New version of our paper (w/ @tzkli.bsky.social) on estimating pre-trends placebos in regression imputation DiD!

3 big changes!

2025 Best Statistical Software Award

Thrilled to share that **fect** has won the 2025 Best Statistical Software Award from the Society of Political Methodology. We're honored!
polmeth.org/statistical-...

To celebrate, we've just released fect v2.0.5 on CRAN & Github 🎉

Bottom line: "Deliverism" itself may not help the incumbent politician/party in today's America. Targeting investments at left-behind communities ≠ winning votes. Politics around place-based (climate) politics seem to be messier than advocates hope 9/9
Full paper: tiny.cc/energycommun....

I refined a two-dimensional RDD method with bootstrap aggregation + delta bootstrap CIs. Simulations show less bias, higher statistical efficiency. 8/9

The finding: The bonus decreased the vote share for Harris by 0.39 percentage points in 2024. (95% CI: -0.78 pp to -0.02 pp) 7/9

There are two sharp cutoffs for eligibility. Areas just above vs. just below these thresholds are "alike" - except one happened to have been eligible for the bonus, one didn't. Great setup for a quasi-experiment! 6/9

So which story wins? I approached this with a focus on Energy Community Tax Credit Bonus: 1/3 extra tax credit for renewable projects in areas with high unemployment + (past or present) fossil fuel jobs. 5/9

But the argument can also go the other way. Maybe renewables threaten local identity. Maybe projects drive up prices in the short term. Maybe benefits are too invisible to matter. The evidence has been mixed 4/9

The "deliverism" theory: Channel green money into struggling towns → ease transition fears → create grateful voters → more support. Sounds reasonable, right? 3/9

The Biden bet: help "left behind" communities → win voters for both the party and further climate action. Such “place-based” policies have become increasingly popular 🔋➡️🗳️2/9

Did the Biden administration's targeted industrial policies help lift the Democratic Party in "left-behind" America? Well, no, and it might've shifted the targeted areas further away from the party, at least in the case of one place-based green subsidy. New working paper 🧵 1/9

Bottom line: "Deliverism" itself may not help the incumbent politician/party in today's America. Targeting investments at left-behind communities ≠ winning votes. Politics around place-based (climate) politics seem to be messier than advocates hope. 9/9

Full paper: tiny.cc/energycommunities

The finding: The bonus decreased the vote share (of the two-party vote) for Harris by 0.39 percentage points in 2024. (95% CI: -0.78 pp to -0.02 pp) 7/9

There are two sharp administrative cutoffs for eligibility. Areas that just met vs. just missed these thresholds are "alike" - except the former happened to have been eligible for the bonus while the latter didn't. Great setup for a quasi-experiment! 6/9

So which story wins? I approached this with a focus on the Energy Community Tax Credit Bonus: 1/3 increase over baseline tax credits for renewable projects in areas with high unemployment + (past or present) fossil fuel jobs. 5/9

But the argument can also go the other way. Maybe renewables threaten local identity. Projects can drive up prices in the short term. Perhaps benefits are too invisible to matter. And empirical evidence has been mixed 4/9

The "deliverism" theory: Channel (green) investments into struggling towns → ease transition fears → create grateful voters → more support. Sounds reasonable, right? 3/9

The Biden bet: help "left behind" communities (yellow areas in the map) → win voters for both the party and further climate action. Such "place-based" policies have become increasingly popular 2/9

"Benchmarking parallel trends violations in regression imputation difference-in-differences" https://osf.io/preprints/socarxiv/ngr3d_v1

New working paper! - zikai.li and I look at the problem of assessing pre-trends using one of the "new" DiD methods - fixed effects regression imputation. tl;dr - don't use the same regression you used to impute post-treatment to also impute pre-treatment. osf.io/preprints/so... (1/14)

Thanks for your interest! I have the data pipeline ready but haven't got the county-level returns yet. Will send the report your way once I have preliminary results!

Thanks, Bit!