How Far Can the Books Be Trusted — Where Discretion Ends, and Whether Reading Them Puts You Ahead

Numbers at the end of the line involve judgment

At the end of the year, items brought in three years ago remain in a company warehouse. Someone has to decide how to record the value of those goods right now. If a client has not paid their credit for half a year, someone also has to decide how much of that amount to write off as uncollectible. Values determined like this gather together to become that year's profit.

In "Stocks — What owning a piece means", the shareholder stood at the back of the line and received what was left. The number recording how much that remainder is is the profit on the income statement, and those books were covered in "Financial Statements — What do the three books measuring a company each ask". This article asks in three directions how much those numbers can be trusted. Where is the scope for judgment to enter, and how large is it? How much do outside mechanisms like rules and audits narrow that scope? And can reading the books put you ahead of others?

This article is written regarding listed companies in Korea and Korea's accounting and auditing systems. Among the studies measuring the scope of judgment, many looked at European or American companies or did not state the country of their sample in the abstract, and such studies are clarified on the spot. And this article does not talk about the value of any company, nor does it discuss what numbers at what amounts make a good company.

Where the Room Is and How It Is Measured

Accountbooks record profit on the day a sale is made, and the cash flow statement separately records the day money actually comes in. The accrual we saw in the 「Financial Statements」 section is the remainder left when you subtract the cash that actually moved from the profit written in that year's accountbook. Decisions to write down the value of goods in a warehouse and decisions to write off unpaid credit sales are both included within this accrual.

Managers moving the size and timing of the profit written in the accountbooks using their own discretion—this is called earnings management. The judgment to set values itself is a necessary task whenever accountbooks are written, and earnings management is the use of that judgment to skew profits toward a target. Outsiders cannot directly see how far the manager used this room, so they measure it with estimates.

Accruals have a portion that arises naturally as a business runs. This is because when sales increase, unpaid credit sales that have not yet been collected also increase alongside them. The portion left after subtracting the amount that can be explained by the company's situation from those accruals—this is called discretionary accruals. The naturally occurring portion is separated out using a model, and what remains is viewed as the spot where judgment moved.

Room exists outside of just the recording method. Rather than the method of writing in accountbooks, twisting actual operating decisions themselves to fit that year's profit—this is called real earnings management. This is done by methods such as slashing prices at the end of the year to push out goods, or postponing money that needs to be spent to the following year to cut that year's expenses. The numbers change without breaking a single accounting rule.

There is also a name for the target that researchers try to measure with such estimates. A term researchers use regarding how faithfully the numbers in accountbooks convey a company's actual situation—this is called earnings quality. Because there is no value that can be measured directly, studies instead measure it using proxies, such as how smoothly earnings were smoothed out or how promptly losses were recorded. The phrases 「quality was high or low」 appearing in this article are all values measured this way.

How estimates are measured changes the results. When gathering 35 studies dealing with the effect of boards of directors and ownership structures on earnings management, the method of measuring discretionary accruals and country-specific governance systems moderated the relationship between earnings management and some governance variables (DOI: 10.1111/j.1467-8683.2009.00753.x · meta-analysis of 35 studies, multiple countries). The fact that research results diverged over whether the chief executive officer also serves as the chairman of the board, or whether the audit committee is independent, was explained by sampling error. The authors conclude that legal and institutional environments must be explicitly considered when looking at the effect of governance on discretionary accruals.

The room for judgment is in two places—inside accruals and inside operating decisions—and outsiders measure both solely through estimates. Therefore, the phrases 「increased or decreased」 that appear later may vary depending on which ruler was used to measure them.

When the Scope Broadens

In the European Union, unlisted companies use the same accounting standards as listed companies. This is because accounting regulations are based on the legal form of a company. When comparing the two groups under the same rules, the level of earnings management was higher in unlisted companies, and in countries with strong legal systems, earnings management was lower in both unlisted and listed companies (DOI: 10.2308/accr.2006.81.5.983 · European companies). The authors interpret these results through the framework that capital market pressure and institutions create incentives to record earnings as one pleases, and they note that the forces of legal systems and capital markets often appear to support each other.

Whether the scope broadens when a shock hits yielded opposite results between two studies. During the pandemic period of 2020-2021, the quality of accounting information measured by discretionary accruals was lower than in the preceding period of 2015-2019 (DOI: 10.35850/KJTA.24.3.01 · domestic academic journal, abstract does not disclose the sample country). Companies that did not disclose the uncertainty of COVID-19 had lower quality than companies that did disclose it, and this phenomenon was prominent in companies with decreased sales. During the same period, the probability of recognizing asset impairments also increased, and that probability was higher in companies that disclosed uncertainty.

Among European listed companies, earnings management significantly decreased during the 2008-2009 financial crisis, and the same trend was confirmed in most of the 16 countries examined (DOI: 10.1016/j.intacc.2014.10.004 · 16 European countries). The level of earnings management was linked to the economic growth rate, and evidence also emerged pointing to the side that country characteristics and market forces influenced the tendency to smoothly adjust earnings but not the quality of accruals (same study).

The two results differ in countries, shocks, and measurers, and their directions are opposite. How much the scope is used varies depending on incentives and circumstances, and which direction it moves in front of a shock has split results.

How Much Did Common Rules Narrow the Range?

European Union listed companies all began writing their books with the same accounting rules in 2005. Accounting standards that multiple countries agree to use in common are called international accounting standards. Korean listed companies have been writing their books using these standards since 2011 (DOI: 10.21737/RAPS.2025.11.30.4.141).

Studies were divided on whether books improved after using the same rules. Companies applying these standards across 21 countries showed less earnings management, recognized losses more timely, and had accounting figures align more closely with stock prices than similar companies using domestic standards (DOI: 10.1111/j.1475-679x.2008.00287.x · 21 countries). However, the authors note they could not be certain whether the difference was thanks to the reporting system changing or to companies' incentives and economic environments changing.

The results were the opposite when comparing companies in 20 countries that mandatorily adopted these standards in 2005 with non-adopting companies similar in law enforcement intensity, industry, and size. In adopting companies, the degree of smoothing earnings evenly and reporting aggressive accruals significantly increased, while the degree of recognizing losses timely significantly dropped. There was no difference in meeting or beating target earnings (DOI: 10.1111/j.1911-3846.2012.01193.x · 20 countries, early adoption effects). These results mostly held in countries with strong law enforcement, and the authors read that those countries' enforcement devices did not offset the early effects of the greater flexibility provided by the new standards.

Both directions sometimes appeared together within a single paper. Comparing EU 15-country listed companies around 2005, many among the five metrics improved. Earnings management toward targets decreased, the magnitude of discretionary accruals shrank, and accrual quality rose. Over the same period, the degree of smoothing earnings evenly actually increased, and the degree of recognizing large losses timely dropped (DOI: 10.1111/j.1467-646x.2010.01041.x · 15 EU countries). The authors of this paper count as their contribution that the observed improvements can be distinguished from changes in managers' incentives, capital market institutions, and the business environment and attributed to the standards themselves.

Research looking at Germany reaches the opposite conclusion. In Germany, companies could freely choose to use these standards even before the mandate, and the improvement of books was limited to companies that had incentives to adopt them voluntarily. Companies that held out against adoption were closer to banks and internal shareholders (DOI: 10.1080/09638180.2015.1009144 · Single country Germany). The authors conclude that reporting incentives overpower accounting standards, and write that it is not justified to infer that the standards themselves raise accounting quality based on improvements from the voluntary adoption period.

Merely making companies use the same rules did not nail down the range. Studies attributing improvements to the standards and studies attributing them to incentives are at odds, and even within studies where improvements appeared, deteriorated metrics appeared alongside them.

The Checkpoint Outside

Listed companies in Korea release their books after having them checked by outside accountants. An accountant or accounting firm that checks from the outside whether a company's submitted books are written according to standards is called an auditor. If the auditor finds errors in the old books, those books are corrected and reissued. The act of correcting and reissuing financial statements that have already been released to the outside is called financial statement restatement.

Let us trace in order the checkpoints that one set of books goes through before coming out to the public this year. Each checkpoint has mechanisms to hold down volatility, and Korea has changed many of them into systems.

The first checkpoint is inside the company. When the External Audit Act was amended in October 2017, listed companies became required to have even the management system operated inside the company to ensure books are written properly—namely, the internal accounting control system—audited by outside accountants. This applies starting in 2019 for companies with assets exceeding 2 trillion KRW, 2020 for 500 billion KRW or more, 2022 for 100 billion KRW or more, and 2023 for smaller companies (KCI ART002871932).

After this audit was introduced, real earnings management decreased significantly, and discretionary accruals did not decrease significantly. Looking by period, in the group with assets of 2 trillion KRW or more that started in 2019, real earnings management did not decrease, while it decreased significantly in the 500 billion to 2 trillion KRW group that started in 2020 (KCI ART002871932). The authors infer that discretionary accruals had likely already decreased due to the preceding review system, and conclude that "it can be seen that the quality of accounting information has improved to a certain extent with the introduction of the internal accounting control system audit." The same authors note that there are limitations to the interpretation because the sample covers 3 years post-introduction and focuses on large companies with assets of 500 billion KRW or more.

People inside the company also hold down volatility. When audit committee members possessed both accounting expertise and industry expertise, financial reporting quality—measured by restatements and discretionary accruals—was better than when they possessed only accounting expertise, and auditor oversight—measured by audit fees and non-audit fees—was also better (DOI: 10.2308/accr-50585 · abstract does not specify the country of the sample). Members with oversight experience also performed better in certain cases when combined with industry expertise. The authors write that industry expertise, when paired with accounting expertise, can improve audit committee oversight.

Core executives directly below the chief executive officer also provide checks and balances. The longer the years remaining until their retirement and the larger their compensation relative to the chief executive officer, the less real earnings management occurred (DOI: 10.2308/accr-51275 · US companies). This effect was stronger when the business was more complex, when the chief executive officer's power was weaker, and after US accounting regulations were significantly strengthened, and weaker when the gains obtained in capital markets from meeting or beating earnings benchmarks were larger. The authors of the preceding meta-analysis of 35 studies also write that board independence and size, and audit committee independence, can constrain earnings management and enhance investor trust (DOI: 10.1111/j.1467-8683.2009.00753.x).

The second checkpoint is choosing the auditor. Korea established a periodic auditor designation system in the amended External Audit Act, whereby authorities designate auditors at fixed intervals. For KOSPI and KOSDAQ listed companies from 2015 to 2022, all three variables—before versus after the system's introduction, whether subject to designation, and the first year of designation—significantly reduced restatements. Over the same period, the frequency of modifying financial items in accounting reports increased (DOI: 10.15843/kpapr.38.1.2024.2.119). The authors interpret the increased modifications as a positive effect that corrects discrepancies with management and enhances the reliability of financial statements, and conclude that these results are empirical evidence that "the periodic auditor designation system can enhance the transparency of accounting reports." However, they disclose that the implementation period was short, preventing the analysis of long-term effects.

One study's perspective is that if competition among auditors in the auditor-selection market is excessive, the audited company may end up in a higher position than the auditor. Across 10,372 firm-years of December-closing corporations in the KOSPI and KOSDAQ markets from 2010 to 2019, higher competition among auditors was associated with lower accounting information quality, an effect that large accounting firms mitigated (KCI ART003033038). The authors interpret this as a result of large accounting firms maintaining appropriate audit quality to reduce litigation risk.

The third checkpoint is when auditors change. Previous studies have reported that revisions to past books increase after an auditor changes. This is because a new auditor may discover errors or treatments with differing views that the previous auditor missed. Across 10,267 firm-years of listed companies from 2014 to 2019, all types of changes showed a significant positive relationship with restatements, with the highest restatement rate occurring when transitioning from a non-large accounting firm to a large accounting firm. The likelihood of restatement was also higher when the change resulted in increased industry expertise of the auditor (DOI: 10.24056/KAJ.2022.03.005). The authors write that when selecting auditors through the designation system, both the type of change and the direction of change in industry expertise must be considered. While the designation study showed decreased restatements and this study showed increased restatements after changes, one looked at the before-and-after of the system while the other looked at the event of an auditor change, and the sample periods also differ.

The final checkpoint is when the books are released to the outside and the market receives those numbers. There are results measuring how the registered auditor system for listed corporations, aimed at improving audit quality, appears at this checkpoint. Comparing the 3-year post-introduction period (2020–2022), excluding the transitional year 2019, with the preceding 3-year period (2016–2018), the degree to which accounting earnings are reflected in firm valuation increased significantly, and that change varied depending on auditor type and firm size (KCI ART003335297). The authors write that this result is consistent with and suggests the possibility of increasing the reliability and usefulness of financial information.

A 2014 literature review considered the measures of how well an audit was performed to be crude. There are many proxies for audit quality with few guidelines on which to choose, and audit quality and financial reporting quality are intertwined, so models separating the two also require better conceptual guidance (DOI: 10.1016/j.jacceco.2014.09.002 · literature review, country not specified). The authors therefore provide a framework that systematically weighs the strengths and weaknesses of each proxy.

Among studies examining Korea's auditing system, the authors of two papers conclude that the internal accounting control system audit improved the quality of accounting information "to a certain extent," and that the periodic auditor designation system "can enhance" transparency. The study on the registered auditor system stops at suggesting the possibility in the same direction. The authors themselves noted limitations next to these conclusions, such as results varying by measurement metric and the short timeframe.

Can Reading Get You Ahead?

There are over 18,000 signals that can be extracted from financial statements. Even after gathering all those signals and filtering out the shares that matched purely by chance, many signals significantly predicted which stocks would yield higher returns among multiple stocks (DOI: 10.1093/rfs/hhx001 · the abstract does not disclose the country of the sample). That predictive power was more prominent after periods when market sentiment was hot, and in stocks where it is hard to correct mispriced values by buying and selling them. The authors consider that this anomaly cannot be attributed to chance and is better explained by pricing errors.

The existence of a signal and the ability to get ahead using that signal are different questions. The places where the predictive power was more prominent were precisely the stocks where it is difficult to correct values through buying and selling.

There are also Korean results pointing in the opposite direction. This study presupposes the anomaly that above-average returns are observed in companies whose stock prices are low compared to the values written in their ledger. The higher the comparability in accounting—meaning the degree to which the same tasks are entered into the ledger in a manner similar to other companies—the more this excess return decreased on average (DOI: 10.30753/emr.2023.41.2.004 · Korea Composite Stock Price Index listed companies from 2011 to 2020). The authors read this as meaning that the easier it is to compare ledgers with one another, the more accurately stock values are priced and the smaller the pricing errors become. However, that mitigating effect was not significant in companies on the side where stock prices are low relative to their ledgers—companies that the authors wrote are placed in the relationship between risk and reward inherent in value investing. This passage by the authors leaves open the reading that the returns of those companies may mingle with the compensation for bearing risk.

There are also Korean results looking at the relationship between ledger signals and the subsequent risk of stock price crashes. The higher the composite score made through financial statement analysis, the lower the subsequent risk of a stock price crash, and that relationship was established through changes in future earnings (DOI: 10.35214/rfis.11.1.202202.001 · Korea Exchange listed companies from 2005 to 2015). The authors write that this result shows the importance of financial statement analysis and the usefulness of accounting information in valuation, which serves as the foundation of investment strategies. This is an average association pooling 11 years of data together, and does not speak to the future of any single company.

The quality of the ledger is already being read by those pricing it. Companies that corrected their business reports showed a significantly lower degree to which the following year's net income was reflected in the stock price, and this relationship appeared only in companies not analyzed by financial analysts (DOI: 10.21737/RAPS.2025.11.30.4.141 · 13,257 Korean listed company-years from 2011 to 2019). The authors interpret that investors use correction disclosures for valuation and evaluate them more negatively when there are no analysts to fill in the lacking information, and they write that the preparation and use of faithful business reports play a practical role in the capital market.

The same goes for those lending money. Borrowers with low accounting quality chose bank loans over public corporate bonds. In bank loans, maturity and collateral conditions became stricter alongside interest rates, whereas in public corporate bonds, only the interest rate conditions became stricter, and the impact on public corporate bond interest rates was 2.5 times that of bank loans (DOI: 10.2308/accr.2008.83.1.1 · the abstract does not disclose the country of the sample). The authors read this as matching the result that banks acquire and process information better.

There is a scope of judgment in the numbers of a ledger, and how much of that scope is used varies depending on incentives and circumstances. The authors of the study looking at Korea's auditing system concluded, along with the limitation of a short period, that the system improved the quality of the ledger to a certain extent and could increase transparency. Signals extracted from the ledger had predictive power, but that power was greater in places that are hard to correct. The easier it was to compare ledgers between companies, the more the excess returns of companies with low stock prices relative to their ledgers decreased on average, and while the authors read this as a decrease in pricing errors, that effect was not significant precisely in the companies on the side where stock prices are low relative to their ledgers. Reading a ledger means receiving the numbers while knowing where the scope lies. The predictive power of the signal appeared in research, and whether one can get ahead of others using that signal remains open, coupled with the condition that its power was prominent in stocks that are hard to correct.

Where to Look Next

What accruals are, and what each of the three ledgers asks, are covered in "Financial Statements — What Do the Three Ledgers That Measure a Company Each Ask".

Whether stock prices already contain everything worth knowing is covered in "Can You Beat the Market", and what happens in the market when what is known inside a company and what is known outside differ is covered in "When the Market Fails".

Why profit remains at the very bottom line can be revisited by turning to "Stocks — What Does It Mean to Own a Piece".

Source Papers

• "Corporate Governance and Earnings Management: A Meta-Analysis" (2009) — DOI: 10.1111/j.1467-8683.2009.00753.x — A meta-analysis of 35 studies covering the effects of boards and ownership structures on earnings management. Variations in results regarding CEO duality and audit committee independence stemmed from sampling error, while discretionary accrual measurement methods and governance systems moderated the relationship between earnings management and some governance variables. The authors emphasize the need to explicitly consider legal and institutional environments, and write that board independence, board size, and audit committee independence can constrain earnings management and enhance investor confidence.
• "The Importance of Reporting Incentives: Earnings Management in European Private and Public Firms" (2006) — DOI: 10.2308/accr.2006.81.5.983 — In the European Union, accounting regulations are based on legal form, so private firms use the same standards as public firms. Earnings management was higher in private firms, stronger legal systems were associated with less earnings management in both, and the power of legal systems and capital markets often reinforced each other.
• "The COVID-19 Pandemic and the Quality of Accounting Information" (2023) — DOI: 10.35850/KJTA.24.3.01 — The quality of accounting information during the pandemic period (2020–2021) was lower than before (2015–2019), lower in firms that did not disclose uncertainty compared to those that did, and prominent in firms with declining sales. The probability of recognizing impairment losses increased during the pandemic period and was higher in firms that disclosed uncertainty (domestic academic journal, abstract does not specify the sample's country).
• "Financial Crisis And Earnings Management: The European Evidence" (2014) — DOI: 10.1016/j.intacc.2014.10.004 — During the 2008–2009 financial crisis, earnings management by European listed firms decreased significantly, a trend shared by most of the 16 countries. It reports a link between the level of earnings management and economic growth rates, and provides evidence that country characteristics and market forces affect income-smoothing tendencies but not accrual quality.
• "Restatement of Business Reports and Value Relevance of Accounting Information" (2025) — DOI: 10.21737/RAPS.2025.11.30.4.141 — 13,257 firm-years of listed companies from 2011 to 2019 following the adoption of K-IFRS. Corporations that restated business reports in the previous period showed a significant negative association with the incremental value relevance of next-period net income, and this appeared only in samples not analyzed by financial analysts. The authors write that investors use restatement disclosures for valuation.
• "International Accounting Standards and Accounting Quality" (2008) — DOI: 10.1111/j.1475-679x.2008.00287.x — IAS-applying firms across 21 countries exhibited less earnings management, recognized losses more in a timely manner, and had higher value relevance than non-US domestic standard-applying firms. The authors write that they cannot be sure whether the results are due to changes in reporting systems or changes in incentives and economic environments.
• "Does Mandatory Adoption of IFRS Improve Accounting Quality? Preliminary Evidence" (2012) — DOI: 10.1111/j.1911-3846.2012.01193.x — Firms in 20 countries that adopted IFRS in 2005 experienced a significant increase in income smoothing and aggressive accrual reporting and a significant decrease in loss recognition timeliness compared to matched non-adopting country firms, but there was no difference in target-earnings achievement. The results held primarily in countries with strong enforcement, and the authors read this as the enforcement machinery of those countries failing to offset the initial effects of IFRS's greater flexibility.
• "The Role of International Financial Reporting Standards in Accounting Quality: Evidence from the European Union" (2010) — DOI: 10.1111/j.1467-646x.2010.01041.x — Comparing before and after the 2005 full adoption of IFRS by listed firms in 15 EU countries using five proxies. While many indicators improved (earnings management toward targets ↓, absolute magnitude of discretionary accruals ↓, quality of accruals ↑), income smoothing increased and large loss recognition became less timely. The authors cite as a contribution that the improvements are attributable to IFRS rather than changes in incentives, institutions, and environments.
• "Incentives or Standards: What Determines Accounting Quality Changes around IFRS Adoption?" (2015) — DOI: 10.1080/09638180.2015.1009144 — Single-country design of Germany. Accounting quality improvements were limited to voluntary adopters who had incentives to adopt, and firms that resisted adoption were closer to banks and internal shareholders. The authors conclude that reporting incentives override accounting standards, and it is not justified to infer from evidence in the voluntary adoption period that IFRS itself improves accounting quality.
• "The Effect of Internal Accounting Control System Audits on the Quality of Accounting Information" (2022) — KCI ART002871932 — Due to the revision of the External Audit Act in October 2017, audits of internal accounting control systems for listed companies were introduced sequentially by asset size. With the introduction of audits, real earnings management decreased significantly while discretionary accruals did not decrease significantly, and the effects varied by period. The authors' conclusion is that 「it can be seen that the introduction of internal accounting control system audits has improved the quality of accounting information to some extent」, and they note that there are limitations in interpretation as the sample is focused on the post-introduction 3-year period and large-scale companies.
• "The Effect of Audit Committee Industry Expertise on Monitoring the Financial Reporting Process" (2013) — DOI: 10.2308/accr-50585 — Viewed through two measures of financial reporting quality (restatements and discretionary accruals) and two measures of auditor monitoring (audit fees and non-audit fees), audit committee members with both accounting and industry expertise outperformed members with only accounting expertise, and in some cases oversight experts also performed better when they combined industry expertise (abstract does not specify the sample's country).
• "Internal Governance and Real Earnings Management" (2015) — DOI: 10.2308/accr-51275 — US firms. Real earnings management decreased as key subordinate executives' horizons (years remaining until retirement) and influence (compensation relative to CEO) were larger. The effect was stronger in complex businesses, weak CEOs, and post-SOX, and weaker when the capital market gains from meeting earnings benchmarks were large.
• "A Study of the Effects of Periodic Auditor Designation System on the Restatements of the Financial Statements" (2024) — DOI: 10.15843/kpapr.38.1.2024.2.119 — KOSPI and KOSDAQ from 2015 to 2022. The pre- and post-introduction of the periodic auditor designation system, whether subject to designation, and first-year designation variables all significantly reduced financial statement restatements, and financial item revisions in accounting reports increased. The authors interpret the increase in revisions as a positive effect that enhances reliability, and conclude that it is empirical evidence that the designation system can increase the transparency of accounting reports. Long-term effects could not be analyzed due to the short implementation period.
• "Audit Market Competition and Accounting Information Quality" (2023) — KCI ART003033038 — 10,372 firm-years of KOSPI and KOSDAQ December-closing corporations from 2010 to 2019. High competition among auditors lowered the quality of accounting information, and large audit firms (Big4) mitigated that effect. The authors interpret this as a result of large audit firms maintaining appropriate audit quality to reduce litigation risk.
• "The Effect of Auditor Change Type on Financial Statement Restatements" (2022) — DOI: 10.24056/KAJ.2022.03.005 — 10,267 firm-years of KOSPI and KOSDAQ from 2014 to 2019. All types of auditor changes had a significant positive relationship with restatements, the Non-Big4→Big4 change had the highest restatement rate, and the likelihood of restatement was also high in changes where industry expertise increased. The authors write that both the type of change and the direction of change in industry expertise must be considered when designating auditors through the designation system.
• "The Introduction of the Auditor Registration System and the Value Relevance of Accounting Information" (2026) — KCI ART003335297 — Comparing the 3 years before (2016–2018) and 3 years after (2020–2022) introduction using a difference-in-differences approach. Following the introduction of the registration system, the value relevance of accounting earnings increased significantly, and differed across groups depending on auditor type and firm size. The authors write that this is consistent with, and suggests the possibility of, enhancing the reliability and usefulness of financial information.
• "A review of archival auditing research" (2014) — DOI: 10.1016/j.jacceco.2014.09.002 — Although there are many audit quality proxies, guidance on selection is scarce, providing a framework to systematically evaluate the strengths and weaknesses of proxies. Reviewing common models separating audit quality and financial reporting quality, it writes that better conceptual guidance is needed (literature review).
• "Fundamental Analysis and the Cross-Section of Stock Returns: A Data-Mining Approach" (2017) — DOI: 10.1093/rfs/hhx001 — Even after filtering over 18,000 signals extracted from financial statements using bootstrapping, many signals remained significant predictors of the cross-section of returns, and predictive power was more prominent after high-sentiment periods and in stocks with large arbitrage constraints. The authors write that anomalies cannot be attributed to chance and are better explained by mispricing.
• "The Role of Financial Statement Comparability in Corporate Valuation: Evidence Based on BM Anomaly" (2023) — DOI: 10.30753/emr.2023.41.2.004 — Korea Composite Stock Price Index listed companies from 2011 to 2020. Firms with high comparability reduced the excess returns of the BM anomaly on average, and the authors read that comparability allows for more accurate valuation of value, mitigating mispricing. That mitigation effect was not significant in high-BM firms (value stocks) placed in the risk-return relationship inherent in value investing.
• "Financial Statement Analysis and Predicting Stock Price Crash Risk: Evidence from Korea Stock Market" (2022) — DOI: 10.35214/rfis.11.1.202202.001 — Korea Exchange listed companies from 2005 to 2015. The composite score of financial statement analysis had a negative relationship with future stock price crash risk, and that relationship was mediated by future earnings changes. The authors emphasize the importance of financial statement analysis and the usefulness of accounting information in valuation.
• "Accounting Quality and Debt Contracting" (2008) — DOI: 10.2308/accr.2008.83.1.1 — Borrowers with low accounting quality preferred bank loans, and in bank loans both price and non-price terms became stricter, whereas in public debt only price terms became stricter. The effect of accounting quality on public debt interest spreads was 2.5 times that of bank loans (abstract does not specify the sample's country).