Abstract
This study examines the dual impact of artificial intelligence (AI) on employment quantity and employment quality in the Iranian banking industry. Using a two-stage empirical framework, the study first constructs an Employment Quality Index (EQI) through Principal Component Analysis (PCA) based on indicators related to skill composition, staff specialization, and technology-related employment structure. It then applies a Panel Seemingly Unrelated Regression (Panel-SUR) model to panel data from 8 Iranian banks over the period 2019–2024. The results show that AI adoption does not have a statistically significant negative effect on employment quantity, suggesting the absence of a strong displacement effect in the sample period. In contrast, AI has a positive and statistically significant effect on employment quality, with an estimated elasticity of 0.28. Robustness checks, including the exclusion of state-owned banks and the use of alternative proxy measures, confirm the stability of the findings. Overall, the evidence suggests that digital transformation in Iranian banking is associated with a shift toward more skill-intensive and higher-value employment structures rather than labor replacement.
Purpose/Aims:
The primary objective of this study is to empirically examine the dual impact of artificial intelligence (AI) adoption on both the quantity and the structural quality of employment in the Iranian banking sector. Specifically, the study seeks to contribute to the ongoing debate between technological displacement and the “reinstatement effect” by investigating whether AI-driven digital transformation reduces the overall workforce or instead reshapes employment toward more specialized and skill-intensive roles. In addition, the research evaluates how technological adoption influences a multidimensional Employment Quality Index (EQI), which captures the evolving skill composition of banking employment. By estimating the elasticity of employment quality with respect to AI adoption, the study aims to provide evidence-based insights for policymakers and banking managers regarding the labor market implications of digital transformation and the growing importance of human–AI complementary skills.
Methodology & Framework:
This study adopts a quantitative empirical framework using panel data from 8 Iranian banks over the period 2019–2024. The methodological approach is implemented in two stages. First, to capture the multidimensional structure of employment quality, a composite Employment Quality Index (EQI) is constructed using Principal Component Analysis (PCA). The index integrates indicators reflecting the skill composition of the workforce, including the share of IT-related personnel, the proportion of employees with postgraduate education, and the relative intensity of IT-related expenditures. All variables are standardized prior to estimation to ensure comparability.
In the second stage, the study employs a Panel Seemingly Unrelated Regression (Panel-SUR) model to estimate the simultaneous effects of AI adoption on two dimensions of employment: employment quantity and employment quality (EQI). This econometric framework allows for correlated error terms across the two equations and therefore provides more efficient estimates than separate panel regressions. The model further includes control variables such as bank size, profitability, workforce reskilling intensity, and task routineness in order to isolate the specific impact of technological transformation. To enhance the reliability of the results, lagged technological variables and bootstrap standard errors are applied, and several robustness checks—including alternative specifications and sample adjustments—are conducted.
Findings:
The empirical results reveal a differentiated impact of AI adoption on the two dimensions of employment in the Iranian banking industry. The Panel-SUR estimates indicate that AI adoption does not have a statistically significant negative effect on employment quantity, suggesting that large-scale technological displacement has not occurred during the sample period. However, AI adoption demonstrates a positive and statistically significant effect on employment quality, with an estimated elasticity of approximately 0.28. This result implies that a 1% increase in AI-related technological adoption is associated with a 0.28% improvement in the Employment Quality Index. The PCA results further indicate that this improvement is mainly driven by changes in the skill composition of the workforce, including a higher share of specialized human capital and greater technological intensity within banking operations. Additional robustness tests confirm the stability of these findings across alternative model specifications.
Discussion:
The findings suggest that the Iranian banking sector is experiencing a labor reconfiguration process rather than a simple technological substitution of labor. Consistent with the task-based framework of technological change, AI appears to complement human labor by automating routine and data-intensive tasks while increasing the relative importance of analytical, technological, and interpersonal skills. The absence of a significant reduction in employment quantity, combined with the positive effect on employment quality, indicates the presence of a “reinstatement effect,” whereby new skill-intensive tasks emerge alongside technological adoption. Consequently, employment structures in the banking sector are gradually shifting from routine administrative functions toward more knowledge-intensive and technology-oriented roles.
Conclusion & Implications:
Overall, the results indicate that AI adoption acts as a catalyst for the qualitative transformation of employment within the Iranian banking industry. Rather than causing widespread job losses, technological progress is associated with an upgrading of workforce skills and a greater emphasis on specialized human capital. These findings carry important implications for both bank managers and policymakers. Financial institutions should prioritize continuous upskilling and reskilling initiatives to prepare employees for increasingly technology-integrated work environments. At the policy level, educational and training systems should place greater emphasis on digital literacy, analytical capabilities, and interdisciplinary skills that complement emerging technologies. In this context, the transition toward AI-integrated banking can be understood not as a threat to employment, but as an opportunity to enhance productivity and foster a more skill-intensive labor market.