This paper analyses the impact of the 2017–2018 personal income tax (PIT) reforms in Latvia on long-term voluntary saving behaviour in tax-favoured instruments, namely the third pension pillar and life insurance policies with a savings component. Using rich administrative microdata from personal income tax declarations covering the period 2010–2023, we study behavioural responses along both the extensive margin (entry and exit) and the intensive margin (contribution levels and income elasticity). We document that the reforms substantially reduced opportunity-seeking behaviour, particularly among high-income individuals, maximizers of PIT refunds, and those aged 55 and above. Entry responses are considerably stronger than exit responses, highlighting pronounced behavioural inertia. Contribution levels increase with income and participation duration, but income elasticity declined after the 2018 reform. The findings underline the importance of policies that broaden initial participation, as wider coverage is likely to translate into higher aggregate retirement savings over time.

What are the economic and welfare effects of liberalizing access to private retirement wealth in the U.S.? In the empirical section, I document that early withdrawals from individually managed accounts are infrequent due to early-withdrawal penalties, but conditional on withdrawal, amounts are sizeable, whereas withdrawals from employer-sponsored retirement plans are virtually absent due to strict withdrawal rules. Furthermore, the minimum withdrawal rules bind for a substantial fraction of retirees. Guided by these facts, I build a rich lifecycle model that features three frictions in accessing retirement wealth: early-withdrawal penalties, restricted access to employer-sponsored retirement savings, and required minimum withdrawals during retirement. I evaluate a reform that removes these frictions, mimicking recent proposals to make retirement accounts more flexible. The cohorts who are alive at the reform implementation gain on average 0.23% (CEV), while cohorts born after the reform experience welfare losses of 0.15%.

Institutions routinely delegate complex measurement to external experts, creating scope for expert judgment and discretion to shape reported information. This paper studies delegated measurement in state and local public pensions, where actuaries value long-term benefit promises that determine reported funding, required contributions, and fiscal risk. I use novel data linking plan-level financials to actuarial firms and valuation assumptions. Leveraging actuarial firm switches and malpractice litigation, I show that actuaries materially affect reported liabilities and funding status. Small actuarial firms, in particular, are associated with more favorable reporting. When an actuarial firm is sued for malpractice, its other public pension clients subsequently report stronger financials and use more aggressive assumptions, consistent with reputational and market incentives shaping expert discretion. Counterfactual exercises suggest that these forces shift aggregate reported liabilities by tens to hundreds of billions of dollars.