Accounting comparability and accruals-based earnings management: Evidence on listed firms in an emerging market

Abstract Users of financial statements are in a better position to evaluate the performance of the firm in question if its accounting system and methods are comparable with those of other firms. Even though accounting comparability is markedly beneficial, the empirical studies on its impact on earnings management are scant, especially in the context of developing countries that have not applied International Financial Reporting Standards (IFRS). In the present study, we examined the association between comparability and accruals-based earnings management, considering the moderating impacts of financial constraints in the context a developing economy. The article used a sample of 502 Vietnamese non-financial listed firms from 2010 to 2019. Three proxies of comparability and two measures of accruals-based earnings manipulation, together with several empirical strategies, are employed. First, the research provides evidence suggesting that comparability reduces firms’ tendency to engage in earnings management. Furthermore, there is evidence suggesting that firms could give more weighting to the comparability with a small number of firms rather than various firms in the same industry. Second, we are the first to identify that financial constraints have a moderating effect: more constrained firms do not reduce earnings manipulation as their accounting comparability improves. Based on research findings, we provided implications to stakeholders and propose directions for future studies.


PUBLIC INTEREST STATEMENT
Financial statements of different firms can be compared if they are produced using the same accounting frameworks and procedures. Therefore, comparable financial statements should make it easier for corporate outsiders to uncover mananagerial unusual manipulation of accruals to falsely impress investors. In this study, we examined whether accounting comparability acts as a governing mechanism, i.e. it helps deter earnings manipulation. Using a research sample of listed non-financial firms in Vietnam, we find that accounting comparability reduces accrualsbased earnings management. Furthermore, we find that the incentive to engage in earnings management when accounting comparability is higher differs depending on the level of financial constraints. To be more specific, constrained firms that have more difficulty in accessing external financing tend to not reduce earnings management even when accounting comparability increases. Based on these findings, we provide several implications to the relevant stakeholders.

Introduction
Financial reports are composed to enable users to evaluate characteristics of a firm's cash flows, including the timing, amount and certainty (J-B. S. Kim et al., 2013). Information is more likely to meet this purpose if it could be compared within the same firm across multiple points in time or with information of other firms in the same period. According to Financial Accounting Standards Board (FASB) (2010), comparability is a particularly important trait of financial information that makes it useful . With comparable financial statements, users have the luxury of better benchmarks to compare and contrast the performance of a firm, which makes it cheaper, faster and more convenient to make informed decisions.
Despite the theoretical appeal of comparability, little effort has been exerted to empirically evaluate the impact of comparability on different economic outcomes (Chen & Gong, 2019;Sohn, 2016). This is due to the abstract feature of comparability which makes it challenging to measure; however, De Franco et al. (2011) pioneered in providing a widely-accepted proxy for accounting comparability by resorting to the idea that comparable accounting systems should reflect similar transactions more similarly. In line with the theoretical expectations, De Franco et al. (2011) find that comparability benefits market players, by favorably increasing forecast accuracy and coverage and lowering dispersion of analyst forecast figures.
Subsequent studies also document the merits of accounting comparability in terms of improving information environment and facilitating the assessment of firm performance (Chen et al., 2018;J. B. Kim et al., 2020;S. Kim et al., 2013;Qingyuan & Lumeng, 2018). Gong et al. (2013) find that managers of firms with lower comparability tend to issue more earnings forecasts in an attempt to compensate for the increased information asymmetry between the firm and outsiders.
Theoretically, by increasing the cost to engage in earnings management through its ability to improve information environment, accounting comparability should limit the tendency to engage in accruals-based earnings manipulation. There have been studies investigating many positive aspects of comparability on information environment and resulting outcomes including cost of equity, cost of debt, stock price crash risk, analyst forecast accuracy, earnings from mergers and acquisitions. Nonetheless, Chen and Gong (2019) and Sohn (2016) claim that there are limited studies on the link between accounting comparability and earnings management. Agency theory and empirical evidence explain that the motivations to misrepresent firm performance through accruals-based earnings management emanate from the insiders' opportunistic behaviors (Alhebri et al., 2020;Asogwa et al., 2019;Elghuweel et al., 2017;Haw et al., 2004;Leuz et al., 2003;Shleifer & Vishny, 1997;Zingales, 1994). Sohn (2016) contributes to the literature by examining the trade-off between accruals-based earnings management and real earnings management as a function of accounting comparability. The author showed that comparability hinders managerial incentives and ability to engage in accruals-based earnings manipulation since the inappropriate use of accruals are prone to be detected more easily, but the management can instead opt for more manipulation of real activities. Chen and Gong (2019) find that investors effectively understand the firms' accruals when they have better financial statement comparability. Those studies suggest that accounting comparability limits opportunistic accruals-based earnings management, since it is challenging for managers to use accruals to distort earnings when accruals become more transparent and information is more easily obtained with higher accounting comparability.
We utilize a sample of all the listed non-financial firms in Vietnam in the period from 2010 to 2019. Vietnamese firms are striving to fully adopt IFRS in order to improve the comparability of financial statements and to be ready for global integration. It is interesting to examine how accounting comparability affects managerial manipulation of earnings figures. This should provide a solid basis for policy-makers to determine whether IFRS should be advocated, since the adoption of IFRS is supposed to enhance accounting comparability.
In short, the mainstream findings of previous studies point to a positive impact of accounting comparability on different aspect of information environment and helps alleviate information asymmetry. In the current study, we seek to contribute to the accounting comparability literature in two ways. Firstly, we contribute to the research strand on the link between financial reporting quality and earnings management. According to Sohn (2016), this empirical strand is, surprisingly, quite limited, even with the practical significance of accounting comparability and the global race to adopt the International Financial Reporting Standards (IFRS). We investigate this important issue in the context of Vietnam, a country that is late in the adoption of IFRS. Vietnam is now under intense and concrete reforms to have IFRS integrated into the local reporting standards in an effort to raise the transparency and comparability of financial statements. This is particularly relevant in the globalization context of this country. Previous studies on accounting comparability and earnings management focus on developed countries and with the application of IFRS (in US in Sohn (2016) and Chen and Gong (2019)), so the comparability level among firms should be high in those countries. On the other hand, in Vietnam, since IFRS has not been mandatorily applied, the comparability level should be low and this could affect the ability of this factor in alleviating accruals-based earnings management. Vietnam, therefore, serves as an important setting to conduct this research to compare with the findings for developed countries.
Secondly, extant studies tend to point to beneficial effects of comparability in improving information environment and incentivizing managers to provide quality information, thus curbing opportunistic earnings management practices. We test whether the impact of comparability on earnings management changes when firms have different levels of financial constraint. Theoretically, financial constraints render firms to face with difficulty in accessing external financing (Korajczyk & Levy, 2003); therefore, the impact of financial statement comparability on earnings management may differ depending on the level of financial constraints that the firms face. However, the debt incentive, which tends to affect mainly those firms with high financing needs, is expected to act as a constraint to the adoption of income decreasing actions, given that firms want to signal their quality to banks (Moreira, 2006). Again, this points to the case when accounting comparability is less likely to manifest its desirable effects on financial statement quality. We also perform several robustness tests to ensure the validity and robustness of our findings.
The paper continues as follows: Sections 2-4 provide research background and a review of prior literature and establish main empirical hypotheses; Section 5 discusses research methodology, covering empirical models, variable description and measurement and methods to test the hypotheses; Section 6 presents data statistics and results of estimation, together with robustness tests of the findings. Section 7 concludes the research, with implications for relevant stakeholders provided.

Background
There have been considerable studies on the influence of accounting comparability on the information environment around the firm. Nonetheless, the literature on the association between accounting comparability and earnings management is scant. Since financial statement comparability allows the collection of more valuable and relevant inputs with smaller costs (Engelberg et al., 2018;De Franco et al., 2011;Kini et al., 2009;Sohn, 2016), it should be of high interest to examine whether this factor could curb managerial incentives to manage earnings.
There has been an international trend to harmonize accounting standards to bring accountability, comparability, transparency and efficiency to users of financial statements in the past two decades (Lee, 2019). Since 1975, the International Accounting Standards Board has a chief role to play through its attempt in establishing a new framework of accounting standards (IFRS). The new standards have quickly been gaining ground to become an internationally acceptable financial reporting standards, because it is supposed to enhance financial statement comparability across firms and countries (Barth et al., 2008) and earnings value relevance (Odoemelam et al., 2019). Among other researchers, Lee (2019) finds that firms tend to report less manipulated earnings and the information environment about the firm improves following the adoption of IFRS. Nonetheless, Lee (2019) documents that a number of jurisdictions only encourage the use of IFRS standards, e.g., Japan and Switzerland, while seven other economies, including Vietnam and India, still hold on to their national standards.
Since 1995, with the membership of The World Trade Organization, Vietnamese firms have realized the critical role of rendering their financial statements comparable with those of international peers (Tran et al., 2019). Policymakers deeply acknowledge the disadvantages of Vietnamese Accounting Standards in reflecting accurately the value of assets and liabilities, compared to IFRS. From 2001, several local accounting standards have been issued based on the modification of standards composed by the International Accounting Standards Board. Basically, Vietnamese Accounting Standards try to adopt IFRS with some modifications to fit with Vietnamese situations.
IFRS Roadmap was recently approved in March 2020, marking an official and seismic milestone in the commitment to enhance the quality and transparency of financial reporting in Vietnam. In the Roadmap, the scope of IFRS adoption has been widened to cover all entities except for small, medium and micro-sized firms. Even though entities may have compulsory or voluntary options to apply IFRS, upcoming financial statements are prone to alter significantly compared to those composed under the current Vietnamese Accounting Standards. In essence, firms are allowed to have more managerial discretion to reflect the substances of financial transactions, rather than be required to follow current local accounting systems.
In Vietnam, cases of fraudulent financial statements have also surfaced and shattered investors' trust, e.g., KSS and JVC corporation (Trinh et al., 2020). Such cases contribute to the reduction in the potential to utilize the external financing channels. In addition to the reported cases, Nguyen and Nguyen (2017) empirically provide evidence of the existence of earnings management practices to inflate the bottom line. Asian Development Bank (ADB) (2014) points to the issue of insufficient regulations and monitoring of reporting mechanisms. Furthermore, Ta et al. (2018) highlight the remarkable divergence between Vietnamese Accounting Standards and International Financial Reporting Standards, i.e., the commonality between the two systems is only 50 per cent, while Vietnamese Accounting Standards suffer from several shortcomings. This could be the conducive environment for firms to abuse with the management of earnings.

Theoretical framework
As mentioned earlier, agency theory explains that the motivations to misrepresent firm performance using accruals-based earnings management emanate from the opportunistic behavior of firm insiders (Haw et al., 2004;Leuz et al., 2003;Shleifer & Vishny, 1997;Zingales, 1994). In the relationship between shareholders and managers, the formers are the agent hired to oversee corporate operations. In the relationship between insiders and fund providers, the formers are the agent. To facilitate the appropriation of the principal's interests, the agent could be opportunistic and use tactics to inflate earnings to obfuscate the principals about the true corporate performance (Leuz et al., 2003). Nevertheless, this practice is costly in the sense that if such deed is uncovered, the agent will have to take disciplinary actions, e.g., the management could be replaced.
According to De Franco et al. (2011) and Sohn (2016), as a firm's accounting system is more similar to that of its industry peers, the cost of collecting and processing its information becomes more affordable, thanks to the availability of information of comparable firms. Comparability allows stock analysts to make comparison about the performance of a firm through time and/or between firms as additional inputs. Furthermore, the managers should also have less room to engage in earnings management, and in fact, studies including J-B.  find that managers are less likely to conceal bad news when their financial statements are more comparable.
As a consequence, a firm's authentic performance is less elusive and can be estimated more accurately when its financial statements are more comparable. This is because the outside stakeholders (analysts, shareholders and debtholders) can compare accounting information of the focal firm and its peers with lower cost and more convenience (Gong et al., 2013;Kini et al., 2009;Engelberg et al., 2016;De Franco et al., 2011). When accounting information is more transparent and comparable, the cost to engage in earnings manipulation should be higher, because earnings management behavior is more likely to be discovered. Therefore, this should increase the risk of being caught, so managers are less prone to opt for earnings inflation in the light of financial statement comparability. In general, this argument is consistent with the view that information environment affects managers' incentives and ability to conceal bad news (Kim and Zhang, 2016;. Sohn (2016) studied the trade-off between real and accruals-based earnings management, the two methods of earnings manipulation usually documented in the literature (Zang, 2012). Sohn (2016) confirms that as a firm's financial statements become more comparable, managers tend to reduce accruals-based earnings management.

Accounting comparability and information quality
There are various studies investigating the implications that accounting comparability have for different types of financial statement users (Chen & Gong, 2019). De Franco et al. (2011) have identified solid proxies for accounting comparability, and use them to examine whether comparability supports analysts in deriving firm value. The authors find that accounting comparability allows investors to infer about the true performance of a firm more efficiently, with lower dispersion and better accuracy in forecasting figures. Through acquiring information from firms with comparable financial statements, investors are able to uncover firm-specific information, especially bad news, even when managers choose not to disclose it (J-B. . As a result, when firms have greater comparability, managers have lower incentives to hoard bad news because of the smaller expected benefits and larger expected costs. In the same vein, S. Kim et al. (2013) examine whether financial statement comparability reduces credit risk. Developing measures related to comparability in debt markets using Moody's adjustments to reported earnings, S. Kim et al. (2013) suggest that comparability enhances information environment and helps investors understand firms' credit risk and ascertain its pricing. J-B.  opine that high comparability disincentivizes managers to hoard bad news, making investors feel more secured about a firm's future crash risk. On the other hand, if firms have low levels of accounting comparability, managers are more prone to provide additional earnings forecasts in order to alleviate information asymmetry (Gong et al., 2013). Choi et al. (2019) further suggest that firm-specific information is more likely to be used by analysts when firms have high comparability. Francis et al. (2016) argue that more similar accounting frameworks between countries facilitate cross-border transactions and generate higher takeover premiums for target firms. Chen et al. (2018) find that acquirers are more likely to earn higher returns, gain better synergies and have higher operating performance when target firms have more comparable financial statements. These results support the view that similar accounting standards lower information gathering and processing costs, and help investors evaluate accruals more effectively.

Accounting comparability and earnings management
Even though there have been numerous studies on the impact of accounting comparability on information environment, the literature strand on its effect on earnings manipulation is scarce. A firm's true financial performance can be accurately evaluated because the firm's accounting comparability allows analysts to gather and process more valuable inputs from financial statements of comparable peers at lower costs (Kini et al., 2009;De Franco et al., 2011;Engelberg et al., 2016). Sohn (2016) states that as accounting comparability improves, accounting information becomes more transparent to outsiders and managers have less room for earnings manipulation, which should undermine the incentives to use accruals-based earnings management.
According to agency theory and opportunism hypothesis, through engaging in earnings management, insiders can obtain private benefits if they succeed in providing misleading figures for financial performance (Bergstresser & Philippon, 2006;Leuz et al., 2003;Louis & Robinson, 2005;Shleifer & Vishny, 1997). Such opportunistic behavior could result in negative repercussions on firm value, thus destroying the interest of creditors and minority shareholders. Nonetheless, if these behaviors are uncovered, disciplinary actions can be implemented against insiders, especially the managers. Managers, therefore, are likely to consider the pros and cons of earnings management. As accounting comparability is better, investors and creditors have more channels and sources to evaluate firm performance. Therefore, the agency cost related to the incentives to hoard bad news and conceal negative performance of the firm should be lower with better accounting comparability. More recently, Zhang et al. (2020) argue that accounting comparability, which markedly lowers the monitoring and governing costs of different types of stakeholders through its ability to handle the issues of information asymmetry as well as agency-related cost, enhances labour investment efficiency.
Consistently with the theoretical view above, J-B.  show that when firms have greater comparability, managers have lower incentives to hoard bad news because of the smaller expected benefits and larger expected costs. Furthermore, in addition to the benefit of enhancing financial reporting quality, Chen and Gong (2019) suggest that investors comprehend accruals more thoroughly as comparability improves. Consequently, on average better accounting comparability should translate to higher costs of accruals-based earnings management due to its higher likelihood of being detected and increases incentives to provide meaningful accruals (Sohn, 2016). All in all, it can be predicted that accruals-based earnings management should reduce as financial statements become more comparable.
Vietnam is an emerging market, and the institutional and governance quality is not high (Asian Development Bank (ADB), 2014). Furthermore, the differences between the local accounting standards and IFRS, which is expected to increase the comparability, are considerable. This should worsen the overall information environment about the firm, and this specific condition should make the impact of accounting comparability more meaningful, i.e., the ability to curb earnings management. Nonetheless, no studies have been conducted to compare and investigate whether the impact of comparability on earnings management remains the same for firms in developed and developing countries.
Based on the above arguments, the following hypothesis is established in the alternative form:

H1: Managers are less likely to engage in accruals-based earnings management when its accounting comparability is higher.
Nonetheless, it is challenging to completely eradicate managerial opportunism in terms of earnings management. As long as conducive environments are in sight, managers may be willing to manipulate earnings to enjoy private benefits in the hope that such manipulation is not prone to be detected. Financial constraints make it riskier to expose the true financial conditions to outsiders, as this may intensify the negative impressions that investors and lenders have. Therefore, firms with constraints are less likely to reduce income-increasing earnings management when their financial statements are more comparable (Moreira, 2006). Consistently, Korajczyk and Levy (2003) document that financially constrained firms tend to encounter higher agency and transaction costs when entering capital markets, compared to their unconstrained peers.
As a consequence, constrained firms are more likely to employ income-increasing earnings management to falsely impress outsiders regarding their performance. In line with this view, Kurt (2018) and Moreira (2006) find that financial constraints are positively related to earnings management. Financial statement comparability should make it more costly to conduct earnings management, but financially constrained firms may find it more compelling to perform earnings management. This warrants an empirical study to investigate whether constrained firms with comparable financial statements tend to employ earnings manipulation.
On the other hand, financially unconstrained firms are in the position to reap the benefit of accounting comparability. De Franco et al. (2011) and J-B.  argue that accounting comparability enables users of financial statements to understand firm value more thoroughly. Unconstrained firms are expected to have more good news and fewer incentives to engage in earnings management to hoard bad news. Therefore, for unconstrained firms, the agency cost related to the incentives to hoard bad news and conceal negative performance of the firm should be lower with better accounting comparability, compared to their constrained counterparts. This should be particularly relevant in Vietnam, with a rather poor information environment as discussed previously. This link has not been examined before. The argument put forward here is in accordance with opportunism hypothesis in Kurt (2018), (Bergstresser & Philippon, 2006;Louis & Sun, 2016).
In conclusion, based on the above arguments, it is expected that financially unconstrained firms are prone to reduce earnings management as their financial statements become more comparable.

Data collection
Data on financial statements are collected from Refinitiv Thomson Reuters for all non-financial listed firms in the two major exchanges in Vietnam from 2010 to 2019. Financial firms are not included in the sample because they follow different accounting frameworks and have different operations. Before 2010, there were few listed entities, so we do not include this period in the study. Extreme values are filtered out by removing the observations at the top and bottom one percentile, as well as unconventional values such as negative equity, in order to reduce the impact of outliers . . . . We also remove firms that only have less than 2 years of observations. The outcome from the data collection and cleaning process comprises 502 firms, totaling 3,401 firm-year observations. Table 1 provides the descriptive statistics for the variables in the model.

Variable construction
Accounting comparability, also referred to as financial statement comparability, is defined as the extent of similarity that similar economic transactions are recorded (Financial Accounting Standards Board (FASB), 2010;De Franco et al., 2011). More similar accounting systems should generate identical numbers when reflecting the same set of economic events. Greater comparability helps enhance the quality and quantity of information since more comparable financial statements serve as benchmarks for others, facilitating the acquisition and processing of information.
The current research employs the approach of De Franco et al. (2011) to empirically measure accounting comparability of firm i in year t. Previous studies that use accounting standards to gauge comparability which can result in several fundamental concerns such as weightings and approaches to take care of variation (De Franco et al., 2011). In principle, similar accounting systems should result in similar outputs, e.g., earnings, for the same economic event. The researchers account for economic events and outputs using stock returns and earnings, respectively. Earnings should be the mapping result of economic events to accounting output.
First, observations of 16 prior quarters are used to estimate the following equation, with earnings being the ratio of net income to market value of equity at the beginning of period. Return in the equation (1) is the stock return for the quarter.

Earnings it = α i + β i *return it + ε it (1)
Assuming firms i and j have the same performance in terms of return, accounting functions are used to predict earnings of firms i and j:

E(Earnings) iit = α i + β i *return it (2) E(Earnings) ijt = α j + β j *return it (3)
The expected earnings of firm i in year t under the accounting mapping function of firm i of firm i's return is E(Earnings) iit , while E(Earnings) ijt is the earnings that is predicted for firm j under the accounting mapping function of firm j to record return of firm i in year t. Firm i's return is used in both equations (2 & 3) to ensure that the same event is accounted for.
First, the average of absolute differences between the two expected earnings caused by firm i and firm j's accounting system is calculated. Then, negative value of the number is taken, so that greater values of CompAcct ijt should indicate higher accounting comparability between the two firms:

jE(Earnings iit )-E(Earnings ijt )| (4)
To construct comparability measure for a firm in each year, we aggregate j values of CompAcct ijt for firm i in year t. Comp4 it is basically the mean of the highest CompAcct values for 4 firms j in year t. CompInd is the median value of CompAcct for all firms that operate in the same industry with firm i in year t. As per construction, higher values of CompAcct and CompInd indicate that the firm has similar accounting mapping functions.
With regard to the measure of opportunistic accruals-based earnings management, we resort to the concept of discretionary accruals, or the abnormal part of total accruals. Total accruals are the result from the subtraction of cash flow from operations from earnings before extraordinary items. We employ the approach Kothari et al. (2005), which modifies Jones (1991) to decompose total accruals into normal part (or the expected part of accruals) and the abnormal part. AEM_Jones indicates the original method of measuring accruals-based earnings management, while AEM_Kothari follows Kothari et al. (2005) which further consider return on assets to derive performance-adjusted discretionary accruals. We do not use the absolute value of AEM because we are more interested in income-increasing techniques, which are more relevant in the case where managers are motivated to falsely impress the outsiders.
Our baseline empirical model is as follows:

AEM it = β 0 + β 1 CompAcct it + β 2 Size it + β 3 ROA it + β 4 Mgrown it + β 5 Big4 it + β 6 LEV it + β 7 CFOA it + β 8 BM it + ε it
Where: AEM is the dependent variable, with two proxies of AEM_Jones and AEM_Kothari. CompAcct is the explanatory variable of interest, with three proxies of Comp4, CompInd and CompAcct, as discussed above. Other explanatory variables include: firm size (Size), book-tomarket ratio (BM), the absolute value of cash flow from operations (CFOA), ROA, leverage (Frankel et al., 2002;Haw et al., 2004;Sohn, 2016). Furthermore, Mgrown is included to control for the effect of managerial ownership, because the alignment of interest between managers and shareholders/creditors may affect the incentive to engage in earnings management. For variable construction, please see the appendix. Table 1 presents the descriptive statistics of variables in the model. The average value of CompAcct has the value of −2.651, and there is a wide variation in this variable, indicating that the sampled firms have various levels of accounting comparability. For CompInd (measure of comparability for firms in the same industry), the standard deviation (mean value) is much lower (higher) than that of CompAcct, suggesting that firms in the same industry have better accounting comparability with each other, compared to the general firms. As expected, Comp4 is the measure of comparability from the four best comparable firms has even lower standard deviation, and higher mean value (−0.009). These statistics suggest that the measures constructed in the paper are properly constructed, in terms of the hierarchy of comparability level.

Empirical results and discussion
On average, ROA is about 7.2 per cent, while LEV is 0.235, indicating that about a quarter of firm assets are financed by debt. CFOA has an average value of 6.1 per cent, or cash flow from operations is rather low compared to the amount of the firm assets. BM has the value of 1.538, or the market equity is rather low as opposed to book value of equity. BIG4 has the value of 0.243 or about a quarter of the sampled firms are audited by auditors of renowned auditing firms in Vietnam, namely KPMG, PriceWaterhouseCooper, Ernst & Young, and Deloitte. CEO ownership is low on average, which could make it unable to have significant impact on corporate earnings management. Table 2 presents the results from the regression of comparability on accruals-based earnings management, with AEM_Jones as the dependent variable. It is clear that all the three proxies of comparability have negative and significant coefficients, implying that financial statement comparability curbs income-increasing earnings management. This result is consistent with the findings from previous studies claiming that comparability helps investors to obtain and process information at lower costs and more efficiently, leading to better exploitation of available information sources (Francis et al., 2015;Chen et al., 2018;Choi et al., 2019;, and is in line with our first hypothesis H1 on the negative association between accounting comparability and earnings management. Sohn (2016) states that as accounting comparability improves, accounting information tends to be more transparent to outsiders and managers have less room for earnings manipulation, which should lessen the incentives to engage in this behavior. Interestingly, we also find that firms with more established auditing firms are less prone to conduct earnings management, as expected. Furthermore, high R-squared values indicate a good fit of the model in explaining the variation of accounting comparability of firms in Vietnam.
In Vietnam, where information asymmetry is still at high level, financial statement comparability is conducive to tackling the situation, allowing investors the access to considerable source of information by referring to firms that have similar financial statements. A firm's financial performance can be accurately evaluated when the firm's accounting comparability allows the analysts to gather and process more valuable inputs from financial statements of comparable peers at lower costs (Kini et al., 2009;De Franco et al., 2011;Engelberg et al., 2016). Source: author's calculation from dataset. *, **, *** denote 1%, 5% and 10% respectively.  Table 3 provides the first robustness test with the dependent variable being AEM_Kothari. This variable is constructed using the approach of Jones (1991) with the modification of the inclusion of return on assets, detailed in Kothari et al. (2005). The results are basically the same as in Table 2. The proxies for accounting comparability have significantly negative coefficients, indicating that as financial statements of a firm are more comparable, its manager is less likely to engage in accruals-based earnings management. Furthermore, high R-squared values are presented. Therefore, this result provides more robust evidence in support of Hypothesis H1, which emphasizes the merits of having comparable financial statements.
The analysis thus far is based on the assumption that accounting comparability is not an endogenous variable. Following the reasoning of Sohn (2016), both earnings management and accounting comparability can be endogenous. For instance, managers can choose accounting methods to facilitate earnings management activities, thus affecting accounting comparability. Such accounting methods should be uncommon or not comparable to allow the manipulation of earnings, in a hope that such earnings cooking is not easily uncovered. Another reason that can contribute to the mutual relationship between accounting comparability and earnings management is that managers may prefer accounting systems or methods that result in highly comparable financial statements to enjoy the benefits of transparency and alleviation of information asymmetry.
To control for the potential endogeneity, we conduct the analysis with the main explanatory variable being one-period lagged values of proxies of accounting comparability. Table 4 presents Source: author's calculation from dataset. *, **, *** denote 1%, 5% and 10% respectively. the regression results which support consistent findings about the favorable impact of comparability on earnings management: more comparable financial statements limit opportunities for managers to reduce earnings manipulation. Therefore, we obtain additional evidence in support of Hypothesis H1.
We aim to seek further robustness to our findings by resorting to 3SLS regression technique, through which we simultaneously estimate models with both earnings management and accounting comparability taking turn in being the independent variable in the model determining the other variable. Table 5 provides the results of 3SLS regression, and it is clear that Comp4 is negatively related to earnings manipulation, which is again consistent with hypothesis H1, even after controlling for the potential endogeneity. The result is similar (not reported here) when we replace AEM_Jones by AEM_Kothari. Nonetheless, we are unable to find robust results when replacing Comp4 by CompAcct (the overall accounting comparability measure) and CompInd (accounting comparability measure constructed using accounting figures from the same industry peers). Sohn (2016) only uses Comp4 in his study. Our results might point out that firms may be particularly interested in comparing their financial statements with those that are most comparable, because those firms can be more relevant sources to investors to gather and process information for evaluation purposes, rather than information from firms in the same industry or firms in general. All in all, our results that accounting comparability reduces earnings manipulation are robust to different approaches (with and without control for potential endogeneity) and various measures of both accounting comparability and earnings manipulation. Source: author's calculation from dataset. *, **, *** denote 1%, 5% and 10% respectively. For the evaluation of Hypothesis H2, we estimate the baseline model for two subsamples based on measures of financial constraint. Table 6 provides estimation results when the sample is subdivided into small and large firms based on the median value of Size variable. Small firms are more financially constrained, and financial constraints are expected to make it challenging for firms to access financial markets (Korajczyk & Levy, 2003). Furthermore, more constrained firms may have higher incentives to opt for earnings manipulation to provide misleading figures for financial performance, in an attempt to have better impression from outsiders.
For small firms, accounting comparability has no significant impact on earnings management. Previous papers frequently find that accounting comparability enables users of financial statements to understand firm value more thoroughly (De Franco et al., 2011;. Constrained firms may find it more compulsory to engage in earnings management (Kurt, 2018), but it can be uncovered easily, especially with more comparable financial statements. Therefore, constrained firms with highly comparable accounting system might resist from engaging in earnings management, because this can be discovered and punished accordingly (Kurt, 2018). The results in Table 6 suggest that the two effects cancel out and comparability shows no effect on earnings management. The result here is consistent with opportunism hypothesis in Kurt (2018), Bergstresser and Philippon (2006) and Louis and Sun (2016). Furthermore, it is consistent with the view of debt incentive from Moreira (2006).
On the other hand, financially unconstrained firms are in the position to reap the benefit of accounting comparability. Unconstrained firms are expected to have more good news and fewer incentives to engage in earnings management to hoard bad news. Therefore, unconstrained firms are more likely to reduce earnings manipulation when they have comparable accounting methods and systems. In general, these results are in line with Hypothesis H2. Table 7 provides regression results for two samples based on ROA. Firms with high levels of ROA are expected to have higher possibility to gain access to external financing, reducing the needs to provide misleading financial statements. Table 7 shows that comparability proxies have significantly negative it is clear that firms that have higher ROA tend to reduce earnings management when they have better accounting comparability. On the other hand, firms with low ROA have more constraints, and their accounting comparability tends to exert no significant effect on earnings management. These results are in support of Hypothesis H2.

Summary and conclusion
Financial reports are meant to help outsiders confer or uncover the true performance of firms, and comparability makes the information worthwhile for this purpose (Financial Accounting Standards Board (FASB), 2010). Previous studies have always praised that comparability has a universal benefit of improving information environment, allowing investors to estimate the value of firms more effectively.
However, literature is still relatively silent on the relationship between comparability and accruals-based earnings manipulation, except for the study of Sohn (2016). This author investigates the trade-off between accruals-based earnings management and real earnings manipulation as a function of accounting comparability. The setting of the study is in the US, which is an advanced country, and there have been no other studies for emerging markets. Furthermore, the study only uses one measure of accounting comparability, which may not help with the inferences of the importance of comparability with particular stakeholders. Finally, prior literature does not Source: author's calculation from dataset. *, **, *** denote 1%, 5% and 10% respectively. distinguish the situations when accounting comparability may not manifest its positive effect, and tends to focus on consolidating evidence on its benefits.
We investigate the link between comparability and accruals-based earnings management using a sample of 509 Vietnamese listed firms from 2010 to 2019. Vietnam's market houses an appropriate setting because the country is still aiming towards the convergence to the standards of International Financial Reporting Standards (IFRS) and applying local standards until the mandatory application in 2025 for all listed firms, which heavily thwarts the comparability of financial statements across firms. Furthermore, as a fledgling market with developing institutions, information asymmetry is still playing a havoc on the ability to access external financing. This is why accounting comparability should play a particularly important role in Vietnam. The present study employs three measures of comparability and two measures of accruals-based earnings manipulation, and several empirical tactics to ascertain our findings.
We contribute to the research strand on the link between financial reporting quality and earnings management. We investigate this important issue in the context of Vietnam, a country that is late in the adoption of IFRS. There have been no studies on the link between comparability and earnings management in a developing country. In a developing country like Vietnam, since IFRS has not been mandatorily applied, the comparability level should be low and this could affect the Source: author's calculation from dataset. *, **, *** denote 1%, 5% and 10% respectively. ability of this factor in alleviating accruals-based earnings management. This is why it is important to test the effect of comparability in this setting.
Our research provides robust evidence confirming comparability reduces the likelihood of firms engaging in earnings management. Interestingly, there is (weak) evidence suggesting that firms tend to reduce earnings management, especially when its comparability with the four highest-comparable firms increases. Meanwhile, this link can be weaker for firms in the same industry or general firms. This finding has not been documented in previous studies.
Second, we identify a factor that can moderate the impact of comparability on earnings management. We find that more constrained firms do not reduce earnings management, while less constrained peers are more willing to reduce earnings manipulation, as comparability is enhanced. This study has consolidated in general the favorable impact of restraining earnings management of comparable financial statements, and suggest that financial constraints modify the above relationship. Therefore, if IFRS improves comparability while Vietnam is still plagued with poor information environment, the commitment to adopt IFRS should be further encouraged. Another implication from this research is that firms might be more interested in making their financial statements comparable with some peers, not average firms in the same industry. Furthermore, the constraint-related finding implies that investors and regulators are to be more vigilant with constrained firms because comparability seems unable to curb the manipulation of earnings.
Future studies can introduce another factor/condition which may alter the relation, which can bring about new implications to relevant stakeholders. Furthermore, extending the sample to firms in both developed and developing markets might ascertain whether accounting comparability has more or less meaning in developing countries.