Dynamic effects

Note

I run the distributed lags and leads models

I save a table for the models with 0, 1, 3, 7 lags
I save a plot for the model with 7 leads and 7 lags

library(tidyverse)
library(here) # path control
library(fixest) # estimations with fixed effects
library(patchwork)
library(modelsummary) # for modelplot
library(knitr)
library(tictoc)
tic("Total render time")

source("fun_dynamic_bins.R")

mem.maxVSize(vsize = 40000)

[1] 40000

data <-
  read_rds(here("..", "output", "data_reports.rds")) %>% 
  filter(covid_phase == "Pre-pandemic")

data <- data %>% 
  mutate(prec_quintile = ntile(prec, 5),
         rh_quintile = ntile(rh, 5),
         wsp_quintile = ntile(wsp, 5))

Lags

Run regressions with 0, 1, 3, and 7 days of lags.

I do it in separate commands, because all at once exhausts the memory.

# baseline for reference
reg_linear_reports <- data %>%
  fepois(reports_dv ~ tmean |
           prec_quintile + rh_quintile + wsp_quintile +
           ageb + year^month + day_of_week + day_of_year,
         cluster = ~ ageb)


reg_lags_1 <- data %>% 
  fepois(reports_dv ~ l(tmean, 0:1) |
           prec_quintile + rh_quintile + wsp_quintile +
           ageb + year^month + day_of_week + day_of_year,
         panel.id = ~ ageb + date,
         cluster = ~ ageb)

reg_lags_3 <- data %>% 
  fepois(reports_dv ~ l(tmean, 0:3) |
           prec_quintile + rh_quintile + wsp_quintile +
           ageb + year^month + day_of_week + day_of_year,
         panel.id = ~ ageb + date,
         cluster = ~ ageb)

reg_lags_7 <- data %>% 
  fepois(reports_dv ~ l(tmean, 0:7) |
           prec_quintile + rh_quintile + wsp_quintile +
           ageb + year^month + day_of_week + day_of_year,
         panel.id = ~ ageb + date,
         cluster = ~ ageb)

# reg_lags_14 <- data %>% 
#   fepois(reports_dv ~ l(tmean, 0:14) |
#            prec_quintile + rh_quintile + wsp_quintile +
#            ageb + year^month + day_of_week + day_of_year,
#          panel.id = ~ ageb + date,
#          cluster = ~ ageb)


etable(reg_linear_reports, reg_lags_1, reg_lags_3, reg_lags_7) %>% 
  kable()

	reg_linear_reports	reg_lags_1	reg_lags_3	reg_lags_7
Dependent Var.:	reports_dv	reports_dv	reports_dv	reports_dv

tmean	0.0274*** (0.0033)	0.0234*** (0.0046)	0.0237*** (0.0046)	0.0237*** (0.0047)
l(tmean,1)		0.0051 (0.0041)	0.0073 (0.0056)	0.0077 (0.0056)
l(tmean,2)			-0.0059 (0.0056)	-0.0062 (0.0057)
l(tmean,3)			0.0059 (0.0041)	0.0057 (0.0055)
l(tmean,4)				0.0027 (0.0056)
l(tmean,5)				-0.0042 (0.0056)
l(tmean,6)				-0.0006 (0.0054)
l(tmean,7)				0.0020 (0.0040)
Fixed-Effects:	——————	——————	——————	——————
prec_quintile	Yes	Yes	Yes	Yes
rh_quintile	Yes	Yes	Yes	Yes
wsp_quintile	Yes	Yes	Yes	Yes
ageb	Yes	Yes	Yes	Yes
year-month	Yes	Yes	Yes	Yes
day_of_week	Yes	Yes	Yes	Yes
day_of_year	Yes	Yes	Yes	Yes
_______________	__________________	__________________	__________________	__________________
S.E.: Clustered	by: ageb	by: ageb	by: ageb	by: ageb
Observations	3,624,826	3,622,403	3,617,557	3,607,865
Squared Cor.	0.01333	0.01333	0.01333	0.01334
Pseudo R2	0.05615	0.05614	0.05614	0.05615
BIC	804,159.2	803,666.8	802,960.2	801,469.3

I register an extraline in fixest to compute cumulative effect, and export the reg table as reg_lags.tex

Lags and leads

Run a model with 7 leads and 7 lags:

reg_leads_lags <- data %>% 
  fepois(reports_dv #reports_dv_day
         ~ f(tmean, -7:7) |
           prec_quintile + rh_quintile + wsp_quintile +
           ageb + year^month + day_of_week + day_of_year,
         panel.id = ~ ageb + date,
         cluster = ~ ageb)

Plot and save as reg_lags_leads.png

Rendered on: 2025-04-03 19:32:19

Total render time: 706.23 sec elapsed