Victor Bennett
@veeembee.bsky.social
Prof U of Utah
Previously: White House Council of Economics of Advisers, Duke, Initialized Capital, USC, Google
Opinions mine, at best
vmbennett.com
Previously: White House Council of Economics of Advisers, Duke, Initialized Capital, USC, Google
Opinions mine, at best
vmbennett.com
Some of this is incentives, some of it is the job changing. I went from limited service to 20+ hours per week when I got tenure!
The lab fields mean that you still have a team working on research on your behalf when you're pulled away, right?
The lab fields mean that you still have a team working on research on your behalf when you're pulled away, right?
October 3, 2025 at 5:44 PM
Some of this is incentives, some of it is the job changing. I went from limited service to 20+ hours per week when I got tenure!
The lab fields mean that you still have a team working on research on your behalf when you're pulled away, right?
The lab fields mean that you still have a team working on research on your behalf when you're pulled away, right?
I think collinearity is connected to "no causation without manipulation." If almost all of X are Y, then do we really need to distinguish the effect of X from the effect from Y? Is Y even philosophically different from X? If it is, maybe the interesting question is why X and Y go together so often
October 1, 2025 at 7:07 PM
I think collinearity is connected to "no causation without manipulation." If almost all of X are Y, then do we really need to distinguish the effect of X from the effect from Y? Is Y even philosophically different from X? If it is, maybe the interesting question is why X and Y go together so often
The example that blew my mind as a ugrad was:
I planned on going on a hike in the desert. One enemy drilled a hole in my canteen. The other poisoned it. Since I would have died anyway no one caused my death.
In social science we often have two things that need to happen, so neither can cause alone!
I planned on going on a hike in the desert. One enemy drilled a hole in my canteen. The other poisoned it. Since I would have died anyway no one caused my death.
In social science we often have two things that need to happen, so neither can cause alone!
August 26, 2025 at 6:46 PM
The example that blew my mind as a ugrad was:
I planned on going on a hike in the desert. One enemy drilled a hole in my canteen. The other poisoned it. Since I would have died anyway no one caused my death.
In social science we often have two things that need to happen, so neither can cause alone!
I planned on going on a hike in the desert. One enemy drilled a hole in my canteen. The other poisoned it. Since I would have died anyway no one caused my death.
In social science we often have two things that need to happen, so neither can cause alone!
Causation, but not just that it isn't correlation.
Measuring it is hard, but often the idea isn't well defined!
Measuring it is hard, but often the idea isn't well defined!
August 26, 2025 at 6:46 PM
Causation, but not just that it isn't correlation.
Measuring it is hard, but often the idea isn't well defined!
Measuring it is hard, but often the idea isn't well defined!
My experience, but YMMV:
Gtools and Parallel (and reghdfej) are so much better at using multiple cores, that my 2020 laptop handles dozens of millions of obs well.
gtools.readthedocs.io/en/latest/
github.com/gvegayon/par...
Gtools and Parallel (and reghdfej) are so much better at using multiple cores, that my 2020 laptop handles dozens of millions of obs well.
gtools.readthedocs.io/en/latest/
github.com/gvegayon/par...
Home - Gtools
Faster Stata for Big Data
gtools.readthedocs.io
June 2, 2025 at 9:02 PM
My experience, but YMMV:
Gtools and Parallel (and reghdfej) are so much better at using multiple cores, that my 2020 laptop handles dozens of millions of obs well.
gtools.readthedocs.io/en/latest/
github.com/gvegayon/par...
Gtools and Parallel (and reghdfej) are so much better at using multiple cores, that my 2020 laptop handles dozens of millions of obs well.
gtools.readthedocs.io/en/latest/
github.com/gvegayon/par...
CA has data on education performance: www.cde.ca.gov/ds/sp/cl/
PIAAC has data on test performance
NCES also has data on majors: nces.ed.gov/programs/coe...
PIAAC has data on test performance
NCES also has data on majors: nces.ed.gov/programs/coe...
California Longitudinal Pupil Achievement Data System (CALPADS) - School Performance (CA Dept of Education)
A longitudinal data system used to maintain individual-level data including student demographics, course data, discipline, assessments, staff assignments, and other data for state and federal reportin...
www.cde.ca.gov
March 25, 2025 at 3:36 PM
CA has data on education performance: www.cde.ca.gov/ds/sp/cl/
PIAAC has data on test performance
NCES also has data on majors: nces.ed.gov/programs/coe...
PIAAC has data on test performance
NCES also has data on majors: nces.ed.gov/programs/coe...
February 19, 2025 at 6:48 PM
2) Second, increasing access to finance for women and racial minorities isn't enough to spur minority entrepreneurship. Finance is important, but people need to know they have mentors and networks that can help them build their business before it is worth looking for resources.
January 28, 2025 at 7:00 PM
2) Second, increasing access to finance for women and racial minorities isn't enough to spur minority entrepreneurship. Finance is important, but people need to know they have mentors and networks that can help them build their business before it is worth looking for resources.
I learned a lot working on this study and it taught me two things in particular.
1) First, one size doesn't fit all. There are different barriers to entrepreneurship faced by women and Black and Hispanic americans
1) First, one size doesn't fit all. There are different barriers to entrepreneurship faced by women and Black and Hispanic americans
January 28, 2025 at 7:00 PM
I learned a lot working on this study and it taught me two things in particular.
1) First, one size doesn't fit all. There are different barriers to entrepreneurship faced by women and Black and Hispanic americans
1) First, one size doesn't fit all. There are different barriers to entrepreneurship faced by women and Black and Hispanic americans
Even though women and racial minorities are less likely to start businesses, it isn't because they don't have ideas or belief in their ideas
Something happens in the funnel between having the idea and starting the business
We find patterns consistent with differences in tapping social networks
Something happens in the funnel between having the idea and starting the business
We find patterns consistent with differences in tapping social networks
January 28, 2025 at 7:00 PM
Even though women and racial minorities are less likely to start businesses, it isn't because they don't have ideas or belief in their ideas
Something happens in the funnel between having the idea and starting the business
We find patterns consistent with differences in tapping social networks
Something happens in the funnel between having the idea and starting the business
We find patterns consistent with differences in tapping social networks
IIRC, this might be because debts less than $500 don't hit your credit score. Because this is known, people in real financial dire straits just won't pay those, so clear them doesn't make life easier.
I'd be curious about bigger debts. I bet the implications would be different.
I'd be curious about bigger debts. I bet the implications would be different.
April 9, 2024 at 4:43 PM
IIRC, this might be because debts less than $500 don't hit your credit score. Because this is known, people in real financial dire straits just won't pay those, so clear them doesn't make life easier.
I'd be curious about bigger debts. I bet the implications would be different.
I'd be curious about bigger debts. I bet the implications would be different.