Digital Exodus or Digital Magnet? How State Data Privacy Laws Reshape the Technology Sector

Abstract

Do state data privacy laws drive technology firms away, or do they reshape the technology sector's composition? Between 2020 and 2025, nineteen U.S.\ states enacted comprehensive consumer data privacy legislation modeled on California's landmark CCPA, though only seven have effective dates with sufficient post-treatment data within our 2015–2024 data window (California, Virginia, Colorado, Connecticut, Utah, Oregon, and Texas); the remaining twelve states' laws either take effect in 2025 or later or provide insufficient post-treatment exposure (Montana, effective October 2024, contributes at most one quarter) and are coded as not-yet-treated. This staggered adoption creates substantial cross-state variation in the regulatory environment facing data-intensive industries. Using a staggered difference-in-differences design with the Callaway-Sant'Anna (2021) estimator, we examine the effect of these laws on technology-sector employment and new business formation. Drawing on quarterly state-level employment data from the BLS Quarterly Census of Employment and Wages and state-level business applications from the Census Bureau's Business Formation Statistics covering 2015–2024, we find evidence that privacy laws reduce employment in data-intensive Software Publishing (NAICS 5112) by approximately 7.7%—driven largely by California's experience as the first mover with the longest post-treatment window—with a modest negative effect on Computer Systems Design (NAICS 5415), while the net effect on total Information Sector (NAICS 51) employment is statistically indistinguishable from zero. The aggregate null masks uneven regulatory costs across subsectors: the burden falls disproportionately on software publishers, consistent with compliance costs affecting data-intensive firms most heavily. Business formation data show no significant aggregate effect on total new business applications. These results suggest that privacy regulation does not destroy technology jobs per se but that regulatory costs fall unevenly across technology subsectors, with data-intensive software publishing bearing the brunt while the broader Information Sector adjusts through compositional reallocation rather than net contraction. Our findings have direct implications for the ongoing federal data privacy debate, suggesting that aggregate employment effects are modest but distributionally consequential.