Plots needed
MAIN TEXT: DATA Summary table (we already have this but will need to rearrange, update numbers. Ashwin will rearrange and coordinate with Kit and Rae to get the relevant numbers updated)
MAIN TEXT: THEORY Ashwin and Ishan working on illustrated figures in this section
MAIN TEXT: RESPONSE FUNCTION ESTIMATES NOTE: All FE regimes include the standard ADM3 FE and dow FE Estimate common response (w/differentiated FE) for ADM0 week FE (Kit) Tables of high- and low-risk responses, with different FEs (Kit) Main model ("ADM0 x week" FE): Common, Low, High columns, with rows for 0,5,10,30,35,40 Low and High responses under following FE: "adm3_id dow_week adm0_id#month#year" "adm3_id dow_week adm0_id#year" "adm3_id dow_week adm0_id#week_fe adm0_id#year" "adm3_id dow_week adm3_id#month#year" . 3 low-risk columns, followed by 3 high-risk columns. Rows for temperature as above.
Place histograms under common-, high- and low-risk response plots (Kit)
MAIN TEXT: RISK SHARE ESTIMATES Table of coefficient estimates, 2 columns- with and without continent FEs (might move to appendix) level curves of income effect (at terciles of temperature) level curves of temperature effect (at terciles of income) WE ALREADY HAVE THE TWO LEVEL CURVE PLOTS FROM THE FED PRESENTATION
MAIN TEXT: POST-PROJECTION (examples of these are in Fed slides) IMPACTS IN MINUTES Impact map at 2099 for low- and high-risk workers Share of high-risk workers map at 2020 and 2099 beta map for 37 C response (averaging across high- and low-risk responses by IR) at 2020 and 2099 overall impact map at 2099 (mins/person/day) when we have uncertainty, we will add kernel density for selected IRs time series of overall impacts (mins/person/day) under no adapt and full adapt IMPACTS IN DOLLARS Time series of monetized global disutility Map of global monetized disutility at 2099 as percent of GDP TBD some other spatial breakdown of monetized global disutility (e.g. countries) DAMAGE FUNCTION End-of-century damage function with GCM/RCP points Extrapolated damage functions beyond 2100 PULSE FIGURE Final damage panel to be updated SCC table
APPENDICES
Functional form comparison (bins etc.) Robustness projection using different high/low shares specification (without continent FEs) Single projection with edge restrictions Point estimate projection for different SSPs Inclusion of China (response estimation and single projection) Interacted model
projection (./generate.sh + yml) -> aggregation (./aggregate.sh + yml) -> extraction (single.py, quantiles.py)