David Johnson
@davidblakejohnson.bsky.social
Stay at home dad moonlighting as an associate professor of economics. Mostly interested in gun policy and fishing.
Most popular words appearing in the names of gun stores (2003-2025).
November 8, 2025 at 10:00 PM
Most popular words appearing in the names of gun stores (2003-2025).
Our cat seems to bring in more friends than she takes out. Saved this fella and (accidentally) scared some kids.
October 27, 2025 at 7:47 PM
Our cat seems to bring in more friends than she takes out. Saved this fella and (accidentally) scared some kids.
No Kings Lees’s Summit. Complete lawlessnes.
October 18, 2025 at 3:53 PM
No Kings Lees’s Summit. Complete lawlessnes.
Applying some of what I learned from @chadtopaz.bsky.social .
Gun dealers isn't the only thing that is low dimensional. Concealed carry permits are well. Which is a little remarkable considering the heterogeneity in state laws and renewal cycles.
Gun dealers isn't the only thing that is low dimensional. Concealed carry permits are well. Which is a little remarkable considering the heterogeneity in state laws and renewal cycles.
October 15, 2025 at 7:20 PM
Applying some of what I learned from @chadtopaz.bsky.social .
Gun dealers isn't the only thing that is low dimensional. Concealed carry permits are well. Which is a little remarkable considering the heterogeneity in state laws and renewal cycles.
Gun dealers isn't the only thing that is low dimensional. Concealed carry permits are well. Which is a little remarkable considering the heterogeneity in state laws and renewal cycles.
You can do the same exercise with actual gun sales! year fixed effects are even more similar across the two outcome variables (gun sales and estimated gun sales).
October 7, 2025 at 12:44 AM
You can do the same exercise with actual gun sales! year fixed effects are even more similar across the two outcome variables (gun sales and estimated gun sales).
ta-da!
Almost parallel!
Which makes perfect sense because not every permit is going to come with a new gun
Almost parallel!
Which makes perfect sense because not every permit is going to come with a new gun
October 2, 2025 at 10:07 PM
ta-da!
Almost parallel!
Which makes perfect sense because not every permit is going to come with a new gun
Almost parallel!
Which makes perfect sense because not every permit is going to come with a new gun
Then do the same with our logged gun sales estimates.
If we are right, then both models should produce similar year fixed effects.
Here is what happens when I try it:
If we are right, then both models should produce similar year fixed effects.
Here is what happens when I try it:
October 2, 2025 at 10:03 PM
Then do the same with our logged gun sales estimates.
If we are right, then both models should produce similar year fixed effects.
Here is what happens when I try it:
If we are right, then both models should produce similar year fixed effects.
Here is what happens when I try it:
August and September are generally pretty average in terms of accidents (February is the low).
Hasn’t been this low since COVID.
Hasn’t been this low since COVID.
September 30, 2025 at 1:44 AM
August and September are generally pretty average in terms of accidents (February is the low).
Hasn’t been this low since COVID.
Hasn’t been this low since COVID.
I ”jokingly“ suggested to my wife we put this guy out.
After she found out about Kimmel, she became very open to it.
After she found out about Kimmel, she became very open to it.
September 18, 2025 at 2:10 AM
I ”jokingly“ suggested to my wife we put this guy out.
After she found out about Kimmel, she became very open to it.
After she found out about Kimmel, she became very open to it.
Got a comment that the accuracy of our gun sales estimates is just due to population (e.g., pop grows -> sales grow).
Big nope there. Correlation between standardized gun sales estimates per capita and NICS gun checks per capita is about .8.
Big nope there. Correlation between standardized gun sales estimates per capita and NICS gun checks per capita is about .8.
September 4, 2025 at 10:42 PM
Got a comment that the accuracy of our gun sales estimates is just due to population (e.g., pop grows -> sales grow).
Big nope there. Correlation between standardized gun sales estimates per capita and NICS gun checks per capita is about .8.
Big nope there. Correlation between standardized gun sales estimates per capita and NICS gun checks per capita is about .8.
If you standardize (within state) our new gun estimates and compare them to standardized Google searches with the word gun, you get this.
August 25, 2025 at 7:40 PM
If you standardize (within state) our new gun estimates and compare them to standardized Google searches with the word gun, you get this.
I can go one step further and look at the first differences (e.g., standardized estimates in Missouri in 2010 minus standardized estimates in Missouri in 2009).
August 23, 2025 at 9:18 PM
I can go one step further and look at the first differences (e.g., standardized estimates in Missouri in 2010 minus standardized estimates in Missouri in 2009).
This isn't just capturing population.
When I standardize (within state) our new gun estimates and NICS checks and plot them all, I get this.
When I standardize (within state) our new gun estimates and NICS checks and plot them all, I get this.
August 23, 2025 at 9:17 PM
This isn't just capturing population.
When I standardize (within state) our new gun estimates and NICS checks and plot them all, I get this.
When I standardize (within state) our new gun estimates and NICS checks and plot them all, I get this.
Just updated @troach.bsky.social and I's gun sales estimates using 2020-2023 data. The estimates are really good.
The raw correlation using NICS checks explicitly for a gun and data from 2004 - 2023 is .85.
Green line 45 degree line; red is ols
The raw correlation using NICS checks explicitly for a gun and data from 2004 - 2023 is .85.
Green line 45 degree line; red is ols
August 23, 2025 at 9:14 PM
Just updated @troach.bsky.social and I's gun sales estimates using 2020-2023 data. The estimates are really good.
The raw correlation using NICS checks explicitly for a gun and data from 2004 - 2023 is .85.
Green line 45 degree line; red is ols
The raw correlation using NICS checks explicitly for a gun and data from 2004 - 2023 is .85.
Green line 45 degree line; red is ols
I am exaggerating, but not by much. Here are standardized searches with the word "Gun" in Washington and Florida using monthly data that runs from 2004 to 2024. These states are 3,000 miles apart.
August 19, 2025 at 10:54 PM
I am exaggerating, but not by much. Here are standardized searches with the word "Gun" in Washington and Florida using monthly data that runs from 2004 to 2024. These states are 3,000 miles apart.
The similarity of monthly NICS checks across states is amazing. Even the 12-month differences are on top of each other.
August 14, 2025 at 4:12 PM
The similarity of monthly NICS checks across states is amazing. Even the 12-month differences are on top of each other.
Monthly gun homicides in the US.
August 11, 2025 at 5:49 PM
Monthly gun homicides in the US.
Here is what you get with a basic approach that includes only month fixed effects to account for seasonality. Peak is about year off from what I expected but this expectation was probably a mistake on my part. The 3 and 4 year lags generally produce similar results.
July 22, 2025 at 9:07 PM
Here is what you get with a basic approach that includes only month fixed effects to account for seasonality. Peak is about year off from what I expected but this expectation was probably a mistake on my part. The 3 and 4 year lags generally produce similar results.
I try to hide my redneck side but can’t suppress the nature boy in me.
Daughter also lets out the twang.
Daughter also lets out the twang.
July 18, 2025 at 6:31 PM
I try to hide my redneck side but can’t suppress the nature boy in me.
Daughter also lets out the twang.
Daughter also lets out the twang.
Look at this guy throwing gang signs
July 9, 2025 at 2:30 PM
Look at this guy throwing gang signs
Our findings suggest that an increase of 1,000 new handguns results in approximately 5 additional homicides in Mexico, with the effect peaking after 4 years.
April 28, 2025 at 6:16 PM
Our findings suggest that an increase of 1,000 new handguns results in approximately 5 additional homicides in Mexico, with the effect peaking after 4 years.
Smith & Wesson Brands, Inc. v. Estados Unidos Mexicanos got me thinking that we could probably test if there was a relationship between US handgun manufacturing and Mexican Homicides. Never expected it to be this obvious.
March 9, 2025 at 8:20 PM
Smith & Wesson Brands, Inc. v. Estados Unidos Mexicanos got me thinking that we could probably test if there was a relationship between US handgun manufacturing and Mexican Homicides. Never expected it to be this obvious.
Am I imagining things or does the yearly total of Mexican and US homicides look more like US handgun manufacturing than either country's homicide counts in isolation?
March 5, 2025 at 3:01 PM
Am I imagining things or does the yearly total of Mexican and US homicides look more like US handgun manufacturing than either country's homicide counts in isolation?
Updated version. US handgun manufacturing and the Mexican Homicide Rate 1949-2020
March 4, 2025 at 8:03 PM
Updated version. US handgun manufacturing and the Mexican Homicide Rate 1949-2020