YT YouTube2Text

YouTube2Text → Transcripts → Pierre-Olivier Gourinchas on the Story Behind the Number

Pierre-Olivier Gourinchas on the Story Behind the Number — Transcript

by IMF · 5,406 words · 773 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:09So, welcome back to another IMF uh video
0:13podcast. Today I have the great pleasure
0:16of sitting down with um outgoing IMF
0:19chief economist Pierre Olivier Goransena
0:23um who's been in the role since 2022 but
0:26is now heading back to UFC Berkeley to
0:29uh continue his uh invaluable research
0:32with this IMF experience in his back
0:36pocket. So uh Pierre Olivier, welcome
0:39back to the podcast.
0:41>> Thank you for having me, Bruce. Yeah.
0:42And I know this is going to be your last
0:44experience as chief economist, but um
0:47I'm hoping that you might consider
0:49coming back sometime soon as Professor
0:52Galen.
0:53>> I'd be happy to.
0:55>> So uh I'd like to just sort of dive in
0:58here with a a question about tariffs
1:01>> um which many economists thought that um
1:04or predicted that would upend the the
1:07global economy. and and it seems that
1:10that really hasn't happened at least to
1:12the extent that many thought it would.
1:14Do you think we overreacted perhaps to
1:17uh the prospect of tariffs?
1:20>> I wouldn't say that we overreacted. I
1:22mean we at the fund I think we were in a
1:24good place when I look back. we were
1:26sort of at the milder end of the
1:28downgrades that um everyone was doing
1:32>> in 2025 in April of 2025 after the
1:35announcement of the set of unilateral
1:37tariff imposed by the US on pretty much
1:39every country out there.
1:41>> Um and then the rest of the year even
1:43compared to our own um reference
1:47forecast as of um April 2025 we've been
1:51revising up through throughout 2025. Um
1:54but there were a number of reasons for
1:55that. Uh things changed. Uh among the
1:58things that changed I think that the
2:00most important one is the tariffs
2:02themselves were significantly scaled
2:04back down compared to what was announced
2:07initially on April 2nd
2:09>> in the Rose Garden. Uh and so of course
2:11if you have a smaller tariff shock then
2:13you're not going to have as big of a
2:14shock to the global economy. And that
2:17has played out. There were a number of
2:18deals that were signed. There were a
2:20number of pauses, exemptions etc. And so
2:23instead of being at 25%, we ended up
2:25closer to 9% 10%.
2:27>> Um the second thing is uh and perhaps we
2:30had factored this in but not enough. So
2:33there I would maybe uh you know go back
2:35and look at the way we did our our
2:37assessment is the incredible amount of
2:39resilience by um businesses, private
2:44sector uh global supply chains adapted
2:47very very quickly, maybe faster than we
2:50thought uh they could. uh and some of it
2:52in anticipation also. So there was a big
2:55trade boom ahead of uh April 2nd. A lot
2:59of countries, a lot of businesses were
3:00anticipating that tariffs might be
3:02coming and after all it was kind of
3:03telegraphed since uh the election
3:06campaign and the since the election of
3:08President Trump. So so a lot of
3:10prepositioning of inventories moving
3:12things around. So there was a big boom
3:14in trade in the early part of 2025.
3:17And then the other thing is there were a
3:18number of tailwinds that were stronger
3:21than we could have predicted back in
3:23April. One of which was the incredibly
3:25accommodating financial conditions.
3:27>> So that helped support the global
3:29economy. There was an investment AI tech
3:32boom. I'm sure we'll talk about some of
3:33that.
3:34>> Uh that also provided a lot of lift to
3:37countries like the US etc. So you put
3:38all of these things together and I would
3:40say that our assessment was right in
3:42terms of the direction of travel. Uh
3:44this was a negative shock. had increased
3:46inflation. It was mostly borne out by US
3:49consumers and businesses. Uh it slowed
3:52down mildly output but not in a
3:54extremely noticeable way because there
3:56were these tailwinds. Um and by early
3:592026, I think this was largely uh in the
4:02rearview mirror.
4:04>> Do you think that there are still things
4:06sort of bubbling beneath the surface
4:08that we need to worry about? Well, one
4:10of one of these would be if we have a
4:12reescalation of um uh tariffs uh between
4:16countries, but the landscape has
4:17changed. The Supreme Court ruling in the
4:20United States with uh restricting uh
4:22heavily the use of AIPA
4:24>> uh um as a basis for imposing tariffs
4:28require means that now if the US wants
4:31to put tariffs on industries or
4:33countries, it has to do an investigation
4:34section 301 and they're in the process
4:36of doing that. That doesn't mean that we
4:38cannot have an escalation of tariffs,
4:40but it's not going to be as widespread.
4:43It's not going to be as dramatic as what
4:45we saw in April 2025. So, I think the
4:48expectation is the tariffs are not going
4:49away. They're going to be replaced uh
4:52over time in a way that will continue to
4:54generate actually a fairly significant
4:57amount of money for the US Treasury. uh
5:00and uh but there's not um in my sense
5:03there isn't really a prospect that it
5:05could be a massive escalation on the
5:07scale of what we saw a bit more than a
5:09year ago.
5:11>> So one of the first uh challenges that
5:13you faced as a chief economist was
5:16inflation and it seems like your
5:19successor may be faced with a little bit
5:21of the same scenario. Uh what do you
5:23make of this recent rise in prices uh
5:26especially in the United States? Well,
5:28so this uh I mean the f first the um we
5:31have an uptick in headline inflation in
5:33the US and a number of other countries
5:35that's due to the uh spike in energy
5:37prices that's directly a consequence of
5:40the developments in the Middle East the
5:42closing of the straight of the uh
5:46disruptions in the supply of uh of oil
5:48and other products uh to the global
5:51market. Uh so that's clearly adding to
5:53inflation pressures but the inflation
5:55pressures in the US um uh I think we are
5:58in a different environment compared to
6:002022 both in the US and other parts of
6:02the world. Uh when I came to the fund uh
6:062022 inflation was already there. It
6:09didn't start with the disruption in
6:11energy markets due to Russia's invasion
6:13of Ukraine. They were already there.
6:15They were already rising. There were
6:16already a number of voices calling for
6:18tightening of monetary policy. And a lot
6:20of that was actually tied to um the
6:24support policies that were put in place
6:25at the time of COVID. They were u in a
6:29big way they were necessary. They were
6:30they were the right policies to put in
6:32place but on the back end they would
6:34create conditions in which there would
6:36be a surge in prices and we had that
6:38surge in prices and ongoing inflation.
6:41So that was kind of the background there
6:43and of course the surge in gas and oil
6:45prices in 2022 that added fuel to that
6:48fire. the fire was already there.
6:50>> Um, we are in a different situation now.
6:51In many countries, inflation had been
6:54stabilized back to central bank targets.
6:59US is a bit of an exception. Inflation
7:00has remained too high. It has not been
7:03coming down to 2% which is the Federal
7:04Reserve target. It's closer to 3% for
7:07headline and even when you look at core
7:10uh one of the measures that the Fed
7:12likes to look at is core PCE that still
7:14is not at 2%. So there was kind of an
7:17ongoing inflation problem that's that
7:18was there but not accelerating. 2022
7:21inflation was accelerating. So we're
7:23starting from a different condition and
7:25I think the 2026 shock may be more in
7:29the nature of I want to say a classic
7:32negative supply shock. You get a surge
7:34in energy price that's going to increase
7:37the price of energy and everything that
7:39uses energy. Whether that translates
7:41into persistent inflation requires other
7:43things. It requires deanchoring of
7:45inflation expectations. It requires wage
7:48price spirals. It requires that maybe
7:51the wrong policies are implemented where
7:53central banks are not responsive or even
7:55fiscal policy in trying to protect
7:57people becomes too expansionary. So it
7:59requires a number of things that
8:01transform what is just a surge in energy
8:03price that eventually is going to be
8:05temporary into something that is an
8:07ongoing problem. And we are not seeing
8:09that just yet. you know, some of the
8:12price increases as you uh just said
8:15there were is about, you know, the
8:18conflict in the Middle East, the
8:19conflict in in Ukraine, um and also the
8:23escalation of uh geopolitical tensions
8:26in the world. How would you uh
8:28characterize the political landscape
8:31today as compared to what it was when
8:34you stepped into the job four years ago?
8:37And and what does that mean for a chief
8:40economist in a global institution
8:43uh that is trying to provide some
8:45guidance to its membership, you know, as
8:48to how to navigate through all this
8:50stuff?
8:51>> Well, I think this has been a defining
8:52feature of uh my time at the IMF. I
8:56mean, I it didn't start in 2022.
8:58Certainly geopolitical tensions have
9:00always been there to some degree and one
9:02could argue that one of the first
9:04tremors was in 2017 2018 when the US
9:08imposed tariffs on China. Uh but
9:10certainly the uh Russian invasion of
9:12Ukraine uh the trade war the uh uh
9:17conflict in the Middle East. All of
9:18these things are I think um a a sign of
9:23a world in which geopolitical forces are
9:27really shaping the world uh we live in.
9:30So it's really a defining feature.
9:32Remember back in 2022 I wrote a piece
9:34for finance and development where I
9:36talked about uh the tectonic plates of
9:39geopolitical forces that you know things
9:41could look quiet for quite some time but
9:43then you get a jolt you get the
9:44equivalent of an earthquake and it
9:46ripples far and wide and I've seen we've
9:48I think we've seen a number of the a
9:50number of these ripples and we we're
9:52going to continue to see that. So the
9:54the world we live in, this shockprone
9:57world, um it's a consequence of this
10:00geopolitical tensions rising. I mean,
10:02not all shocks are the result of
10:04geopolitical tensions, but many of the
10:06ones we've had in the last four years
10:08are the consequence of these uh these
10:11tensions. And I think that's something
10:13that uh that's that's a reality of the
10:15of the world uh that we're in. I don't
10:18think this is going away anytime soon.
10:20And my biggest worry is that some of it
10:23might manifest itself in the form of
10:26economic conflict. A trade war is a form
10:28of economic conflict, but some of these
10:30manifest themselves in the form of
10:31actual conflict wars. And that can have
10:35uh much larger impact and consequences
10:37on livelihoods and welfare. And maybe
10:39more specifically uh I mean what do you
10:42think is the impact of of this
10:44disengagement
10:46uh between economic powers you know in
10:48that geopolitical
10:51those geopolitical tensions are driving
10:53some disengagement and I'm thinking
10:55mostly about the US and China what what
10:57do you think is the impact of that
10:59>> well the the the background here is
11:02really that the world is becoming
11:04multipolar
11:05>> uh we've lived through most of the
11:09post-war period. I mean, we had blocks
11:11for a while. Of course, there was a
11:12Soviet block and there was a western
11:14block. Uh the western block was uniolar,
11:18was dominated by the US. Then the period
11:21after the collapse of the Soviet Union,
11:23the world became truly uniolar. There
11:25was back Americana if you want. There
11:28was a form of acceptance of the norms
11:31and rules and institutions that were
11:33developed by the west and were
11:35championed by by the US. But I think
11:37since 20002 2010 and the rise of China
11:42as a major economic power uh in the
11:45world there is um the necessity of a
11:49rearrangement of how power is shared how
11:53how everything is organized and I think
11:55a lot of the um earthquakes and tremors
11:59etc these tectonic plates are related to
12:01that uh reorganization of the global
12:04order and so that's the multipolarity
12:06that is emerging. Uh the world is we are
12:09witnessing the birth of this
12:10multipolarity and it's it's painful and
12:12and and we are seeing conflict of
12:15various kinds emerging as a result of
12:16that. Now the biggest challenge going
12:19forward is how we deal with this
12:21multipolarity. Take it as a given. We're
12:23not going to roll it back. Take it as a
12:25given and organize how we work, how we
12:28uh function as a global economy, as a
12:31system in that multipolar world. uh so
12:34we can keep it integrated or it can
12:36become more fragmented. That's the
12:37choice. The choice is not whether we
12:39stay multipolar or not. That is not
12:40something that we control. We want other
12:43countries, we want other powers to be
12:45there and emerge. That's part of what it
12:47means to be catching up and to be
12:48growing and becoming developed economies
12:50is you have your seat at the table. Uh
12:53and so the question is how we arrange
12:55that new world.
12:57>> Absolutely. Um so shifting gears a
13:00little here. Uh AI is something that
13:04sort of uh took off uh in the middle of
13:07your uh term as chief economist and it
13:10certainly as you alluded to a little
13:12earlier it's driving a huge tech boom.
13:16Uh but some economists would argue that
13:20the AI factor is uh is being exaggerated
13:24in in the economy. Um so is like to what
13:28extent is AI supporting the economy
13:31today and and is it possible that um you
13:33know the promise of AI is actually
13:36masking uh you know deeper problems
13:39within the global economy.
13:41>> Well there there are two sides to this.
13:44Um one is the I would want to say this
13:47is the the demand side. So, the fact
13:50that AI at this point has tremendous
13:54potential, we all agree on this. Um, you
13:58know, we use it, I'm sure you use it, I
14:00use it, we're we're odd. And let's just
14:02step back and think that four years ago,
14:05this didn't exist in that shape. And the
14:10way it is already helping us is is
14:12tremendous in in a lot of the things we
14:14do. Um but it is not yet something to
14:17your point it's not yet something that
14:19we see as generating tremendous
14:21productivity gains. So when we look at
14:24the economies where you have the biggest
14:26deployment of AI and of course the US is
14:28one of them but other advanced economies
14:30as well. Um we are not seeing I mean
14:32we're seeing strong productivity in the
14:34US but I don't think we can ascribe it
14:36to the most recent development of AI.
14:37There is productivity in services. There
14:39is productivity in various things. It
14:41may be related to tech more broadly. But
14:43not clear that it's related to what
14:44happened with the development of large
14:47uh language models and chat GPTs and all
14:50the subsequent models that everyone
14:52thinks of when they think about AI that
14:54has not yet seeped into uh the fabric of
14:57the economy at a level where you can see
14:59it in the national statistics. It may
15:01happen. Everyone hopes it will happen. I
15:04pretty confident it will to some degree.
15:06And there's a lot of discussion about
15:07how big it could be, but we haven't seen
15:09it just yet. But what we're seeing very
15:11clearly right now is the investment boom
15:15that is related to AI. So that is in
15:18anticipation of these productivity
15:19gains. There is tremendous investment,
15:21data centers, hyperscalers, uh the
15:24demand for chips is off the charts. That
15:27is uh giving a boost to the economy just
15:30on the demand side. It's not on
15:32productivity side, it's on demand side.
15:34just tremendous amount of investment in
15:35that sector. It's sucking in capital
15:38from around the world. You can see
15:40foreign investors very excited about
15:42having a share of the pie, you know,
15:44investing in some of these companies are
15:45involved in AI. So, we're seeing all
15:47that excitement that generates um you
15:50know, stock market gains that generates
15:52wealth, that wealth generates
15:54consumption, that's another layer in
15:56terms of the demand side. So I think
15:58right now the effects are mostly on that
16:01demand side and that's certainly
16:02boosting US economies boosting economies
16:05in Asia that are very plugged into the
16:07AI tech supply chain. Uh maybe less so
16:10in other parts of the world are not as
16:12not as involved. The promise is
16:14eventually it's going to transform the
16:16way we do things,
16:18>> the way we produce knowledge, the way we
16:21um invent things, the way we um uh
16:24process things. um and we're going to
16:26we're going to become better at doing
16:28things and that's what productivity
16:29gains mean. But I think we don't have
16:31really a good sense of what shape this
16:34will take. Uh we don't have a good sense
16:36of what shape this will take when we
16:37look at labor markets. What kind of jobs
16:39are going to be displaced? I mean we
16:41hear stories about well maybe AI means
16:43we don't need coders anymore. Maybe
16:45maybe not. Maybe we need coders they're
16:47just going to become incredibly more
16:48productive at what they're doing and
16:49therefore we want more coders. Um, most
16:53importantly, I don't think we have any
16:55idea what kind of jobs are going to be
16:58emerging as a result of AI. Those are
17:00jobs that don't exist anymore. The the
17:02statistics I like to quote is from David
17:05Otter's work at MIT. Basically argues
17:07that most of the jobs that
17:10you know are exist now in advanced
17:13economy like the US or European
17:15countries didn't exist 40 years ago. So
17:19if you roll back the clock 40 50 years
17:20and you ask well what's what will happen
17:22to the jobs you might be worried but in
17:24fact there will be all kinds of new jobs
17:25that will emerge and so I think we're
17:27going to have something a little bit
17:28like that we just have no idea yet
17:30processes will be will be transformed
17:32the same thing can be said about who
17:34will gain from AI that will very much
17:36depend on the market structure the
17:38degree of competition are the gains
17:40going to be passed through through lower
17:42prices to the users of the technology
17:44the businesses that are adopting it or
17:46the individuals who are augmenting their
17:48own um capacities to do things with AI
17:52or will it be captured by the companies
17:55that have done all these investments
17:56that depends very much on regulation,
17:58market structure, competition. And one
18:00of the things here that is particularly
18:02important that goes back to the
18:04discussions we're having about geconomic
18:05fragmentation
18:07>> is um these many of these frontier
18:10models are incredibly expensive to
18:12develop and train. these billions of
18:15dollars of investment we're talking
18:16about there are really to just train
18:18these models super high fixed costs may
18:21be very low marginal cost
18:24that's an argument for increasing
18:27returns in scale effects if you have
18:30very very high fixed cost you want to
18:32have a very big market because you have
18:33to recoup those costs the world we're
18:36getting into is a world in which access
18:38to AI could become part of the tensions
18:41could be weaponized could be turned
18:44countries may decide that they don't
18:45want to be reliant on AI systems that
18:47are provided by another country. They
18:49want to develop their own may be
18:50legitimate but it means the scale at
18:53which you operate is much smaller
18:54because now it's the US or it's Europe
18:57or it's China and you don't have the
18:59ability to scale things to achieve
19:02efficiency. So there there are economic
19:04costs to geocconomic fragmentation that
19:06will also impact the way the AI
19:08technology is going to deploy. It's
19:10going to scale is going to be
19:11implemented. So just sticking with AI
19:13for a minute, um I mean as we
19:15increasingly depend on it to sort of do
19:18a lot of our thinking for us, um do you
19:22worry what the impact might be on our
19:23own cognitive skills down the road? I
19:26mean, what do you think all this means
19:28for the world of research, which is a
19:31world that that you've lived in most of
19:34your career?
19:35>> Well, at a very personal level, uh I'm
19:38kind of excited. Um I I will tell you
19:41why. And one of the tremendous
19:45um joys of the job I've been doing the
19:48last four years is you know I'm
19:50surrounded by 100 plus economists,
19:54PhD economists, talented researchers,
19:57incredibly active and uh innovative and
20:00uh creative. Um and so whenever I have a
20:05question or something I want to work on,
20:06something I want to explore, there are
20:08people in the research department or
20:10elsewhere in the fund are happy to
20:12engage and that's incredibly good for
20:14me. And I was always, you know, when I
20:16was thinking about my after I leave the
20:18fund, I would be then I would go back to
20:20my office, I would have maybe one
20:21research assistant and I would have my
20:23computer and I would be having to do all
20:25of this myself and it would be a
20:27tremendous loss of productivity from my
20:29end. And now I'm thinking, well, yes,
20:31it's true. It's still going to be a
20:32tremendous loss of productivity, but at
20:34least I get AI maybe will help me a
20:36little bit to compensate a bit. So I
20:37look at this as something that will
20:39soften the the the the lending if you
20:42want um and and as a plus. Now on the
20:46cognitive side, I think um so one
20:50anecdote um people will read when
20:53calculators come in.
20:54>> Yeah.
20:54>> Uh we're not going to know our log
20:56tables anymore. who are not going to
20:58know how to multiply things. I think the
21:00world adapted to having calculators,
21:02even having calculators in the
21:03classrooms and even having them when
21:05students are taking exams. People were
21:07worried when Google came on Google
21:09effect. Oh, we're not going to memorize
21:11stuff. We won't remember when, you know,
21:13Magna Carta happened. We won't remember
21:15because you can just Google it. Okay,
21:18yes, fine. But you also have a library
21:20of every old human knowledge that is at
21:21your fingertips and that's tremendously
21:23useful. I think we're going to have kind
21:25of the same adaptation when it comes to
21:27AI. Some of the processes will probably
21:29degrade just like we don't remember or
21:31log tables or maybe we're not so good
21:34with dates and things like that, but
21:36there are some things that we will still
21:38be very very good at doing and they will
21:40be augmented by by AI. I kind of like
21:42the way Luis Garano for instance is
21:44thinking about these things. He's he's
21:46talking about what he calls messy jobs.
21:48A lot of tasks cannot be codified. they
21:51cannot be uh put in an algorithm. Even
21:53AI is not going to be very good at doing
21:55them. A lot of these things require
21:57human interactions. They require kind of
21:58the soft touch and humans are kind of
22:01needed for that. So we are going to see
22:04a displacement of activities. Um I'm not
22:07saying it's going to be easy.
22:08Dislocations when they happen are never
22:10easy. So so I think the the um sectoral
22:13reallocation, the labor market
22:15disruptions, these are important things
22:17we should not underestimate. There is a
22:19tendency sometimes to see the pie
22:22growing and thinking okay this is great
22:24let's not worry about distributional
22:25effects but what will happen is
22:28certainly there will be jobs that are
22:29destroyed there will be livelihoods that
22:31are impacted and that can be complicated
22:34and if it happens very fast then it can
22:36be very complicated so the speed the
22:39scale those are things we want to be
22:41careful about but the direction of
22:43travel and where we're going to be going
22:45I don't think that it will make us
22:48dumber in um systemic sense. I don't
22:51think it will make us unemployed in a
22:53systemic sense. I think we'll end up
22:55having productivity gains.
22:57>> So, you've studied many things at a very
23:00deep level, but um and for a long time,
23:03and one of those things is dollar
23:05dominance. Um, you know, given what you
23:08you've experienced in your role as chief
23:10economist here, do you see a day uh
23:14anytime soon where uh you know, the
23:17reserve currency of choice will be
23:20something other than the dollar?
23:23>> Well, certainly not quite yet. I mean,
23:25the uh interlocking forces that are
23:29putting the dollar in the position it is
23:31now are still very very strong. uh and I
23:34don't think they are uh changing in any
23:37meaningful way. I mean there are small
23:40changes but really nothing um that is
23:42too worrisome. But if I come back to the
23:45question of the multipolar world
23:48>> I think ultimately that is the key
23:52question. A multipolar world that is a
23:55fragmented world cannot rely on just the
24:00dollar as the dominant currency whether
24:02for payments for reserves for anchoring
24:06one's own currency for issuance in
24:09international markets invoicing all of
24:12these things a fragmented world
24:14fragments and if it fragments on trade
24:16it will fragment on all these other
24:18dimensions as well. So the only way for
24:22the dollar to ultimately retain its
24:26dominant status in uh the world
24:30>> Mhm.
24:31>> which is really very very strong right
24:32now is if that multipolar world remains
24:36an integrated world. It has to remain
24:38integrated on trade. It has to remain
24:40integrated on payments on access to
24:44dollar assets. uh it has to be to remain
24:47integrated on um asset issuance etc. And
24:52those are policy choices. That's one of
24:55the reason why it's important to make
24:58sure we remain very uh integrated. Um
25:02and that also implies that um
25:07any attempt at weaponizing finance
25:11because of course if you have the
25:14dominant instrument and you have the
25:16dominant infrastructure
25:18it could be tempting to try to use that
25:20for leverage but the ultimate
25:22consequences of this would be more
25:24fragmentation and then losing that
25:26capacity. So I think this is where the
25:29discussion is and to me it is
25:31particularly interesting to see that
25:34while we've seen attempts at
25:36fragmentation in trade with tariffs and
25:39and the like there's been very little in
25:43terms of financial fragmentation.
25:46uh there have been financial sanctions
25:48against entities of course countries
25:50like the US have the ability to target
25:52and it's legitimate to be able to do
25:54that but nothing that would be
25:56systematic
25:58and I think that's also because there is
26:00a sense that this if it does unravel
26:04will have much bigger consequences so to
26:07answer your question I think it will
26:10remain a policy choice it will reflect
26:12whether we really make the efforts to
26:15keep an integrated structure with
26:17multilateralism,
26:19cooperation,
26:20uh rules-based system or whether we
26:23fragment further.
26:26What have you uh learned in the in the
26:29past four years about forecasting?
26:33And and with that in mind, where do you
26:35think this is all going? Like all the
26:37technology, um you know, all the trade
26:40tensions, all you know, the weakening of
26:43of u institutions in some cases. Um
26:47where is this all going?
26:50>> Well, so first on forecast, I mean, I
26:52could start by paraphrasing Yogi Bar.
26:54you know, forecasting is hard,
26:55especially if you're thinking about the
26:57future. So, um I I've learned a lot
27:02about um that side. You know, when you
27:04when you come from academia, um
27:07forecasting is like the poor parent of
27:10um academic research.
27:11>> Uh no one really does for it's not
27:14viewed as glamorous. It's not viewed as
27:16uh interesting sort of predicting the
27:18next growth number for the US or
27:20anything like that. Um so I came to this
27:22with almost zero experience and
27:26>> um certainly being a bit concerned about
27:28the fact that I would have to be the
27:29voice of the fund presenting what our
27:31forecast is when I myself uh thought
27:33that I didn't have a lot of creds uh in
27:35that space. uh but we have fantastic
27:38teams and I learned uh by working with
27:40them and I guess the most important
27:42thing for me and this is what I learned
27:45from um the work with the WEO and and uh
27:48the world economic outlook is a lot of
27:51forecasting isn't about the number
27:54whether it's going to be 2.7 2.8 eight
27:572.9 it's really about the story it's
28:00really about how you come to that what
28:02are the forces that are driving that
28:04number you know we're talking about
28:06terrorists versus AI financial
28:08conditions it's when you put these
28:10things in the mix and you try to unpack
28:12what's the direction of travel what is
28:13dominating and then from there what's
28:16useful with forecasting is not so much
28:17the number itself either is what it
28:21means for policy at the end of the day
28:24you know or I I think our role at the
28:26fund is not to come up with some oracle
28:29about where the global economy is going,
28:31but it's also about saying okay well if
28:34we go in that direction this is what if
28:36you're country A you should be doing
28:38this this and this and if you're country
28:40B well maybe for you the situation is a
28:42bit different you should think about
28:43doing that that and that and you can
28:44only make that assessment if you
28:46understand the underlying forces so a
28:48lot of the exercise is around that and
28:51for instance when you know in April we
28:53decided we didn't have a baseline
28:55because the proximity of the war in the
28:58Middle East to our round of projections
29:01was you know it was very very close. So
29:04we didn't have a lot of time to digest
29:06and think about what was happening. So
29:08we decided to have three scenarios but
29:10we didn't organize the scenarios in
29:12terms of what if the shock is bigger
29:14this way or big bigger that way. We
29:16organized it by thinking okay well what
29:18configuration of shocks would lead us to
29:21a different policy advice in a different
29:23space where for instance monetary policy
29:25would really have to tighten because now
29:27the shock is impacting inflation in the
29:29way that central banks can't ignore or
29:32what if financial conditions become so
29:34degraded that now we have to think
29:36differently about both fiscal and
29:38monetary policy and so the scenarios
29:39where entry points into policy
29:42discussions as opposed to just numbers
29:44that we want to compare and one is
29:45bigger and one is small. And I think
29:47that way of thinking about forecasting
29:49to me is much richer and much more um
29:53related I think to my own comparative
29:56advantage which is to kind of think
29:57about underlying drivers.
29:59>> So any words of wisdom for your
30:01successor like is is there anything that
30:03you wish you would have known walking
30:05into this job?
30:07>> I don't think I have words of wisdom and
30:09I I think the the fund is a tremendous
30:11place. uh I think uh when I came in I
30:13received uh incredible help in thinking
30:16and learning and I think he or she will
30:19be in the same position and I wish that
30:21person uh you know to have a run that is
30:24as fun uh and enriching as the one I had
30:28lia
30:30has been the uh chief economist at the
30:32IMF since 2022 and is now heading back
30:36to academia in sunny California. Um,
30:40thanks for everything you've done here.
30:41I know people will miss having you
30:43around and I wish you all the best.
30:45>> Thank you very much, Bruce.

About this transcript

This page contains the full transcript of Pierre-Olivier Gourinchas on the Story Behind the Number by IMF, generated from the public captions YouTube serves with the video. The transcript has 5,406 words across 773 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.