Working Papers
Working Papers
Wage and Tax-Bracket Erosion: Where Inflation Losses Fall (Draft and slides coming soon)
Abstract:
This paper examines how inflation affects real after-tax earnings through the interaction of wage erosion and tax-bracket erosion. When tax thredholds are imperfectly indexed to inflation, bracket erosion can amplify the transmission of wage erosion to after-tax earnings while increasing government tax revenues. Because of the nonlinearity of the tax schedule, this amplification varies across the earnings distribution. A tractable model shows that the distributional incidence of inflation depends on two factors: workers’ positions relative to tax thresholds and the relative degree of wage and tax-bracket indexation. As a result, for a given tax schedule, the ranking of inflation losses across earnings levels can reverse depending on whether wage or bracket erosion dominates. To characterize these effects, I introduce two novel measures of the incidence and dispersion of inflation-induced losses. The model reveals a policy trade-off between protecting real after-tax earnings, limiting the dispersion of inflation losses, and preserving tax revenues. In particular, full tax-bracket indexation reduces tax revenues but makes inflation regressive. By contrast, partial tax-bracket indexation that exceeds wage indexation can achieve distributional neutrality at a substantially lower fiscal cost.
Keywords: tax-threshold indexation, wage indexation, tax policy, inflation incidence.
Presented at:
7th Naples Summer School of Economics and Finance, Poster session; 6th Sailing the Macro Workshop 2026, Poster session; Ph.D. NSE internal seminar; Ph.D. LBS internal seminar; 5th Ph.D. and Post-Docs NSE Workshop; Bank of Italy Conference - 37th SUERF Colloquium, Poster Session.
Work in Progress
Time-Varying Policy (with Andrea Ferrara and Francesco Simone Lucidi)
Other Research
Are You Series? Mining Temporal Rules From Financial Time Series (with Livia Blasi, Eleonora Laurenza, Luigi Bellomarini, Emanuel Sallinger) (Draft coming soon)
Abstract:
Understanding conditional co-movement among financial and macroeconomic time series is essential for risk assessment and financial stability, yet its identification depends on assumptions about market regimes and temporal relationships. We propose a neuro-symbolic framework for discovering verifiable co-movement rules before committing to an econometric specification. Using specification mining, we represent seven financial phenomena as reusable DatalogMTL templates, fit their temporal parameters to data, and evaluate them out of sample with independent symbolic verification. To improve scalability, a neural surrogate trained on synthetic data accelerates correlation labeling while preserving measurable accuracy. Experiments on real and simulated financial data identify robust conditional relationships, expose regime-dependent findings, and show low false-alarm rates, with detection power improving on longer samples. Our approach complements econometric analysis with explicit, reusable, and empirically verifiable temporal rules.
Keywords: temporal reasoning, rule mining, DatalogMTL, financial co-movement, knowledge distillation.