When does employee turnover matter? Analyzing the role of organizational memory in the federal IT workforce

Abstract This paper proposes a theory of organizational memory predicting that task knowledge retained in organizational structures (i.e., agency centralization and task standardization) buffers the disruptive effects of employee turnover. Using a dataset of information technology (IT) projects implemented by 94 US federal agencies between FY2015 and FY2021, estimates from fractional panel probit models reveal that both centralized authorities under chief information officers and task standardization reduce project delays associated with IT workforce turnover; they are particularly effective in mitigating knowledge loss from voluntary and managerial position turnover. This evidence highlights the contingent nature of the turnover–performance relationship, suggesting that government agencies with high turnover may experience fewer delays when organizational structures retain task knowledge.