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

Global Firms in Large Devaluations

View more insights in Social Sciences - Economics - Microeconomics

View more figures from Boston University

View the page of Joaquin Blaum

Read the Study

Joaquin Blaum discusses the consequences of firms’ joint import and export decisions in the context of large devaluations.

Image courtesy of interviewee. May 9, 2024

Generating Workspace

Export as PDF

This video video discusses the implications of currency devaluations in international macroeconomics, particularly focusing on how economies adjust their import and export behaviors. The central premise is that during a devaluation, as foreign goods become relatively more expensive, countries tend to increase their overall import intensity, meaning they spend a larger share of their import expenditure on foreign inputs, despite the relative quantities of those inputs declining.

The author presents a study that examines 43 significant devaluations between 1970 and 2011, primarily observing that a 1% devaluation correlates with a 0.65% decrease in the relative quantity of foreign inputs but an increase in the aggregate import share. This video video phenomenon is attributed to the characteristic of firms that import and export simultaneously, revealing a robust correlation between these activities at the firm level.

The study argues that traditional trade models, which assume a quick decrease in the aggregate import share following a price increase for foreign inputs, do not accurately reflect real-world data. Instead, it posits a new model accounting for firms that are both importers and exporters, highlighting that devaluation can lead to an expansion in import-intensive firms.

Importantly, the findings suggest that while the aggregate import share may rise, the quantities of foreign inputs do not rise enough to offset their increased prices. This video video leads to a nuanced understanding of trade behavior post-devaluation, which is further illustrated through the experiences of Mexico and Indonesia.

Additionally, the findings indicate that the presence of global firms—those that are engaged in both importing and exporting—shapes responses in aggregate trade flow and can result in a more limited improvement in a country’s trade balance after a devaluation compared to traditional models without global firms. This video video complexity suggests that the global behavior of firms is critical to comprehending trade dynamics in the context of currency valuation changes.

Language:

Export as PDF

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

00:00.185 --> 00:02.925
One of the central and perhaps oldest questions in

00:02.925 --> 00:06.245
international macroeconomics relates to the consequences

00:06.305 --> 00:09.005
of currency devaluations, in particular, understanding

00:09.065 --> 00:11.205
how economies substitute between foreign

00:11.345 --> 00:14.845
and domestic goods during the valuations is the subject

00:14.905 --> 00:15.965
of a vast literature

00:15.985 --> 00:18.965
and has implications for many outcomes of interest, such

00:18.965 --> 00:22.285
as the response of aggregate trade flows that is exports

00:22.285 --> 00:23.765
and imports their difference.

00:24.185 --> 00:26.005
Um, that is the trade balance which matters

00:26.065 --> 00:27.925
for the country's national, you know, aggregate savings

00:28.425 --> 00:30.925
as well as domestic production costs, prices,

00:30.965 --> 00:32.405
productivity, et cetera.

00:33.155 --> 00:35.935
Now, at the same time, in recent decades,

00:35.935 --> 00:38.655
the world has seen a dramatic change in the nature

00:38.835 --> 00:41.175
of international trade, with the rise

00:41.175 --> 00:42.375
of global value chains

00:42.375 --> 00:45.495
and the fragmentation of production across countries so

00:45.495 --> 00:46.815
that the majority of the goods

00:46.815 --> 00:49.935
that are traded about two thirds in value are not final

00:50.025 --> 00:52.495
goods, but actually unfinished, goods.

00:52.555 --> 00:55.935
Uh, that is intermediate inputs that cross multiple,

00:55.935 --> 00:57.215
international borders

00:57.225 --> 01:00.015
until richer consumers somewhere, somewhere in the globe.

01:00.715 --> 01:03.375
Now, it turns out that this increased participation

01:03.375 --> 01:06.575
of countries in global trade, as importers

01:06.575 --> 01:09.335
and exporter actually hold at the firm level

01:09.685 --> 01:12.775
with the very same firms that rely on imported inputs

01:12.775 --> 01:15.455
to produce also exporting, their products.

01:16.185 --> 01:18.805
Now, this firm level correlation between importing

01:18.805 --> 01:21.645
and exporting is one of the, most robust features

01:21.705 --> 01:23.565
of modern international trade data.

01:24.245 --> 01:28.065
And so my study basically revisits the classic question

01:28.165 --> 01:31.385
of the consequences of the valuations with a focus on,

01:31.385 --> 01:33.505
the pattern of substitution,

01:33.925 --> 01:37.425
taking into account this new nature of international trade,

01:37.485 --> 01:39.825
namely, the, you know,

01:39.845 --> 01:42.905
firm's global participation in international trade as both,

01:43.045 --> 01:44.585
importers and exporters.

01:44.765 --> 01:47.785
The key empirical finding of my study is that

01:48.045 --> 01:49.345
during the valuations,

01:49.685 --> 01:52.505
as foreign goods become more expensive relative

01:52.505 --> 01:55.505
to domestic ones, the economy's overall import

01:55.505 --> 01:56.745
intensity rises.

01:57.095 --> 02:01.385
That is the expenditure in foreign inputs as a share

02:01.485 --> 02:04.665
of total import expenditure of the country goes up.

02:05.505 --> 02:07.385
I established this fact using a sample

02:07.385 --> 02:10.905
of 43 large devaluations that took place between 1970

02:11.125 --> 02:12.785
and 2011 2011.

02:13.215 --> 02:15.865
That includes the well-known effects, e events,

02:15.865 --> 02:19.145
of Argentina, 2002, the East Asian countries of,

02:20.045 --> 02:21.865
97, 98, et cetera.

02:22.485 --> 02:24.585
Now, in these devaluations, on average,

02:24.965 --> 02:27.105
the real exchange rate basically plummets.

02:27.105 --> 02:29.825
It depreciates by about 25% on impact

02:30.205 --> 02:33.585
and remains persistently depreciated for several years, even

02:33.585 --> 02:36.985
after, after eight years, it is still 15% below

02:37.685 --> 02:39.865
its pre-evaluation level, so large

02:39.925 --> 02:41.225
and highly persistent events.

02:41.645 --> 02:44.265
At the same time, the aggregate imported input share

02:44.705 --> 02:46.145
increases by about 20%

02:46.485 --> 02:50.025
and remains 15% above its pre devaluation level,

02:50.635 --> 02:51.785
again, after eight years.

02:52.645 --> 02:55.025
And in a battery of robustness checks, I show

02:55.025 --> 02:57.265
that the rise in the aggregate import chair is not

02:57.265 --> 02:58.665
driven by shocks

02:59.045 --> 03:00.925
Commonly associated to the valuations such

03:00.925 --> 03:02.685
as financial crisis, recessions,

03:03.065 --> 03:04.565
and, you know, many other confounders.

03:05.745 --> 03:08.565
Now, does this rise in the aggregated import share mean

03:08.565 --> 03:12.205
that the quantities of foreign inputs, increase relative

03:12.205 --> 03:13.405
to the quantities of domestic inputs?

03:13.625 --> 03:15.565
No. The relative quantities

03:15.565 --> 03:18.045
of foreign inputs fall consistent

03:18.115 --> 03:20.965
with downward lobbying demand, relative demand curve, which

03:20.965 --> 03:23.285
what we, which is what we all learn in, equal 1 0 1,

03:23.425 --> 03:25.445
but they do not fall by enough

03:25.665 --> 03:29.085
to offset the rise in the relative import prices making

03:29.115 --> 03:30.365
this import share go up.

03:31.035 --> 03:33.725
More specifically, I find that a 1% devaluation

03:33.725 --> 03:38.085
of the real exchange rate is associated with a point 65%

03:38.605 --> 03:40.445
decrease in relative import quantities.

03:40.715 --> 03:42.565
This means that the elegant elasticity

03:42.565 --> 03:45.045
of substitution at the country level between domestic

03:45.065 --> 03:47.445
and foreign inputs, known in the literature

03:47.445 --> 03:50.485
as the arming on elasticity, is actually lower than unity.

03:51.465 --> 03:52.565
Uh, and why do I care?

03:53.035 --> 03:56.045
Well, in modern quantitative trade models,

03:56.115 --> 03:59.285
that are widely used in, in, in, you know, for policy making

03:59.285 --> 04:02.485
and so on, they typically assume values for this elasticity

04:02.485 --> 04:04.965
that are much larger and in particular, larger than one.

04:05.265 --> 04:06.765
So they would imply a decrease

04:06.765 --> 04:07.805
in the aggregate import share.

04:07.825 --> 04:09.525
So it's in sharp, in star contrast.

04:09.985 --> 04:13.205
So my study also shows that this finding is linked

04:13.465 --> 04:16.445
to firm level patterns, in particular to the prevalence

04:16.445 --> 04:17.845
of imported exporter firms.

04:18.385 --> 04:21.045
So using micro data for Mexico in 1995,

04:21.045 --> 04:23.085
and for Indonesia in 1998, I showed

04:23.085 --> 04:26.645
that these devaluations are characterized by an expansion

04:27.105 --> 04:30.005
of import intent in firms That is firms that

04:30.005 --> 04:33.285
before the devaluation spent a high fraction of their,

04:33.305 --> 04:36.165
you know, total budget for inputs in, in imported inputs,

04:36.595 --> 04:38.205
they actually gain market share.

04:38.855 --> 04:40.875
And it is this pattern of reallocation

04:40.945 --> 04:43.155
that quantitatively explains the increase in

04:43.155 --> 04:44.235
the import share.

04:45.145 --> 04:47.285
Now, why would intense importers,

04:47.315 --> 04:51.005
grow their market shares when the price of the, you know,

04:51.105 --> 04:52.565
inputs they rely on so much,

04:52.565 --> 04:54.405
those are the foreign inputs goes up?

04:54.995 --> 04:57.045
Well, I show that the answer is related

04:57.385 --> 04:59.885
to the exporting activities of these importers.

05:00.175 --> 05:04.165
After the valuations, exporters strongly, grow, right?

05:04.165 --> 05:07.325
They gain market share. And these exporters actually are,

05:07.665 --> 05:09.725
you know, highly, importing intensive firms.

05:09.945 --> 05:12.245
As mentioned earlier, existing,

05:12.245 --> 05:14.805
quantitative models typically feature values

05:14.865 --> 05:17.245
of the armington elasticity that are greater than one,

05:17.295 --> 05:19.245
which are based on empirical studies

05:19.275 --> 05:21.205
that exploit low frequency

05:21.505 --> 05:23.485
or rather permanent variation in prices.

05:23.865 --> 05:26.125
For example, estimates of the trade elasticity

05:26.775 --> 05:30.725
based on variation in tariffs, taxes to imports

05:30.745 --> 05:33.245
or prices typically range between two

05:33.265 --> 05:36.045
and 10 approaches based on high frequency,

05:36.045 --> 05:37.165
higher frequency variation.

05:37.165 --> 05:40.725
Time series variation in contrast often yield low estimates

05:40.725 --> 05:44.285
of the arm elasticity, say at quarterly frequency.

05:44.285 --> 05:46.565
However, the valuations I study feature large

05:46.565 --> 05:48.245
and persistent changes in creative prices.

05:48.265 --> 05:49.445
So they don't really speak to that.

05:49.825 --> 05:51.165
Um, time series variation,

05:51.165 --> 05:52.645
high frequency time series variation.

05:53.625 --> 05:55.125
In terms of theory,

05:55.155 --> 05:56.765
perhaps it's not surprising given the

05:56.765 --> 05:58.830
Importance of input trade in world trade.

05:58.830 --> 06:00.845
There is a recent literature that developed models

06:01.105 --> 06:03.725
of input trade, with firm heterogeneity,

06:04.185 --> 06:06.605
but it has typically abstracted from the firm level

06:06.605 --> 06:08.485
correlation between importing and exporting.

06:08.485 --> 06:10.405
So they typically don't have exports, actually.

06:10.585 --> 06:13.165
Uh, so this includes my own previous work with clearly large

06:13.165 --> 06:15.085
and micro peters, as well as other papers.

06:15.585 --> 06:17.645
Now, these frameworks are firm level importing,

06:17.645 --> 06:20.565
predict a decrease in the aggregate import share when the,

06:20.585 --> 06:23.205
price of foreign inputs increases, which as,

06:23.345 --> 06:26.485
as it should be clear now, it's, you know, the opposite of

06:26.555 --> 06:28.525
what my empirical findings are in this study.

06:28.945 --> 06:30.005
So why do these,

06:30.325 --> 06:32.885
previous models predict high substitution

06:32.885 --> 06:34.525
between domestic and foreign inputs?

06:34.985 --> 06:36.365
Uh, in other words that the relative

06:36.365 --> 06:37.605
import quantities fall by a lot.

06:38.395 --> 06:40.455
The main reason, if these bullets predict

06:40.975 --> 06:43.655
a sharp contraction of intense importers,

06:43.655 --> 06:45.735
which are more exposed to the increase in the relative

06:45.735 --> 06:47.175
price of foreign inputs.

06:47.635 --> 06:48.975
And, but I've shown in the Mexican

06:49.115 --> 06:51.375
and Indonesian cases that this pattern of relocation

06:51.435 --> 06:53.895
of these models is at odds, with the,

06:53.895 --> 06:56.015
with the experience of Mexico and Indonesia,

06:56.015 --> 06:58.695
where actually intensive importers grow, they don't contract

06:58.835 --> 07:01.695
to account for the rise in the import intensity

07:01.875 --> 07:04.655
and the expansion of import intensive firms observed

07:04.655 --> 07:06.615
after the valuations, I proposed a theory

07:06.905 --> 07:09.815
where firms can participate in international markets, both

07:09.835 --> 07:11.935
as importers and exporters.

07:11.935 --> 07:13.935
The model essentially bridges existing models

07:13.955 --> 07:17.575
of input trade, which typically abstract, from exporting

07:17.575 --> 07:20.335
with a canonical model of firm level exporting.

07:20.435 --> 07:22.175
Namely the famous Malitz model.

07:22.605 --> 07:25.375
Importing is a mechanism to reduce production costs

07:25.675 --> 07:28.855
as imported inputs are complementary with domestic inputs

07:28.915 --> 07:31.935
and may feature lower prices or higher quality.

07:32.725 --> 07:34.265
Now, importing,

07:34.265 --> 07:36.225
however, is limited by fixed costs,

07:36.225 --> 07:39.105
which accrue per country the firm imports from, so

07:39.105 --> 07:40.585
that larger firms,

07:40.585 --> 07:43.225
because of these fixed costs import from more countries and

07:43.225 --> 07:44.785
therefore will end up being more importing.

07:44.945 --> 07:48.545
Tennessee exporting on the other hand, is a mechanism

07:48.685 --> 07:50.985
to increase the firm's revenue sales, and

07:50.985 --> 07:52.865
therefore ends up increasing the returns

07:52.885 --> 07:55.785
to importing a given reduction in the unit.

07:55.855 --> 07:58.055
Cost from importing associated

07:58.055 --> 08:01.015
with importing is gonna be more profitable if it can be

08:01.015 --> 08:03.215
applied to more units, as is the case with exporters

08:03.215 --> 08:04.255
that are large and sell a lot.

08:04.985 --> 08:08.075
This induces a positive association naturally in the model

08:08.185 --> 08:10.315
between firm size, import orientation

08:10.335 --> 08:11.435
and export orientation.

08:12.415 --> 08:14.315
Now, a devaluation,

08:14.315 --> 08:16.235
brings about two effects in the model.

08:16.685 --> 08:19.235
First, as the price of foreign inputs increases,

08:19.385 --> 08:22.235
firm substitute their expenditure, you know,

08:22.235 --> 08:23.955
from imported towards domestic inputs.

08:24.455 --> 08:26.115
And so there is an increasing in cost,

08:26.375 --> 08:27.995
but it's gonna be more pronounced for the firms

08:27.995 --> 08:29.115
that are more important intensive,

08:29.525 --> 08:30.995
which makes them contract.

08:31.605 --> 08:34.185
Second, there is an increase in foreign demand

08:34.285 --> 08:36.225
as domestic prices and,

08:36.705 --> 08:40.465
measured in tradables the US dollars, basically,

08:40.465 --> 08:42.785
false, that makes exporting more profitable.

08:43.805 --> 08:45.335
This effect is gonna be more beneficial

08:45.435 --> 08:47.215
for intense exporters,

08:47.225 --> 08:49.295
which are somehow efficient exporters.

08:49.345 --> 08:51.455
Think of firms that have a low fixes

08:51.455 --> 08:53.655
of exporting per country, in the model.

08:53.915 --> 08:54.935
Now, the overall effect

08:54.935 --> 08:56.405
of the devaluation depends on the

08:56.405 --> 08:57.485
balance between these two forces.

08:57.625 --> 08:59.965
The first, the substitution effect that I first mentioned

08:59.965 --> 09:02.645
and the scale effect I later mentioned in the study.

09:02.925 --> 09:05.645
I show that the strength of these two effects can be trace

09:06.235 --> 09:08.365
back to, two simple objects.

09:08.365 --> 09:11.565
You can actually measure, the pre-evaluation, import

09:12.255 --> 09:14.845
share, and the export share so

09:14.845 --> 09:17.725
that these two firm level objects fully summarize all

09:18.005 --> 09:20.085
relevant information about the effects of a devaluation.

09:20.505 --> 09:23.165
Now, can the devaluation lead to the expansion

09:23.165 --> 09:24.245
of import intensive firms

09:24.425 --> 09:27.365
and an increase in the I import share that, that we do,

09:27.365 --> 09:28.805
that I documented in the data in the study?

09:29.345 --> 09:32.685
Yes, as long as the scale effects from exporting are strong

09:32.685 --> 09:34.965
enough, are strong, and there is enough correlation

09:34.965 --> 09:36.005
between importing and exporting,

09:36.065 --> 09:38.565
but ultimately it's gonna be a quantitative matter.

09:38.585 --> 09:41.885
It will depend on the entire distribution, of,

09:41.885 --> 09:45.365
firm size and import and export orientation model parameters

09:45.365 --> 09:46.565
and the full structure of the model,

09:47.175 --> 09:51.165
which is why I consider a quantitative exercise in an

09:51.165 --> 09:54.645
application to the Mexican devaluation of 1995, I showed

09:54.645 --> 09:56.925
that the model can quantitatively come to terms,

09:56.925 --> 09:59.125
with the patterns, of the data.

09:59.785 --> 10:01.685
So I choose model parameters.

10:01.685 --> 10:03.605
That is, I calibrate the model so

10:03.605 --> 10:07.085
that it replicates the distribution of firm, size

10:07.265 --> 10:08.605
and firm level import

10:08.605 --> 10:10.965
and export orientations, in the cross section

10:11.025 --> 10:12.565
before, the evaluation.

10:12.985 --> 10:15.125
So technically this, this requires

10:15.125 --> 10:17.605
that firms are heterogeneous in at least three dimensions,

10:17.625 --> 10:20.085
namely their efficiency, how good they are producing,

10:20.225 --> 10:21.885
as well as the fixes of importing

10:22.185 --> 10:24.445
and exporting, which control the firm's comparative

10:24.445 --> 10:25.925
advantage in importing and exporting.

10:26.065 --> 10:27.085
And you can have all kinds

10:27.085 --> 10:28.805
of cross correlation between these three things.

10:29.305 --> 10:31.325
Uh, that's how it works. Uh, so crucially,

10:31.325 --> 10:34.005
the model will replicate the strong positive association

10:34.005 --> 10:36.885
between import and export intensities in the data.

10:37.065 --> 10:39.325
That's kind of the robust feature of the,

10:39.325 --> 10:40.965
international trade data I mentioned earlier.

10:41.695 --> 10:43.595
So in a counterfactual devaluation

10:43.905 --> 10:47.395
that mimics the decline in the price of domestic inputs

10:47.455 --> 10:49.795
and wages, measuring tradables.

10:49.905 --> 10:53.595
Then in Mexico, the model predicts an increase in the area

10:53.675 --> 10:54.755
imported input share.

10:54.775 --> 10:57.075
That's good. That's what I had, in, in, in the,

10:57.075 --> 10:58.515
in the main, empirical findings.

10:58.695 --> 11:00.995
Now, at the same time, the model predicts a widespread

11:01.275 --> 11:02.915
reduction in firm level import share.

11:02.915 --> 11:04.235
Like the entire distribution

11:04.235 --> 11:05.715
of firm level import shares is kind

11:05.715 --> 11:08.155
of the marginal distribution shifts to the left in a way.

11:08.415 --> 11:10.555
So how, how does the aggregated import share go up?

11:10.705 --> 11:11.995
Well, because firm size

11:11.995 --> 11:14.435
and import intensity become more correlated

11:14.695 --> 11:17.995
as ex andante intense importers gain market share.

11:18.455 --> 11:22.035
And in turn, this relocation is linked to the expansion

11:22.295 --> 11:24.835
of ex andante intense exporters, which tend

11:24.835 --> 11:26.475
to be intense importers.

11:27.305 --> 11:30.525
Now, if instead we consider a standard model of importing,

11:30.945 --> 11:32.805
which abstracts from exporting

11:32.945 --> 11:36.005
or a version of the model of this, study,

11:36.215 --> 11:38.285
where firms can either import or export,

11:38.345 --> 11:39.605
but not both at the same time.

11:40.215 --> 11:41.645
These models,

11:41.695 --> 11:44.765
these alternative models will predict a sharp reduction in

11:44.765 --> 11:45.805
the aggregating import share,

11:46.185 --> 11:48.445
and they will feature this stark pattern of relocation

11:48.505 --> 11:52.765
by which intense importers sharply contract. Um, this means

11:53.075 --> 11:55.285
That the presence of global firms, that these firms

11:55.285 --> 11:57.285
that both import and export is crucial

11:57.425 --> 12:00.125
to reproduce the absurd pattern of substitution

12:00.125 --> 12:03.165
and reallocation, that we see in, the valuations.

12:03.265 --> 12:06.005
So what are the implications of firms, global behavior?

12:06.465 --> 12:09.045
Uh, I think there are many, but in the study I focus on too,

12:09.665 --> 12:11.965
the response of aggregate trade flows

12:12.465 --> 12:13.565
and of the trade deficit.

12:14.305 --> 12:17.245
So the presence of global firms matters for the response

12:17.245 --> 12:19.125
of exports and aggregate exports, in particular,

12:19.235 --> 12:22.685
because the changes in the sourcing strategies like take,

12:22.785 --> 12:26.125
of, from how many countries firms are importing,

12:26.305 --> 12:29.565
of exporters, they're gonna map into changes in the

12:29.685 --> 12:32.805
production cost of exporters, which could lead to larger

12:32.945 --> 12:35.005
or smaller reaction of their exports.

12:35.005 --> 12:36.045
It's gonna depend on,

12:36.265 --> 12:38.965
on whether the exporters become more important intensive

12:39.025 --> 12:40.885
or less important intensive, and anything can happen.

12:41.065 --> 12:43.165
Now, in the Mexican application that I,

12:43.165 --> 12:44.605
that I do in the study, I actually find

12:44.605 --> 12:46.405
that extending intense exporters,

12:46.755 --> 12:49.485
they actually end up increasing their import intensity

12:50.025 --> 12:52.085
fueled by the scale effects,

12:52.375 --> 12:54.685
which dominate the substitution, leading

12:54.705 --> 12:58.325
to a larger response of aggregate export exports compared

12:58.325 --> 12:59.965
to a model, without global firms.

13:00.725 --> 13:01.785
But what about imports?

13:02.325 --> 13:03.905
The adjustment of total imports

13:04.045 --> 13:07.105
to the valuation is radically different between a model

13:07.225 --> 13:09.105
with global firms and without.

13:10.195 --> 13:12.175
As global firms grow in size

13:12.195 --> 13:14.975
and import intensity, they demand greater quantities

13:14.975 --> 13:15.975
of imported inputs,

13:15.975 --> 13:18.575
and that leads to actually an increase in the levels of,

13:18.635 --> 13:20.095
in imports of inputs.

13:20.115 --> 13:21.215
So we have a devaluation

13:21.215 --> 13:22.575
that makes imported inputs more expensive,

13:22.575 --> 13:23.655
and the actual, not the relative,

13:23.655 --> 13:26.895
but the actual level of, of, of imports,

13:26.895 --> 13:28.015
of inputs goes up.

13:28.435 --> 13:30.735
And that's in stark contrast with the standard models.

13:30.735 --> 13:32.055
Without global firms that are gonna,

13:32.055 --> 13:35.215
they're gonna predict a decrease in total imports.

13:35.435 --> 13:38.215
Now, this higher growth of imports positive,

13:38.215 --> 13:41.495
but actually higher, with global firms is so strong

13:41.725 --> 13:45.335
that it more than upsets the larger growth in exports

13:45.335 --> 13:48.255
and leads to a lower reduction in the trade deficit visit.

13:48.725 --> 13:51.375
That is a lower improvement in the trade balance.

13:51.805 --> 13:54.215
What does this mean? It means that the presence

13:54.215 --> 13:56.975
of global firms reduces the country's ability

13:57.075 --> 13:59.575
to improve its external position,

13:59.575 --> 14:01.215
through a currency devaluation.

00:00 - One of the central and perhaps oldest questions in international macroeconomics relates to the consequences of currency devaluations, in particular, understanding how economies substitute between foreign and domestic goods during the valuations is the subject of a vast literature

00:15 - and has implications for many outcomes of interest, such as the response of aggregate trade flows that is exports and imports their difference. that is the trade balance which matters for the country's national, you know, aggregate savings as well as domestic production costs, prices, productivity, et cetera. Now, at the same time, in recent decades, the world has seen a dramatic change in the nature of international trade, with the rise

00:41 - of global value chains and the fragmentation of production across countries so that the majority of the goods that are traded about two thirds in value are not final goods, but actually unfinished, goods. that is intermediate inputs that cross multiple, international borders until richer consumers somewhere, somewhere in the globe. Now, it turns out that this increased participation of countries in global trade, as importers

01:06 - and exporter actually hold at the firm level with the very same firms that rely on imported inputs to produce also exporting, their products. Now, this firm level correlation between importing and exporting is one of the, most robust features of modern international trade data. And so my study basically revisits the classic question of the consequences of the valuations with a focus on, the pattern of substitution, taking into account this new nature of international trade,

01:37 - namely, the, you know, firm's global participation in international trade as both, importers and exporters. The key empirical finding of my study is that during the valuations, as foreign goods become more expensive relative to domestic ones, the economy's overall import intensity rises. That is the expenditure in foreign inputs as a share of total import expenditure of the country goes up.

02:05 - I established this fact using a sample of 43 large devaluations that took place between 1970 and 2011 2011. That includes the well-known effects, e events, of Argentina, 2002, the East Asian countries of, 97, 98, et cetera. Now, in these devaluations, on average, the real exchange rate basically plummets. It depreciates by about 25% on impact and remains persistently depreciated for several years, even

02:33 - after, after eight years, it is still 15% below its pre-evaluation level, so large and highly persistent events. At the same time, the aggregate imported input share increases by about 20% and remains 15% above its pre devaluation level, again, after eight years. And in a battery of robustness checks, I show that the rise in the aggregate import chair is not driven by shocks

02:59 - Commonly associated to the valuations such as financial crisis, recessions, and, you know, many other confounders. Now, does this rise in the aggregated import share mean that the quantities of foreign inputs, increase relative to the quantities of domestic inputs? No. The relative quantities of foreign inputs fall consistent with downward lobbying demand, relative demand curve, which what we, which is what we all learn in, equal 1 0 1,

03:23 - but they do not fall by enough to offset the rise in the relative import prices making this import share go up. More specifically, I find that a 1% devaluation of the real exchange rate is associated with a point 65% decrease in relative import quantities. This means that the elegant elasticity of substitution at the country level between domestic and foreign inputs, known in the literature as the arming on elasticity, is actually lower than unity.

03:51 - and why do I care? Well, in modern quantitative trade models, that are widely used in, in, in, you know, for policy making and so on, they typically assume values for this elasticity that are much larger and in particular, larger than one. So they would imply a decrease in the aggregate import share. So it's in sharp, in star contrast. So my study also shows that this finding is linked to firm level patterns, in particular to the prevalence

04:16 - of imported exporter firms. So using micro data for Mexico in 1995, and for Indonesia in 1998, I showed that these devaluations are characterized by an expansion of import intent in firms That is firms that before the devaluation spent a high fraction of their, you know, total budget for inputs in, in imported inputs, they actually gain market share. And it is this pattern of reallocation that quantitatively explains the increase in

04:43 - the import share. Now, why would intense importers, grow their market shares when the price of the, you know, inputs they rely on so much, those are the foreign inputs goes up? Well, I show that the answer is related to the exporting activities of these importers. After the valuations, exporters strongly, grow, right? They gain market share. And these exporters actually are, you know, highly, importing intensive firms.

05:09 - As mentioned earlier, existing, quantitative models typically feature values of the armington elasticity that are greater than one, which are based on empirical studies that exploit low frequency or rather permanent variation in prices. For example, estimates of the trade elasticity based on variation in tariffs, taxes to imports or prices typically range between two and 10 approaches based on high frequency,

05:36 - higher frequency variation. Time series variation in contrast often yield low estimates of the arm elasticity, say at quarterly frequency. However, the valuations I study feature large and persistent changes in creative prices. So they don't really speak to that. time series variation, high frequency time series variation. In terms of theory, perhaps it's not surprising given the

05:56 - Importance of input trade in world trade. There is a recent literature that developed models of input trade, with firm heterogeneity, but it has typically abstracted from the firm level correlation between importing and exporting. So they typically don't have exports, actually. so this includes my own previous work with clearly large and micro peters, as well as other papers. Now, these frameworks are firm level importing, predict a decrease in the aggregate import share when the,

06:20 - price of foreign inputs increases, which as, as it should be clear now, it's, you know, the opposite of what my empirical findings are in this study. So why do these, previous models predict high substitution between domestic and foreign inputs? in other words that the relative import quantities fall by a lot. The main reason, if these bullets predict a sharp contraction of intense importers,

06:43 - which are more exposed to the increase in the relative price of foreign inputs. And, but I've shown in the Mexican and Indonesian cases that this pattern of relocation of these models is at odds, with the, with the experience of Mexico and Indonesia, where actually intensive importers grow, they don't contract to account for the rise in the import intensity and the expansion of import intensive firms observed after the valuations, I proposed a theory

07:06 - where firms can participate in international markets, both as importers and exporters. The model essentially bridges existing models of input trade, which typically abstract, from exporting with a canonical model of firm level exporting. Namely the famous Malitz model. Importing is a mechanism to reduce production costs as imported inputs are complementary with domestic inputs and may feature lower prices or higher quality. Now, importing,

07:34 - however, is limited by fixed costs, which accrue per country the firm imports from, so that larger firms, because of these fixed costs import from more countries and therefore will end up being more importing. Tennessee exporting on the other hand, is a mechanism to increase the firm's revenue sales, and therefore ends up increasing the returns to importing a given reduction in the unit. Cost from importing associated

07:58 - with importing is gonna be more profitable if it can be applied to more units, as is the case with exporters that are large and sell a lot. This induces a positive association naturally in the model between firm size, import orientation and export orientation. Now, a devaluation, brings about two effects in the model. First, as the price of foreign inputs increases, firm substitute their expenditure, you know,

08:22 - from imported towards domestic inputs. And so there is an increasing in cost, but it's gonna be more pronounced for the firms that are more important intensive, which makes them contract. Second, there is an increase in foreign demand as domestic prices and, measured in tradables the US dollars, basically, false, that makes exporting more profitable. This effect is gonna be more beneficial

08:45 - for intense exporters, which are somehow efficient exporters. Think of firms that have a low fixes of exporting per country, in the model. Now, the overall effect of the devaluation depends on the balance between these two forces. The first, the substitution effect that I first mentioned and the scale effect I later mentioned in the study. I show that the strength of these two effects can be trace

09:06 - back to, two simple objects. You can actually measure, the pre-evaluation, import share, and the export share so that these two firm level objects fully summarize all relevant information about the effects of a devaluation. Now, can the devaluation lead to the expansion of import intensive firms and an increase in the I import share that, that we do, that I documented in the data in the study? Yes, as long as the scale effects from exporting are strong

09:32 - enough, are strong, and there is enough correlation between importing and exporting, but ultimately it's gonna be a quantitative matter. It will depend on the entire distribution, of, firm size and import and export orientation model parameters and the full structure of the model, which is why I consider a quantitative exercise in an application to the Mexican devaluation of 1995, I showed that the model can quantitatively come to terms, with the patterns, of the data.

09:59 - So I choose model parameters. That is, I calibrate the model so that it replicates the distribution of firm, size and firm level import and export orientations, in the cross section before, the evaluation. So technically this, this requires that firms are heterogeneous in at least three dimensions, namely their efficiency, how good they are producing, as well as the fixes of importing

10:22 - and exporting, which control the firm's comparative advantage in importing and exporting. And you can have all kinds of cross correlation between these three things. that's how it works. so crucially, the model will replicate the strong positive association between import and export intensities in the data. That's kind of the robust feature of the, international trade data I mentioned earlier. So in a counterfactual devaluation

10:43 - that mimics the decline in the price of domestic inputs and wages, measuring tradables. Then in Mexico, the model predicts an increase in the area imported input share. That's good. That's what I had, in, in, in the, in the main, empirical findings. Now, at the same time, the model predicts a widespread reduction in firm level import share. Like the entire distribution of firm level import shares is kind

11:05 - of the marginal distribution shifts to the left in a way. So how, how does the aggregated import share go up? Well, because firm size and import intensity become more correlated as ex andante intense importers gain market share. And in turn, this relocation is linked to the expansion of ex andante intense exporters, which tend to be intense importers. Now, if instead we consider a standard model of importing, which abstracts from exporting

11:32 - or a version of the model of this, study, where firms can either import or export, but not both at the same time. These models, these alternative models will predict a sharp reduction in the aggregating import share, and they will feature this stark pattern of relocation by which intense importers sharply contract. this means That the presence of global firms, that these firms that both import and export is crucial

11:57 - to reproduce the absurd pattern of substitution and reallocation, that we see in, the valuations. So what are the implications of firms, global behavior? I think there are many, but in the study I focus on too, the response of aggregate trade flows and of the trade deficit. So the presence of global firms matters for the response of exports and aggregate exports, in particular, because the changes in the sourcing strategies like take, of, from how many countries firms are importing,

12:26 - of exporters, they're gonna map into changes in the production cost of exporters, which could lead to larger or smaller reaction of their exports. It's gonna depend on, on whether the exporters become more important intensive or less important intensive, and anything can happen. Now, in the Mexican application that I, that I do in the study, I actually find that extending intense exporters, they actually end up increasing their import intensity

12:50 - fueled by the scale effects, which dominate the substitution, leading to a larger response of aggregate export exports compared to a model, without global firms. But what about imports? The adjustment of total imports to the valuation is radically different between a model with global firms and without. As global firms grow in size and import intensity, they demand greater quantities

13:14 - of imported inputs, and that leads to actually an increase in the levels of, in imports of inputs. So we have a devaluation that makes imported inputs more expensive, and the actual, not the relative, but the actual level of, of, of imports, of inputs goes up. And that's in stark contrast with the standard models. Without global firms that are gonna,

13:32 - they're gonna predict a decrease in total imports. Now, this higher growth of imports positive, but actually higher, with global firms is so strong that it more than upsets the larger growth in exports and leads to a lower reduction in the trade deficit visit. That is a lower improvement in the trade balance. What does this mean? It means that the presence of global firms reduces the country's ability to improve its external position, through a currency devaluation.

Faculti AI can make mistakes