The Impact of IFRS 9 Adoption on the Financial Performance and Sustainability of Romanian Credit Institutions

ABSTRACT

This paper aims to quantify the impact of the policies established by IFRS (International Financial Reporting Standards) 9 regarding the recognition and measurement of financial instruments and derivatives on the sustainability and financial performance of Romanian credit institutions. The results show a positive significant impact on performance when the evaluation is made at fair value through profit and loss and a negative one when is made through elements of other comprehensive income. While the expected credit loss policy is essential to underpin prudential policy, it also brought some financial instability for the sampled institutions.

KEYWORDS:

• Credit institution
• evaluation methods of assets
• expected credit loss
• financial performance and sustainability
• IFRS 9

JEL CLASSIFICATION:

• M41
• G21

Introduction

The current global context characterized by intensive international capital transfers and economic transactions calls for uniform regulation of accounting, so that an optimal level of comparability and transparency of data is ensured and high-quality accounting information is delivered to stakeholders (Grosu, Socoliuc, and Hlaciuc 2017). The development of the International Financial Reporting Standards (IFRS) responded to this specific need of the global economic environment. Over time, these standards have proved their usefulness and effectiveness and have gained international recognition, being of interest even for economic cooperation blocks like BRICS (Brazil, Russia, India, China, and South Africa), whose member countries are working on the development of their accounting legislation based on the IFRS provisions (Melega 2022). Currently, these standards are one of the most widely used conceptual frameworks for financial reporting along with the United States General Accepted Accounting Principles - US GAAP. This is driven by their contribution to ensuring the quality of disclosures (Malo-Alain, Aldoseri, and Melegy 2021) and minimizing the information gap between investors and investees (Achim and Tiron-Tudor 2018), as well as by their ongoing support to boost financial performance through various accounting policy recommendations (Abdullah and Tursoy 2021).

The core business of credit institutions is based on “taking deposits or other repayable funds from the public and granting credits for its own account” (EBA 2020, 2). Therefore, it is based on financial instruments that are represented by legal arrangements involving any form of monetary value, most of which ensure an efficient transfer of capital between the involved parties. IFRS 9 focuses on the accounting policies and treatments for the recognition, measurement, and impairment of such financial instruments, bringing significant changes that have a direct impact on the results of credit institutions. When an entity recognizes or evaluates a financial asset, it must consider its business model for asset management and the contractual characteristics of the cash flows that it generates (Deloitte 2023). Under IFRS 9, the financial instruments are classified as measured at amortized cost, at fair value through other comprehensive income, and at fair value through profit and loss (IFRS Foundation 2022b). Before the use of IFRS 9, the recognition of credit losses under IFRS 39 would have been considered”once a financial asset is impaired or there is objective evidence that a loss will be incurred” (Hansen, Charifzadeh, and Herberger 2023). In times of crisis, this situation could be problematic, as the value of provisions may not cover all the generated losses and the banks would struggle to keep their financial balance and collect new capital. This is why IFRS 9 requires to be recognized an expected credit loss that reflects an”unbiased and probability-weighted amount that is determined by evaluating a range of possible outcomes” (ACCA 2022a), using a three-staged classification based on the risk level of the loans.

Considering all these aspects, this paper aims to assess the impact of these changes on the financial performance and sustainability reported through the annual financial statements of Romanian credit institutions. The main objectives of the research are focused on: O₁ - identifying the medium-term impact of different methods of valuation and recognition of financial instruments under IFRS 9 on the financial performance and sustainability of credit institutions; O₂ - identifying the medium-term impact of expected credit loss recognition on the financial performance and sustainability of the credit institutions.

Literature Review and Hypotheses Construction

The results of the scientific papers considered to be the most significant for the aim and objectives of the present research are presented in Table 1 below, together with the developed hypotheses.

Table 1. Meta-analysis of literature.

Authors, year	Scope	Results	Impact
Ozili (2023)	to present a sustainable loan loss provisioning system that aligns bank loan loss provisioning with the sustainable development goals;	the proposed sustainable loan loss provisioning system would align bank loan loss provisions with sustainable development goals by adjusting loan loss provisioning estimates to reflect the environmental benefits and costs of borrowers’ business activities;	significant impact because it projects future potential instruments to financially constrain companies to comply with a certain level of sustainability;
Kund and Rugilo (2023)	to analyze the implications of the change from IAS 39 to IFRS 9 regarding the resilience of banks, with attention to the “cliff-effect”* and “front-loading”** effects in times of crisis;	the “cliff-effect” of IAS 39 has been weakened under IFRS 9, which indicates the potential of the staging model to enhance the financial stability of the banking sector in the future, but at the cost of “front-loading” expected credit losses; the impairments grew excessively at the beginning of the COVID-19 crisis, exerting a stronger impact on financial stability, but the gap between the two accounting standards narrowed with time;	significant impact considering that the authors have built a model able to mitigate the impact of crises on the impairments of expected losses;
Percevic (2022)	to identify and analyze how the financial instruments were previously classified and measured under IAS 39 and reclassified and measured under IFRS 9 in the Croatian banking sector;	IFRS 9 implementation had an impact on the financial position and performance of the Croatian banking sector due to the recognition of impairment losses and loss allowances according to the expected loss model established in IFRS 9 and re-measurement of financial assets at fair value through profit and loss;	significant impact, considering that it highlights the main factors of IFRS 9 implementation that negatively influenced the financial performance and position of credit institutions;
Egbon-Aghaleleghian and Oziegbe (2022)	to compare the firm performance obtained by manufacturing companies in Nigeria when using fair value versus amortized cost measurement;	there are significant differences between the impact of the historical cost accounting valuation technique and the fair value technique on firm performance; it was found that profit after tax and return on equity have a positive correlation that is stronger in the context of the fair value method than in the context of the amortized cost method, therefore there is a difference between the outcome of their implementation;	significant impact because it outlines the differences in the performance results, considering the evaluation method that was applied;
Aladwan (2022)	to find if the transition from the historical cost method to the fair value method has any impact on performance results;	it was identified a direct association between fair value evaluation and the value of stock price and also, the inclusion of the cumulative change in fair value in total assets or equity improved the relationship between ROA and ROE and share price;	significant impact because it proves that the transition from the amortized cost method to the fair value evaluation method can generate earnings that will positively influence the performance of companies;
Zang, Kabir, and Scott (2022)	to assess the impact of equity financial assets (EFA) presentation at fair value in other elements of other comprehensive income (OCI) under IFRS 9;	the percentage of firms using the OCI presentation for EFA remains stable, while firms choosing to present the EFA at fair value in profit and loss (FVPL) increased significantly during the adoption year for the financial sector;	significant impact because it proves that there is a difference when the same financial assets are analyzed at fair value but are presented in different locations (PL/OCI);
Ãollaku Ahmeti, and Aliu (2021)	to examine the transition effect of IFRS 9 on the level of assets balance, allowance for loan losses, and capital of banks in Kosovo.	the effect of IFRS 9 on the loan loss allowance for the analyzed banks is negative and presents a significant increase in the level of loan allowance, having a direct impact on their financial position and performance; the early recognition of credit losses is reflected in the reduction of ROA and ROE indicators;	a significant impact of the study because it clearly shows that the new expected credit loss model introduced by IFRS 9 harms the financial performance and position of banks;
Hameedi et al. (2021)	to explore the effect of IFRS on the financial performance of Iraqi commercial banks and to investigate the value significance of financial performance statements;	the adoption of IFRS has a positive and significant effect on the rate of return on equity; the adoption of IFRS has a statistically positive effect on the value relevance of the book value of equity, and also a positive, but not statistically significant impact on the value relevance of accounting earnings per share;	significant impact considering that was identified a positive impact of IFRS adoption, generating an upward trend in important indicators of financial performance.
Brito and Judice (2021)	to analyze the advantages and disadvantages of classifying a financial asset at amortized cost and fair value; defining a bi-objective model to find efficient allocations;	the higher the proportion of the investment allocated at fair value, the greater the variability of the annual income and the overall annual result; if the management decides to allocate an investment entirely at fair value, a higher average annual income and average annual comprehensive income will be obtained than if it is decided to allocate the investment entirely at amortized cost.	significant impact because it highlights the best option to evaluate financial assets in terms of recorded results.
Bui, Le, and Huy (2020)	to identify the scope, investigate the benefits, and assess the challenges of IFRS implementation;	IFRS implementation increases the quality of financial information, reduces investment risks, increases market efficiency, and attracts foreign direct investment;	significant impact considering that all its findings are drivers of financial performance;

This table presents some of the main papers that support our research hypotheses enlisted in the following part of the paper. It highlights the objectives, the results, and the impact of the selected papers, considering their contribution to the literature and the relevance and applicability of their results.

*which refers to sudden increases in impairments.

**the earlier recognition of losses may adversely impact bank resilience through lower capital levels.

Elaborated by the authors.

Regarding the impact of the IFRS implementation on financial performance, the results of the study conducted by Bui, Le, and Huy (2020) have highlighted the fact that it reduces investment risk by increasing market efficiency, and encourages foreign investment flows which leads to an increase in the individual financial performance of the credit institutions. In their study, Hameedi et al.(2021) have come to the same conclusion, highlighting the positive impact on indicators such as ROE, earnings per share, and the book value of equity. Not far from these findings are the results of Chouaibi and Mutar (2024), who have stated that the higher the trend of IFRS implementation in the listed Iraqi banks, the higher the increase in the quality of the reported profits. Twiah and Oyewo (2024) have stated that the implementation of IFRS is significantly and positively correlated to an increased level of foreign investment, which represents an important source of funding for banks. Moreover, Eiler, Miranda-Lopez, and Tama-Sweet (2022) have identified a negative correlation between earnings management and IFRS adoption that implicitly generates an increase in the transparency of the financial accounting disclosure process (Setiawan et al. 2020). This fact can be considered as another indicator that the practice of IFRS principles leads to a real increase in the financial performance reported by the financial statements. This results because management’s behavior of manipulating revenue by applying accounting policies that, in the short term, generate high revenue values is more effectively controlled.

H₁: The implementation of the IFRS framework is leading to an increase in the financial performance and sustainability of credit institutions.

IFRS 13 defines fair value as”the price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date” (IFRS Foundation 2022a). When financial instruments are measured at amortized cost, their value does not change based on market fluctuations, thus presenting less investment risk. On the other hand, when financial instruments are measured at fair value, they are subject to market fluctuations, thus presenting a higher investment risk but also the possibility to obtain potential gains over their lifetime. This situation does not take place when they are measured at amortized cost. According to the findings of Andrzejewski, Dunal, and Ozga (2018), the gains or losses that have been generated by the reevaluation or reclassification of the financial assets according to the categories required by IFRS 9, will adjust the retained earnings or equity item by showing a decrease in asset value by an average of 0.53% as a result of these accounting treatments. The findings of the study conducted by Brito and Judice (2021) on some categories of financial assets in the context of the banking investment portfolio suggest that the higher the proportion of the investment allocated at fair value, the higher the annual income (which means an increase in the return on the investment), as well as the variability of the overall result, which implies a higher risk of the investment. Moreover, Ltaief and Molla (2023) argue that the evaluation of financial assets at fair value through other comprehensive income positively influences the firm value of banks, while those evaluated through profit and loss or at amortized cost would harm it. The above-mentioned findings are also supported by the results of Benetti et al, (2024), who have concluded that the financial assets that have been evaluated through other comprehensive income, had the most significant influence on the data in the financial statements, while the adjustments of those which have been evaluated through profit and loss do not cause considerable changes in the reported financial data. Therefore, the use of the fair value method is considered to be more cost-effective than if it was decided to allocate the entire investment at amortized cost, as stated by Egbon-Aghaleleghian and Oziegbe (2022).

H₂: The impact of applying financial asset valuation methods – under IFRS 9 - on the financial performance and sustainability of credit institutions is different depending on the implemented method.

IFRS 9 also changes the classification of the impairment of the financial instruments from realized credit losses to expected credit losses (Pastiranova and Witzany 2021), calculated in three steps: 1 – theoretically, all loans are classified in this category, 2 - when the credit risk associated with the financial instrument has increased significantly since the initial recognition; 3 - when the loss is incurred (Delgado-Vaquero, Morales-Díaz, and Zamora-Ramírez 2020). According to the results of the study conducted by Percevic (2022), the implementation of the provisions of IFRS 9 has had a more significant impact on the financial position and performance of Croatian credit institutions through the recognition of the adjustments for expected credit losses, than through the revaluation of financial assets at fair value through profit or loss. Another study which has been conducted by Percevic and Ercegociv (2022) has highlighted the fact that the increase in the impairment losses during the pandemic year 2020 did not have a significant impact on the financial position of the Croatian banks, but rather a strong negative impact on their financial performance. These findings are also consistent with those of Ãollaku, Ahmeti, and Aliu (2021) who pointed out that the recording of adjustments for the expected credit losses is reflected in the decreases of the ROE and ROA indicators, thus influencing the performance of credit institutions. Mahieux, Sapra, and Zhang (2022) have also identified a reduction of the analyzed bank’s equity, driven by the recognition of expected losses provisions which have had an impact on the bank’s capital ratio. The same idea has also been highlighted by Loew, Schmidt, and Thiel (2019) who have pointed out that the overall impact on equity caused by the adoption of IFRS 9 ranged from minus 26.6% to plus 4.9%, which was largely due to the recording of impairments, while the impact of the classification and valuation of financial instruments has been positive for many of the analyzed banks.

H₃: Recognition of expected credit loss adjustments has a significant negative impact on the financial position and performance of credit institutions.

In addition to the fact that IFRS 9 has a clear impact on the financial performance and sustainability of credit institutions, its application also leads to an improved communication process with different categories of stakeholders, by increasing their transparency and trust in banks, as well as decreasing the level of skepticism toward their activity (Jassem, Razzak, and Sayari 2021).

Materials and Methods

The target group of our research refers to the credit institutions in an emerging country, namely Romania. According to the study conducted by Apetri, Mihalciuc, and Grosu (2015), before 2015, the performance indicators of the Romanian banking system were negatively impacted by the effects of the 2008 financial crisis. The present paper aims to identify whether the implementation of IFRS 9 had a positive impact on the financial performance and sustainability of Romanian credit institutions. For the statistical analysis, a database was created with the main economic and financial indicators of the sample, which were manually collected from their financial statements and annual reports published in the period 2018–2021. The initial sample consisted of 23 credit institutions, according to the Register of Credit Institutions, Part I, issued by the National Bank of Romania (BNR, NBR 2022b), but due to the inaccessibility of the financial statements or the lack of various indicators necessary for the pursued econometric analysis, the final sample consists of 20 credit institutions.

To design the econometric models that consider the impact of applying IFRS 9 provisions on the recognition and measurement of financial instruments using the three established methods, as well as the recognition of expected credit loss adjustments, the indicators presented in Table 2 below were considered.

Table 2. Presentation of econometric models and dependent and independent variables.

Econometric model description	Variable setting and description
Model 1: $\mathrm{FVEvRPL}={\mathit{\alpha }}_1+{\mathit{\beta }}_{11}\ast \mathrm{CCE}+{\mathit{\beta }}_{12}\ast \mathrm{TLG}+{\mathit{\beta }}_{13}\ast \mathrm{IA}+{\mathit{\beta }}_{14}\ast \mathrm{TE}$ $+{\mathit{\beta }}_{15}\ast \mathrm{Pr}+{\mathit{\beta }}_{16}\ast \mathrm{CD}+{\mathit{\beta }}_{17}\ast \mathrm{Inc}+{\mathit{\beta }}_{18}\ast \mathrm{Exp}$ $+{\mathit{\beta }}_{19}\ast \mathrm{FVEvAPL}+{\mathit{\epsilon }}_1$	Dependent	FVEvRPL - Fair Value Evaluation Result through Profit and Loss
	Independent	CCE - Cash and cash equivalents
		TLG - Total Loans Granted
		IA - Intangible Assets
		TE - Total Equity
		Pr - Provisions
		CD - Clients’ Deposits
		Inc - Income
		Exp - Expenditures
		$\mathit{FVEvAPL}$ - Fair Value Evaluated Assets through Profit and Loss
Model 2: $\mathrm{FVEvROCI}={\mathit{\alpha }}_2+{\mathit{\beta }}_{21}\ast \mathrm{CCE}+{\mathit{\beta }}_{22}\ast \mathrm{TLG}+{\mathit{\beta }}_{23}\ast \mathrm{IA}+{\mathit{\beta }}_{24}\ast$ $\mathrm{TE}+{\mathit{\beta }}_{25}\ast \mathrm{Pr}+{\mathit{\beta }}_{26}\ast \mathrm{CD}+{\mathit{\beta }}_{27}\ast \mathrm{Inc}+{\mathit{\beta }}_{28}\ast$ $\mathrm{Exp}+{\mathit{\beta }}_{29}\ast \mathrm{FVEvAOCI}+{\mathit{\epsilon }}_2$	Dependent	FVEvROCI - Fair Value Evaluation Result through Other Comprehensive Income
	Independent	CCE - Cash and cash equivalents
		TLG - Total Loans Granted
		IA - Intangible Assets
		TE - Total Equity
		Pr - Provisions
		CD - Clients’ Deposits
		Inc - Income
		Exp - Expenditures
		FVEvAOCI - Fair Value Evaluated Assets through Other Comprehensive Income
Model 3: $\mathrm{FSI}={\mathit{\alpha }}_3+{\mathit{\beta }}_1\cdot \mathrm{ROE}+{\mathit{\beta }}_2\cdot \mathrm{ROA}+{\mathit{\beta }}_3\cdot \mathrm{RMA}+{\mathit{\beta }}_4\cdot \mathrm{FVEvRPL}+$ ${\mathit{\beta }}_5\cdot \mathrm{FVEvROCI}+{\mathit{\beta }}_6\cdot \mathrm{AI}+{\mathit{\epsilon }}_3$	Dependent	FSI (Financial Sustainability Index)* - computed as: loans granted/deposits attracted
	Independent	ROE (Return on Equity) - computed as: net profit/equity
		ROA (Return on Assets) - computed as: net profit/total assets
		RMA (Return of main activity) - computed as: operating income/operating expenses
		FVEvRPL** - computed in the first econometric model
		FVEvROCI** - computed in the second econometric model
		AI (Adjustment index)** - computed as: expected loan loss adjustments/gross loan exposure

This table presents the calculation formulas of the econometric models, as well as the dependent and independent variables considered in their construction, for each one of them the corresponding acronym being established. From this point, this is the way they will be referred to in this paper.

*on the basis that the main activity of a credit institution is of a financial nature – their lending capacity – the ratio of loans granted to deposits attracted was chosen as the dependent variable of model 3, expressing the level of financial sustainability since the greater the lending capacity of a banking institution, the more it can be said that the activity it carries out has continuity and is therefore sustainable.

**except for the flagged variables of model 3, the indicators are calculated according to the aggregate indicator formulas for credit institutions presented by the BNR, NBR 2022a.

***other indicators representative of the financial sustainability of credit institutions, such as solvency or liquidity, could not be calculated due to the lack or difficulty in collecting the necessary data.

Elaborated by the authors.

The selection of the variables for the design of the econometric models is motivated by their importance in the financial statements of the credit institutions and also in their collaboration with present or potential investors, considering that they reflect the financial position and performance. Moreover, the selected indicators entered under the impact of IFRS 9 provisions regarding the methods of recognition and evaluation of financial instruments, and therefore this is the best way to determine the impact of IFRS 9 implementation. The CCE variable presents the liquidity level of the sampled institutions, their ability to pay off their current liabilities with the most liquid assets, and could also reflect their investment policies: to hold the CCE for long-term development plans (risk-averse) or to invest them in riskier opportunities but with higher possible return. Moreover, the CCE can be generated by financial assets evaluated at fair value through profit and loss or other comprehensive income, depending on the conditions met to be classified in either one of them. The TLG and CD variables are significant indicators for any credit institution as they reflect the level of its main activity: taking deposits and granting credits, which are also under the impact of IFRS 9 provisions. The IA variable is also an important indicator for the present or potential investors, as its fair value can reflect the expected future economic benefits that will flow in the business, showing the institutions’ potential for performance. The TE variable reflects the value attributable to the owners of a business – an important reference for investment decisions – and its relevance in the context of our research is given by its components. The Pr indicator is directly addressed by IFRS 9 and reflects the prudential policy of the analyzed banks and their capacity to face difficult financial challenges without deteriorating their financial position. The inclusion of Inc and Exp variables was essential, as they directly reflect the dynamics of the operational activity and are linked with evaluation results and with banks’ performance. Regarding the selection of the variables for the third model, it was made based on their relevance in reflecting banks’ performance in the context of our research and also based on data availability for computing.

Some of the limitations in designing these econometric models refer to the limited number of variables considered, as there may be many other indicators that could have a relevant impact on the selected dependent variables, as the effects of IFRS 9 provisions could spread to other balance sheet elements. By including solely the amount of the expected credit loss provisions in the statistical modeling and not the total amount of provisions – as it is -, a more specific result for the present study would have been generated, but because these data were not available for the entire sample of banks, they could not be taken into account. Given the relevance of the independent variables selected for the analysis, it is expected that the econometric models will be statistically validated and that significant correlations between variables will be identified.

For the statistical processing of the collected economic and financial indicators, necessary to obtain and validate the FVEvRPL, FVEvROCI, and FSI variables, the SPSS Statistics software, v26 was used.

Results and Discussions

As mentioned above, the present statistical analysis aims at building econometric models of multiple linear regression for each of the valuation methods of financial instruments promoted by IFRS 9, namely: at fair value through profit and loss, at fair value through other comprehensive income, and at amortized cost. This is an important step to observe their impact on the financial sustainability index of Romanian credit institutions, subsequently modeled.

Therefore, the following general multiple linear regression is considered to construct and validate the FVEvRPL and FVEvROCI variables:

(1) $Y_i=\mathit{\alpha }+{\sum }_{j=1}^9{\beta }_{\mathit{ij}}{X}_{\mathit{ij}}+{\epsilon }_i,\mathrm{i}=1,..,\mathrm{n}.$(1)

where n is the custom sample size for each model.

Model no. 1. The following statistical model is considered for the construction of the FVEvRPL variable and the independent variables taken into account in its calculation are highlighted below:

(2a) $\mathbf{FVEvRPL}={\alpha }_1+{\beta }_{11}\ast \mathit{CCE}+{\beta }_{12}\ast \mathit{TLG}+{\beta }_{13}\ast \mathit{IA}+{\beta }_{14}\ast \mathit{TE}+{\beta }_{15}\ast \mathit{Pr}+{\beta }_{16}\ast \mathit{CD}+{\beta }_{17}\ast \mathit{Inc}+{\beta }_{18}\ast \mathit{Exp}+{\beta }_{19}\ast \mathit{FVEvAPL}+{\epsilon }_1$(2a)

It should be noted that the indicator of losses from financial assets measured at fair value through profit and loss could not be taken into account as a variable in this econometric model as it did not materialize in the financial statements of the analyzed credit institutions, or it exhibited very insignificant values. In most cases, earnings from financial assets valued at fair value through profit and loss account (which was selected as the dependent variable of the model) were shown, which is consistent with the results obtained by Aladwan (2022), Brito and Judice (2021) and by Andrzejewski, Dunal, and Ozga (2018), thus confirming H₂: The impact of applying financial asset valuation methods – according to IFRS 9 - on the financial performance and sustainability of credit institutions is different depending on the implemented method.

Table 3 above shows a very strong correlation level of 0.937 between the variables considered in the model construction and R Square – the determination ratio – shows that the variations of the independent variables explain 87.8% of the variations of the FVEvRPL dependent variable.

Table 3. Model 1 Summary.^b

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.937^a	.878	.805	38406.803

This table presents the summary of the first econometric model, highlighting the level of correlation between the following dependent and independent variables.

^aPredictors: (Constant), FVEvAPL, Pr, Exp, CCE, IA, TE, CD, Inc, TLG

^bDependent Variable: FVEvRPL

Computed using SPPS Statistics v26 software.

Table 4 above shows the validation of the designed model, as the Sig. value does not exceed 0.05 points, the significance threshold at which the statistical analysis was performed. It can also be seen that the model was designed based on 25 (df: 24 + 1) observations, which means that the FVEvRPL indicator taken as the dependent variable was available for 25 observations out of the 80 that make up the database (20 banks, 4 years each).

Table 4. Model 1 ANOVA^a test.

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	159359520365.611	9	17706613373.957	12.004	.000^b
	Residual	22126237851.029	15	1475082523.402
	Total	181485758216.640	24

This table presents the results of the ANOVA test performed for the first econometric model to identify if there are significant differences between the variance of the means of the analyzed sample. We can see that the F value is quite high, which means that between-group variation is larger than the within-group variation, making it more likely that these differences would reflect differences at the population level regarding the analyzed correlation. The ANOVA test was made considering the same dependent and independent variables.

^aDependent Variable: FVEvRPL

^bPredictors: (Constant), FVEvAPL, Pr, Exp, CCE, IA, TE, CD, Inc, TLG

Computed using SPPS Statistics v26 software.

Table 5 shows the correlation coefficients that will be used to fill in the regression equation 2a as follows:

(2b) $\mathbf{FVEvRPL}=21378.873+0,026\ast \mathit{CCE}+0,008\ast \mathit{TLG}-0.195\ast \mathit{IA}+0,04\ast \mathit{TE}-0,283\ast \mathit{Pr}-0,009\ast \mathit{CD}+0,098\ast \mathit{Inc}-0,256\ast \mathit{Exp}+0,052\ast \mathit{FVEvAPL}$(2b)

Table 5. Model 1 coefficients^a.

Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	21378.873	17165.584		1.245	.232
	CCE	.026	.004	1.816	6.549	.000
	TLG	.008	.004	1.470	1.871	.081
	IA	−.195	.213	−.295	−.916	.374
	TE	.040	.011	1.678	3.522	.003
	Pr	−.283	.044	−1.151	−6.406	.000
	CD	−.009	.002	−2.882	−3.974	.001
	Inc	.098	.046	1.573	2.132	.050
	Exp	−.256	.049	−2.172	−5.198	.000
	FVEvAPL	.052	.041	.274	1.268	.224

This table presents the significance (Sig.) of the correlation between the dependent variable FVEvRPL and the independent variables listed, as well as the nature of this correlation: positive or negative, considering the value of B.

Computed using SPPS Statistics v26 software.

Regarding the order of influence of the independent variables on the FVEvRPL dependent variable, it is as follows: CCE, Pr, Exp, CD, TE, Inc, TLG, FVEvAPL, and IA. We can estimate that the significant importance of cash and cash equivalents in the construction of the FVEvRPL variable is determined by the positive cash flow generated by financial assets. In this category can be included contracts for price differences – CFD: gold, silver, shares, bonds; FUTURES contracts; forward contracts; options, etc., which are valued at fair value through the profit and loss account, meaning that as these favorable differences increase, the FVEvRPL variable will also increase and will be directly reflected through an income item in the profit and loss account. The second variable with the most significant impact on the dependent variable is Pr, which has a statistically significant negative correlation with it, a fact that can be explained by the very context in which provisions are recognized: a certainty of a future expenditure based on a present obligation whose total value is still unknown. Therefore, as the value of FVEvRPL increases, there is no reason to recognize elements of Pr, as the present does not suggest any certainty of future expenditure, but rather an income. It can also be observed a significant negative correlation between the CD variable and the dependent variable, which can be explained by the fact that, usually, a bank’s liabilities are evaluated at amortized cost and therefore it can’t have any positive or negative impact on the variable FVEvRPL. As for the significant impact of expenses and income on the construction of the FVEvRPL variable, this can be understood from the very accounting mechanism of the fair value through the profit and loss method, which is shown below Figure 1.

Figure 1. Accounting mechanism for the evaluation method at fair value through the profit and loss.

The figure presents a straight line that represents the life of an asset evaluated at fair value through profit and loss from year N when it enters at input value (Iv) to year N+e when it goes out from the company’s management at output value (Ov). It can be seen that the value of assets changes in each year of their life.

Elaborated by the authors.

where the segment represents the life of the asset measured at fair value through profit or loss (FVEvAPL);

N: year of the FVEvAPL’s entry into management; N+e: year of the FVEvAPL’s exit from the management;

Iv: FVEvAPL input value; O_v : FVEvAPL output value;

V₁, V₂, V₃, V₄, ……: the fair value at which the asset is valued in each of the financial years over its entire lifetime up to year N+e.

Therefore, the recognition of income or expense resulting from fair value (F_v) measurement through the profit and loss account (PLA) is based on the following mechanism:

if:

1. F_v > I_v, this results in an income recognized in the PLA, generating an increase in the result;

2. F_v < I_v, results in an expense recognized in the PLA, generating a decrease in the result;

3. F_v = I_v, no changes to the asset’s value are recorded.

An income increase from the fair value valuation of an asset will cause an increase in the FVEvRPL dependent variable, and an increase in expenses from the valuation of an asset will cause a decrease in the same variable. As already mentioned, these variations are recognized in the retained earnings, which is a component of equity, meaning that an increase in the FVEvRPL will also lead to an increase in TE. This aspect is also demonstrated in the analyzed sample by the high level of correlation significance (0.003) between the TE variable and the dependent variable. This can be generated by the fact that during the period under review, earnings from fair valuation of assets were mainly reflected in the retained earnings, a component of TE. Also, as the volume of investments in FVEvAPL increases, their profitability increases, which in our case is reflected by the increase of the FVEvRPL variable, supported also by the results obtained by Brito and Judice (2021) in their paper. The statistically insignificant correlation of IA with the dependent variable can be explained by the fact that the depreciation of some intangible assets evaluated at amortized cost that is recognized in profit and loss, did not have a significant impact on FVEvRPL.

Model no. 2. The following econometric model is considered for the construction of the FVEvROCI variable. The independent variables taken into account in its calculation are highlighted in the equation:

(3a) $\mathbf{FVEvROCI}={\alpha }_2+{\beta }_{21}\ast \mathit{CCE}+{\beta }_{22}\ast \mathit{TLG}+{\beta }_{23}\ast \mathit{IA}+{\beta }_{24}\ast \mathit{TE}+{\beta }_{25}\ast \mathit{Pr}+{\beta }_{26}\ast \mathit{CD}+{\beta }_{27}\ast \mathit{Inc}+{\beta }_{28}\ast \mathit{Exp}+{\beta }_{29}\ast \mathit{FVEvAOCI}+{\epsilon }_2$(3a)

As with the construction of the FVEvRPL variable, the indicator of losses from financial assets measured at fair value through other comprehensive income could not be considered as a variable in this econometric model, as no mention of its value was identified in the financial statements of the analyzed credit institutions, or it presented very insignificant values. In the majority of cases, earnings from financial assets measured at fair value through other comprehensive income (which was selected as the dependent variable of the model) were recorded, which is consistent with the results obtained by Aladwan (2022), Brito and Judice (2021) and Egbon-Aghaleleghian and Oziegbe (2022). This result confirms again H₂: The impact of applying financial asset valuation methods – under IFRS 9 - on the financial performance and sustainability of credit institutions is different depending on the implemented method.

Table 6 above shows a correlation level of 0.782 points, which is lower than the one recorded in the regression of the first model, the variation of the independent variables explaining in proportion of 61.2% the variation of the FVEvROCI dependent variable. This means that in addition to the considered independent variables, there are other factors with a significant influence on the value of FVEvROCI.

Table 6. Model 2 summary.^b

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.782^a	.612	.478	121208.974

This table presents the summary of the second econometric model, highlighting the level of correlation between the following dependent and independent variables.

^aPredictors: (Constant), FVEvAOCI, IA, CCE, Exp, Pr, TE, TLG, Inc, CD

^bDependent Variable: FVEvROCI

Computed using SPPS Statistics v26 software.

Table 7 shows the validation of the designed model by the fact that the Sig. value does not exceed 0.05 points, which is the statistical significance threshold established in the econometric modeling that was performed. It can also be seen that the model was designed based on 36 (df: 35 + 1) observations, which means that the FVEvROCI indicator taken as the dependent variable was available for 36 observations out of the 80 that constitute the database (20 banks, 4 years each). The fact that the value of the FVEvROCI indicator is available for more observations than for the value of the FVEvRPL indicator, may suggest that this method is more cost-effective – considering that it generated more earnings – and that it is used more often by credit institutions in the recognition and valuation of financial assets. This can be attributed to the fact that most of them are based on a business model whose objective is achieved both through the collection of contractual cash flows (excluding payments of principal and interest on the principal amount due) and through the sale of financial assets.

Table 7. Model 2 ANOVA^a test.

Model		Sum of Squares	df	Mean Square	F	Sig.
1	Regression	602336728926.906	9	66926303214.101	4.555	.001^b
	Residual	381982001166.733	26	14691615429.490
	Total	984318730093.638	35

This table presents the results of the ANOVA test performed for the second econometric model to identify if there are significant differences between the variance of the means of the analyzed sample. We can see that the F value is considerably lower than in the first case, the means of the groups clustering together more tightly than the within-group variability, meaning that the analyzed population is not strongly different regarding the analyzed correlation. The ANOVA test was made considering the same dependent and independent variables.

^aDependent Variable: FVEvROCI

^bPredictors: (Constant), FVEvAOCI, IA, CCE, Exp, Pr, TE, TLG, Inc, CD

Computed using SPPS Statistics v26 software.

Table 8 shows the correlation coefficients that will be used to fill in the regression equation 3a as follows:

(3b) $\mathbf{FVEvROCI}=-2487.215-0,010\ast \mathit{CCE}-0,027\ast \mathit{TLG}-0,887\ast \mathit{IA}+0,084\ast \mathit{TE}+0,195\ast \mathit{Pr}+0,015\ast \mathit{CD}+0,009\ast \mathit{Inc}+0,033\ast \mathit{Exp}-0,012\ast \mathit{FVEvAOCI}$(3b)

Table 8. Model 2 coefficients.^a

Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	−2487.215	34599.178		−.072	.943
	CCE	−.010	.013	−.314	−.744	.464
	TLG	−.027	.013	−2.439	−2.122	.043
	IA	−.887	.564	−.641	−1.572	.128
	TE	.084	.034	1.669	2.474	.020
	Pr	.195	.440	.212	.444	.661
	CD	.015	.012	2.396	1.237	.227
	Inc	.009	.155	.076	.060	.953
	Exp	.033	.143	.153	.232	.818
	FVEvAOCI	−.012	.012	−.703	−1.022	.316

This table presents the significance (Sig.) of the correlation between the dependent variable FVEvROCI and the independent variables listed, as well as the nature of this correlation: positive or negative, considering the value of B.

Computed using SPPS Statistics v26 software.

Regarding the influence order of the independent variables on the FVEvROCI dependent variable, it is as follows: TE, TLG, IA, CD, FVEvAOCI, CCE, Pr, Exp, and Inc. It can be seen that in this model, the order of influence of the independent variables on the dependent variable is different. Among the most significant changes, it can be mentioned that TE occupies the first place in the current model, while in the previous one, it occupies the 5^th place. Also, the Inc and Exp were among the first ones in the previous model, while in the current model, they occupy the last places. The relationship of the TLG variable with the dependent variable is a significant negative one, which suggests that the analyzed population of banks does not frequently apply the evaluation of loans granted through elements of other comprehensive income – and therefore does not impact the value of FVEvROCI -, but rather through elements of profit or loss, as in the previous model was proved. In the case of an evaluation through other comprehensive income of the TLG, it can be appreciated that the fair value adjustments following the recognition date are not positive ones. The correlation of the dependent variable with the IA variable is not statistically significant and is negative, which suggests that the revaluation of IA does not have a considerable contribution in constructing the FVEvROCI variable. These results are sustained by the fact that the intangible assets with finite life (which are the majority) are evaluated at amortized cost with recognition of depreciation in profit or loss and just those with indefinite life – like goodwill or brand names – are annually evaluated at fair value through other comprehensive income. Regarding the correlation of the CD with the dependent variable, this is not statistically significant, considering that the deposits are usually evaluated at amortized cost with a reflection in profit and loss elements. To better understand the influence of these variables on FVEvROCI, the accounting mechanism of the fair value evaluation method through other comprehensive income will be presented below Figure 2.

Figure 2. Accounting mechanism for the evaluation method through other comprehensive income.

The figure presents a straight line that represents the life of an asset evaluated at fair value through other comprehensive income from year N when it enters at input value (Iv) to year N+e when it goes out from the company’s management at output value (Ov). It can be seen that the value of assets changes in each year of their life.

Elaborated by the authors.

where the segment represents the life of the asset measured at fair value through other comprehensive income (FVEvAOCI);

• N: year when the FVEvAOCI’s entries into management; N+e: year when the FVEvAOCI exits from the management;

• I_v: FVEvAOCI input value; O_v: FVEvAOCI output value;

• V₁, V₂, V₃, V₄, ……: the fair value at which the asset is valued in each of the financial years over its entire life until year N+e.

Therefore, the recognition of fair value (F_v) earnings or losses through other comprehensive income (OCI) is based on the following mechanism:

if

1. F_v > I_v, this results in earnings recognized in equity items, thus generating an increase in the revaluation reserve;

2. F_v < I_v, this results in a loss recognized in equity items, thus generating a decrease in the revaluation reserve;

3. F_v = I_v, there is no change in the revaluation reserve.

These revaluation earnings or losses will remain recognized in the revaluation reserve account until the revalued financial instrument that represents financial assets is derecognized. Only then will the positive or negative revaluation difference be recognized in the income or expense accounts – which will affect the income statement – or will remain in equity items such as retained earnings in the case of equity financial instruments. Therefore, we can appreciate that it is obvious why the variables Inc and Exp have the least significance in the construction of the FVEvROCI variable. We can also see that an increase in the value of this variable will also lead to an increase in equity items, which means that there is a positive correlation between the dependent variable and the TE variable, highlighted also in Table 8. As for the correlation between the dependent variable and the CCE independent variable, it is an inversely proportional one. This derives from the fact that financial instruments that are valued at fair value through OCI are not held for trading, which means that no earnings can be realized from their sale at fair value that would be reflected in an increase in cash and cash equivalents through the receipt of the appropriate consideration.

The design of the econometric model having the earnings on financial assets valued at amortized cost (EFAEvAC) as the dependent variable was not possible, as their value is not subject to the influence of market fluctuations, their value remaining constant throughout their lifetime. This means that it is not possible to register potential earnings resulting from their valuation, because they are kept at a minimum level of risk, rather than for an increase in their profitability. The accounting mechanism of the amortized cost method can be seen below Figure 3:

Figure 3. Accounting mechanism for the evaluation method at amortized cost.

The figure presents a straight line that represents the life of an asset evaluated at amortized cost from year N when it enters at input value (Iv) to year N+e when it goes out from the company’s management also at input value (Iv). It can be seen that the value of assets does not change in any year of their life.

Elaborated by the authors.

where the segment represents the useful life of the asset evaluated at amortized cost (AEvAC);

N: year of entry into the management of the AEvAC; N+e: year of exit from the management of the AEvAC;

Iv: input value of AEvAC;

Therefore, because AEvACs cannot generate earnings as a result of their revaluation, it was not possible to build an econometric model on the same concept as for fair value valuations. Therefore, as Andrzejewski, Dunal, and Ozga (2018) and Brito and Judice (2021) stated, we can estimate that by applying valuation methods at fair value, an increase in the profitability of the assets thus valued is determined, having a direct impact on the result, which is not possible by applying the method of valuation at amortized cost. Although in this context we can’t talk about earnings or losses resulting from evaluation with a direct impact on a bank’s result, there are some situations when the financial assets evaluated at amortized cost can generate an impact on the profitability of a bank, through the expenses generated by the recognition of provisions of expected credit losses or of expenses for incurred losses. Based on these considerations, H2 is also confirmed: The impact of applying financial asset valuation methods on the financial performance and sustainability of credit institutions is different depending on the implemented method.

Model no. 3. For the design and validation of the FSI, the following multiple linear regression is considered, where the considered independent variables are highlighted:

$\mathit{FSI}={\alpha }_3+{\beta }_1\cdot \mathit{ROE}+{\beta }_2\cdot \mathit{ROA}+{\beta }_3\cdot \mathit{RMA}+{\beta }_4\cdot \mathit{FVEvRPL}+{\beta }_5\cdot \mathit{FVEvROCI}+$

(4a) ${\beta }_6\cdot \mathit{AI}+{\epsilon }_3$(4a)

Table 9 shows the existence of a correlation level of 0.675 between the analyzed variables, a lower level than in the first two modeling runs, but which nevertheless highlights its significance in determining the financial sustainability index. The variation ratio R Square shows that the variation of the independent variables explains 45.5% of the variation of the FSI dependent variable. This means that in addition to the considered variables, there are other factors with significant influence on its value, such as liquidity and solvency indicators, and the degree of coverage of non-performing loans, whose absence from this regression was motivated in the research methodology section of the paper. Also, the constructed model is validated by the fact that the Sig. value did not exceed the statistical significance threshold, which was set at 0.05.

Table 9. Model 3 summary.^b

Model	R	R Square	Adjusted R Square	Std. Error of the Estimate
1	.675^a	.455	.313	.092099004038906

This table presents the summary of the third econometric model, highlighting the level of correlation between the following dependent and independent variables.

^aPredictors: (Constant), AI, ROA, FVEvRPL, FVEvROCI, RMA, ROE

^bDependent Variable: FSI

Computed using SPPS Statistics v26 software.

The regression coefficients needed to complete the multiple linear regression 4a are shown in Table 10 and they will be used to fill it in as follows:

(4b) $\mathit{FSI}=0,812-1,233\ast \mathit{ROE}+7,210\ast \mathit{ROA}-0,007\ast \mathit{RMA}+0,003\ast \mathit{FVEvRPL}-0,045\ast \mathit{FVEvROCI}-1,144\ast \mathit{AI}$(4b)

Table 10. Model 3 coefficients.^a

Model		Unstandardized Coefficients		Standardized Coefficients	t	Sig.
		B	Std. Error	Beta
1	(Constant)	.812	.131		6.221	.000
	ROE	−1.233	.753	−.671	−1.637	.115
	ROA	7.210	6.954	.479	1.037	.311
	RMA	−.007	.090	−.025	−.074	.942
	FVEvRPL	.003	.025	.031	.132	.896
	FVEvROCI	−.045	.024	−.468	−1.862	.075
	AI	−1.144	1.076	−.195	−1.063	.299

This table presents the significance (Sig.) of the correlation between the dependent variable FSI and the independent variables listed, as well as the nature of this correlation: positive or negative, considering the value of B.

Computed using SPPS Statistics v26 software.

Therefore, the influence order of the independent variables on the FSI dependent variable is as follows: FVEvROCI, ROE, AI, ROA, FVEvRPL, and RMA. It can be seen that there is an inverse correlation between the ROE variable and the FSI index. This situation can materialize in the case of recognition of a loss or adjustment related to the financial asset (loans granted) through OCI, which decreases the equity elements but does not influence the net profit, thus causing an increase in ROE and a decrease in FSI. An increase in ROE may also occur in the case of the reversal of the income from expected loan loss provisions when these have not been used due to the non-materialization of the risk for which they were established. In this context, the FSI variable is not significantly reduced.

There is also a significant negative correlation between AI and FSI, which means that the higher the amount of expected loan loss adjustments, the more the value of FSI is decreased by adjusting the value of loans granted by the impaired value at the time the risk materializes. This implicitly leads to a decrease in the results and financial performance of the credit institutions. This result is also consistent with the research conducted by Percevic (2022), Pastiranova and Witzany (2021), and Ãollaku, Ahmeti, and Aliu (2021) who point out that the transition phase from the recognition of realized loan loss adjustments to the recognition of expected loan loss adjustments has a significant influence on the recognition of additional loan impairment. This context can cause a certain level of instability and reconsolidation of capital in a short and medium-term perspective. Following these considerations, the H3 is confirmed: Recognition of expected credit loss adjustments has a significant negative impact on the financial position and performance of credit institutions.

It can be seen that the correlation between ROA and FSI is a positive one, which means that as the return on assets increases, the financial sustainability of credit institutions will also increase. In the case of the present research, this fact is materialized by the increase in the return on financial assets valued at fair value (H₂), by recording earnings generated by this category of assets. This is in contradiction with the results obtained by Ãollaku, Ahmeti, and Aliu (2021) who demonstrated a decrease in the ROA indicator as a result of the implementation of IFRS 9 provisions for the recognition and valuation of financial instruments. The correlation of FSI with RMA is negative but quite insignificant. However, future cash flows are also affected in the case of impaired loans, as interest income is calculated based on the net book value and loss adjustments, which would result in a decrease in operating interest income (ACCA Global 2022b).

Regarding the impact of the FVEvRPL and FVEvROCI variables, we can see that their behavior toward the FSI dependent variable is different, with FVEvRPL showing a positive correlation and FVEvROCI a negative one. This highlights the fact that a valuation of financial assets at fair value through PLA positively influences the level of financial sustainability of credit institutions, while the valuation of financial assets at fair value through OCI does not have a positive impact on FSI. This can be determined by the fact that, in the case of valuation through PLA, valuation earnings are recognized each financial year in the profit and loss account. In contrast, in the case of valuation through OCI, the generated earnings are carried in equity accounts until the assets subject to these gains are derecognized, thus making it impossible to immediately reflect the return on these assets in the profit and loss account. The level of significance of the FVEvRPL and FVEvROCI variables in the calculation of the FSI may also be influenced by the number of statistical observations on which they were constructed, FVEvRPL being constructed based on 25 statistical observations and FVEvROCI based on 36 statistical observations.

Given these considerations, we can see a partial confirmation of H₁: The implementation of the provisions of the IFRS framework leads to an increase in the financial performance and sustainability of credit institutions. This is because, among all three variables (FVEvRPL, FVEvROCI, and AI) that materialized as a result of the implementation of the provisions and requirements of IFRS 9, only one positively influences the financial sustainability of credit institutions, namely FVEvRPL. The other two show a negative correlation with the FSI index.

Conclusions

As can be seen in the results section of the paper, all working hypotheses H₁, H₂, and H₃ were totally or partially confirmed by the results obtained in the research. Therefore, a significant impact of the fair value valuation method through OCI on equity items was found, through the establishment of valuation reserves, with a direct impact on the financial results of credit institutions. It could also be identified that the valuation of financial instruments at fair value leads to an increase in their profitability (implicitly in the ROA indicator), generating related earnings. However, only the ones coming from the valuation at fair value through PLA have a positive impact on the financial sustainability of credit institutions, given the possibility of their immediate recognition in the financial result of the year. This is not possible in the case of the earnings from financial assets valued through OCI, which can be assessed as deferred income until the asset valued is derecognized. Last but not least, the negative impact of adjustments for expected credit losses on the financial position and performance of Romanian credit institutions was also highlighted, which in the medium term creates some instability, but in the long term, reduces the possibility of large and sudden losses materialization. Considering these aspects, we have to mention that one of our paper’s limitations is that it does not present the effects of the recognition of provisions for expected credit losses in the long term for the banks’ profitability, as well as of the chosen evaluation method of financial instruments, as the analysis is made for the first five years after the mandatory implementation of IFRS 9 by banks. We consider that the selected period of analysis is sufficient to consolidate and validate the obtained results, but to get more in-depth results, a longer period of analysis would be recommended and an expansion of the population of banks to those from other emerging economies would be welcomed.

As regards the contribution and originality of the present work, it is because – to the best of our knowledge – there are not many approaches in the literature on the impact of IFRS 9 on the financial performance of credit institutions from the perspective of methods of recognition and measurement of financial assets, thus adding value to the literature. Also, the paper has some policy implications. Therefore, based on the obtained results, the authors recommend that institutional managers and accounting policymakers apply the fair value method of valuation for financial instruments through PLA (when they meet the conditions to be so valued), especially when they want an immediate reflection of the result of the valuation of these items in the profitability indicators of the institution. This method is also recommended if the banks are in a period of development in which they are looking for various sources of funding, given the high interest of investors in financial sustainability and performance indicators and reported financial position. The choice of this method may also be motivated by the objectivity it offers in presenting the result of the financial year, thus making the information more accessible to all categories of investors or other stakeholders. On the other hand, the fair value evaluation method through other comprehensive income is more difficult for stakeholders to understand because only those with a certain level of specialized training can fully understand its mechanism. Also, this method is recommended when the bank has no immediate need to increase reported profitability and can afford a delayed recognition of valuation gains in the profit and loss account. As a result of the recognition of income or gains, expenses or losses from the valuation of financial instruments, in the case of both methods, the overall value of capital is influenced, the determining factor in the choice of one of the two methods being the interest linked to the immediate reflection or not of the generated return.

The interpretation of the results was made through the prism of the interpreters’ knowledge, experience, and reasoning, so this process could have unintentionally fallen under the influence of more or less subjective factors, which is why the results were discussed between the authors to find the most relevant interpretation of them in the context of the research and to eliminate any potential interpretation biases. Regarding some other limitations of the present research, they may refer to the relatively small number of the sample, constructed only from credit institutions in Romania, from which were eliminated those whose financial reporting was not accessible or did not provide relevant and necessary information for the calculation of the selected aggregate indicators. Also, another limitation of the paper can be considered the fact that the indicator of losses from financial assets measured at fair value through PLA and OCI could not be taken into account as a variable, as it did not materialize in the financial statements of the analyzed credit institutions or it presented very insignificant values.

Given the complex nature of financial performance and sustainability evaluation, a future research direction could be to consider other external factors specific to the period of the IFRS 9 implementation and observe their contribution to influencing the financial performance and sustainability of credit institutions, along with the changes introduced by IFRS 9. Among the factors that could be taken into consideration, we can mention the lack of predictability of the tax legislation specific to the Romanian economy, the standard of living of the population, the number of bankruptcies among economic operators – which can significantly influence the level of provisions for expected credit losses and the recognition of actual losses -, inflation rates, the level of ROBOR (interest rate on deposits placed) and other equally important factors. We believe that such research, combined with an extension of the analysis period to highlight the long-term impact of IFRS 9 provisions, would illustrate a much clearer, broader, and more nuanced picture of the impact of IFRS 9 implementation on the performance of Romanian credit institutions.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Veronica Grosu

Veronica Grosu is an accounting Ph.D. Full Professor and head of the Accounting, Audit, and Finance Department at Ștefan cel Mare University from Suceava, Romania. Her teaching, research and interest areas focus on international accounting, management accounting, financial accounting, audit, and finances. She holds a Ph.D. degree in Accounting from the West University of Timișoara.

Liliana Ionescu-Feleagă

Liliana Ionescu-Feleagă is an accounting Ph.D. Full Professor and Dean of the Faculty of Accounting and Management Informatics at the Academy of Economic Studies in Bucharest, Romania. Subjects like IFRS implementation, financial accounting, environmental accounting, consolidated accounting, intellectual capital, and corporate governance represent her teaching and research areas of interest. She holds a Ph.D. degree in Accounting from the Academy of Economic Studies in Bucharest.

Anamaria-Geanina Macovei

Anamaria-Geanina Macovei is a Ph.D. Lecturer in the Department of Accounting, Audit, and Finance at Ștefan cel Mare University from Suceava, Romania. Her teaching and research interests focus on mathematics, financial and actuarial mathematics, mathematical analysis, operational research, and econometrics. She holds a Ph.D. degree in Mathematics from Babes-Bolyai University in Cluj-Napoca.

Marius-Sorin Ciubotariu

Marius-Sorin Ciubotariu is a Ph.D. Lecturer in the Department of Accounting, Audit, and Finance at Ștefan cel Mare University from Suceava, Romania. His teaching and research interests focus on financial accounting, financial accounting expertise, price and cost calculation, audit, and internal control systems. He holds a Ph.D. degree in Accounting from Stefan cel Mare University in Suceava.

Elena Hlaciuc

Elena Hlaciuc is a Ph.D. Full Professor in the Department of Accounting, Audit, and Finance at Ștefan cel Mare University from Suceava, Romania. Her teaching and research interests focus on financial accounting, managerial accounting, creative accounting, financial management, and evaluation of organizations. She holds a Ph.D. degree in Economics from Alexandru Ioan Cuza University in Iași.

Marian Socoliuc

Marian Socoliuc is a Ph.D. Associate Professor in the Department of Accounting, Audit and Finance at Ștefan cel Mare University from Suceava, Romania. His teaching and research interests focus on subjects like accounting policies and options, doctrine and ethics of the accounting profession, forensic and extrajudicial accounting expertise, and public accounting. He holds a Ph.D. degree in Accounting from the West University of Timișoara.

Corina Petrescu

Corina Petrescu is a Ph.D. student in Accounting at Faculty of Economics, Administration and Business at Ștefan cel Mare University from Suceava, Romania. Her main research focuses on accounting and finance, taxation, managerial accounting, and cost calculation.