Popular in Social Sciences

Books

City Time: On Being Sentenced to Rikers Island

Independent Scholar David A. Campbell

Culture and Society

The City as a Growth Machine: Toward a Political Economy of Place

New York University Harvey Moltoch

Economics

Gender and Economics

University of California Santa Barbara Shelly Lundberg

American Politics

Keeping the Republic

Boston College Marc Landy

Banks and Policy Organisations

Greening the financial sector

European Investment Bank Matteo Ferrazzi

British Politics

Entrepreneurial thinking: the politics and practice of policy impact

University of Sheffield Matthew Flinders

Culture and Society

Carceral Entrapments: Views from the Prison/Street Interface in India

Indian Institute of Technology Delhi Mahuya Bandyopadhya

Cutting the Innovation Engine

View more insights in Social Sciences - Economics

View more figures from U.S. Census Bureau

View the page of Elisabeth Ruth Perlman

Read the Study

Elisabeth Ruth Perlman shows how federal funding cuts in university research impact innovation outputs. Decreased publications, increased patents, and shift from federal to private funding.

Image courtesy of interviewee. December 21, 2023

Generating Workspace

Export as PDF

The research question addresses how a decline in federal funding impacts university research outputs. The authors believe university research is vital for the U.S. economy and innovation, and point out that a significant decline in federal funding has been observed over the last 40 years, leading universities to seek private sector funding as a means of diversification and stability.

The study differentiates itself by focusing on individual researchers rather than aggregate data, exploring how reductions in federal funding influence their outputs, such as patenting, publishing, and entrepreneurship. The methodology involves an event study design to analyze the effects of federal funding shocks and relate them to changes in researchers' output.

Findings indicate a persistent decline in federal funding leads to a decrease in researchers' individual funding levels. However, there is also a corresponding increase in private funding, suggesting that researchers are able to recover some funding losses through private sources. This video video shift leads to a complex outcome where researchers become less likely to engage in high-tech entrepreneurship but more likely to patent their work.

The authors suggest that federal funding is conducive to basic research, while private funding may better support applied research, influencing how research results are commercialized. Ultimately, they conclude that federal funding cuts change the landscape of research outputs and commercialization pathways, emphasizing a shift towards private sector appropriation and engagement over independent entrepreneurship.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:05.615 --> 00:07.235
The research question we have is,

00:07.335 --> 00:10.675
how does a decline in federal funding affect the outputs

00:10.675 --> 00:11.755
of university research?

00:12.175 --> 00:14.515
So, we're interested in this questions for several reasons.

00:15.135 --> 00:16.555
You know, one, we believe

00:16.585 --> 00:19.955
that university research is very important

00:19.975 --> 00:24.835
to the US economy, to the innovation of new, ideas

00:25.055 --> 00:27.035
and new technologies in the US economy.

00:27.465 --> 00:29.075
It's a large percentage of the r

00:29.075 --> 00:31.035
and d that happens in the US economy.

00:31.895 --> 00:33.435
Two, a large percentage

00:33.495 --> 00:37.155
of university research is impacted by federal funds.

00:37.415 --> 00:40.555
And we've seen a significant decrease in this percentage

00:40.555 --> 00:42.795
of university research impacted

00:43.055 --> 00:46.395
by federal funds over the last 40 years, as well

00:46.395 --> 00:49.355
as a significant decrease in the federal share of r

00:49.355 --> 00:52.035
and d spending over the last 40 years as a whole.

00:52.895 --> 00:56.935
Universities have responded to this by thinking about

00:57.555 --> 01:01.095
how can they get funding that's not federal government,

01:01.095 --> 01:03.655
because they're very reliant on federal government funding.

01:04.155 --> 01:06.735
And many universities have explicitly said, you know,

01:06.735 --> 01:10.215
we want to shift from being just reliant on federal funding

01:10.755 --> 01:12.775
to funding in the private sector

01:13.035 --> 01:16.015
to private companies giving grants to universities

01:16.015 --> 01:17.175
to engage in r and d.

01:17.395 --> 01:19.215
And they've explicitly looked

01:19.275 --> 01:22.135
to shift their funding from federal funding

01:22.595 --> 01:25.095
to private funding because they think this diversified

01:25.095 --> 01:28.095
funding portfolio will be more stable.

01:28.435 --> 01:31.295
Now, we don't engage in this question of stability,

01:31.355 --> 01:34.775
but instead we ask how individual researchers

01:34.775 --> 01:37.655
that are supported by this funding are impacted

01:37.795 --> 01:39.735
by declines in federal funding

01:40.115 --> 01:42.095
and possible increases in private

01:42.095 --> 01:43.695
funding that result from that.

01:44.345 --> 01:47.175
There are a number of papers where we know

01:47.285 --> 01:51.295
that federal funding, you know, increases r and d

01:51.355 --> 01:53.015
and increases private spillovers.

01:53.485 --> 01:57.735
What those papers show is that you have more federal funding

01:57.995 --> 02:00.055
and you get more of innovation outcomes

02:00.085 --> 02:01.215
that you would expect.

02:01.325 --> 02:05.015
Patents and papers. Um, as well as that,

02:05.125 --> 02:08.335
when you see increases in federal funding in specific areas,

02:08.475 --> 02:10.495
you also see sort of this crowd in effect

02:10.495 --> 02:11.815
that there's more private funding.

02:12.195 --> 02:15.975
But these papers that exist, this sort

02:15.975 --> 02:18.695
of look at this are all in the aggregate level.

02:18.955 --> 02:20.735
So what our paper does that's different

02:21.555 --> 02:24.295
is we're actually following the careers of individuals

02:24.295 --> 02:25.575
who are sponsored by this grant.

02:25.675 --> 02:29.455
So it's not just what happens when a field gets

02:29.605 --> 02:31.335
more, funding.

02:31.525 --> 02:32.815
Does that field or

02:32.815 --> 02:35.455
that university also get more private funding?

02:35.805 --> 02:39.375
Does that field or that university also produce more

02:39.375 --> 02:40.415
papers or patents?

02:40.835 --> 02:43.815
But how are the individuals that are supported

02:43.815 --> 02:47.535
by these grants affected in producing more

02:47.535 --> 02:48.575
papers or patents?

02:48.635 --> 02:51.975
And then we are also able to look directly through use

02:51.975 --> 02:55.735
of census data at individuals,

02:56.135 --> 02:59.645
engaging in entrepreneurship, which we define as,

02:59.775 --> 03:04.325
being at a new high tech age zero high tech company.

03:04.785 --> 03:06.805
Uh, and so, you know, that's something

03:06.805 --> 03:09.405
that previous papers really have not been able to get at.

03:09.465 --> 03:12.525
But because through census data we can follow people's

03:12.525 --> 03:14.925
careers over time, we're able to actually

03:16.135 --> 03:17.715
follow them in that way.

03:18.295 --> 03:21.195
The data we're using in the paper is based on

03:21.705 --> 03:26.235
some really large efforts to collect administrative data

03:26.455 --> 03:28.235
and bring administrative data together.

03:28.695 --> 03:31.475
So we start off with a project that's out of the University

03:31.475 --> 03:34.555
of Michigan, called umetrics,

03:34.935 --> 03:38.995
and that collects administrative data on research grants

03:39.055 --> 03:42.475
and awards for, in our case, 22 universities.

03:42.475 --> 03:44.635
They're continuing to try to expand this project,

03:44.855 --> 03:48.235
so I think they're actually up to 30 some universities now,

03:48.725 --> 03:52.795
which has individual level grant payments for all employees.

03:53.455 --> 03:57.435
And so then the census takes these grant individual level

03:57.435 --> 03:58.675
grant payments in

03:59.375 --> 04:03.995
and, links them to our data on individuals.

04:04.375 --> 04:06.275
We also know attributes of the grants.

04:06.535 --> 04:07.555
So in specific,

04:07.655 --> 04:11.275
the thing we use in the the paper are CFDA codes.

04:11.275 --> 04:13.475
That's catalog of Federal Domestic Assistance,

04:14.005 --> 04:16.955
which is a way that the US government categorizes

04:17.695 --> 04:20.755
its grants into roughly to speak of fields.

04:21.415 --> 04:23.075
Uh, so then we can link that

04:23.095 --> 04:26.435
to data on the CFDA codes as a whole.

04:26.455 --> 04:28.475
So they aggregate data in that field.

04:29.555 --> 04:32.895
On the other side, we can bring these individuals into

04:32.895 --> 04:37.175
census and we can link to IRS W2 forms,

04:37.265 --> 04:40.695
which have, you know, the individual, how much they're paid

04:41.155 --> 04:42.455
and who pays them,

04:42.475 --> 04:44.855
or at least what federal tax ID is paying them.

04:45.355 --> 04:49.695
Um, we can bring it to the LEHD, the longitudinal

04:50.255 --> 04:54.935
employer household dataset, which is from UI records,

04:55.315 --> 04:56.455
state UI records,

04:56.995 --> 05:00.175
and again, is linking an individual's employer

05:00.205 --> 05:03.735
through a tax ID to the individual so you know

05:03.735 --> 05:06.175
how much they're paid and where they're, they work, as well

05:06.175 --> 05:09.615
as census data on these establishments.

05:09.755 --> 05:12.575
So knowing what industry the establishments are, knowing

05:12.575 --> 05:15.815
what year these establishments were founded, we're also able

05:15.815 --> 05:19.055
to link these individuals to their patents.

05:19.315 --> 05:21.055
And that's done through the Metrics project

05:21.595 --> 05:23.975
and their publications, which is also done

05:23.975 --> 05:25.135
through the metrics projects.

05:25.235 --> 05:28.055
So we're really able to bring this huge amount

05:28.315 --> 05:30.295
of different administrative

05:30.675 --> 05:34.395
and sort of, you know, large comprehensive sources

05:34.495 --> 05:37.955
of data together to look at these different outcomes.

05:38.615 --> 05:41.555
The main methodology we use here is an event study.

05:42.015 --> 05:44.675
So we specifically look for,

05:45.655 --> 05:47.635
CFDA codes.

05:47.635 --> 05:51.355
So I mentioned the, that's roughly a field that have had

05:51.965 --> 05:54.915
large shocks to funding in them.

05:55.415 --> 05:57.685
So we, this is sort of, you know,

05:57.685 --> 06:00.205
we select based on their characteristics.

06:00.265 --> 06:04.405
So we look for ones that have had at least a 40% decrease

06:04.985 --> 06:07.165
in federal funding or in funding,

06:08.045 --> 06:09.285
'cause they're all federal grants.

06:09.785 --> 06:11.765
And then an increase again,

06:12.065 --> 06:14.325
and we say we don't want you to have more than

06:14.965 --> 06:18.725
a 20% increase or decrease in the two years

06:19.265 --> 06:21.645
before that 40% dip.

06:22.185 --> 06:26.565
Um, and we want you to have at least returned,

06:26.865 --> 06:28.765
you know, 80% in the year after.

06:28.865 --> 06:30.165
So it really is a shock.

06:30.425 --> 06:32.205
So we're looking at something that

06:33.305 --> 06:35.845
we believe is quite idiosyncratic.

06:36.065 --> 06:38.005
We did a lot of work to try

06:38.005 --> 06:40.085
to look at why these shocks happened

06:40.385 --> 06:43.085
and, you know, most of them

06:43.735 --> 06:46.605
don't necessarily bear mentioning in the public record,

06:46.785 --> 06:48.085
so it was really hard to get data.

06:48.185 --> 06:50.445
But the ones we could find, it seems to be

06:50.995 --> 06:53.845
like something quite random that's uncorrelated with,

06:53.955 --> 06:55.245
with the tech fields.

06:55.745 --> 07:00.045
And we also did some testing to see if these shocks seem

07:00.045 --> 07:02.925
to be related to overall trends within these fields

07:02.985 --> 07:06.365
of patenting and, high tech startups.

07:06.365 --> 07:09.085
And they do not. So we use the shocks

07:09.085 --> 07:10.365
to create an event study.

07:10.425 --> 07:13.845
So we look at researchers who are paid on these CFDA codes,

07:14.745 --> 07:16.725
as well as controlled CFDA codes,

07:16.855 --> 07:20.365
which have had no more than a 20% variation,

07:20.595 --> 07:21.725
over the timeframe.

07:22.785 --> 07:24.285
And we say

07:24.695 --> 07:29.565
after this shock, what happens to your being at,

07:30.155 --> 07:34.005
high tech, age zero company to patenting

07:34.145 --> 07:35.165
and to publication?

07:35.585 --> 07:38.365
So it's a very simple event study set up.

07:39.145 --> 07:43.465
I had noted before that we wanted to understand the effect

07:43.465 --> 07:45.905
of these federal funding cuts on the funding

07:46.525 --> 07:47.945
of individual researchers.

07:48.005 --> 07:50.025
And that we saw, one, we saw when one

07:50.025 --> 07:52.345
of these fund federal funding cuts came in,

07:52.345 --> 07:54.345
that individual researchers saw

07:54.905 --> 07:57.385
a decrease in their funding levels that was persistent.

07:57.845 --> 08:01.625
But we were also interested in the question of whether this,

08:01.695 --> 08:03.305
whether people were able to,

08:03.725 --> 08:07.825
as university presidents had talked about substitute private

08:07.825 --> 08:09.145
funding for federal funding.

08:09.885 --> 08:12.285
So we asked the question of

08:12.285 --> 08:15.125
after one of these funding cuts, did we see

08:15.865 --> 08:17.445
an increase in private funding?

08:17.985 --> 08:21.245
And we do in fact, find an increase in private funding

08:21.695 --> 08:22.965
after one of these cuts.

08:23.385 --> 08:28.045
So total funding decreases by about 15%,

08:28.535 --> 08:32.445
which is about a 28% decrease in federal funding,

08:32.785 --> 08:36.085
but a 15% increase in private funding.

08:36.625 --> 08:38.725
So it's something like they're able

08:38.725 --> 08:40.445
to get 10 cents on the dollar

08:40.785 --> 08:44.725
of lost federal funding made up for in private funding.

08:46.115 --> 08:51.015
So we think that this shift from, federal funding

08:51.075 --> 08:54.975
to private funding is very important for what's

08:55.205 --> 08:56.805
Actually going on in these mechanisms,

08:56.875 --> 09:01.205
because we would expect from a pure loss of money

09:01.595 --> 09:04.085
that you would see all of the things decrease,

09:04.115 --> 09:06.685
that you would see less patenting, less publications,

09:07.105 --> 09:09.085
and fewer high tech entrepreneurship.

09:09.145 --> 09:13.405
You know, you're just less able to, engage in research,

09:14.145 --> 09:17.125
but the shift suggests that there's something more nuanced.

09:17.625 --> 09:21.605
And so in this, this, more nuanced shift,

09:21.625 --> 09:25.045
the question is why do you have this increase in patenting?

09:25.465 --> 09:27.525
And so one way to think about this would be

09:27.525 --> 09:30.405
that perhaps federal private funding is

09:31.145 --> 09:33.165
better at supporting applied research,

09:33.165 --> 09:35.485
whereas federal funding is better at supporting basic

09:35.885 --> 09:38.105
research, or at least,

09:38.105 --> 09:40.505
private funding is bad at supporting basic research.

09:40.925 --> 09:42.145
And so this would, you know,

09:42.215 --> 09:45.985
give you a reason why patenting would increase,

09:46.255 --> 09:47.625
even though funding decrease

09:47.625 --> 09:49.665
that people are doing more applied research.

09:50.365 --> 09:53.665
But this wouldn't explain the decrease in,

09:54.375 --> 09:55.705
high tech entrepreneurship

09:55.705 --> 09:59.425
because high tech entrepreneurship is going to be, you know,

09:59.455 --> 10:00.505
very applied as well.

10:00.505 --> 10:03.745
These are things that people think can, they can privatize

10:03.805 --> 10:05.145
and, and turn into companies.

10:05.445 --> 10:07.905
And so we think that what's going on is

10:07.905 --> 10:09.745
that this is an appropriation story.

10:10.125 --> 10:14.385
The government gives funding with a very open notice.

10:14.455 --> 10:16.545
They explicitly want the researchers

10:16.545 --> 10:19.305
and the universities to retain control of the things

10:19.375 --> 10:20.945
that they produce so

10:20.945 --> 10:23.465
that they can have spillover effects in the US economy.

10:23.695 --> 10:25.505
This is the goal of the federal government.

10:25.965 --> 10:28.585
Uh, they believe the public has paid for this research,

10:28.585 --> 10:30.465
and the public should in some sense get

10:30.465 --> 10:31.585
access to this research.

10:32.535 --> 10:35.475
In contrast, when a private company is funding research,

10:35.475 --> 10:38.835
they're doing it because they want those r and d outputs.

10:38.835 --> 10:41.795
They want presumably to commercialize those r and d outputs.

10:41.975 --> 10:43.835
And so they're interested in appropriating

10:43.835 --> 10:45.115
those r and d outputs.

10:45.455 --> 10:48.475
And so this explains why you would see fewer people taking

10:48.615 --> 10:52.235
the ideas that are produced on private, grants

10:52.735 --> 10:55.595
and turning them into their own companies

10:55.855 --> 10:59.435
and why you would see more of those ideas patented.

10:59.435 --> 11:01.995
Because patenting is, you know, intellectual property.

11:01.995 --> 11:05.715
It's a way of creating property that these,

11:05.945 --> 11:08.155
private companies can appropriate.

11:08.655 --> 11:13.475
And we see, in sort of our, our talking to

11:13.985 --> 11:15.035
tech transfer offices

11:15.495 --> 11:20.395
and to researchers that a very common thing, is

11:20.535 --> 11:22.635
for private companies

11:22.695 --> 11:27.235
to explicitly write into their grant contracts that they,

11:27.815 --> 11:29.755
you know, want the researchers to patent

11:29.945 --> 11:32.995
that they will maintain control of the patents.

11:33.175 --> 11:37.155
And so they're actually putting conditions on these grants

11:37.545 --> 11:40.555
that they're negotiating with these, these researchers

11:40.825 --> 11:44.075
that they will create these appropriable,

11:44.075 --> 11:45.195
pieces of property.

11:45.615 --> 11:48.475
Um, which again, then presumably are less free to go

11:48.475 --> 11:51.155
to high tech companies founded by the researchers

11:51.255 --> 11:54.005
and more likely to flow into these large companies

11:54.305 --> 11:57.285
to be commercialized, through that method.

11:57.945 --> 12:01.765
So we found that these federal funding cuts have a large

12:01.765 --> 12:03.605
effect on individual researchers

12:03.605 --> 12:06.925
and what individual researchers do in specific,

12:07.475 --> 12:10.045
when you have less federal funding, they're less likely

12:10.105 --> 12:13.925
to found, high tech companies,

12:14.195 --> 12:15.605
they're more likely to patent

12:15.945 --> 12:17.405
and they're less likely to publish.

12:18.025 --> 12:22.685
And so the larger picture here is that this, both

12:22.685 --> 12:26.005
because it decreases funding, decreases people's output,

12:27.225 --> 12:31.125
but it shifts the way that this r

12:31.125 --> 12:33.165
and d is able to get out into the world

12:33.185 --> 12:34.685
and is able to be commercialized.

12:35.025 --> 12:36.965
So federal funding, as I said

12:36.965 --> 12:40.085
before, explicitly leaves it open for the universities

12:40.145 --> 12:44.655
and the researchers to, commercialize themselves,

12:44.875 --> 12:47.215
to found companies to sort of do

12:47.215 --> 12:49.255
what they want in, in that way.

12:49.525 --> 12:52.575
Whereas when the federal funding is cut

12:52.595 --> 12:55.455
and researchers shift to private funding, they

12:56.075 --> 13:00.655
now have their research outcomes go into these companies

13:01.115 --> 13:02.975
to be commercialized through these companies.

13:03.435 --> 13:06.175
And I don't think it's clear which one is a better way

13:06.175 --> 13:07.175
of commercialization.

13:07.595 --> 13:10.615
On the one hand, we believe that, you know,

13:10.725 --> 13:14.415
more innovative ideas are better situated in small companies

13:14.415 --> 13:16.175
that can really run with them.

13:16.755 --> 13:19.735
Uh, unlike the way large companies sort of.

00:05 - The research question we have is, how does a decline in federal funding affect the outputs of university research? So, we're interested in this questions for several reasons. You know, one, we believe that university research is very important

00:19 - to the US economy, to the innovation of new, ideas and new technologies in the US economy. It's a large percentage of the r and d that happens in the US economy. Two, a large percentage of university research is impacted by federal funds. And we've seen a significant decrease in this percentage of university research impacted by federal funds over the last 40 years, as well as a significant decrease in the federal share of r

00:49 - and d spending over the last 40 years as a whole. Universities have responded to this by thinking about how can they get funding that's not federal government, because they're very reliant on federal government funding. And many universities have explicitly said, you know, we want to shift from being just reliant on federal funding to funding in the private sector to private companies giving grants to universities to engage in r and d. And they've explicitly looked

01:19 - to shift their funding from federal funding to private funding because they think this diversified funding portfolio will be more stable. Now, we don't engage in this question of stability, but instead we ask how individual researchers that are supported by this funding are impacted by declines in federal funding and possible increases in private funding that result from that. There are a number of papers where we know

01:47 - that federal funding, you know, increases r and d and increases private spillovers. What those papers show is that you have more federal funding and you get more of innovation outcomes that you would expect. Patents and papers. as well as that, when you see increases in federal funding in specific areas, you also see sort of this crowd in effect that there's more private funding. But these papers that exist, this sort

02:15 - of look at this are all in the aggregate level. So what our paper does that's different is we're actually following the careers of individuals who are sponsored by this grant. So it's not just what happens when a field gets more, funding. Does that field or that university also get more private funding? Does that field or that university also produce more papers or patents?

02:40 - But how are the individuals that are supported by these grants affected in producing more papers or patents? And then we are also able to look directly through use of census data at individuals, engaging in entrepreneurship, which we define as, being at a new high tech age zero high tech company. and so, you know, that's something that previous papers really have not been able to get at. But because through census data we can follow people's

03:12 - careers over time, we're able to actually follow them in that way. The data we're using in the paper is based on some really large efforts to collect administrative data and bring administrative data together. So we start off with a project that's out of the University of Michigan, called umetrics, and that collects administrative data on research grants and awards for, in our case, 22 universities. They're continuing to try to expand this project,

03:44 - so I think they're actually up to 30 some universities now, which has individual level grant payments for all employees. And so then the census takes these grant individual level grant payments in and, links them to our data on individuals. We also know attributes of the grants. So in specific, the thing we use in the the paper are CFDA codes. That's catalog of Federal Domestic Assistance, which is a way that the US government categorizes

04:17 - its grants into roughly to speak of fields. so then we can link that to data on the CFDA codes as a whole. So they aggregate data in that field. On the other side, we can bring these individuals into census and we can link to IRS W2 forms, which have, you know, the individual, how much they're paid and who pays them, or at least what federal tax ID is paying them. we can bring it to the LEHD, the longitudinal

04:50 - employer household dataset, which is from UI records, state UI records, and again, is linking an individual's employer through a tax ID to the individual so you know how much they're paid and where they're, they work, as well as census data on these establishments. So knowing what industry the establishments are, knowing what year these establishments were founded, we're also able to link these individuals to their patents. And that's done through the Metrics project

05:21 - and their publications, which is also done through the metrics projects. So we're really able to bring this huge amount of different administrative and sort of, you know, large comprehensive sources of data together to look at these different outcomes. The main methodology we use here is an event study. So we specifically look for, CFDA codes. So I mentioned the, that's roughly a field that have had

05:51 - large shocks to funding in them. So we, this is sort of, you know, we select based on their characteristics. So we look for ones that have had at least a 40% decrease in federal funding or in funding, 'cause they're all federal grants. And then an increase again, and we say we don't want you to have more than a 20% increase or decrease in the two years before that 40% dip.

06:22 - and we want you to have at least returned, you know, 80% in the year after. So it really is a shock. So we're looking at something that we believe is quite idiosyncratic. We did a lot of work to try to look at why these shocks happened and, you know, most of them don't necessarily bear mentioning in the public record, so it was really hard to get data.

06:48 - But the ones we could find, it seems to be like something quite random that's uncorrelated with, with the tech fields. And we also did some testing to see if these shocks seem to be related to overall trends within these fields of patenting and, high tech startups. And they do not. So we use the shocks to create an event study. So we look at researchers who are paid on these CFDA codes, as well as controlled CFDA codes,

07:16 - which have had no more than a 20% variation, over the timeframe. And we say after this shock, what happens to your being at, high tech, age zero company to patenting and to publication? So it's a very simple event study set up. I had noted before that we wanted to understand the effect of these federal funding cuts on the funding of individual researchers.

07:48 - And that we saw, one, we saw when one of these fund federal funding cuts came in, that individual researchers saw a decrease in their funding levels that was persistent. But we were also interested in the question of whether this, whether people were able to, as university presidents had talked about substitute private funding for federal funding. So we asked the question of after one of these funding cuts, did we see

08:15 - an increase in private funding? And we do in fact, find an increase in private funding after one of these cuts. So total funding decreases by about 15%, which is about a 28% decrease in federal funding, but a 15% increase in private funding. So it's something like they're able to get 10 cents on the dollar of lost federal funding made up for in private funding. So we think that this shift from, federal funding

08:51 - to private funding is very important for what's Actually going on in these mechanisms, because we would expect from a pure loss of money that you would see all of the things decrease, that you would see less patenting, less publications, and fewer high tech entrepreneurship. You know, you're just less able to, engage in research, but the shift suggests that there's something more nuanced. And so in this, this, more nuanced shift, the question is why do you have this increase in patenting?

09:25 - And so one way to think about this would be that perhaps federal private funding is better at supporting applied research, whereas federal funding is better at supporting basic research, or at least, private funding is bad at supporting basic research. And so this would, you know, give you a reason why patenting would increase, even though funding decrease that people are doing more applied research.

09:50 - But this wouldn't explain the decrease in, high tech entrepreneurship because high tech entrepreneurship is going to be, you know, very applied as well. These are things that people think can, they can privatize and, and turn into companies. And so we think that what's going on is that this is an appropriation story. The government gives funding with a very open notice. They explicitly want the researchers

10:16 - and the universities to retain control of the things that they produce so that they can have spillover effects in the US economy. This is the goal of the federal government. they believe the public has paid for this research, and the public should in some sense get access to this research. In contrast, when a private company is funding research, they're doing it because they want those r and d outputs. They want presumably to commercialize those r and d outputs.

10:41 - And so they're interested in appropriating those r and d outputs. And so this explains why you would see fewer people taking the ideas that are produced on private, grants and turning them into their own companies and why you would see more of those ideas patented. Because patenting is, you know, intellectual property. It's a way of creating property that these, private companies can appropriate. And we see, in sort of our, our talking to

11:13 - tech transfer offices and to researchers that a very common thing, is for private companies to explicitly write into their grant contracts that they, you know, want the researchers to patent that they will maintain control of the patents. And so they're actually putting conditions on these grants that they're negotiating with these, these researchers that they will create these appropriable, pieces of property.

11:45 - which again, then presumably are less free to go to high tech companies founded by the researchers and more likely to flow into these large companies to be commercialized, through that method. So we found that these federal funding cuts have a large effect on individual researchers and what individual researchers do in specific, when you have less federal funding, they're less likely to found, high tech companies, they're more likely to patent

12:15 - and they're less likely to publish. And so the larger picture here is that this, both because it decreases funding, decreases people's output, but it shifts the way that this r and d is able to get out into the world and is able to be commercialized. So federal funding, as I said before, explicitly leaves it open for the universities and the researchers to, commercialize themselves, to found companies to sort of do

12:47 - what they want in, in that way. Whereas when the federal funding is cut and researchers shift to private funding, they now have their research outcomes go into these companies to be commercialized through these companies. And I don't think it's clear which one is a better way of commercialization. On the one hand, we believe that, you know, more innovative ideas are better situated in small companies that can really run with them.

13:16 - unlike the way large companies sort of.

Faculti AI can make mistakes