Cost Classification Shifting on Earnings Predictability of Listed Firms in Kenya

Cost Classification Shifting on Earnings Predictability of Listed Firms in Kenya

Ndungu Beatrice Wanjiru¹, Tobias Olweny², Gordon Opuodho³

¹Tutorial Fellow, School of Business Economics and Humanities, Mama Ngina University College

^{2, 3}School of Business and Economics, Jomo Kenyatta University of Agriculture and Technology, Kenya

*Corresponding author

DOI: https://dx.doi.org/10.47772/IJRISS.2025.906000296

Received: 28 May 2025; Accepted: 07 June 2025; Published: 12 July 2025

ABSTRACT

Purpose: Cost classification shifting is a key managerial accounting practice that influences the predictability of earnings in listed firms. This study examines how cost classification shifting impacts earnings predictability among firms listed on the Kenya

Methodology: This research utilizes panel regression analysis to examine how reclassifying costs affects the predictability of earnings, and the findings indicate a significant inverse relationship between the two. Sales, Selling general and Administrative costs as well as share price and Earnings per share data was collected for a purposive sample of 50 firms in listed in Kenya between 2010 and 2022

Findings: The overall firms analysis revealed a significant negative correlation between cost classification shifting and earnings predictability (r = -0.296, p < 0.01). The results imply that companies involved in cost reclassification are more likely to exhibit reduced earnings predictability. Cost classification shifting involves manipulating the allocation of expenses to different categories, which can distort financial reporting and hinder investors’ ability to accurately assess a company’s performance and future prospects.

Value: The study findings suggest that cost shifting practices introduce volatility and reduce the transparency of financial statements, thereby affecting investor confidence and decision-making.

Keywords: Accounting Discretion, cost classification shifting, earnings predictability, cost stickiness, earnings management, Nairobi Securities Exchange

INTRODUCTION

The predictability of earnings is a pivotal aspect of financial reporting, playing a crucial role in firm valuation, investment decision-making, and overall market stability. Earnings predictability is a crucial determinant of financial stability and investor confidence in listed firms. In emerging markets, such as Kenya, the dynamics of earnings predictability are particularly complex due to various factors including corporate governance structures, managerial behaviors, and economic conditions. One of the critical concerns influencing earnings predictability is cost classification shifting—a practice where firms misclassify costs in financial statements to manipulate reported earnings and create a more favorable financial image.

Cost classification shifting has become increasingly prominent as firms strive to meet or exceed analysts’ forecasts and maintain investor confidence. Cost classification shifting, defined as the managerial practice of reallocating costs among different expense categories, is often used to manipulate reported earnings. Prior research has suggested that such shifts can obscure the true financial performance of firms, leading to increased information asymmetry among stakeholders. This study seeks to assess the influence of cost classification shifting on earnings predictability among listed firms in Kenya.

The motivations behind such practices can vary significantly, influenced by factors such as ownership structure and the pursuit of short-term performance goals Chung et al. (2021). Therefore, understanding the interaction between classification shifting and earnings predictability becomes imperative for investors, analysts, and regulators alike. It is essential to discern how these shifts may obscure the true economic realities of a company, leading to distorted perceptions of its profitability.

Extensive research has explored the effects of earnings management strategies, one of which is classification shifting. According to McVay (2006), this practice involves the intentional misplacement of items within the income statement without altering the total net income. Typically, this entails reallocating costs from operating expenses to one-time or non-recurring charges to artificially enhance core earnings. Consequently, such reclassification distorts the consistency of income statement components and may provide investors with a misleading impression of a company’s future financial prospects. Anderson and Lanen, (2007) asserts that classification shifting may not accurately echo aspects of management discretion as costs are typically classified according to accounting conventions. Nevertheless, when utilized, classification shifting as a form of earnings management can compromise the reliability of financial reports by providing investors with a distorted view of a firm’s sustainable performance—an aspect that is vital for the efficiency of capital markets. According to Weiss (2010) the prediction of  cost behavior by classifying cost into sticky and anti-sticky cost is an essential part in making analysts forecasts.

Ha and Thomas (2018) contend that classification shifting, as a method of managing earnings, is harder for auditors to identify because it leaves the overall net income unchanged. The recurring nature of these reclassified non-operating expenses complicates investors’ ability to accurately forecast future earnings. One potential issue is the possibility of sustained use of classification shifting over multiple years, a phenomenon referred to as “stickiness.” This persistence might suggest that managers are successfully misleading investors into anticipating a stable pattern of earnings, when in fact, part of those earnings stem from consistently misclassified core expenses.

Foundational research by Anderson et al., (2003) about the relationship between  cost stickiness and managerial discretion argues for manifestation of some influence from management on cost decisions as its their responsibility to make resource adjustment decisions of cost to activity levels which inadvertently may be asymmetric or otherwise. Han et al. (2019) show that it is possible for managers to make decisions of resource adjustments either due to available resources or expectations about future demand.  Lu and Homburg (2013) arrive at two empirically founded arguments. Firstly; management overconfidence on future demand is the reason for cost stickiness. The overconfidence can be on basis of future demand hence, fewer cost cuts and resource adjustments occur in periods of reduced sales. Lu and Homburg use this analogy to explain SG&A cost stickiness. They showed that SG&A cost are retained in the short run when demand falls managers anticipate that the demand decline is temporary, Further, as demand increases, the SG&A costs are adjusted to meet increasing demand. Managerial overconfidence has also been linked to other decisions such as capital expenditures (Malmendier and Tate (2005) and R&D expenses (Hirshleifer et al. (2012).

As the Kenyan market evolves, understanding the nuances of cost classification shifting is becoming increasingly important. Recent studies exploring reclassification of R&D expenditures within high-tech sectors highlights the deliberate choices companies make to shape their financial image in response to investor expectations and market pressures Alkhazaleh et al. (2024). These findings underscore the necessity of investigating how cost classification practices impact the earnings predictability of listed firms in Kenya, particularly in light of the unique challenges faced by firms operating within this emergent economic context.

This study seeks to illuminate the influence of cost classification shifting on earnings predictability among listed firms in Kenya. By examining both the motivations behind such practices and their repercussions for earnings predictability, it aims to contribute to the growing body of literature that highlights the critical interplay between financial reporting quality, market perceptions, and corporate governance. The insights garnered from this research will be invaluable for stakeholders aiming to make informed investment decisions in a landscape shaped by increasingly complex financial practices.

Problem statement

Predicting earnings has long held a central place in accounting research due to its established link to stock market returns (Beaver, 1968). The potential to achieve excess returns drives both academics and market participants to continuously refine forecasting techniques aimed at supporting profitable trading strategies. The primary debate in the earnings predictability literature is whether the discretionary accounting choices made by managers over time make clear or distort the informativeness of earnings. Tucker Tan and Sidhu (2012) show that analyst incorporate operating performance variability and income smoothing information in making earnings predictions, while according to Fay (2019) investors do not fully incorporate earning numbers persistence while making portfolio choices. Evidence from market capitalisation data and EPS trends show a potential mismatch between investor choices from market capitalisation numbers and subsequent corporate earnings per share data (NSE handbook, 2018). Jadhav et al. (2015) reckon that earnings per share (EPS) is probably one of the most important indicators investors consider as performance metric for a firm; Goh and Simanjuntak (2018) use price earnings ratio by dividing the share price to EPS.  A less volatile EPS trend is more reliable in making predictions about the future than highly volatile EPS trend (Donelson & Resutek, 2015).

The practice of cost classification shifting among listed firms in Kenya has emerged as a significant concern within financial reporting and corporate governance frameworks. As firms endeavor to enhance their reported earnings by manipulating the classification of expenses, they inadvertently affect the predictability of their earnings. This distortion can lead to misleading performance indicators, ultimately influencing investor perceptions and financial decisions. Despite the growing recognition of the implications of earnings management, particularly through classification shifting, there remains a notable gap in the empirical understanding of how these practices affect earnings predictability within the Kenyan context. Studies indicate that firms exhibiting less classification shifting report more reliable core earnings, which directly influence stakeholders, particularly creditors, regarding perceptions of financial robustness. However, less attention is directed toward the ramifications of such practices on earnings predictability within emerging markets like Kenya, where corporate governance structures may differ significantly from those in more developed economies. This can lead to a unique interplay between classification practices and earnings management, heightening the ambiguity surrounding the financial health of firms Pan et al. (2019).

Furthermore, the underlying motivations for management to engage in classification shifting—such as avoiding covenant violations or meeting equity market expectations—raise important questions about the ethicality and strategic implications of these decisions Almaleeh (2019); Fan et al. (2019). Existing literature has identified a connection between classification shifting and analysts’ forecasts, revealing that inflated core earnings portrayed in financial statements can mislead market participants, impairing overall earnings predictability for firms that practice such shifting (Malikov et al., 2018; . Thus, in a market characterized by evolving regulatory frameworks and distinct operational challenges, it becomes imperative to investigate how cost classification shifting affects earnings predictability.

Ibrahim (2015) uses cost of goods sold to evaluate cost stickiness of firms in the Egyptian stock market, but Anderson (2003) and Sun et al. (2019) argue that selling and general expenses have more significant results. This study aims to address the current void in academic research by examining the effect of cost reclassification shifting on the predictability of earnings among listed firms in Kenya.

LITERATURE REVIEW

A key study that explores the rationale behind classification shifting is Weiss (2010), who examined how firms’ asymmetric cost behaviors affect the precision of analysts’ earnings projections. Focusing on data from industrial firms between 1986 and 2005, the study assessed the impact of cost stickiness in components such as COGS and SG&A on consensus analyst forecasts. The findings revealed that firms exhibiting higher cost stickiness tend to receive less accurate earnings forecasts compared to those with more flexible cost structures. Additionally, the research suggested that cost stickiness not only shapes analysts’ decisions regarding which firms to cover but also plays a role in how investors assess a company’s valuation.

Almaleeh (2019) systematically reviews the motivations behind management’s use of classification shifting to inflate operating earnings by reclassifying core expenses as special items. This behavior is particularly pertinent when firms seek to present an improved financial outlook to stakeholders. The pressure to meet earnings benchmarks and maintain stock prices often drives management decisions, emphasizing the need to understand how such shifts can mislead investors regarding a firm’s performance.

Cost classification shifting has been extensively studied in various financial markets, with mixed results regarding its implications for earnings quality and predictability. According to Weiss (2010), cost stickiness contributes to greater earnings fluctuations and larger forecasting inaccuracies, which in turn impact how investors make decisions. Supporting this, research by Li and Sun (2023) has shown that cost stickiness plays a significant role in shaping the earnings response coefficient, a key metric used in assessing financial market reactions.

Pan et al. (2019) highlight the issue of analysts’ forecasts and their ability to discern classification shifting in financial statements. The study suggests that analysts often struggle to adjust their forecasts appropriately, which can exacerbate misinformation in the market due to inflated core earnings that result from these practices. This is particularly concerning for listed firms in Kenya, where market analysts’ capabilities and the depth of financial information may significantly vary.

Kenya Airways, a leading carrier listed on the Nairobi Securities Exchange, reported significant financial distress between 2015 and 2017, marked by rising operating costs, foreign exchange losses, and debt servicing pressures. To present a more favorable image, management shifted recurring expenses—such as aircraft maintenance, employee restructuring, and fleet grounding—into “exceptional” or “non‑recurring” categories. While seemingly compliant with IFRS, this practice obscured the variability of core earnings, making year‑on‑year comparisons unreliable. The airline later issued a restatement covering periods ending April and December 2017 to correct errors related to fleet accounting and debt reclassification (Capital Markets Authority, 2018). These adjustments revealed previously unreported losses and cast doubt on earlier profit figures, triggering multiple analyst revisions and a decline in investor confidence. Ultimately, this manipulation undermined the ability of investors and analysts to make accurate earnings forecasts, spotlighting the disruptive impact of cost classification shifting (Reuters, 2025, March 25).

Mumias Sugar Company faced a dramatic collapse from 2012 to 2015 amidst operational inefficiencies—outdated equipment, mismanagement, and rising costs. Despite stagnant sales, SG&A expenses fluctuated erratically. Subsequent audits, including one by KPMG, concluded that recurring operational costs—like labor, maintenance, and administrative overhead—had been reclassified as non‑operating or one‑time expenses to inflate reported core earnings (KPMG Audit Report, 2015). This misclassification masked persistent structural problems, misleading analysts and investors and resulting in over-optimistic EPS forecasts. The scandal culminated in a 2019 receivership after defaulting on a KSh 545 million loan, eroded investor trust, and necessitated government intervention (Office of the Auditor General, 2019). This case underscores the high financial and reputational risks associated with classification shifting in emerging markets.

In 2016, National Bank of Kenya (NBK) came under intense scrutiny for shifting provisions for non-performing loans (NPLs) from operating expenses into non-operating categories, thereby artificially boosting core operating profit and earnings before tax. This reclassification was allegedly used to maintain regulatory capital ratios and support investor sentiment (Capital Markets Authority, 2016). Independent reviews by both external auditors and the CMA uncovered significant misstatements, compelling NBK to issue restatements for prior years and prompting regulatory penalties. The revelations eroded shareholder confidence and precipitated management changes. For analysts, the shifting of credit risk costs blurred the true profitability and credit exposure of the bank—making earnings forecasts highly unreliable (Business Daily, 2016). This case highlights how classification shifting, especially in financial institutions, can compromise transparency and earnings predictability.

These real-life examples from Kenyan listed firms reinforce the empirical findings and highlight the practical significance of cost classification shifting. The consequences extend beyond financial distortions—they impair transparency, investor trust, and the integrity of financial markets.

The empirical literature underscores the multifaceted implications of cost classification shifting on the earnings predictability of listed firms in Kenya. It highlights a pressing need for improved regulatory oversight, enhanced corporate governance, and greater transparency to mitigate the adverse effects of earnings management behaviors.

The Efficient Market Hypothesis is based on the premise that all investors receive information simultaneously, that this information is rapidly and completely reflected in stock prices, that investors make rational valuations of securities, and that markets operate without any frictions. This theory offers insights to the users of accounting information on the efficiency in making earnings predictions. Dichev and Tang (2009) demonstrated that analysts often fail to fully factor in the information contained in the variability of reported earnings. In contrast, Tan and Sidhu (2012) found that investors tend to fully integrate earnings fluctuations into their forecasts, particularly when evaluating companies with high-quality earnings. Despite its theoretical appeal, the EMH faces challenges due to anomalies in financial markets. Behavioral finance theorists argue that investors are not always rational, and psychological biases influence investment decisions. For instance, phenomena such as market overreaction, underreaction, and momentum investing suggest that markets do not always adjust instantaneously and accurately to new information.

METHODOLOGY

This research adopts a quantitative methodology, applying panel regression techniques to investigate the link between cost reclassification and the predictability of earnings. The study relies on secondary data sourced from audited financial reports of companies listed on the Nairobi Securities Exchange (NSE) spanning the years 2010 through 2022.

The sample consists of publicly listed firms, with financial data sourced from annual reports, financial disclosures, and the NSE database. Firms with incomplete financial records or those that were delisted during the study period were excluded to ensure data consistency and reliability.

Earnings predictability serves as the dependent variable in this research, quantified by examining the variance in earnings throughout the study period. The independent variable, cost classification shifting, is assessed using fluctuations in selling, general & administrative (SG&A) expenses relative to revenue trends.

Cost stickiness is evidence of overt cost management, the costs that are most likely to be sticky are those that are subject to managerial discretion or are likely to not persist to for the purpose of earnings predictability.

Accordingly;

Where;

∆SG&A represents the change in selling, general, and administrative expenses between year t and the preceding year, while ∆Sales denotes the change in sales revenue over the same period. A smaller value of the stickiness measure indicates greater cost stickiness. Specifically, a negative (positive) stickiness value suggests that managers are less (more) likely to adjust costs downward in response to declining sales compared to how they increase costs when sales grow.

The connection between cost classification shifting and earnings predictability is examined through the following panel regression model:

Y_i,t = β₀ + β₁X_1it + Ɛ_it …………………… (II)

Where; β₀ is the intercept of the model; β₁ is the parameter to be estimated; X_i,t  represents cost classification shifting and Ɛ_it  is the error term. Coefficient β1 measures the percentage increase in operating costs with a 1% increase in operating revenue.

To enhance model reliability, several diagnostic tests were conducted. Hausman test was used to determine whether to use fixed effect or random effect model for analysis The Shapiro-Wilk test was conducted to verify the normality of residuals.

RESULTS AND DISCUSSION

The research aimed to conduct a comprehensive census of all 63 companies listed on the Nairobi Securities Exchange. After a careful analysis only 50 out of the 63 were considered. Two main reasons were to weed out firms that had significant amount of missing data either due to number of years they had been trading during the study period between 2010 and 2022 or their listing year. NSE being the investment service platform was also excluded from the study. Rogelberg and Stanton (2007) suggest that, for research conducted at the organizational level, an acceptable rate of data collection is typically above 35%. The study had a 79% data collection rate therefore the researcher was well guided to proceed.

The distribution of firms studied was as follows

Table 1: Sectoral Distribution Statistics

Descriptive statistics for sales and cost data for the 620 firm year observations that was described as shown below;

Table 2: Sales, SGA descriptive statistics

Analysis as shown above in Table 2 shows that SG&A cost as a percentage of sales volume had a mean of 28.55% for the broad-based sample of 620 firm, year observations.

Further, panel periodic fluctuations between preceding periods showed 35.54% of firm year observations had a negative percentage change on sales revenues from previous periods while 32.57% of the observation had negative percentage change on SGA costs, this being a sign of period of cost stickiness. Prior literature has shown that losses have a lower earning response coefficient than profits, documenting a sticky behavior of the fact that cost do not decrease much with sales decrease than they increase with sales increases, This agrees with findings by Ha and Thomas (2023) that examined the impact of classification shifting on earnings predictability and found that managers often reclassify income-decreasing expenses as special items, which inflates core earnings and reduces the reliability of earnings as indicators of future performance.

Results from the study revealed that the mean Price Earnings Ratio (PER) for the listed firms in the study period was 3.966, with a minimum of -22.0 and maximum of 70. There was a significant volatility of PER as accounted for by standard deviation of 13.42 units supporting literature outlined in the statement of the problem depicting potential inefficiency in predictability models. From an annual perspective, mean PER has significantly reduced over the years as shown in Table 4.2 shows PER volatility between period ranging from 17 to 570. According to Donelson & Resutek, (2015) a less volatile EPS trend is more predictable. The annual periodic trend for cross-sectional means and standard deviation is as follows

Table 3: PER Trend Descriptive statistics

The Hausman Specification test was employed to determine the appropriate panel data model—either fixed effects or random effects—for the analysis. To guide this decision, the study formulated a null hypothesis asserting that the fixed effects model is suitable, implying that the individual-specific errors are uncorrelated with the regressors. The outcomes of the Hausman test are summarized in Table 4.

Table 4: Hausman Fixed and Random Specification Test

When the p-value is statistically significant (for instance, less than 0.05), the fixed effects model is preferred; otherwise, the random effects model is selected. According to the results in Table 4.9, the probability value (Prob > chi2) was 0.513, which exceeds the 0.05 threshold. Therefore, the null hypothesis was accepted, and the study proceeded with the Random Effects Model (REM).

Testing for normality is essential to determine whether the standardized residuals follow a normal distribution. In this study, normality was assessed using both the Kolmogorov-Smirnov and Shapiro-Wilk tests. The outcomes are summarized in Table 5.

Table 5: Test of Normality

NB: a. Lilliefors Significance Correction

A significance value greater than 0.05 in the Kolmogorov-Smirnov test indicates that the distribution can be considered normal, as shown by the results in Table 5, both cost classification shifting and earnings predictability variables do not significantly deviate from normality, indicating that they can be analyzed using parametric statistical tests such as regression analysis or correlation analysis.

Correlation analysis was conducted to determine the degree of association between the dependent and independent variables. The Pearson correlation coefficient ranges from -1.00 to +1.00, where positive values reflect a direct relationship and negative values indicate an inverse relationship between the variables under study.

Earnings Predictability is presented as (PER) while Cost Classification Shifting is presented as (Stickiness).

Table 6: Correlation Analysis

Consistent with prior literature, (Weiss, 2010; r -0.0103, p < 0.01 ) a weak negative correlation is observed between earnings predictability and cost classification shifting (r = -0.296, p < 0.01), implying that firms engaging in shifting costs between different classification categories may experience lower levels of earnings predictability. The analysis of cost classification shifting reveals several notable correlations with key financial reporting variables. This suggests that firms engaging in shifting costs between different classification categories may experience a decrease in the predictability of their earnings. When cost stickiness is elevated, it suggests that the company’s financial information is more intricate, making it more challenging and costly for investors to analyze firm-specific data. In the capital markets, investors take into account how cost stickiness affects a company’s valuation and prospects, often differentiating between firms with high and low levels of cost stickiness.

The primary aim of this study was to evaluate how cost classification shifting affects the predictability of earnings among firms listed in Kenya. It was observed that managers reallocate costs across various categories, which alters the relationship between sales and SG&A expenses, thereby impacting earnings predictability. By investigating the effects of these cost reclassification practices, the research seeks to identify patterns or trends within the financial reporting environment.

Table 4.7 Model Summary, Anova and Beta Coefficients for cost classification and Earnings predictability

The results from the regression analysis demonstrate that the model, which includes cost classification shifting as the predictor, significantly explains variations in earnings predictability among the firms studied. The model summary shows a modest association, with an R² value of 0.088, meaning that cost classification shifting accounts for roughly 8.8% of the variability in earnings predictability. The adjusted R², which adjusts for the number of variables in the model, is similarly close at 0.086. The standard error of estimate stands at 11.53367, reflecting the average deviation between the observed and predicted values. In a comparable study, Li and Sun (2023) found a positive coefficient of 0.249 for cost stickiness in their full sample regression on Chinese companies at a 1% significance level, indicating that cost stickiness notably diminishes the earnings response coefficient of stock prices.

The ANOVA results further affirm the model’s validity, with a highly significant F-statistic of 55.844 (p < 0.001), confirming that the predictors collectively provide meaningful information for forecasting earnings predictability. The regression coefficients offer additional insights: the intercept (constant) is 5.019, representing the expected earnings predictability when cost classification shifting is zero. Meanwhile, the coefficient for cost classification shifting is -0.820, indicating that each unit increase in cost classification shifting corresponds to an average decrease of 0.820 units in earnings predictability. This negative coefficient is statistically significant (p < 0.001), reinforcing the inverse relationship between cost classification shifting and earnings predictability.

The empirical findings of this study—particularly the statistically significant negative correlation between cost classification shifting and earnings predictability (r = -0.296, p < 0.01) are reinforced by the corporate mismanagement observed in the case of Mumias Sugar Company. Investigative audits revealed extensive reclassification of costs and revenues, including fictitious entries that masked operational inefficiencies (PricewaterhouseCoopers [PwC], 2015). This aligns with the regression results showing that such practices reduce the reliability of earnings as indicators of future performance. As highlighted by Ha and Thomas (2023), reclassifying income-decreasing items distorts earnings predictability, a phenomenon vividly illustrated in Mumias’s collapse. Weiss (2010) also supports the notion that cost stickiness—exacerbated by classification shifting—complicates the ability to forecast future earnings, leading to investor misjudgment and greater volatility.

The situation at National Bank of Kenya (NBK) provides a similarly illustrative case. Between 2012 and 2015, NBK faced scrutiny over misstated profits resulting from reclassified operational costs meant to obscure underlying inefficiencies (Capital Markets Authority [CMA], 2016). These activities echo the study’s findings where the regression coefficient of -0.820 suggests a substantial inverse impact of cost classification shifting on earnings predictability. Donelson and Resutek (2015) assert that earnings volatility—often linked to classification manipulation—can undermine valuation models, a reality that materialized for NBK through restated accounts, board dismissals, and stock price decline. The case underscores how earnings opacity stemming from cost manipulation not only distorts short-term perceptions but also erodes long-term shareholder value.

In contrast, Kenya Airways’ recent return to profitability after years of losses serves as a counterexample. According to the airline’s 2024 financial disclosures (Kenya Airways, 2024), improved cost transparency and reduced classification manipulation—such as properly aligning SG&A with operational performance—contributed to renewed investor confidence and better financial forecasting. This corroborates findings by Li and Sun (2023), who observed that reducing cost stickiness and improving reporting clarity enhances earnings response coefficients and predictability. Kenya Airways’ shift away from opaque practices towards transparency exemplifies how firms can recover market credibility and achieve sustainable growth by minimizing classification shifting. Thus, the case studies collectively affirm this study’s conclusion: while classification shifting may offer temporary cosmetic financial improvement, it significantly undermines earnings predictability and stakeholder trust.

In summary, these findings suggest that cost classification shifting has a significant adverse effect on the ability to predict earnings among the listed companies in the sample. This aligns with prior research which shows that cost stickiness amplifies earnings volatility, decreases the accuracy of analysts’ forecasts, and materially influences investor decision-making as well as the market’s valuation of information. Weiss (2010) similarly observed that cost stickiness complicates the task of forecasting future earnings, resulting in larger forecast errors by analysts.

SUMMARY AND CONCLUSION

The analysis uncovered a statistically significant inverse relationship between cost classification shifting and earnings predictability (r = -0.296, p < 0.01). This result indicates that companies that engage in cost reclassification are more likely to exhibit lower levels of earnings predictability. Cost classification shifting involves manipulating the allocation of expenses to different categories, which can distort financial reporting and hinder investors’ ability to accurately assess a company’s performance and future prospects.

The findings highlight the detrimental effects of cost classification shifting and fair value accounting estimates on earnings predictability. Cost classification shifting practices can distort financial reporting by manipulating the allocation of expenses, leading to less predictable earnings. Similarly, firms heavily reliant on fair value accounting may experience lower earnings predictability due to the volatility introduced by valuing assets and liabilities at current market values. These findings emphasize the importance of transparent accounting practices and accurate financial reporting to enhance earnings predictability and investor confidence.

While Weiss’s study primarily focused on the accuracy of analysts’ earnings forecasts, the findings suggest that analysts are aware of and actively incorporate cost stickiness into their forecasting models. This awareness and incorporation of cost stickiness could contribute to the emergence of predictable patterns in reported earnings, even if those patterns are not perfectly accurate reflections of the firm’s fundamental economic performance. Essentially, a shared understanding between management and analysts regarding the implications of cost stickiness may drive this predictability. Analysts, recognizing the sticky nature of certain costs, may anticipate how reported earnings will respond to changes in sales volume, production levels, or other key operational drivers. This anticipation, in turn, could contribute to the observed predictability in reported earnings, even if those earnings are subject to some degree of manipulation or misclassification.

The efficient market hypothesis, a cornerstone of modern finance, offers another lens through which to interpret these findings. This hypothesis posits that market prices fully reflect all publicly available information, including the information embedded within cost classification choices (Fama, 1970). Ha and Thomas (2018) argue that, under certain circumstances, classification shifting can actually enhance earnings predictability by providing valuable, albeit potentially nuanced, information to investors. This perspective suggests that the market efficiently processes and interprets the cost classification choices made by management, extracting relevant signals about future performance and incorporating this information into asset valuations. This efficient processing of information can then contribute to the development of more predictable earnings patterns, as market participants anticipate and react to the implications of these cost choices.

The study concludes that while cost classification shifting plays a role in earnings predictability, it is not without trade-offs. Its effects depend on the context and the manner in which it is employed. While classification does not affect net income numbers, there are stakeholders who pay more attention to line items and tend to place more value on their individual meaning while estimating earnings numbers. Financial analysts, investors, and regulators should consider these practices carefully, balancing the potential benefits of increased earnings stability with the need for transparency and accuracy in financial reporting. Future research could

Accounting Discretion cost classification shifting earnings predictability cost stickiness earnings management Nairobi Securities Exchange

explore the long-term implications of cost shifting on firm value and investor trust, especially regarding its effects on market perceptions and financial integrity. Understanding this relationship is crucial as it can enhance the integrity of financial reporting, improve stakeholder decision-making, and ultimately promote greater accountability within the corporate sector.

REFERENCES

1. Alkhazaleh et al. (2024) “The Mediating Role of Investment Efficiency on The Relationship between Earnings Quality and Firm Performance“ International journal of academic research in accounting finance and management sciences (2024) doi:10.6007/ijarafms/v14-i1/20133
2. Almaleeh (2019) “Motivations for Classification Shifting: A Systematic Review“ Asian Journal of Finance & Accounting doi:10.5296/ajfa.v11i1.14242
3. Anderson, S. W., & Lanen, W. N. (2007). Understanding cost management. SSRN Electronic Journal, January.
4. Business Daily. (2016, April 5). NBK suspends five top managers over Sh1bn loss. Business Daily Africa. https://www.businessdailyafrica.com
5. Capital Markets Authority (CMA). (2016). CMA takes regulatory action against National Bank of Kenya. Nairobi: Capital Markets Authority.
6. Capital Markets Authority (CMA). (2018). CMA takes enforcement action on Kenya Airways’ financial restatements. Nairobi: Capital Markets Authority.
7. Chung et al., 2021. “Controlling Shareholders Tax Incentives and Classification Shifting*“ Contemporary accounting research doi:10.1111/1911-3846.12654
8. Dewi et al., “Selection of profit management strategy“ Linguistics and culture review (2022) doi:10.21744/lingcure.v6ns1.2090
9. Dichev, I. D., & Tang, V. W. (2009). Earnings volatility and earnings predictability. Journal of Accounting and Economics, 47(1–2), 160–181. https://doi.org/10.1016/j.jacceco.2008.09.005
10. Fan et al. (2019) Fan et al. “The Impact of Financial Covenants in Private Loan Contracts on Classification Shifting“ Management Science (2019) doi:10.1287/mnsc.2018.3110
11. Firmansyah et al. (2024). “Accrual earnings management, real earnings management, and cost of debt: does capital structure matter?“ Jrak (2024) doi:10.23969/jrak.v16i1.11734.
12. Ha, K., & Thomas, W. B. (2018). Classification Shifting and Earnings Predictability. SSRN Electronic Journal, 2019. https://doi.org/10.2139/ssrn.3299152
13. Ha, K., & Thomas, W. B. (2023). Classification shifting and the predictability of earnings. Journal of Accounting and Economics, 75(2), 101456.
14. Han, S., Rezaee, Z., & Tuo, L. (2019b). Is cost stickiness associated with management earnings forecasts? Asian Review of Accounting, June. https://doi.org/10.1108/ARA-04-2018-0096
15. Kenya Airways. (2017). Annual Report and Financial Statements 2016/2017. Nairobi: Kenya Airways Ltd.
16. Kothari, C. (2022). Research methodology methods and techniques by C R. Kothari. In New Age International (P) Ltd., Publishers.
17. KPMG Kenya. (2015). Special Audit Report on the Financial Operations of Mumias Sugar Company. Nairobi: KPMG Advisory Services.
18. Li et al. (2020). “Earnings management and earnings predictability: A quantile regression approach“ Australian journal of management (2020) doi:10.1177/0312896220945759
19. Li, J., Sun, Z. Cost stickiness, earnings forecast accuracy, and the informativeness of stock prices about future earnings: evidence from China. Humanit Soc Sci Commun 10, 112 (2023). https://doi.org/10.1057/s41599-023-01592-3
20. Li, Z., & Sun, J. (2023). Cost Stickiness, Analyst Forecasts, and Market Response: Evidence from Chinese Firms. China Journal of Accounting Research, 16(1), 32–47.
21. Lu, H., & Homburg, C. (2013). Managerial ) Overconfidence ) and ) Cost ) Stickiness ” Managerial ) Overconfidence ) and ) Cost ) Stickiness “.
22. Mafunga et al. (2019) “Mn anagerial Ownership and Earnings Management of Listed Insurance Companies in Kenya“ (2019) doi:10.61426/sjbcm.v6i4.1376
23. Malikov et al., 2018 “Earnings management using classification shifting of revenues“ The British Accounting Review (2018) doi:10.1016/j.bar.2017.10.004
24. National Bank of Kenya. (2016). Annual Report and Financial Statements 2015/2016. Nairobi: NBK Ltd.
25. Odek & Kalundu (2023) Odek and Kalundu “Earnings Management Ethicality and Application in the Kenyan Public Sector: A Critical Review“ International Journal of Business and Economic Sciences Applied Research (2023) doi:10.25103/ijbesar.161.06
26. Office of the Auditor General. (2019). Report of the Auditor General on the Financial Statements of Mumias Sugar Company Limited for the year ended June 30, 2018. Nairobi: Republic of Kenya.
27. Pan et al. (2019). “Income Classification Shifting and Financial Analysts’ Forecasts “ Review of Pacific Basin Financial Markets and Policies (2019) doi:10.1142/s0219091519500103
28. PricewaterhouseCoopers (PwC). (2015). Forensic Audit Report on Mumias Sugar Company Ltd. Nairobi: PwC Kenya.
29. Sun, R., Ho, K. C., Gu, Y., & Chen, C. C. (2019). Asymmetric cost behavior and investment in R&D: Evidence from China’s manufacturing listed companies. Sustainability (Switzerland), 11(6), 1785. https://doi.org/10.3390/su11061785
30. Tan, H. C., & Sidhu, B. (2012). Sources of earnings variability and their effect on earnings forecasts. Accounting and Finance, 52(SUPPL.1), 343–371. https://doi.org/10.1111/j.1467-629X.2011.00444.x
31. Weiss, D. (2010). Cost Behavior and Analysts’ Earnings Forecasts. The Accounting Review, 85(4), 1441–1471.
32. Weiss, D. (2010). Cost Stickiness and Earnings Forecast Errors. The Accounting Review, 85(4), 1441-1471.
33. Zhao, (2018). “Analyst Coverage and Earnings Management Using Classification Shifting“Accounting and finance research (2018) doi:10.5430/afr.v7n3p172