JR analysis of California Proposition 40: Overview · Open Output (here) · Open Analysis: Report and Slides · Open Materials Also in this project: Comparison overview · BSZ analysis overview

What the Billionaire Tax Act is worth to the state, by the JR method

Jaros and Rauh (JR; NBER, September 2026; the first version was by Rauh, Jaros, Kearney, Doran and Cosso) estimate the Act's fiscal effect as its one-time wealth tax revenue minus the present value of income tax lost for ever when billionaires leave: NPV = WT − f·C / r, where NPV is the net present value, WT the wealth tax collected once, f the share of the income tax base that leaves, C the annual California income tax that base pays, and r the discount rate. They report the average over 100,000 random draws of those inputs. This page recomputes that average, and the share of draws below zero, for every combination of seven inputs. It is this project's open output: all 2,187 combinations are precomputed, so the outcomes update as you click.

Defined as the one-time wealth tax revenue minus the present value of California income tax lost for ever to billionaires who leave (NPV = WT − f·C / r), averaged over the authors' Monte Carlo draws: 100,000 random runs of their model.

The inputs

How this grid differs from the paper. Each cell is the exact expectation of the authors' formula (a closed form, since the draws are independent uniforms) and an exact share of negative outcomes (numerical integration, accurate to about 0.01 points), not a new random draw. At the main estimate's setting this gives The gap is not an error: the paper prints one seeded run of 100,000 draws (a run started from a fixed seed number, so it gives the same draws every time). Monte Carlo error on the mean is about ±0.5 ($B), four standard errors (one standard error is the printed SD, 37.6, over √100,000, about 0.12). The labels after each input name (data, research, guesswork) follow the report's every input, by origin table. The litigation dial and the growth dial are readings the paper's text states but its code does not implement; see mismatches 3 and 5.

Outcomes

Outcomes for this setting

Exact expectations for the setting you chose, next to the NBER setting (the main estimate); the first row is the authors' printed mean, from their seeded run. $B are billions of dollars, present value at the time the tax is collected.
Outcome	This setting

¹ The main estimate is the authors' own headline: the mean NPV in Jaros and Rauh, NBER (the National Bureau of Economic Research) chapter c15504, September 2026 (PDF p.1).