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

Two-Sided Market Power in Firm-to-Firm Trade

View more insights in Social Sciences - Economics - International Trade - Macroeconomics

View more figures from University of Southern California

View the page of Monica Morlacco

Read the Study

Monica Morlacco develop a quantitative theory of prices in firm-to-firm trade with bilateral negotiations and two-sided market power.

Image courtesy of interviewee. March 25, 2024

Generating Workspace

Export as PDF

In 2018, the United States introduced tariffs of 10-50% on various imports, igniting a trade war with key trading partners, which responded with retaliatory tariffs. This video video event highlighted a shift towards protectionist trade policies and raised concerns about their impact on consumer prices and welfare. Notably, tariffs imposed during the Trump administration resulted in price increases equating to the rate of the tariffs, meaning consumers suffered the full burden of these costs, contradicting traditional trade theories that suggested incomplete price pass-through.

Research aims to explore why the pass-through from these tariffs was so complete and why conventional theories failed to predict this outcome. The study emphasizes two-sided heterogeneity and market power in global production networks, noting that modern trade primarily involves firm-to-firm transactions rather than direct sales to consumers.

A novel theoretical framework is presented, focusing on price-setting dynamics in firm-to-firm relationships, where firms have substantial market power on both sides of the transaction. This video video approach leads to two main predictions: (1) bilateral markups vary significantly across relationships, and (2) the pass-through of tariffs may be complete for smaller importers while being incomplete for larger ones due to the dynamics of demand and marginal costs.

The research utilizes detailed data on firm-to-firm transactions, particularly in the context of Trump tariffs, to empirically test the model’s predictions. Results indicate that US importers generally exhibit greater bargaining power compared to foreign exporters, a departure from conventional assumptions, and that foreign exporters' marginal costs are responsive to output changes.

Lastly, the study concludes that the overall pass-through from Trump tariffs was around 40%, attributed mainly to adjustments in foreign exporters' marginal costs. This video video finding questions previous work suggesting higher pass-through rates and highlights the significance of analyzing price dynamics at the firm level to understand broader economic implications. The research contributes to the literature on trade policy and pricing, emphasizing the relevance of global value chains in current international trade scenarios.

Language:

Export as PDF

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

00:00.145 --> 00:03.245
In 2018, the US implemented a series

00:03.345 --> 00:07.765
of import tires ranging from 10 to 50% on a series

00:07.865 --> 00:09.205
of products and countries.

00:09.795 --> 00:14.205
This move, initiated what's been known famous, as,

00:14.345 --> 00:16.045
the trade war between the US

00:16.065 --> 00:18.565
and major trade partners who responded

00:18.565 --> 00:21.205
by imposing retaliatory tires on US exports.

00:21.745 --> 00:24.765
Now, this historic event, while somewhat unexpected,

00:25.315 --> 00:27.845
reflected the shift in the global landscape

00:27.935 --> 00:31.645
where protection in trade policies are now gaining traction

00:32.145 --> 00:35.365
and may potentially, change the nature of trade

00:35.465 --> 00:36.725
as we've known them for decades.

00:37.465 --> 00:40.045
So it's now very important to understand

00:40.185 --> 00:44.325
and reevaluate perhaps, the effect of trade policies on,

00:44.585 --> 00:46.365
prices and consumer welfare.

00:47.025 --> 00:50.565
Now, a striking finding emerged from studies of, the,

00:50.565 --> 00:51.845
the trade war that is,

00:51.845 --> 00:54.325
that the Trump tariff had a complete pass through on prices.

00:54.865 --> 00:59.205
Put it simply, if the US imposed a tariff by say 10%,

00:59.275 --> 01:02.085
then the US import price as well as the consumer prices

01:02.465 --> 01:04.565
of the same product went up by 10%.

01:04.955 --> 01:08.805
This also meant that the US consumers board the entire cost

01:08.985 --> 01:11.205
of tariffs imposed during the Trump administration.

01:11.945 --> 01:14.365
Now, this is surprising from the point of view

01:14.365 --> 01:15.805
of conventional trade theories,

01:15.855 --> 01:19.045
which would've predicted the pass to be incomplete instead.

01:19.705 --> 01:22.965
So against this backdrop, our research seeks

01:22.965 --> 01:24.445
to answer two questions.

01:24.865 --> 01:27.765
The first is, why was the pass through

01:27.765 --> 01:28.845
of Trump party so high?

01:28.905 --> 01:30.525
If we can, explain why.

01:30.785 --> 01:33.485
And secondly, why did economists getting wrong?

01:34.105 --> 01:37.445
So, are conventional trade theories accurate

01:37.625 --> 01:40.845
or, able to predict the dynamics

01:40.845 --> 01:44.165
of prices following shocks or such as static shocks?

01:44.225 --> 01:47.285
Or is there need for reassessment prior literature

01:47.345 --> 01:48.765
as extensively dealt with?

01:49.105 --> 01:51.725
Um, trying to understand the pastoral shocks on prices,

01:52.015 --> 01:54.445
especially in the context of, in the, in the context

01:54.585 --> 01:58.845
of exchange rate shocks, we contribute to a smaller sort

01:58.845 --> 02:00.805
of subset of this literature that tries

02:00.805 --> 02:02.845
to understand the behavior of aggregate prices

02:03.265 --> 02:04.845
by looking at firm behavior.

02:04.945 --> 02:07.685
So what firms does at the much more granular level,

02:08.185 --> 02:10.205
and specifically our contribution

02:10.205 --> 02:12.645
to this literature is emphasizing the importance

02:12.665 --> 02:14.365
of two-sided heterogeneity

02:14.425 --> 02:15.885
and market power in the context

02:15.985 --> 02:18.845
of global production networks, which is something

02:18.845 --> 02:21.245
that prior literature has typically overlooked,

02:21.625 --> 02:24.525
but we deal them being important for understanding prices.

02:25.265 --> 02:28.165
Now, traditional trade theories,

02:28.635 --> 02:30.245
operate under the assumptions

02:30.315 --> 02:33.685
that exporters from one country sell directly

02:34.025 --> 02:36.685
to end consumers to in a destination country,

02:36.745 --> 02:39.045
and they sell final goods to final consumers.

02:39.435 --> 02:43.285
However, modern trade involves predominantly,

02:43.405 --> 02:44.645
a global value chains

02:44.665 --> 02:47.805
and the exchange of intermediate goods within firm

02:47.845 --> 02:51.045
to firm relationships rather than firm, firm

02:51.105 --> 02:52.125
to final consumers.

02:52.665 --> 02:55.325
And these fundamental shifts in the nature

02:55.345 --> 02:57.165
of international trade, which has to do

02:57.165 --> 02:59.730
with the secular rise of global production networks

02:59.730 --> 03:01.045
and the secular fragmentation

03:01.045 --> 03:04.725
of production carries very important implication for our,

03:04.785 --> 03:07.285
our, our understanding of prices and price changes.

03:08.245 --> 03:11.265
Now, in, in, in conventional trade theories,

03:11.375 --> 03:13.025
when exporter export sell

03:13.045 --> 03:15.985
to price staking consumers in the destination countries

03:16.045 --> 03:18.425
and to many consumers in the destination countries,

03:18.875 --> 03:20.265
conversely, in firm

03:20.265 --> 03:23.505
to firm trade exporters from one countries only sell

03:23.605 --> 03:24.705
to a limited number

03:24.925 --> 03:27.145
of importers in the destination countries.

03:27.205 --> 03:28.465
And this has to do with the fact

03:28.465 --> 03:30.985
that global value chains involve high fixed costs

03:31.045 --> 03:32.705
and, search costs.

03:32.705 --> 03:35.625
And so these relationships tend to be, the network

03:35.805 --> 03:37.705
of relationships tend to be sparse,

03:38.085 --> 03:39.905
and relationships tend to be repeated.

03:40.285 --> 03:42.385
Uh, so the firms tend to operate on, tend

03:42.385 --> 03:45.465
to transact on a repeated basis for prices.

03:45.735 --> 03:47.065
This has important implications

03:47.065 --> 03:50.505
because the prices within firm to firm relationships tend

03:50.505 --> 03:52.585
to be negotiated on a bilateral basis,

03:53.705 --> 03:56.985
challenging the conventional market clearing condition on,

03:57.085 --> 03:58.465
on traditional price models.

03:59.105 --> 04:01.225
Secondly, a second important feature

04:01.285 --> 04:04.465
of global value chains is that, firms on both sides

04:04.465 --> 04:06.425
of the transaction are dominant players.

04:06.535 --> 04:08.385
They're very large, and both sides

04:08.385 --> 04:10.905
of the transaction have market power over prices.

04:11.445 --> 04:13.545
And so there's two sided market power

04:13.545 --> 04:15.745
and not just on on white side of the market as,

04:15.745 --> 04:17.065
as it conventionally believed.

04:18.035 --> 04:20.615
Now, we, the, the, the same features also

04:21.235 --> 04:23.255
affect the way we understand pass through.

04:23.755 --> 04:26.415
In conventional models, exporter sets prices

04:26.555 --> 04:28.655
and they set the markup over the marginal cost.

04:28.965 --> 04:32.255
When a tariff hits the, the price, the marginal cost

04:32.255 --> 04:34.655
of the exporter to shipping the goods to the destination,

04:34.655 --> 04:35.775
country increases.

04:36.115 --> 04:39.735
And the exporter has the power to adjust the markup downward

04:39.955 --> 04:43.375
to avoid large losses in demand in the destination country.

04:43.435 --> 04:45.975
And that's what leads to incomplete posturing,

04:45.975 --> 04:47.135
conventional trade theories.

04:47.565 --> 04:49.295
However, in a scenario

04:49.305 --> 04:51.175
where there's bilateral bargaining power

04:51.175 --> 04:54.615
and take the extreme case where the importer set prices,

04:54.615 --> 04:56.535
for simplicity, then the import,

04:56.715 --> 04:58.855
the importer would set the markup as low

04:58.855 --> 05:00.415
as possible, say a markup to zero.

05:00.995 --> 05:03.175
And then so prices equal marginal cost.

05:03.515 --> 05:06.015
So when the, the tariff hits the marginal cost

05:06.015 --> 05:08.375
of delivering the good increases, the price has

05:08.375 --> 05:09.655
to increase one-to-one,

05:09.745 --> 05:13.135
which can explain like why the pasture can be complete in a

05:13.295 --> 05:14.495
scenario where there's two sided

05:14.495 --> 05:16.815
or, you know, two sided bargaining power.

05:17.395 --> 05:20.655
Now our research sort of builds on this simple intuition

05:20.675 --> 05:23.535
and delves into the nuances of, of, of these dynamics.

05:23.875 --> 05:26.335
Uh, in order to explain, the dynamics

05:26.335 --> 05:29.175
of pasture within global value chains, the main contribution

05:29.175 --> 05:31.495
of this paper is to develop some novel theory

05:31.635 --> 05:33.975
of prices in firm to FM relationships.

05:34.675 --> 05:38.095
So we take the network of firm tofe relationship

05:38.195 --> 05:40.055
as given this is a primitive of the model.

05:40.165 --> 05:41.895
It's something that we can observe in the data.

05:42.115 --> 05:45.215
So we just take this snapshot of firm to FM relationship

05:45.235 --> 05:48.335
as given and focus instead on the price decision problem

05:48.685 --> 05:51.935
between a foreign exporter and a US importer.

05:52.275 --> 05:55.455
Now, US importers buy inputs from all the

05:55.455 --> 05:56.655
exporters in their network

05:56.795 --> 05:57.795
Network. They combine,

05:57.795 --> 05:59.365
combine these inputs in production,

05:59.365 --> 06:02.325
they produce a product, then they then sell downstream.

06:02.905 --> 06:04.005
On the exporter side,

06:04.225 --> 06:06.725
the exporters produce their differentiated variety

06:06.745 --> 06:07.765
of the foreign input.

06:07.915 --> 06:09.045
They sell this product

06:09.185 --> 06:11.165
to all their importers in their network.

06:11.945 --> 06:14.045
And another assumption that we make,

06:14.135 --> 06:16.685
which is actually generalizing what's been done

06:16.685 --> 06:18.165
traditionally in the literature, is

06:18.165 --> 06:21.125
to allow the marginal cost of the exporter

06:21.125 --> 06:24.045
to increasing total output, which is something

06:24.235 --> 06:26.885
that makes sense in the context of global value chains,

06:26.895 --> 06:29.485
where there's a lot of relationship, specific investment,

06:29.755 --> 06:31.285
meaning that a fraction

06:31.285 --> 06:35.325
of inputs in the short run is fixed due to these investments

06:35.325 --> 06:37.685
that take a long time to, to materialize.

06:38.025 --> 06:39.805
And so leading to marginal cost

06:39.805 --> 06:41.645
to be increasing in output in the short term,

06:41.795 --> 06:43.325
this is gonna be important later on

06:43.325 --> 06:44.325
to understand the results.

06:45.235 --> 06:48.535
We then take the, the problem of of price setting decision

06:48.535 --> 06:49.615
between the quarter

06:49.635 --> 06:53.935
and exporter, the price solve the, the Nash product of the,

06:53.935 --> 06:56.775
of despair, meaning they saw the, the,

06:56.775 --> 06:59.095
this Nash product is the product of the gains from,

06:59.165 --> 07:00.695
from trade for the importers

07:00.695 --> 07:01.975
and exporter as weighted

07:02.035 --> 07:04.695
by their bargaining power in the relationship

07:05.035 --> 07:07.535
as in a standard, nash, problem.

07:08.355 --> 07:12.735
And what, and the, the gains from trade is the extra profit

07:12.965 --> 07:14.855
that the importer and exporter make

07:14.915 --> 07:17.415
by participating in the relationship versus not

07:17.735 --> 07:18.815
participating in that relationship,

07:18.835 --> 07:20.055
but carrying transaction

07:20.055 --> 07:22.375
with all their existing counterparts.

07:23.195 --> 07:26.415
And, and another thing that is important to mention is

07:26.415 --> 07:27.815
that what makes the, an assumption

07:27.815 --> 07:30.895
that makes this problem tractable is to assume

07:30.895 --> 07:32.455
that the price is the n in Nash

07:32.455 --> 07:34.095
bargaining solution for despair.

07:34.095 --> 07:37.855
Meaning it's a NH solution for the bilateral pair, assuming

07:37.855 --> 07:39.735
that the spare is in a n equilibrium,

07:39.735 --> 07:42.575
meaning taking every other negotiated outcome

07:42.605 --> 07:43.655
elsewhere in the network.

07:43.795 --> 07:45.135
As, as exos as given,

07:46.675 --> 07:49.655
the model delivers two sets of predictions.

07:49.915 --> 07:52.695
The first set is on the behavior of price levels

07:52.835 --> 07:54.175
and in particular the markups.

07:54.555 --> 07:56.615
And the second set of prediction is on the behavior

07:56.615 --> 08:00.015
of price changes following a shop that is a past elasticity.

08:00.555 --> 08:03.975
In terms of markups, we find that the markup, first

08:03.975 --> 08:06.895
of all, the, the each exporter set a markup

08:07.035 --> 08:08.335
to individual importers.

08:08.335 --> 08:10.015
So the price, the markup is bilateral.

08:10.365 --> 08:11.855
It's not an individual markup

08:11.855 --> 08:14.295
to all the importers in the exporters network,

08:14.555 --> 08:17.215
but the price, the markup varies, across firm

08:17.235 --> 08:18.855
to firm pairs, which is something

08:18.855 --> 08:20.375
that we do observe in the data.

08:20.435 --> 08:22.295
So a prediction that is born by the data

08:22.435 --> 08:24.575
and that traditional model wouldn't,

08:24.575 --> 08:25.975
be able to rationalize.

08:26.675 --> 08:28.295
Um, we then, we find

08:28.295 --> 08:30.855
that this bilateral markup can be understood

08:30.855 --> 08:32.535
as a weighted average between

08:32.535 --> 08:34.295
what we call an oligopoly markup,

08:34.305 --> 08:37.695
which reflect exporting pricing behavior like unilateral

08:37.695 --> 08:41.095
pricing behavior, and an oligo markdown which reflect

08:41.245 --> 08:43.495
importers unilateral price setting behavior.

08:43.875 --> 08:45.335
And the weight attached

08:45.335 --> 08:48.095
to the oligopoly oligo forces is proportional

08:48.115 --> 08:50.055
to the bargaining power of the exporter

08:50.055 --> 08:51.535
and importer respectively.

08:52.455 --> 08:55.725
A similar related prediction is that the markup

08:55.725 --> 08:58.725
or price should increase in the exporter's market share

08:58.865 --> 09:01.765
as long as the exporter has positive bargaining power is in

09:01.765 --> 09:03.885
the negotiations and should decrease

09:03.905 --> 09:05.125
in the importer's market share.

09:05.145 --> 09:08.245
So the market should be lower for large importers as long

09:08.245 --> 09:10.525
as importers have some bargaining power in the negotiation,

09:10.525 --> 09:12.365
which is a prediction where we can directly

09:12.365 --> 09:13.445
test in the data.

09:14.225 --> 09:16.645
Now the same, the behavior of market levels

09:16.865 --> 09:19.325
and the price levels explain, sort

09:19.325 --> 09:21.405
of help us understanding what's going on when we

09:21.405 --> 09:22.525
think about price changes.

09:23.465 --> 09:26.125
Um, again, in order to understand pass through,

09:26.125 --> 09:27.965
let's start from the conventional case

09:28.025 --> 09:29.525
of exporter setting prices.

09:29.995 --> 09:31.605
When the exporter set prices,

09:32.225 --> 09:35.085
the conventional wi wisdom applies that there is a tar,

09:35.485 --> 09:36.485
increases the marginal cost.

09:36.485 --> 09:39.285
There is downward markup adjustment incomplete pass through,

09:39.285 --> 09:40.845
just like in a simple example.

09:41.625 --> 09:46.045
So, the we, we learn two sets of, findings.

09:46.045 --> 09:48.085
So, so we learn two new things from our theory.

09:48.625 --> 09:51.925
The first thing is that other things equal the pass

09:51.925 --> 09:54.445
through increases in the bargaining power of the importers.

09:54.545 --> 09:57.125
And here the intuition is just like the simple example

09:57.155 --> 09:59.925
mentioned earlier, the as importers at prices,

10:00.065 --> 10:02.165
the markup goes down there is lower than one

10:02.165 --> 10:03.405
markup adjustment possible.

10:03.545 --> 10:05.525
And so the pass through tend to be more incomplete.

10:05.655 --> 10:08.085
Other things equal second.

10:08.425 --> 10:09.485
And this is,

10:09.535 --> 10:11.885
where the model gets a little bit more complicated.

10:12.185 --> 10:13.565
We do find that the past through,

10:13.565 --> 10:17.125
even when importer importers set prices can be both

10:17.525 --> 10:18.685
complete and incomplete.

10:19.185 --> 10:22.325
It is complete for small importers, importers that account

10:22.345 --> 10:24.485
for a tiny share of the exporters output

10:24.705 --> 10:28.165
for these importers, marginal costs is somewhat constant,

10:28.385 --> 10:31.245
and so the standard intuition of marginal cost pricing

10:31.305 --> 10:33.165
and full pass through apply for them.

10:33.595 --> 10:36.525
However, when importers are large, they account

10:36.525 --> 10:38.725
for a large share of the exporter output.

10:39.465 --> 10:42.525
And so, if the, if the tariff increases the price,

10:42.825 --> 10:44.365
the importers demand goes down,

10:44.745 --> 10:48.045
and as the importer demand goes down, total output produced

10:48.045 --> 10:49.485
by the exporter also goes down

10:49.485 --> 10:50.885
and so does the marginal cost.

10:51.065 --> 10:53.285
And this pushes the pass through to be incomplete.

10:53.785 --> 10:56.645
So overall, in order to understand pass through,

10:56.915 --> 10:58.645
just like markups, there's a lot

10:58.645 --> 11:00.725
of heterogeneity across firm to firm pairs in

11:00.755 --> 11:01.925
what the pass through could be.

11:02.585 --> 11:05.965
And, and so the model say that in order

11:05.965 --> 11:07.725
to understand aggregate password, one needs

11:07.725 --> 11:10.045
to understand what's going on at the very granular level

11:10.225 --> 11:13.805
and take a bottom up approach where, you know, in order

11:13.805 --> 11:16.325
to under understanding the aggregate means understanding

11:16.385 --> 11:17.845
how we aggregate from firm level

11:18.125 --> 11:19.445
behavior to aggregate behavior.

11:20.785 --> 11:23.325
We just, sorry, on the data.

11:24.315 --> 11:27.695
Uh, the second contribution of this paper is

11:27.695 --> 11:30.535
that we can test this model prediction on,

11:30.595 --> 11:31.735
on the data directly.

11:32.075 --> 11:35.295
And to do so, we construct a novel dataset,

11:35.725 --> 11:38.575
that merges reach information on firm

11:38.595 --> 11:41.775
to firm freight transaction for US importers

11:41.775 --> 11:42.895
and foreign exporter.

11:42.895 --> 11:45.855
This comes from the lifted dataset of the US census,

11:46.225 --> 11:47.655
which has information on price

11:47.655 --> 11:50.375
and quantities for firm to firm transaction dating back

11:50.375 --> 11:53.055
to 92, to 2019 or 2020.

11:53.785 --> 11:55.925
We merged this information with,

11:56.235 --> 11:58.845
information on firm level characteristics, both

11:58.865 --> 12:00.685
for importers and US importers,

12:00.685 --> 12:03.965
which come from the US census, of manufacturers

12:04.025 --> 12:07.205
or the longitudinal business dynamics datasets with,

12:07.265 --> 12:09.085
and also information on foreign exporters,

12:09.085 --> 12:11.365
which we obtained from the orbit dataset.

12:11.945 --> 12:15.645
Uh, finally, we also integrate information on trade policy.

12:15.785 --> 12:18.925
In particular, we focus on Trump tar as our case study.

12:19.305 --> 12:23.965
Uh, we use tariff, data, from, PA Obama quarters

12:23.965 --> 12:26.125
that, that work on returns to protectionism.

12:26.505 --> 12:29.445
Our empirical analysis, does several things.

12:29.935 --> 12:33.445
First of all, we confirm, we try, try

12:33.445 --> 12:36.245
to validate the model's prediction in what's called a

12:36.245 --> 12:39.125
reduced form approach, meaning simple regressions

12:39.305 --> 12:41.125
of prices and market shares.

12:41.385 --> 12:44.165
We do see that bilateral prices in particular bilateral

12:44.195 --> 12:45.205
markups tend

12:45.205 --> 12:48.605
to be higher in relationships involving exporters

12:48.875 --> 12:50.205
with higher market share

12:50.425 --> 12:53.485
and tend to be lower in relationship is involving importers

12:53.485 --> 12:57.285
with a larger buyer share, as predicted by a model

12:57.295 --> 12:59.565
where there's both bilateral bargaining.

12:59.585 --> 13:02.165
So both sides of the market has some bargaining power in

13:02.165 --> 13:04.805
negotiations as well as a model where the, on average,

13:05.185 --> 13:07.365
the marginal cost of the exporters tend

13:07.365 --> 13:08.925
to increase in total output.

13:09.665 --> 13:12.845
We then validate this prediction structurally.

13:12.905 --> 13:15.405
So we structurally estimate the model in the model scale

13:15.405 --> 13:18.085
elasticities and, and,

13:18.265 --> 13:20.605
and we find something which is kind of surprising,

13:20.605 --> 13:21.685
well, not surprising, but novel.

13:22.025 --> 13:23.205
The first thing we find is

13:23.205 --> 13:26.685
that on average in US importers ha the,

13:26.685 --> 13:29.125
the US importers bargaining power is about four times

13:29.225 --> 13:32.285
as large as the bargaining power, of foreign exporters.

13:32.705 --> 13:35.565
And again, remember that in, in conventional trade theories,

13:35.565 --> 13:38.285
the assumption is that foreign exporters have full

13:38.285 --> 13:41.165
bargaining power and importers have zero bargaining power.

13:41.225 --> 13:44.965
So what we estimate in our model is very different than

13:45.115 --> 13:47.485
what traditional theories, assume.

13:48.065 --> 13:49.725
And secondly, we also find

13:49.725 --> 13:52.965
that on average the exporter's marginal costs are very

13:52.965 --> 13:54.605
elastic to total output.

13:55.025 --> 13:57.085
Um, and, which implies

13:57.085 --> 13:59.525
that the exporter supply elasticity is,

13:59.665 --> 14:00.885
is quite, is is large.

14:01.145 --> 14:04.245
And, again, which is something that conventional models,

14:04.705 --> 14:07.485
have typically overlooked with without.

14:07.515 --> 14:11.485
With these estimates, we then use our model to try speak to,

14:11.705 --> 14:13.765
the, the consequences,

14:13.945 --> 14:16.325
of t both at the bilateral pair level

14:16.345 --> 14:17.365
and at the aggregate level.

14:17.785 --> 14:20.965
The first exercise that we do is a sort of a goodness

14:20.965 --> 14:22.765
of fit test of the model.

14:23.465 --> 14:27.925
Um, so we take, we start in 2017, we observe the network

14:27.985 --> 14:29.285
of firm to firm relationships,

14:29.585 --> 14:32.205
and then we hit the economy with the Trump tariffs.

14:32.465 --> 14:35.165
And we use the model to predict what's the effect

14:35.185 --> 14:39.045
of the Trump tariffs on the, on the prices at, at the firm

14:39.065 --> 14:40.445
to firm level.

14:41.025 --> 14:43.885
And so what the model predicts should be the change in price

14:44.025 --> 14:45.325
in, in its relationship

14:45.625 --> 14:47.725
and compare the model predicted price

14:47.755 --> 14:48.765
with the observed price.

14:48.785 --> 14:50.885
So press price change between 2017

14:50.945 --> 14:53.885
And 2018, and we cannot reject the null

14:53.995 --> 14:55.845
that the model predicted price change

14:55.905 --> 14:58.685
and the observed price change are equal to each other,

14:58.935 --> 15:02.645
which lends credibility to the model forecasting ability.

15:03.425 --> 15:05.605
So then the last thing we do is to say, okay,

15:05.625 --> 15:08.205
we take the model, what's the aggregate pass through

15:08.265 --> 15:10.085
of Trump tires bottom up.

15:10.585 --> 15:12.485
And the model says, well, the aggregate pass

15:12.485 --> 15:13.845
through is about 40%,

15:13.895 --> 15:15.805
which is much lower than a hundred percent.

15:15.875 --> 15:16.885
It's very incomplete.

15:17.145 --> 15:19.645
And the model attributes about a hundred percent

15:19.665 --> 15:23.125
of this incompleteness to downward cost adjustment

15:23.705 --> 15:26.005
of foreign, of foreign exporting, meaning

15:26.585 --> 15:28.365
the US importers on the one hand,

15:28.485 --> 15:30.125
a very large bargaining power.

15:30.225 --> 15:33.565
So they set prices, which other things equal push pass

15:33.565 --> 15:35.445
through to be larger than expected.

15:35.865 --> 15:39.165
But on the other hand, they also have a lot of oligo power,

15:39.235 --> 15:42.405
meaning they face a very elastic marginal cost schedule.

15:42.705 --> 15:45.005
And has the demand of US importers

15:45.605 --> 15:46.765
decrease due to the tariff?

15:46.765 --> 15:50.125
The marginal cost should also adjust downward, which is

15:50.125 --> 15:53.085
what explain the very incomplete passthrough.

15:53.175 --> 15:55.205
Where do we stand? Uh, what are we learn

15:55.425 --> 15:57.565
and what does our model have have

15:57.565 --> 15:59.885
to say about the incidents of Trump tariff?

16:00.185 --> 16:03.485
So on the one hand, our research proposes a new pricing

16:03.585 --> 16:05.725
theory of global value chains,

16:06.065 --> 16:10.445
and we show that the model is able to forecast the, the,

16:10.445 --> 16:11.725
the show the effect

16:11.725 --> 16:14.925
of trade policy on prices at the firm to firm level.

16:15.385 --> 16:18.125
So this is a, a price theory of global value chains,

16:18.125 --> 16:20.525
which is important, and as global value chains are gaining

16:20.765 --> 16:21.965
traction is, and they account

16:21.985 --> 16:24.085
for a large share of international trade.

16:24.275 --> 16:27.365
However, we do acknowledge that there is a discrepancy,

16:27.365 --> 16:29.125
which we, we between what the,

16:29.465 --> 16:31.285
our model say the aggregate incidents

16:31.285 --> 16:32.885
of Trump tariff should be,

16:32.985 --> 16:36.445
and what other studies of aggregate price changes have,

16:36.445 --> 16:38.965
have, have shown the about pass through.

16:38.965 --> 16:41.525
So we, we predict that the pass is very incomplete.

16:41.815 --> 16:43.765
These existing, studies show

16:43.765 --> 16:45.405
that the password was actually very complete.

16:45.945 --> 16:48.085
We do attribute this discrepancy

16:48.185 --> 16:50.605
to different sample selection in our studies

16:50.825 --> 16:52.205
for estimation purposes.

16:52.425 --> 16:55.005
Uh, we need to focus on a restricted sample

16:55.105 --> 16:57.285
of importers on a restricted sample of

16:57.845 --> 16:59.925
relationships which are repeated.

17:00.265 --> 17:04.525
And also and importantly, we do have to drop relationship

17:04.525 --> 17:08.645
between related party transaction means involving,

17:08.685 --> 17:10.845
relationship between firms that are related

17:10.945 --> 17:13.165
by a business relationship, which account

17:13.165 --> 17:14.925
for about 50% of total import.

17:14.925 --> 17:17.005
So we are dropping a, a big chunk

17:17.065 --> 17:18.965
of total imports in the United States.

17:19.265 --> 17:21.125
And the reason why we dropped this is be

17:21.125 --> 17:24.285
because we do believe that our theory applies for arms,

17:24.355 --> 17:26.885
arms length relationship rather than related party trade,

17:26.935 --> 17:30.125
where prices are not set by these sort of dynamics,

17:30.125 --> 17:31.765
but may reflect transfer pricing

17:31.905 --> 17:34.405
and other forces that we, that our model silent about.

17:34.985 --> 17:37.925
We are currently working on trying to pin down, which

17:37.965 --> 17:39.845
of this feature is most important

17:40.185 --> 17:41.965
for explaining the discrepancy.

17:42.465 --> 17:45.445
Uh, but we do believe it has to do with sample selection.

17:45.665 --> 17:48.685
But overall, our study contributes to the literature of pass

17:48.685 --> 17:50.805
through, providing a comprehensive understanding

17:50.805 --> 17:52.565
of prices in firm to term relationship.

17:52.565 --> 17:53.885
And we believe it,

17:54.025 --> 17:57.685
it is an important study helping policy makers understand

17:57.685 --> 18:00.485
the consequences of trade policy in the context of,

18:00.485 --> 18:01.565
of global value chains.

00:00 - In 2018, the US implemented a series of import tires ranging from 10 to 50% on a series of products and countries. This move, initiated what's been known famous, as, the trade war between the US and major trade partners who responded

00:18 - by imposing retaliatory tires on US exports. Now, this historic event, while somewhat unexpected, reflected the shift in the global landscape where protection in trade policies are now gaining traction and may potentially, change the nature of trade as we've known them for decades. So it's now very important to understand and reevaluate perhaps, the effect of trade policies on, prices and consumer welfare. Now, a striking finding emerged from studies of, the,

00:50 - the trade war that is, that the Trump tariff had a complete pass through on prices. Put it simply, if the US imposed a tariff by say 10%, then the US import price as well as the consumer prices of the same product went up by 10%. This also meant that the US consumers board the entire cost of tariffs imposed during the Trump administration. Now, this is surprising from the point of view of conventional trade theories, which would've predicted the pass to be incomplete instead.

01:19 - So against this backdrop, our research seeks to answer two questions. The first is, why was the pass through of Trump party so high? If we can, explain why. And secondly, why did economists getting wrong? So, are conventional trade theories accurate or, able to predict the dynamics of prices following shocks or such as static shocks? Or is there need for reassessment prior literature

01:47 - as extensively dealt with? trying to understand the pastoral shocks on prices, especially in the context of, in the, in the context of exchange rate shocks, we contribute to a smaller sort of subset of this literature that tries to understand the behavior of aggregate prices by looking at firm behavior. So what firms does at the much more granular level, and specifically our contribution to this literature is emphasizing the importance

02:12 - of two-sided heterogeneity and market power in the context of global production networks, which is something that prior literature has typically overlooked, but we deal them being important for understanding prices. Now, traditional trade theories, operate under the assumptions that exporters from one country sell directly to end consumers to in a destination country, and they sell final goods to final consumers.

02:39 - However, modern trade involves predominantly, a global value chains and the exchange of intermediate goods within firm to firm relationships rather than firm, firm to final consumers. And these fundamental shifts in the nature of international trade, which has to do with the secular rise of global production networks and the secular fragmentation of production carries very important implication for our,

03:04 - our, our understanding of prices and price changes. Now, in, in, in conventional trade theories, when exporter export sell to price staking consumers in the destination countries and to many consumers in the destination countries, conversely, in firm to firm trade exporters from one countries only sell to a limited number of importers in the destination countries. And this has to do with the fact

03:28 - that global value chains involve high fixed costs and, search costs. And so these relationships tend to be, the network of relationships tend to be sparse, and relationships tend to be repeated. so the firms tend to operate on, tend to transact on a repeated basis for prices. This has important implications because the prices within firm to firm relationships tend to be negotiated on a bilateral basis,

03:53 - challenging the conventional market clearing condition on, on traditional price models. Secondly, a second important feature of global value chains is that, firms on both sides of the transaction are dominant players. They're very large, and both sides of the transaction have market power over prices. And so there's two sided market power and not just on on white side of the market as, as it conventionally believed.

04:18 - Now, we, the, the, the same features also affect the way we understand pass through. In conventional models, exporter sets prices and they set the markup over the marginal cost. When a tariff hits the, the price, the marginal cost of the exporter to shipping the goods to the destination, country increases. And the exporter has the power to adjust the markup downward to avoid large losses in demand in the destination country. And that's what leads to incomplete posturing,

04:45 - conventional trade theories. However, in a scenario where there's bilateral bargaining power and take the extreme case where the importer set prices, for simplicity, then the import, the importer would set the markup as low as possible, say a markup to zero. And then so prices equal marginal cost. So when the, the tariff hits the marginal cost of delivering the good increases, the price has

05:08 - to increase one-to-one, which can explain like why the pasture can be complete in a scenario where there's two sided or, you know, two sided bargaining power. Now our research sort of builds on this simple intuition and delves into the nuances of, of, of these dynamics. in order to explain, the dynamics of pasture within global value chains, the main contribution of this paper is to develop some novel theory of prices in firm to FM relationships.

05:34 - So we take the network of firm tofe relationship as given this is a primitive of the model. It's something that we can observe in the data. So we just take this snapshot of firm to FM relationship as given and focus instead on the price decision problem between a foreign exporter and a US importer. Now, US importers buy inputs from all the exporters in their network Network. They combine, combine these inputs in production,

05:59 - they produce a product, then they then sell downstream. On the exporter side, the exporters produce their differentiated variety of the foreign input. They sell this product to all their importers in their network. And another assumption that we make, which is actually generalizing what's been done traditionally in the literature, is to allow the marginal cost of the exporter

06:21 - to increasing total output, which is something that makes sense in the context of global value chains, where there's a lot of relationship, specific investment, meaning that a fraction of inputs in the short run is fixed due to these investments that take a long time to, to materialize. And so leading to marginal cost to be increasing in output in the short term, this is gonna be important later on to understand the results.

06:45 - We then take the, the problem of of price setting decision between the quarter and exporter, the price solve the, the Nash product of the, of despair, meaning they saw the, the, this Nash product is the product of the gains from, from trade for the importers and exporter as weighted by their bargaining power in the relationship as in a standard, nash, problem. And what, and the, the gains from trade is the extra profit

07:12 - that the importer and exporter make by participating in the relationship versus not participating in that relationship, but carrying transaction with all their existing counterparts. And, and another thing that is important to mention is that what makes the, an assumption that makes this problem tractable is to assume that the price is the n in Nash bargaining solution for despair.

07:34 - Meaning it's a NH solution for the bilateral pair, assuming that the spare is in a n equilibrium, meaning taking every other negotiated outcome elsewhere in the network. As, as exos as given, the model delivers two sets of predictions. The first set is on the behavior of price levels and in particular the markups. And the second set of prediction is on the behavior of price changes following a shop that is a past elasticity.

08:00 - In terms of markups, we find that the markup, first of all, the, the each exporter set a markup to individual importers. So the price, the markup is bilateral. It's not an individual markup to all the importers in the exporters network, but the price, the markup varies, across firm to firm pairs, which is something that we do observe in the data. So a prediction that is born by the data

08:22 - and that traditional model wouldn't, be able to rationalize. we then, we find that this bilateral markup can be understood as a weighted average between what we call an oligopoly markup, which reflect exporting pricing behavior like unilateral pricing behavior, and an oligo markdown which reflect importers unilateral price setting behavior. And the weight attached

08:45 - to the oligopoly oligo forces is proportional to the bargaining power of the exporter and importer respectively. A similar related prediction is that the markup or price should increase in the exporter's market share as long as the exporter has positive bargaining power is in the negotiations and should decrease in the importer's market share. So the market should be lower for large importers as long as importers have some bargaining power in the negotiation,

09:10 - which is a prediction where we can directly test in the data. Now the same, the behavior of market levels and the price levels explain, sort of help us understanding what's going on when we think about price changes. again, in order to understand pass through, let's start from the conventional case of exporter setting prices. When the exporter set prices,

09:32 - the conventional wi wisdom applies that there is a tar, increases the marginal cost. There is downward markup adjustment incomplete pass through, just like in a simple example. So, the we, we learn two sets of, findings. So, so we learn two new things from our theory. The first thing is that other things equal the pass through increases in the bargaining power of the importers. And here the intuition is just like the simple example mentioned earlier, the as importers at prices,

10:00 - the markup goes down there is lower than one markup adjustment possible. And so the pass through tend to be more incomplete. Other things equal second. And this is, where the model gets a little bit more complicated. We do find that the past through, even when importer importers set prices can be both complete and incomplete. It is complete for small importers, importers that account

10:22 - for a tiny share of the exporters output for these importers, marginal costs is somewhat constant, and so the standard intuition of marginal cost pricing and full pass through apply for them. However, when importers are large, they account for a large share of the exporter output. And so, if the, if the tariff increases the price, the importers demand goes down, and as the importer demand goes down, total output produced by the exporter also goes down

10:49 - and so does the marginal cost. And this pushes the pass through to be incomplete. So overall, in order to understand pass through, just like markups, there's a lot of heterogeneity across firm to firm pairs in what the pass through could be. And, and so the model say that in order to understand aggregate password, one needs to understand what's going on at the very granular level and take a bottom up approach where, you know, in order

11:13 - to under understanding the aggregate means understanding how we aggregate from firm level behavior to aggregate behavior. We just, sorry, on the data. the second contribution of this paper is that we can test this model prediction on, on the data directly. And to do so, we construct a novel dataset, that merges reach information on firm to firm freight transaction for US importers

11:41 - and foreign exporter. This comes from the lifted dataset of the US census, which has information on price and quantities for firm to firm transaction dating back to 92, to 2019 or 2020. We merged this information with, information on firm level characteristics, both for importers and US importers, which come from the US census, of manufacturers or the longitudinal business dynamics datasets with,

12:07 - and also information on foreign exporters, which we obtained from the orbit dataset. finally, we also integrate information on trade policy. In particular, we focus on Trump tar as our case study. we use tariff, data, from, PA Obama quarters that, that work on returns to protectionism. Our empirical analysis, does several things. First of all, we confirm, we try, try to validate the model's prediction in what's called a reduced form approach, meaning simple regressions

12:39 - of prices and market shares. We do see that bilateral prices in particular bilateral markups tend to be higher in relationships involving exporters with higher market share and tend to be lower in relationship is involving importers with a larger buyer share, as predicted by a model where there's both bilateral bargaining. So both sides of the market has some bargaining power in negotiations as well as a model where the, on average,

13:05 - the marginal cost of the exporters tend to increase in total output. We then validate this prediction structurally. So we structurally estimate the model in the model scale elasticities and, and, and we find something which is kind of surprising, well, not surprising, but novel. The first thing we find is that on average in US importers ha the, the US importers bargaining power is about four times

13:29 - as large as the bargaining power, of foreign exporters. And again, remember that in, in conventional trade theories, the assumption is that foreign exporters have full bargaining power and importers have zero bargaining power. So what we estimate in our model is very different than what traditional theories, assume. And secondly, we also find that on average the exporter's marginal costs are very elastic to total output. and, which implies

13:57 - that the exporter supply elasticity is, is quite, is is large. And, again, which is something that conventional models, have typically overlooked with without. With these estimates, we then use our model to try speak to, the, the consequences, of t both at the bilateral pair level and at the aggregate level. The first exercise that we do is a sort of a goodness of fit test of the model.

14:23 - so we take, we start in 2017, we observe the network of firm to firm relationships, and then we hit the economy with the Trump tariffs. And we use the model to predict what's the effect of the Trump tariffs on the, on the prices at, at the firm to firm level. And so what the model predicts should be the change in price in, in its relationship and compare the model predicted price with the observed price.

14:48 - So press price change between 2017 And 2018, and we cannot reject the null that the model predicted price change and the observed price change are equal to each other, which lends credibility to the model forecasting ability. So then the last thing we do is to say, okay, we take the model, what's the aggregate pass through of Trump tires bottom up. And the model says, well, the aggregate pass through is about 40%,

15:13 - which is much lower than a hundred percent. It's very incomplete. And the model attributes about a hundred percent of this incompleteness to downward cost adjustment of foreign, of foreign exporting, meaning the US importers on the one hand, a very large bargaining power. So they set prices, which other things equal push pass through to be larger than expected. But on the other hand, they also have a lot of oligo power,

15:39 - meaning they face a very elastic marginal cost schedule. And has the demand of US importers decrease due to the tariff? The marginal cost should also adjust downward, which is what explain the very incomplete passthrough. Where do we stand? what are we learn and what does our model have have to say about the incidents of Trump tariff? So on the one hand, our research proposes a new pricing theory of global value chains,

16:06 - and we show that the model is able to forecast the, the, the show the effect of trade policy on prices at the firm to firm level. So this is a, a price theory of global value chains, which is important, and as global value chains are gaining traction is, and they account for a large share of international trade. However, we do acknowledge that there is a discrepancy, which we, we between what the, our model say the aggregate incidents

16:31 - of Trump tariff should be, and what other studies of aggregate price changes have, have, have shown the about pass through. So we, we predict that the pass is very incomplete. These existing, studies show that the password was actually very complete. We do attribute this discrepancy to different sample selection in our studies for estimation purposes. we need to focus on a restricted sample

16:55 - of importers on a restricted sample of relationships which are repeated. And also and importantly, we do have to drop relationship between related party transaction means involving, relationship between firms that are related by a business relationship, which account for about 50% of total import. So we are dropping a, a big chunk of total imports in the United States. And the reason why we dropped this is be

17:21 - because we do believe that our theory applies for arms, arms length relationship rather than related party trade, where prices are not set by these sort of dynamics, but may reflect transfer pricing and other forces that we, that our model silent about. We are currently working on trying to pin down, which of this feature is most important for explaining the discrepancy. but we do believe it has to do with sample selection. But overall, our study contributes to the literature of pass

17:48 - through, providing a comprehensive understanding of prices in firm to term relationship. And we believe it, it is an important study helping policy makers understand the consequences of trade policy in the context of, of global value chains.

Faculti AI can make mistakes