How AI and technology’s skill-bias impact the delayed retirement desire in China

Abstract

Abstract China has introduced delayed retirement policies to sustain its pension system, threatened by population aging. We present theories to argue that such policies, if applied universally across the private and public sectors, may ironically exacerbate the problem by increasing rent-seeking, which decreases output and tax revenue. Moreover, the theory explains our empirical findings that the policy’s popularity varies significantly between the private and the public sectors. We model the optimal retirement age as a function of human capital and occupation. The administrative authority and opportunities for human discretion empower public-sector managers to extract rents from producers. The universal delayed retirement policy raises the prospect of higher rent extraction and, hence, lowers incentives to produce, implying a divergence in the popularity of this policy between rent-seekers and producers and reducing funds for pensions. We model how, through AI-enabled e-governance, the government can reduce human discretion, thereby reducing rent-seeking and boosting production to sustain pensions. We calibrate our model to China’s economy. The simulations of the calibrated model align well with the empirical findings regarding the diversity in the desire for delayed retirement, illustrate how artificial intelligence (AI) influences optimal retirement ages, and show how the optimal use of AI changes as technology advances.