Unemployment is a major concern in developing countries, exacerbated by global crises and events like the COVID-19 pandemic. The concept of 'employment pressure,' encompassing negative emotional states related to job insecurity and socioeconomic uncertainty, is crucial. Technological changes, particularly the digital transformation driven by the Fourth Industrial Revolution, significantly impact employment. While some studies highlight potential job displacement due to automation, others emphasize the creation of new jobs and opportunities. The existing literature lacks a comprehensive understanding of the mechanisms through which digital transformation influences employment pressure. China's national smart city pilot (SCP) policy, implemented in phases between 2012 and 2014, provides a valuable natural experiment to study this relationship. This study investigates whether SCP effectively impacts employment pressure and explores the underlying mechanisms, bridging the gap in current research by integrating both macro and micro-level perspectives.

The impact of digital transformation on employment is a complex issue. Studies by Autor et al. (2003) and others suggest both positive (creation effects) and negative (substitution effects) impacts on labor demand. Some research indicates that digital technologies primarily displace middle-skilled workers, while others argue that new technologies create new jobs, particularly for high-skilled individuals. However, there's limited research on the mechanisms through which digital transformation affects employment pressure. This paper contributes by examining the multifaceted impact of smart city policies on employment, combining macro and micro effects into a single analytical framework, and investigating specific policy effects rather than just general digital transformation.

This study uses a difference-in-differences (DID) model with panel data from 2003-2019, comparing employment in Chinese cities designated as smart city pilot program participants (treatment group) with non-pilot cities (control group). City-level data includes employment numbers, population, wages, GDP per capita, and retail sales. Firm-level data from the CSMAR database encompasses employment, revenue growth, costs, financial ratios, and tax expenses for listed companies. The core explanatory variable is the interaction between smart city pilot status and a post-policy time dummy. Control variables account for factors affecting employment. To address potential endogeneity, a parallel trend test using an event study approach is conducted. A placebo test involves artificially shifting the policy implementation time. Additional robustness checks include sample trimming, inclusion of benchmark variables (to account for city characteristics), PSM-DID, and controlling for joint fixed effects. Further analysis explores the mechanism of impact through configuration optimization, technological upgrading, siphoning effects, factor substitution, and efficiency improvement. Spatial econometric models are used to analyze spatial spillover effects. The study further investigates the heterogeneity of effects across city size, location, human capital, government intervention, digital infrastructure, firm age, ownership, industry, factor intensity, and worker education.

The DID analysis shows that the SCP policy significantly increases employment in pilot cities (7.43% at the city level and 16.9% at the firm level), alleviating employment pressure. Parallel trend and placebo tests confirm the validity of the findings. Robustness checks across various model specifications maintain this significance. Mechanism testing reveals that SCP enhances total factor productivity (TFP), particularly for technology-intensive firms, and boosts innovation outputs (patents). Spatial econometric models show a positive direct effect and negative indirect effect on employment in neighboring cities (siphoning effect). Factor substitution effects are present, particularly affecting technology-intensive firms and high-wage workers. SCP stimulates urban economic agglomeration, industrial structure transformation (especially digitalization), and regional innovation. Heterogeneity analysis indicates stronger positive employment effects in large cities, western cities, cities with low education levels and low government intervention, cities with high digital infrastructure, older firms, foreign and private firms, tertiary sector firms, technology-intensive firms, and highly educated workers. Further analysis demonstrates that SCP promotes digital transformation in firms and, surprisingly, leads to reduced average worker wages, particularly in labor-intensive firms.

The findings confirm that China's smart city pilot policy effectively alleviates employment pressure. The positive impact surpasses the job displacement caused by automation, reflecting the creation effect of new jobs. The mechanism analysis highlights the importance of configuration optimization, technological upgrading, and the positive spillover effects of economic agglomeration and regional innovation. The heterogeneous effects emphasize the need for targeted policies considering local contexts and characteristics. The unexpected finding of wage reduction suggests the need for further investigation into the distributional consequences of digital transformation, advocating for policies that ensure the benefits of the digital economy are shared more equitably.

This study demonstrates the significant positive impact of China's smart city pilot policy on employment pressure. The detailed mechanism analysis and heterogeneity testing provide valuable insights for policy design. Future research could focus on refining the measurement of employment pressure, exploring the long-term impacts of the policy, and investigating policy adjustments to mitigate negative effects like wage reductions. Cross-country comparisons can also be done to check the generalizability of the findings.

The study relies on publicly available data, potentially missing information on informal employment and underemployment. The focus on listed companies in the firm-level analysis might not fully represent the employment landscape of all firms. Future research could expand the data set to include informal sectors and smaller firms to obtain a more comprehensive view. The time frame of the study might not fully capture long-term effects.