Electronic payment methods are defined as the exchange of payment values from buyer to seller through an online payment channel, i.e., the conversion of financial transactions from a traditional environment to an electronic environment without the need to visit a financial institution ( Sakanko and David, 2019:144). Some define it as an innovative payment mechanism that uses electronic media instead of cash. Electronic payment systems are described as electronic systems for transferring payment values from the payer to the recipient via electronic payment methods ( Mohammed, Ismael, 2024: 29).

They also enable users of e-commerce applications to exchange money electronically instead of using coins, paper money, or paper checks. Online sellers offer easy, fast, and secure methods for customers to make payments for their products ( Abu Shaala, 2024: 381).

Advantages of electronic payment methods from the perspective of banks: 1.

Strengthening the relationship between the bank and its customers, highlighting that the speed and low cost of electronic transfers are key to ensuring customer satisfaction and reinforcing their confidence in banking operations ( Mohammed, 2025: 242).

Figure 1 shows the advantages that banks gain from applying electronic payment methods:

Liquidity risk refers to an institution’s inability to meet its financial obligations when due or pay them at a higher cost, as well as its inability to finance increased assets without having to liquidate assets at unfair prices or resort to high-cost sources of funding ( Ahmed, 2013: 206).

Liquidity risk is generally defined as the risk of not being able to provide cash when needed ( Bessis, 2015: 4).

It is also possible that a company or institution may face difficulty in meeting its short-term obligations due to a lack of cash liquidity. In other words, it is the risk of not being able to convert assets into cash quickly and at the required value to cover financial obligations ( Ghdhaib, 2024: 114).

These risks negatively affect a bank’s ability to both maintain the provision of its services and carry out its operations, as well as to take advantage of the opportunities offered by the banking environment. Liquidity risk arises as a result of internal and external factors ( Al-Hashemi & Al-Rifai, 2022: 106). Internal factors include weak liquidity planning within the bank, which leads to a mismatch between assets and liabilities in terms of maturity, the sudden transformation of some contingent liabilities into actual liabilities, or the difficulty of converting liquid balances due to investments and uses of similar degrees. External factors include a recession or economic stagnation affecting the national economy, which leads to the failure of some projects and their inability to meet their obligations to the bank’s creditors on the due date, or severe crises affecting the financial markets.

1.

Deepening and Developing Capital Markets and Facilitating Access to Finance:

Electronic platforms enable banks and companies to access capital markets quickly and easily (such as issuing and purchasing commercial paper or short-term bonds electronically). This provides alternative and liquid sources of funding beyond traditional deposits, reducing liquidity risks and contributing to financial stability ( Kasri et al., 2022: 3).

In this section, we will attempt to measure the relationship between the dependent variable (Liquidity Risk Index, represented by liquid assets relative to total assets) and the first independent variable (total electronic cards) and the second independent variable (number of electronic wallets) using a linear model and monthly values for the period (2017-2023). The sample included the entire Iraqi banking sector, and data published in the Central Bank’s bulletins were used. To measure the relationship between the dependent variable and the independent variables, the ARDL co-integration model was used because it provides efficient estimates, especially with small sample sizes. Furthermore, it does not require the time series to be integrals of the same order and can be applied regardless of whether the variables are integrals of order I(0), I(1), or a combination of both, provided that none of the variables are integrals of order I(2). Therefore, the relationship can be tested using the following function: Y = β 0 + β 1 X 1 + β 2 X 2 + u t (1)

Y = Liquidity Risk Index, represented by liquid assets to total assets (dependent variable)

X ₁ = Total electronic cards (first independent variable)

X ₂ = Number of electronic wallets (second independent variable)

u _t : Random error term for the model.

Based on an economic analysis perspective, the increased use of electronic payment methods will lead to a decrease in banking liquidity risks, i.e., there is an inverse relationship between the two.

We observe from the Phillips-Perron test that the time series was stationary when taking only first differences and with a constant term, The t-value we obtained was above the critical t-value at the 1%, 5%, and 10% significance levels. Furthermore, the results according to the augmented Dickey-Fuller (ADF) test did not differ significantly from those obtained using the PP test, further supporting our conclusions. This argument is strengthened by the fact that the p-values remained below 5%. So, with this evidence in hand, we decided to adopt the alternative hypothesis (H1: B ≠ 0), which states that the series can be considered stationary. Table 1 illustrates this:

After conducting stationarity tests for the time series of the research variables, it became appropriate to use the co-integration methodology according to the ARDL model and estimate the equilibrium relationship in the short and long term. This is because the time series of the variables under study are stationary at the first difference; if they were stationary at the second difference, we could not use the ARDL model. Therefore, this model is the best for measuring and analyzing the relationship between the study variables.

Before applying the ARDL model, the appropriate lag period for each model was selected by using auto-regression for each variable, one after the other, until the model that achieved the best results in the criteria (HQ, SC, AIC, FPE, LR) was obtained and selected. The optimal lag period was a single lag interval (Lag = 1).

To test for the long-term equilibrium relationship between the dependent and independent variables (the presence of co-integration), the F-statistic was calculated

Table 2 indicates that the calculated F-statistic value is greater than the critical values at the significance levels of (2.5%, 5%, 10%). This means that there is a co-integration relationship between the study variables, i.e., a long-term equilibrium relationship that runs from the independent to the dependent variable using the Bounds Test, as shown in Table 2.

Testing the auto-correlation between errors: This is done through the Lagrange factor for the serial correlation between the residues (Brush-Godfrey), as shown in Table 5 since the results of this test show the absence of auto-correlation between the errors or residues. Since the calculated F value is (0.072) and the probability greater than 5% is (0.78), and since the probability of Obs*R-squared is (0.77), we accept the null hypothesis stating that there is no auto-correlation between errors.

Auto-regressive Conditional Variance Instability Test (ARCH): From Table 5 we can see that the residuals do not suffer from the variance difference problem, as the results showed that the calculated F value was (0.329) with a probability value of (0.856), which is greater than 5%. Also, the probability of Obs*R-squared was (0.033), which is also greater than 5%. This means accepting the null hypothesis, which states that there is no variance difference problem.

To verify that the data had not undergone structural changes, both the cumulative sum test of repeated data (CUSUM) and the CUSUM of squares were performed.

Looking at Figure 2, the two tests show a degree of stability and consistency between the long-term and short-term parameters, as the cumulative sum of recurrent residuals (cusum) test statistic fell within the critical limits at a significance level of 5%.

While the cumulative sum of squares test showed that part of the test statistics appeared outside the critical limits during 2021-2022, this reflects the existence of structural changes as a result of the return to life after the Corona crisis and the increase in the number of banks participating in electronic payment operations.