Working papers, computation notes and open-source code from Risk & Return. Papers link to SSRN and code to GitHub; every item also has a permanent archive with a DOI on Zenodo.
Nari Han
Code, data and saved outputs for a computation note on Great Britain's 2024 seat totals under a factor-based first-past-the-post model. A Poisson-binomial recursion is paired with a hybrid quasi-Monte Carlo and Gauss–Hermite integral, and the note documents two errors that a standard error bar did not catch. A fixed case study, not a forecast.
Nari Han
Rebuilds the latest UK rigid-central (RC) inflation figure from public data. Its gap to headline inflation comes mostly from item coverage and the aggregation method. Public output-gap measures do not back the claim that RC tracks the gap better, and RC does not improve out-of-sample forecasts of headline inflation.
Numerical tests that a heterogeneous-agent New Keynesian (HANK) code should pass at limits with known answers and at exact aggregate identities, built on sequence-jacobian (SSJ) 1.0.0. Each test reports how the error shrinks toward its limit and the observed convergence order, and the model-independent checks can be reused on other HANK codes.
Nothing here is peer reviewed unless stated. Code and notes are research material, not investment advice.