ICT usage increases workforce geographical diversity

Abstract

The adoption of information and communication technology (ICT) by rural-urban migrants is reshaping job-search mobility, significantly shaping city-level workforce geographical diversity. This study provides compelling evidence of ICT's impact by examining China's cities. We introduce the rural-urban migrant workforce Geographical Diversity Index (GDI), a metric that captures the mobility patterns of 20 million migrant workers across Chinese cities from Q1 2019 to Q4 2023. This study highlights how ICT usage shapes migration dynamics and connectivity across geographic spaces, with implications for labor mobility and urban inclusivity. Using panel vector autoregression models, we establish a causal relationship between ICT usage and GDI, revealing heterogeneous impacts: large cities and male workers benefit more from ICT usage than small cities and female workers. While ICT-driven diversity enhances labor productivity, it also increases migrant workers' job-hunting travel distances, contributing to higher carbon emissions. These findings underscore the dual role of ICT as a facilitator of inclusivity and a source of sustainability challenges, offering critical insights for policymakers aiming to leverage digital tools for equitable and sustainable urban development.

Keywords: ICT usage; panel vector autoregression (PVAR); rural–urban migrants; workforce geographical diversity.

Fig. 1.

Mechanism of ICT’s impact on rural–urban migrant workforce geographical diversity. The different-colored icons represent migrant workers from various regions. (A) Low ICT Usage: In a low ICT usage context, rural–urban migrants have limited job-search areas (yellow regions), restricting them to nearby small or medium-sized cities. As a result, workforce geographical diversity remains relatively localized and limited to smaller urban areas. (B) High ICT Usage: with increased ICT usage, the job-search areas of rural–urban migrants expand significantly (larger yellow regions), enabling access to a wider range of urban opportunities. Migrants are more likely to choose large cities with better job prospects, enhancing workforce geographical diversity in large cities. However, this dynamic reduces workforce geographical diversity in smaller cities as they attract fewer migrants.

Fig. 2.

Index of rural–urban migrant workforce geographical diversity. (A) The full panel dataset is used to estimate the effects of distance and GDP on rural–urban migrants’ employment choices Eq. 1. The estimated parameters for distance and GDP are then applied to predict employment choices for each cross-section Eq. 3. (B) The differences between observed and predicted employment choices are used to estimate the influence of nondistance and non-GDP factors, capturing both spatial and temporal dimensions ($B_{\mathit{ij}}$, Eq. 4). (C) The impact of nondistance and non-GDP factors on workforce geographical diversity is reflected in the area of Φ Eqs. 5 and 6.

Fig. 3.

Spatiotemporal distribution of GDI. (A–E) Spatial distribution of GDI in the first quarter of each year: (A) 2019, (B) 2020, (C) 2021, (D) 2022, and (E) 2023. Darker blue colors indicate higher levels of GDI. Gray areas indicate missing data.

Fig. 4.

The Impact of ICT on GDI. (A) PVAR regression results for the full sample;(B) PVAR regression results for cities in the top 50% by GDP; (C) PVAR regression results for cities in the bottom 50% by GDP; (D) PVAR regression results for males under 30; (E) PVAR regression results for females under 30; (F) PVAR regression results for males aged 30 to 60; (G) PVAR regression results for females aged 30 to 60.

Fig. 5.

PVAR regression coefficients for randomly selected subsamples of cities. (A) Results from 1,000 iterations of sampling 30 cities at a time. (B) Results from 1,000 iterations of sampling 50 cities at a time. (C) Results from 1,000 iterations of sampling 100 cities at a time.

Fig. 6.

Assessment of social effects of migrant geographical diversity. (A) The relationship between GDI and TFP. (B) The relationship between GDI and carbon emission.