accruals are calculated, we have to split them into their non-discretionary and discretionary components. Non-discretionary accruals are aimed to improve the informational content of financial statements. According to the Jones [24] model, total accruals are affected by the firm’s usual business (which can affect non-cash current assets and liabilities) and by fixed assets (which can affect the depreciation expense). Consequently,
Regarding the expected signs for
Hence, the value of (
where
Similarly, and as stated earlier, our second proxy measure of discretionary accruals also follows a cross-sectional model based on the Jones [24] model as described by Dechow et al. [23] as:
The coefficient estimates from Eq. (4) are used to estimate the firm-specific non-discretional accruals as:
where Δ
Similar to the first measure, we also compute discretionary accruals in their absolute values.
In our models, the firm market performance as a dependent variable is computed through a number of alternative measures. First, we use the market return (
where
where
For
where
The other independent variables correspond to control variables entered into the model in order to avoid problems of misspecification. The first control variable corresponds to the leverage at book value (
where
For country-level variables we use the Worldwide Governance Index2 (
Therefore, our estimation model would take the following form:
where
4. Empirical results
4.1. Univariate analysis
For the empirical results we proceed in two parts. As a starting point, in order to make the empirical analysis significant, we have to test the null hypothesis that the mean values of the discretionary accruals measures are statistically significant from zero. Previous literature suggests that managers of companies with weak governance structures have greater discretion to engage in opportunistic earnings management [29]. A similar situation is observed when the regulatory environment does not efficiently constrain management’s flexibility to misrepresent financial results [30]. The p-values reported in Table 1 suggest that the mean values of our alternative measures of discretionary accruals are significantly different from zero. In accordance with the previous literate, therefore, this preliminary finding may be used as evidence that listed firms in our sample opportunistically manipulate their financial reports.
Variables | Obs | Mean | Std. error | Std. dev. | |
---|---|---|---|---|---|
DACC1 | 9635 | 0.0217 | 0.0002 | 0.0239 | 0.0000 |
DACC2 | 9635 | 0.0260 | 0.0004 | 0.0385 | 0.0000 |
DACC3 | 9635 | 0.0276 | 0.0003 | 0.0257 | 0.0000 |
Additionally, we split the sample into two big groups depending on the average value of the GOVINDEX variable by country. Although not reported, the average values were negative for all the countries, except for Chile and Brazil. Consequently, we can state that for the group of countries comprised of Chile and Brazil the transparency and corporate governance rules are relatively more efficient than for Argentina, Colombia, Mexico and Peru. Thus, the sample slip corresponds to these two groups, namely ‘Transparent Countries’ for those countries with relatively better transparency and corporate governance, and ‘Opaque Countries’ corresponding to those with relatively worse transparency and corporate governance. As observed in Table 2, the mean difference test was applied to verify if the extent of discretionary accruals as a measure of financial overstatement is the same across the two groups. The null hypothesis is that there is no difference in discretionary accruals between the two groups and the alternative hypothesis is that discretionary accruals are greater in the group of countries with relatively weaker transparency and governance. As tabulated, DACC1 is statistically greater in the set of countries which are less transparent and where corporate governance is weaker (e.g. Argentina, Colombia, Mexico and Peru, all of which we call the ‘Opaque Countries’) than in the set of countries such as Chile and Brazil where institutional transparency and corporate governance are better, and which we term the ‘Transparent Countries’. This is shown to provide evidence that more financial statement manipulation is present when institutions and governance are weaker.
Variables | Countries | Obs | Mean | Std. error | Std. dev. | Difference | Ha: diff > 0 |
---|---|---|---|---|---|---|---|
DACC1 | Opaque countries | 4002 | 0.0224 | 0.0004 | 0.0223 | 0.0012 | 0.0070 |
Transparent countries | 5633 | 0.0212 | 0.0003 | 0.0249 | |||
DACC2 | Opaque countries | 4002 | 0.0262 | 0.0005 | 0.0316 | 0.0002 | 0.3770 |
Transparent countries | 5633 | 0.0259 | 0.0006 | 0.0428 | |||
DACC3 | Opaque countries | 4002 | 0.0280 | 0.0004 | 0.0261 | 0.0006 | 0.1120 |
Transparent countries | 5633 | 0.0273 | 0.0003 | 0.0254 |
The descriptive statistics in Table 3 show that for our three measures of market performance (e.g. MP1, MP2 and MP3) there are positive average values for the companies included in the sample. Additionally, a typical company finances its total assets with about 48.73% of debt. In our sample, companies achieve an average rate of return of 4.27% on total assets. Finally, the average indicator of transparency and quality of corporate governance (GOVINDEX) is only 0.2146 with a maximum coefficient of 1.2482, which is still far away from its theoretical maximum achievable designated to be 2.5 [21].
Variable | Mean | Std. dev. | Min | Max |
---|---|---|---|---|
MP1 | 0.1272 | 0.4908 | −0.8700 | 1.9984 |
MP2 | 8.7660 | 3.2880 | 0.0010 | 16.9760 |
MP3 | 6.4865 | 2.0207 | 0.0058 | 11.9799 |
DACC1 | 0.0217 | 0.0239 | 0.0000 | 0.3838 |
DACC2 | 0.0260 | 0.0385 | 0.0000 | 0.7007 |
DACC3 | 0.0276 | 0.0257 | 0.0000 | 0.2895 |
LEV | 0.4873 | 0.2349 | 0.0072 | 0.9467 |
SIZE | 6.1524 | 2.0842 | −2.0887 | 13.2225 |
ROA | 0.0427 | 0.0945 | −0.4515 | 0.4948 |
RISK | 7.2761 | 5.3480 | 0.1667 | 33.0204 |
GOVINDEX | 0.2146 | 0.5851 | −0.6658 | 1.2482 |
The matrix of correlation coefficients is exhibited in Table 4. As would be expected, there is a high correlation for some measures of performance, such as the 0.533 correlation coefficient between MP2 and MP3. Similarly, high correlations are observed between the measures of discretionary accruals. On the other hand, we do not observe relatively high levels of correlation between the explanatory variables, with the exception of the correlation between the firm size (SIZE) and its leverage (LEV) (e.g. significant correlation of 0.408) and between firms’ default risk (RISK) and the level of debt (LEV) (e.g. significant correlation of 0.691).3 These slightly high correlations might eventually cause problems of multicollinearity in the regression estimates. Nevertheless, as reported in the subsequent regression tables, the variance inflation factor (VIF) test allows us to accept the hypothesis of the inexistence of this econometric problem.
Variables | MP1 | MP2 | MP3 | DACC1 | DACC2 | DACC3 | LEV | SIZE | ROA | RISK |
---|---|---|---|---|---|---|---|---|---|---|
MP2 | 0.040 | 1.000 | ||||||||
(0.000) | ||||||||||
MP3 | 0.030 | 0.533 | 1.000 | |||||||
(0.004) | (0.000) | |||||||||
DACC1 | −0.028 | 0.025 | −0.065 | 1.000 | ||||||
(0.007) | (0.019) | (0.000) | ||||||||
DACC2 | −0.033 | 0.015 | −0.068 | 0.934 | 1.000 | |||||
(0.001) | (0.168) | (0.000) | (0.000) | |||||||
DACC3 | −0.036 | −0.036 | −0.121 | 0.859 | 0.839 | 1.000 | ||||
(0.001) | (0.001) | (0.000) | (0.000) | (0.000) | ||||||
LEV | −0.014 | 0.016 | 0.358 | −0.125 | −0.080 | −0.103 | 1.000 | |||
(0.165) | (0.135) | (0.000) | (0.000) | (0.000) | (0.000) | |||||
SIZE | 0.031 | 0.528 | 0.965 | −0.085 | −0.074 | −0.136 | 0.408 | 1.000 | ||
(0.003) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | ||||
ROA | 0.184 | 0.115 | 0.012 | −0.003 | −0.005 | −0.029 | −0.274 | 0.004 | 1.000 | |
(0.000) | (0.000) | (0.266) | (0.744) | (0.626) | (0.005) | (0.000) | (0.709) | |||
RISK | 0.044 | −0.015 | −0.238 | −0.041 | −0.020 | −0.0362 | −0.691 | −0.284 | 0.334 | 1.000 |
(0.000) | (0.147) | (0.000) | (0.000) | (0.050) | (0.001) | (0.000) | (0.000) | (0.000) | ||
GOVINDEX | −0.050 | 0.535 | −0.088 | −0.065 | 0.060 | 0.042 | −0.176 | −0.115 | 0.057 | 0.139 |
(0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) | (0.000) |
4.2. Multivariate analysis
Concerning the multivariate analysis, we interpret the outcomes of the model (12) for the whole sample according to the regression estimates shown in Table 5. This table includes nine regressions for our three alternative measures of the dependent variable (e.g. MP1, MP2 and MP3) explained by our three one-period lagged independent variables as measures of the quality of the financial reports and earnings manipulation (e.g. DACC1, DACC2 and DACC3). As a starting point for the interpretation of the coefficient estimates in our regression analysis, a battery of diagnostic tests is used to ensure the validity of our results in Tables 5 and 6. Robust standard errors were used in all the regression estimates. According to the Wald test, all the independent variables are jointly significant at the standard confidence levels. As mentioned above, the variance inflation factor (VIF) reported at the bottom of Tables 5 and 6 confirm that collinearity does not skew our estimation results since the VIFs are greater than 2. Regarding the moment conditions, the Hansen over-identification tests did not reject the over-identifying restrictions, meaning that we accept the null hypothesis of validity of the instruments in our estimations. Additionally, the AR(2) test proves the lack of second-order serial correlation. Consequently, our results are not biased by a possible incorrect choice of instruments or by autocorrelation and are robust, according to the standard diagnostic tests for panel data.
Variables | MP1 | MP1 | MP1 | MP2 | MP2 | MP2 | MP3 | MP2 | MP3 |
---|---|---|---|---|---|---|---|---|---|
DACC1t−1 | 2.1660 | 6.2290*** | 2.1726*** | ||||||
(1.3985) | (6.0160) | (4.6340) | |||||||
DACC2t−1 | 6.0404*** | 5.1725*** | 0.3698 | ||||||
(3.6418) | (9.2875) | (1.6330) | |||||||
DACC3t−1 | 6.6370*** | 0.5416 | 0.5009** | ||||||
(3.3999) | (1.0670) | (−2.3382) | |||||||
LEV | −2.7174 | −3.8098 | 9.3240 | 3.9353*** | 5.5361*** | 1.5004*** | 1.7003*** | 1.5353*** | 1.5901*** |
(−0.7238) | (−0.9337) | (0.7826) | (11.4338) | (15.2080) | (5.6679) | (10.3436) | (10.6859) | (11.4474) | |
SIZE | −0.5093 | −0.3941 | −1.2559 | 1.1193*** | 1.2469*** | 1.0573*** | 0.9666*** | 0.9670*** | 0.9491*** |
(−1.0917) | (−0.7404) | (−1.3556) | (25.8565) | (28.5999) | (29.8487) | (72.3698) | (69.9110) | (82.4417) | |
ROA | 2.2433 | 8.8811*** | 7.1137*** | 0.0855 | 0.7537** | 2.5251*** | 0.0410 | 0.4219*** | 0.5651*** |
(0.4507) | (2.6688) | (4.2203) | (0.2407) | (−2.2683) | (8.2591) | (0.3049) | (3.0741) | (4.4991) | |
RISK | 0.6709** | 0.7802*** | 0.5878 | −0.0101 | 0.0571*** | 0.0554*** | 0.0182*** | 0.0029 | 0.0145** |
(2.0610) | (2.6424) | (0.6086) | (−1.3095) | (5.6442) | (4.7087) | (2.9191) | (0.4227) | (2.4912) | |
Constant | −0.6884 | −3.6828 | −7.4544 | 3.7492*** | 4.1199*** | 3.1807*** | −0.5662*** | −0.3507*** | −0.3625*** |
(−0.1412) | (−0.6646) | (−0.5519) | (13.6469) | (14.1623) | (12.9888) | (−4.5775) | (−2.6674) | (−3.3534) | |
Observations | 8848 | 8848 | 8848 | 8965 | 8965 | 8965 | 9608 | 9608 | 9608 |
Number of iden | 886 | 886 | 886 | 908 | 908 | 908 | 895 | 895 | 895 |
Wald test | 10.860*** | 15.337*** | 12.138*** | 577.000*** | 242.250*** | 671.023*** | 112.337*** | 62.253*** | 332.902*** |
AR(2) | 0.202 | 0.170 | 0.148 | 0.423 | 0.396 | 0.477 | 0.860 | 0.510 | 0.578 |
Hansen test | 49.950 | 112.740 | 96.312 | 276.400 | 161.599 | 130.220 | 325.322 | 161.830 | 147.520 |
VIF | 1.73 | 1.03 | 1.29 | 1.15 | 0.98 | 1.29 | 1.18 | 1.17 | 1.30 |
Variables | MP1 | MP1 | MP1 | MP2 | MP2 | MP2 | MP3 | MP3 | MP3 |
---|---|---|---|---|---|---|---|---|---|
DACC1t−1 | 7.059** | 1.6018*** | 1.7355* | ||||||
(0.9151) | (5.7779) | (1.8141) | |||||||
DACC1t−1*SYS | −4.9041* | −0.4964*** | −0.8013*** | ||||||
(−0.7209) | (−12.9280) | (−5.6391) | |||||||
DACC2t−1 | 3.5026*** | 0.7596* | 5.3086*** | ||||||
(3.2505) | (9.2927) | (5.2294) | |||||||
DACC2t−1*SYST | −1.9704*** | −0.6997** | −4.2790*** | ||||||
(−3.1764) | (−15.7794) | (−6.9984) | |||||||
DACC3t−1 | 2.0647*** | 0.4148*** | 3.2447*** | ||||||
(3.0704) | (9.9215) | (3.7770) | |||||||
DACC3t−1*SYST | −1.1733*** | −0.1247* | −2.7271* | ||||||
(−2.7950) | (−11.9847) | (−3.6554) | |||||||
LEV | −1.6711 | −10.1574 | −3.8463 | 4.3005*** | 5.2190*** | 0.4523 | 1.6498*** | 1.7214*** | 1.6522*** |
(−0.2296) | (−1.6045) | (−0.2746) | (11.3608) | (14.3559) | (1.2470) | (10.4653) | (11.5689) | (10.5313) | |
SIZE | −0.7768 | 0.6852 | 0.3946 | 1.0962*** | 1.1786*** | 1.0024*** | 0.9617*** | 0.9405*** | 0.9377*** |
(−0.8334) | (0.7225) | (0.2922) | (25.2228) | (27.2567) | (25.3953) | (64.1562) | (63.9162) | (75.0594) | |
ROA | 9.0907 | 17.2895** | 42.4076** | −0.2355 | 0.6890** | 2.4617*** | 0.0619 | 0.3948** | 0.5256*** |
(0.8169) | (2.1722) | (2.0678) | (−0.6157) | (−2.1810) | (6.4494) | (0.4093) | (2.3561) | (3.4411) | |
RISK | 0.2831* | 0.8423** | 1.1136 | −0.0159 | 0.0709*** | 0.0629*** | 0.0143** | 0.0029 | 0.0109 |
(0.5280) | (2.4286) | (1.0804) | (−1.4665) | (6.2080) | (4.0482) | (2.1930) | (0.3744) | (1.5524) | |
Constant | 2.5543 | −9.2934 | −16.1757 | 4.0527*** | 4.5408*** | 3.1268*** | −0.4891*** | −0.2535* | −0.2807** |
(0.2844) | (−1.1314) | (−0.9620) | (13.4278) | (15.6612) | (10.2658) | (−3.9125) | (−1.8302) | (−2.2376) | |
Observations | 8848 | 8848 | 8848 | 8965 | 8965 | 8965 | 9608 | 9608 | 9608 |
Number of iden | 886 | 886 | 886 | 908 | 908 | 908 | 895 | 895 | 895 |
Wald test | 762.470*** | 527.150*** | 649.033*** | 112.934*** | 270.732*** | 174.122*** | 133.826*** | 84.724*** | 137.590*** |
AR(2) | 0.998 | 1.124 | 1.490 | 1.859 | 1.460 | 1.385 | 0.994 | 1.137 | 1.280 |
Hansen test | 153.300 | 166.920 | 211.965 | 193.836 | 103.570 | 140.242 | 178.103 | 148.049 | 195.624 |
VIF | 1.121 | 1.205 | 1.380 | 1.094 | 1.183 | 1.124 | 1.150 | 0.902 | 1.661 |
4.2.1. Discussion of results of the whole sample
Concerning the findings, the results systematically show that higher manipulation of financial reports (DACC1, DACC2 or DACC3) leads to greater firm market performance (MP1, MP2 and MP3). Although in regressions (1), (6) and (8) our measures of discretionary accruals are not statistically significant, the direction of the relationship is still positive (e.g. see Table 5). For instance, in the second regression, we observe that an increase by one percentage point in our first one-period lagged measure of discretionary accruals (DACC1
Our results also support the Lee et al. [9] model, where firms with higher accounting performance over-report earnings by a larger amount when looking for greater price responsiveness or market performance. In the Lee et al. [9] model, managers manage earnings to influence the stock price. This is a plausible explanation for our results. We suggest that under a rational setting that is free of market frictions, where information is symmetrically distributed and where there is complete alignment of interest between managers and shareholders, there is no room for managers to opportunistically manage earnings to increase market performance. Under these conditions, the market would be able to discriminate and choose a separating equilibrium as suggested by Akerlof [18], by rationally discounting for the over-statement of earnings. This supports the idea that buyers are guided by earnings but are unaware that earnings are inflated by the generous use of accruals, and that this is a consequence of individual biases, wrong perceptions and misuse of heuristics in making their financial decisions. Thus, investors are misled to pay too high a price [31], which triggers a greater change in stock price, enterprise value and Tobin’s Q. Hence, when there is a lack of transparency and the agency conflict is not efficiently minimized, managers take advantage of their discretionary power to artificially boost the market performance of the company. The major motivation behind this is basically the improvement of contractual conditions and reward with better compensation packages. Or as stated in terms of Teoh et al. [31], managers manage earnings to exploit market credulity. This idea is consistent with investors naively extrapolating earnings without fully adjusting for the potential manipulation of reported earnings [32]. Consequently, the previous findings allow us to accept our research hypothesis that there will be a positive relationship between opportunistic manipulation of earnings and firms’ market performance.
Concerning the control variables included in the model specifications, we observe that leverage (LEV) does not impact on the stock price change (MP1), but it has a positive relationship with our two other alternative measures of market performance, namely the enterprise value (MP2) and the proxy for Tobin’s Q (MP3). As suggested by the capital structure literature, debt can be efficiently used to undertake profitable investment projects that the market interprets as positive growth opportunities by pushing up the market valuation [33]. Similarly, the size of the company (SIZE) is also positively related to its performance. According to our findings, larger firms take advantage of economies of scale and this dimension is rewarded with a premium in market valuation. The return on assets (ROA) is also positively associated with the market performance. Consequently, there is a direct correspondence between the bottom-line net income and the firms’ stock performance. Regarding our last control variable, the default risk (RISK) is found to be negatively associated with market performance. As mentioned earlier, by construction, the RISK variable increases as the default risk decreases, and consequently, the results reported in Tables 5 and 6 must be interpreted in the opposite direction. Hence, our findings suggest that the market discounts prices when the firm is approaching bankruptcy as suggested by the literature [28, 34].
4.2.2. Discussion of results by levels of governance
In this section on multivariate analysis, we aim to study the impact of different levels of transparency and efficiency of country-level governance systems on the relationship between earnings management and the firms’ market performance. As has been widely supported in previous literature, governance systems in the Latin American region are comparatively weaker than in other more developed economies such as the US or Europe [35–37]. Consequently, such opaqueness in the financial markets and the weaker protection of investor rights determines how actively managers over-state financial information disclosed to the markets [38]. Moreover, differences in governance systems have also been observed across Latin American countries [39].
To disentangle this issue we add to our estimation model (12) a variable that allows us to control for cross-country differences in governance systems and transparency. To do so, we create a dummy variable (SYS) based on the subsamples of ‘Transparent Countries’ and ‘Opaque Countries’ described in Section 4.2. This dummy variable takes the value 1 if the country belongs to the subsample of ‘Transparent Countries’ and zero for the subsample of ‘Opaque Countries’. Afterwards, we interact the SYS variable with our alternative measures of discretionary accruals and create a multiplicative variable, for instance, DACC
The results reported in Table 6 indicate, on the one hand, that the one-period lagged variable of discretionary accruals is always positively related to market performance. On the other hand, the interacted or multiplicative variables between discretionary accruals and transparency and governance efficiency (see for instance DACC
In all the subsequent regression estimates of Table 6, we observe that the impact on any measure of market performance (MP1, MP2 or MP3) as a consequence of a change in the discretionary accruals is systematically greater in the group of ‘Opaque Countries’ than in the group of ‘Transparent Countries’. This may be used as robust evidence that managers take more advantage of market myopia when institutional settings are endowed with weaker governance systems and where greater gaps of information exist between insiders and outsiders. In other words, although we subscribe to previous literature on the fact that governance systems are relatively weak in the Latin American region [40], we also recognize that there are still some intraregional differences in transparency and governance, as supported by our findings. Thus, in more transparent financial systems and where the right of shareholders is relatively better protected, the impact on market performance caused by opportunistic manipulation of financial reports is not as large as in contexts of less transparency and governance. We can derive out of this finding that the market is fooled in order to increase the firm’s valuation by mispresenting the financial information. And even more, the weaker the governance systems across countries in the Latin American region, the greater the changes will be to boost firm value by misleading the market towards making wrong investment decisions.
Finally, findings concerning the control variables listed in Table 6 remain consistent with those previously interpreted based on Table 5. Thus, we can conclude that our overall findings are robust to a battery of alternative test specifications and controls, as well as to elaborate dependent and independent variables.
5. Conclusions and final remarks
The main goal of this chapter has been to measure the impact of earnings management and reporting on market performance. We have sought to examine this phenomenon in a holistic way. Far from a purely statistical correlation analysis, we have sought to examine the phenomenon in light of theories that support this from a management point of view, in an attempt to merge the two together.
The two major theories we have applied include agency theory and social cognitive theory. According to Eisenhardt [12], agency theory is particularly effective when coupled with complementary perspectives. We have therefore created a theoretical model that serves to illustrate our operational model, by showing how this process happens as a whole. While it does describe our two mediating variables, financial reporting quality and investment decision making, we conceptually consider these to be a ‘black box’ that then allows us to focus more on the relationship of the two variables in the extremes of the model, earnings management and stock performance. Overall, our approach seeks to show how both a cognitive and an agency approach can be used together to demonstrate how a firm’s earnings quality can impact on its market performance.
A number of policy recommendations are derived from our findings. First, regulators and those who set accounting standards may find these results useful for assessing the levels of discretion that should be permitted to corporate managers for adjusting their financial reports. Second, our results suggest that individual investors will behave more rationally and be more aware in their investing decisions if the impact of discretionary accruals on the stock price is made more apparent. Overall, we argue that there is a clear need for more transparent financial markets and enhanced corporate governance measures.
References
- 1.
Wu S-W, Lin F, Fang W. Earnings management and investor's stock return. Emerging Markets Finance & Trade. 2012; 48 :129-140 - 2.
Arellano M. Sargan's instrumental variables estimation and the generalized method of moments. Journal of Business & Economic Statistics. 2002; 20 (4):450-459 - 3.
La Porta R, Lopez-De-Silanes F, Shleifer A. The economic consequences of legal origins. Journal of Economic Literature. 2008; 46 (2):285-332 - 4.
La Porta R, Lopez-De-Silanes F, Shleifer A, Vishny R. Investor protection and corporate governance. Journal of Financial Economics. 2000; 58 (1-2):3-27 - 5.
Akisik O. Accounting regulation, financial development, and economic growth. Emerging Markets Finance & Trade. 2013; 49 (1):33-67 - 6.
Mokoaleli-Mokoteli T, Taffler R, Agarwal V. Behavioural bias and conflicts of interest in analyst stock recommendations. Journal of Business & Accounting. 2009; 36 (3):348-418 - 7.
Agarwal S, Chomsisengphet S, Liu C, Ghon Rhee S. Earnings management behaviors under different economic environments: Evidence from Japanese banks. International Review of Economics & Finance. 2007; 16 (3):429-443 - 8.
Nagar V. Discussion of “performance, growth and earnings management”. Review of Accounting Studies. 2006; 11 (2):335-337 - 9.
Lee C-WJ, Li LY, Yue H. Performance, growth and earnings management. Review of Accounting Studies. 2006; 11 (2):305-334 - 10.
Davis JH, Schoorman FD, Donaldson L. Toward a stewardship theory of management. The Academy of Management Review. 1997; 22 (1):20-47 - 11.
Jensen MC, Meckling W. Theory of the firm: Managerial behaviour, agency cost and ownership structure. Journal of Financial Economics. 1976; 3 (4):305-360 - 12.
Eisenhardt KM. Agency theory: An assessment and review. The Academy of Management Review. 1989; 14 (1):57-74 - 13.
Bandura A, Adams NE, Beyer J. Cognitive processes mediating behavioral change. Journal of Personality and Social Psychology. 1977; 35 (3):125-139 - 14.
Kahneman D, Tversky A. On the psychology of prediction. Psychological Review. 1973; 80 (4):237-251 - 15.
De Neys W. Heuristic bias, conflict, and rationality in decision-making. In: Glatzeder B, Goel V, Müller A, editors. Towards a Theory of Thinking: Building Blocks for a Conceptual Framework. Berlin, Heidelberg: Springer Berlin Heidelberg; 2010. pp. 23-33 - 16.
Sklad M, Diekstra R. The development of the heuristics and biases scale (HBS). Procedia—Social and Behavioral Sciences. 2014; 112 :710-718 - 17.
Shanteau J. Cognitive heuristics and biases in behavioral auditing: Review, comments and observations. Accounting, Organizations and Society. 1989; 14 (1-2):165-177 - 18.
Akerlof G. The market for “lemons”: Quality uncertainty and the market mechanism. The Quarterly Journal of Economics. 1970; 84 (3):488-500 - 19.
Saona P, Vallelado E. Firms’ capital structure under Akerlof’s separating equilibrium. Spanish Journal of Finance and Accounting. 2012; XLI (156):471-495 - 20.
Leuz C, Nanda D, Wysocki PD. Earnings management and investor protection: An international comparison. Journal of Financial Economics. 2003; 69 (3):505-527 - 21.
Kaufmann D, Kraay A, Mastruzzi M. The worldwide governance indicators: Methodology and analytical issues. Hague Journal of the Rule of Law. 2011; 3 (2):220-246 - 22.
Arellano M, Bover O. La econometría de datos de panel. Investigaciones Económicas (Segunda Época). 1990; 14 (1):3-45 - 23.
Dechow PM, Sloan RG, Sweeney AP. Detecting earnings management. Accounting Review. 1995; 70 (2):193-225 - 24.
Jones JJ. Earnings management during import relief investigations. Journal of Accounting Research. 1991; 29 (2):193-228 - 25.
Gabrielsen G, Gramlich JD, Plenborg T. Managerial ownership, information content of earnings, and discretionary accruals in a non-US setting. Journal of Business Finance & Accounting. 2002; 29 (7-8):967-988 - 26.
Cohen DA, Dey A, Lys TZ. Real and accrual-based earnings management in the pre- and post-Sarbanes-Oxley periods. The Accounting Review. 2008; 83 (3):757-787 - 27.
Perfect S, Wiles K. Alternative constructions of Tobin's Q: An empirical comparison. Journal of Empirical Finance. 1994; 1 (3-4):313-341 - 28.
Altman EI. An emerging market credit scoring system for corporate bonds. Emerging Markets Review. 2005; 6 (4):311-323 - 29.
Dechow PM, Sloan RG, Sweeney AP. Causes and cnsequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research. 1996; 13 (1):1-36 - 30.
Davidson R, Goodwin-Stewart J, Kent P. Internal governance structures and earnings management. Accounting & Finance. 2005; 45 (2):241-267 - 31.
Teoh SH, Welch I, Wong TJ. Earnings management and the long-run market performance of initial public offerings. The Journal of Finance. 1998; 53 (6):1935-1974 - 32.
Teoh SH, Welch I, Wong TJ. Earnings management and the underperformance of seasoned equity offerings. Journal of Financial Economics. 1998; 50 (1):63-99 - 33.
Turk Ariss R. Legal systems, capital structure, and debt maturity in developing countries. Corporate Governance: An International Review. 2016; 24 (2):130-144 - 34.
Altman EI. Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of Finance. 1968; 23 (4):589-609 - 35.
Chong A, Lopez-De-Silanes F. Investor Protection and Corporate Governance: Firm-level Evidence Across Latin America: Stanford Economics and Finance/Stanford University Press ,Washington, D.C.; 2007 - 36.
Lefort F, Walker E. Corporate governance: Challenges for Latin America. Revista ABANTE. 2000; 2 (2):99-111 - 37.
La Porta R, Lopez-De-Silanes F, Shleifer A. Corporate ownership around the world. The Journal of Finance. 1999; 54 (2):471-517 - 38.
Sáenz González J, García-Meca E. Does corporate governance influence earnings management in Latin American markets? Journal of Business Ethics. 2014; 121 (3):419-440 - 39.
Castro Martins H, Schiehll E, Soares Terra PR. Country-level governance quality, ownership concentration, and debt maturity: A comparative study of Brazil and Chile. Corporate Governance: An International Review. 2016:n/a-n/a - 40.
Saona P, San Martín P. Determinants of firm value in Latin America: An analysis of firm attributes and institutional factors. Review of Managerial Science. 2016:1-48
Notes
- The latest update took place in September 2014. Information can be downloaded from www.govindicators.org.
- The latest update took place in September 2015. Information can be downloaded from www.govindicators.org.
- Although the tabulated correlation is negative, its interpretation is in the opposite direction as a consequence of the construction of the RISK variable where the firm risk increases as the variable decreases.