Economic Risks Versus Media Alarmism with Philipp Carlsson-Szlezak

28:08Did we learn something from the financial crisis about the lack of fiscal stimulus and maybe the pendulum swung too far the other way? What's your takeaway from that? No, for sure. Look, I think two crises were very different. You had in 2008 damaged balance sheets, not just in the banking system, but households. Their balance sheets had to be repaired. Households had to dig themselves out of that hole, had to rebuild their wealth. And that would have called for more intervention than what we got in 2008. In 2020, I think policymakers, politicians, they had internalized that learning. So they went extra large on the fiscal side.

28:47And that hole that COVID created was basically filled with fiscal stimulus. As you know, it's widely believed and accepted that this was extremely big, too much perhaps. And so we had an overshoot in certain consumption areas, particularly in the good space. There was an overshoot in consumption. It pushed up demand. Together with supply crunches, it pushed up inflation in an idiosyncratic and more cyclical way, not structural, but tactical way. And so I think, yes, policymakers did learn something, and they were risk averse, so they went extra large. So you said the financial crisis, clearly a shock.

29:32The other thing's not as much as a shock. And we've had plenty of false alarms. How do you define what a true shock or crisis is? And what do you put in the category of false alarms or things that are genuine, but just don't rise to the level as described? Yeah, there are two things to consider. One is sort of the new cycle level. We have a constant doomsaying about supposed things that could lead to recession or otherwise downgrade the economy. Just the last few years, we went numerous. For example, consumers were supposed to run out of cash and consumers were not going to keep up their spending.

30:14We had lots of false alarms about the labor market, even last summer. We had last summer in August, there was somewhat of a panic because supposedly the labor market was going to be very soft and very weak. So we have these new cycle false alarms stories that often are rooted in a data point that is noteworthy, that is interesting, that does signify risk, but we're extrapolating from the data point to conclusions that don't hold up. That is one category of false alarms. The other category is where you have real crises, but the question is, are they going to have structural impact? Are they going to have a long-term impact on the economy?

30:50Are they going to downgrade the economy's capacity? So 2008 does qualify. 2008 left an indelible mark on the U.S. economy, but 2020 didn't in terms of performance and output. We've regained the trend output that we were on, the path we were traveling on pre-COVID. We've come back to that trend output path. It has not left the kind of permanent mark on economic performance that you saw after 2008. So in that sense, we need to differentiate between what is a likely shock that will pass and that we can fix versus what is something that changes the structural composition, structural setup of the economy durably.

31:32Those are two very different types of situations. That sounds like a usable framework for distinguishing between real crises and, do I call it media alarmism? Or, you know, everybody is blaming the media these days, especially with this administration. But there has been a fairly relentless negativity, especially in social media. What's the best framework for, you know, separating the wheat from the chaff? Well, typically when we see knee-jerk reactions and doomsday stories, they're taking a data point and then they're extrapolating usually on the basis of a model. So, I mean, think about the inevitable recession.

32:16Even Larry Summers, people like that, they came out and said, look, to bring down wage growth, to bring down inflation, you need, I don't know, five years of unemployment at this and that level. Why? He threw out 10%. Well, 10 % for one year or 5 % for five years. So you had different configurations, but they were all based on basically the Phillips curve. This was all a Phillips curve take on the economy. Which was a great model 50 years ago, wasn't it? Yeah, it described the UK and certain other countries empirically quite well. It wasn't ever really a model and a theory. It was more of a description of empirical facts.

32:51But certainly it was useful for a window. It's still useful as an instrument to think about dynamics, right? But it was basically used as the truth. You know, there's an input and there's an output. And my model gives me the truth if I give it certain inputs. And then, well, what happens, we're extrapolating data points often outside the range of empirical facts. The models are only trained on historical facts. You know, you can't make up data points to train your model. So when a crisis hits, likely you get data points that were not empirically known in the past. So what does the model do? It extrapolates outside its historical empirical range.

33:32And then you get these kind of point forecasts that just don't work. I mean, case in point, in 2008, unemployment goes up to around 10%, right? And it takes almost the whole 2010s, a full decade almost, to bring down this very high unemployment rate. So in COVID, when unemployment shoots up to 14%, what does the model do? It says, well, if it takes a decade to bring down 10 % unemployment, it will take even longer to bring down 14 % of unemployment, right? And that is exactly this kind of limitation of the model-based approach. Empirically, you never had 14 % unemployment. So if the model extrapolates from past data points, it's going to go off the tracks.

34:13And that's exactly what happened in that instance. So the underlying flaw built into most models is that the future will look like the past. And as we've learned, that often is not the case. It's always idiosyncratic. Look, the U.S. economy since the Second World War has only seen a dozen recessions. Now, each of those recessions is totally idiosyncratic. And even if they had a lot of commonalities, 12 is not a sample size that a natural scientist would consider large enough to build sort of an empirical model around, right? Each of these crises or each of these recessions was idiosyncratic. And the idiosyncrasy demands much more than a simple model or even a sophisticated model.

34:55It demands the eclectic view across many, many drivers. And that comes down to judgment. There isn't an output in an Excel sheet or a Python model or anything. In the end, it comes down to human judgment. And I think that is something we lose sight of way too often. You very much strike me as a fan of Professor George Box. All models are wrong, but some are useful. Tell us a little bit about how models can be useful. Well, they're always a good starting point. Even the Phillips curve has a lot of validity to think about what might be happening. They're always a sketch of reality. But the moment we're translating that from a sketch and a map into something that is hardwired in a quantified model, and the moment we then expect that the output will resemble anything like the truth, we're sort of denying the reality of this.

35:50It just doesn't work that way. Look, I'm not the first person to make that point. In fact, Hayek, Keynes, von Mises, they've long basically trashed economics for saying you're too gullible and you're too naive about the constant nature of these variables. They've long pointed out that you don't have what the natural sciences provide, which is stability in all these relations of variables. You don't have that in economics. And there's an anecdote that we pick up in the book. When Hayek receives the Nobel Prize in 1974, he actually uses his acceptance speech, or I think it was a dinner speech he gave right after being awarded the prize.

36:33He uses that speech to say, look, you shouldn't do this prize in economics. You should have never done the Nobel Prize in economics. But if you must have this prize, at least ask the recipients to swear an oath of humility. because unlike physicists and in chemistry and other natural sciences, economists have a big microphone, right? Policymakers listen to them. Politicians listen. Public listens to them. But they don't have that certainty of analysis. They don't have that stability in their models. So they're going to go off the tracks all the time. So at least ask them to be humble about what they're doing.

37:09And I think that is a good reminder of the long history of recognizing the limits of model-based approaches through the eyes of some of the leading thinkers in the space.

37:45and the politics that shape global tech markets. We do this all every weekday, then bring you the most important conversations and analysis in our podcast. Search for Bloomberg Tech on YouTube, Apple, Spotify, or anywhere else you listen. Join us every afternoon on your commute home and stay ahead of the tech news cycle. That's the Bloomberg Tech Podcast. I'm Caroline Hyde in New York. And I'm Ed Ludlow in San Francisco. Subscribe today, wherever you get your podcasts. So let's talk a little bit about a lot of the false alarms and faux crises. So many economists got 2022 wrong, 2023, 2024. They were expecting a recession.

38:28It never showed up. Why is that? It starts with the master model mentality that we call out in the book, where we place too much trust in models. So the Phillips curve was essentially used by many forecasters and commentators. Define the Phillips curve for the lay reader who may not be feeling like that. Yeah, the Phillips curve is an old theory going back middle of the last century, describing the relationship between wage growth and unemployment. So the idea is that you trade off the two variables. and that led commentators like Larry Summers to say, to bring inflation under control, you would need either many years of high unemployment or a sharp recession, 10 % unemployment for a year to reset the inflation picture.

39:13In other words, in layperson's terms, a soft landing is impossible, right? And this is what fed into the inevitable recession that was the dominant received wisdom the last few years. Now, you know, these things are good starting points. They have validity historically and a lot of empirical data. But in the end, it's idiosyncratic. It's a very idiosyncratic constellation of drivers and risks. And so it was in the last few years. So let's look at that for a moment. One of these master models was also interest rate sensitivity. We think interest rates go up, and that eats into disposable incomes for households.

39:46But in reality, mortgages in the US, unlike in Canada, mortgages are long-term, didn't actually take a big bite out of disposable income. Mostly fixed rate. Exactly. Very long-term fixed rate, low. and most of them were done at low rates because we had low rates for a long time. Contrast that with the flexible contracts and mortgages in Canada where they lost a lot of disposable income. That wasn't the case here. Same thing about interest rate sensitivity in the corporate sector. You know, the textbook tells you interest rates go up and investment will fall. But does it? You know, when you do the empirical analysis for whatever window, you'll see a very flimsy correlation between interest rates and CapEx.

40:25Firms invest when they have a narrative to do so, when they see a return on the investment. And if they believe the investment is beneficial to them, they'll do it whether the interest rate is 2%, 3%, or 4%. And just look at what happened the last few years. You had a lot of narrative and belief in worthwhile investments, data centers, software. So with or without higher interest rates, firms are going to do that. particularly also because a lot of our investment has shifted away from you know fixed structures physical investment to intellectual property software type of investment which has a much higher rate of depreciation so a bridge or road will be good for 30 40 years but software is maybe three or four years so you constantly have to invest just to stand still just to keep the stock of investment in the space to keep it steady you constantly have to run faster just to to maintain And so there is a lot of idiosyncratic drivers that led to very different outcomes from what was predicted from a model-based Phillips-Curve type approach to reading that context.

41:33So a lot of highly regarded economists like Larry Summers kind of reminded me of the Paul Graham quote, all experts are experts in the way the world used to be. And we're seeing a lot of that in that. So not only did people get the recession calls wrong for the past couple of years, what have we had, two months of recessions in the past 15 years? Are we in a post-recession economy now? You can still get recessions, but I think we've become better at fighting them. So this is the topic of stimulus. There are two different types of stimulus that we describe in the book across three chapters. And we differentiate between what we call tactical stimulus, which is just to smooth the cycle, accelerate growth in between recessions, maybe de-risk the cycle when necessary, versus existential stimulus, which is when policymakers, politicians step in when the economy is truly at risk of a structural break.

42:35Those two types of stimulus, they're evolving differently. I think the tactical kind is more challenged going forward. It was very easy when inflation was below target. It was very easy when interest rates were very, very low. There was little cost to the Fed put. You could do that. There wasn't sort of an inflation risk associated with it. That's different now, and I think that will remain different now that we're skewed to the upside in terms of inflation where interest rates are likely to be higher for much longer. But the existential type of stimulus, the ability to step up when it's needed, I think that is still very strong.

43:08And if you have another shock or a crisis or a recession, I think we'll be able to deploy stimulus effectively still. So we said earlier all recessions are not homogeneous. They're all idiosyncratic and unique. But one of the things you mentioned in the book that kind of intrigued me, we shouldn't conflate recession intensity and recovery. Explain what that means. Yeah. When COVID hit, we had extreme data prints. Unemployment is sort of exhibit A of the story. Unemployment went to 10 % in 2008, but it went to 14 % in 2020, right? So the intensity, the sudden collapse of activity was much more pronounced in COVID than it was in 2008.

43:54GDP also much worse during the first few months of COVID than GSM. All variables. And we have a chart early in the book that shows the fifth to 90th percentile of historical experience of these variables. And COVID is like far outside that historical range. So you get data prints that you're not used to, that the models don't know. The models were trained on data points that were simply not experienced until they happened in COVID. Now, all of that fed into extreme intensity was equated with this will be a very long and difficult recovery. why the 10 % unemployment rate led to many years of recovery in the 2010s.

44:33So now if the unemployment rate is even higher, it's going to take even longer to work it down to a level that is a good economy again. But that wasn't the case. 2020 wasn't about a balance sheet recession. It wasn't about banks repairing their balance sheets. It wasn't about households repairing the balance sheet. We took care of that with stimulus. And therefore, the ability to recover was much faster, much stronger. There were other idiosyncratic factors. Essentially, what was underestimated was the ability to adapt of society. You know, society's found ways to work around the virus. The pathway to a vaccine was faster.

45:09So there were a lot of things that were underestimated. You know, it kind of reminds me of the Y2K fear that when there's a little bit of a fear of panic, the expected crisis may not show up because we're taking steps to avoid it. We don't know, was Y2K a false alarm or did the fear lead us to make sufficient changes to avoid problems? I honestly can't answer that question. I'm wondering how you look at crises in terms of do some of the fear-mongering and some of the media absolute extremism lead to government action that prevents the worst-case scenario from happening? It's possible that it shapes the perception of policymakers and politicians.

46:01But I think the reality is on the ground. The variables that are visible and measurable, the unemployment rate, GDP growth, imports, exports, all of that was under pressure. I think that is more telling for those who take decisions than what public discourse does. Is public discourse particularly fearful and a lot of angst pervades how we think about the economy? Does that spur action? Maybe. That's part of it. So we don't know, as you rightly say, what would have been in a counterfactual world. But essentially, when the economy is genuinely in trouble, I think the willingness to act on the stimulus side is very strong.

46:42So let's talk about some of those metrics. You have an image in the book scanning the recession barcode. So tell us about that and the history of U.S. recessions, which seem to have been more frequent and more intense. You go back a century, they were depressions, not even recessions. Tell us about how this has changed over the past, I don't know, a couple of hundred years. Yeah. So if you do a very long run chart of recessions in the US economy and you shade each recession as a bar, what you get is a barcode, an image that looks a bit like a barcode, but it thins out as you move to the right.

47:24So you had recessions very frequently 100 years ago and further back. The economy was constantly in recession. Essentially half the time it was in recession. Banking panics all the time. Yeah, but also real economy. The economy was very agrarian. A bad harvest could drag down performance of the economy. So there were a lot of shocks. But yes, there were also banking crises and things like that. And what we identify in the book is a recession risk framework. We say, look, all recessions come in one of three flavors. They're either real economy recessions, which is when investment and consumption drop abruptly and pull GDP growth down.

48:00So that's the real economy type of recession. The second is a policy era when policymakers get it wrong. They raise interest rates too fast or too high, which only you ever know exposed whether it was the right thing to do. So it's a very tricky thing to do. And the third type of recession is the most pernicious kind. It's a financial recession when something blows up in the financial system, like 2008. And what we're showing in this chapter of the book over the long run, the composition of these two drivers has changed over the last 40 years. The real economy recessions, they really took a backseat because the economy calmed down, the volatility calmed down, services play a bigger role in the economy today, so they're less volatile than physical production.

48:46But also policymakers just got better at managing the cycle. So, you know, policy errors kind of also lost a lot of share, if you will, in the overall prevalence of recessions. But when you think about what has given us the biggest headaches, it was 2008, a financial recession, and dot-com, in a way, is also a financial type of recession. So the share and the risk from financial blowups is significant, if you look at it in recent history. And that doesn't mean that the next recession will be that type, but its share of the risk spectrum is relatively high. So what should we be listening to when we hear economists discussing various risks?

49:28What are the red flags that, hey, maybe this is a little too doom and gloomy for our own portfolio's best interests? Yeah, I think the litmus test for me is often what would it take for a certain outcome, a certain doomsday outcome to actually come to pass? Not just will it happen and what would be the damage, but walk me through the conditions that actually lead us to the precipice and then make us fall off that macroeconomic cliff. We need to talk about drivers, causes. We need to talk about their probabilities and their constellations. So it's not good enough to say the model says the recession will happen.

50:09Walk us through exactly what is the confluence of headwinds that together make that credible. Right. It's it's it's it's more than the point forecast. Really, really kind of intriguing. I also noticed that I'm not an economist, but when I listen to economists talk about the possibility of a black swan or the possibility of this event, it it's almost as if there won't be any intervening actions either by the market or the policymakers. Tell us a little bit about that. What was George Soros' phrase? Reflexivity. That when certain events happen, there are going to be natural reactions that just prevent this extrapolation to infinity or to zero, as the case may be.

50:59Yeah. I mean, this is back to the topic of stimulus, first and foremost. 2008 came as a big surprise because the models in the early part of the 2000s, they didn't even really look at the financial sector as a risk driver. They kind of assumed the financial system away. And then when the problem brewed in the financial system itself, the models were kind of blind to that. And the reaction couldn't be gauged if you didn't have a view of that. And the reaction really depended on stimulus. And stimulus is about politics. It is about policy. It's not about economics, first and foremost. It's about political economy.

51:36It's about people coming together and fighting crises. And so I think that remains the case that the idiosyncrasy happens before the crisis. The drivers are idiosyncratic. But the moment a crisis starts, a shock hits, what happens as a reaction is also idiosyncratic. It is political. It is about society. It's about choices. It's not stuff that you can model in a rigid, natural science way. So let's talk about something that clearly wasn't in the models, forget 20 years ago. They weren't in the models five years ago or even three years ago. And that's the impact of artificial intelligence on our economy, on the labor pool, and on productivity.

52:20How do you look at a giant structural change like AI? How do you put this into context as to what it might mean across all these different areas within both traditional economic modeling and the real world? You know, we've had productivity growth the last few decades. even though often the narrative is productivity growth is really, really low. We've had productivity growth, just not in services. But in the physical economy, there's been pretty decent productivity growth even the last 20 years, where we didn't have productivity growth with services because we didn't have the technology to move that part of the economy along.

53:01Now, why is that? Essentially, productivity growth goes up when technology displaces labor. That is really the definition of productivity growth. Both you need to produce the same with less labor inputs or more with the same labor inputs. But either way, technology, whether we like it or not, is about the displacement of labor. And we weren't able to do that in the service economy. Now, with AI, I think you have a better chance of doing this. At least the promise is very strong that this will work. But I think we're getting ahead of ourselves. And I'm not saying that now. We've published on this over the last few years, even as COVID hit and even before AI, when the Zoom economy was sort of this dominant narrative.

53:42It's a hard slog to do this. It happens over years and it's little by little. It's not a flip of the switch. It happens very incrementally. And I don't think AI will turbocharge GDP growth. It is a lift to growth over the medium term, but there are many little obstacles. There are many little things that need to fall into place for people to really adopt the technology and for this to little by little give us a tailwind. So it's not an abrupt step change. It's something that is credible, something we need to work through, and then it will show impact over a 10-year frame, 15-year frame. So let me push back a little bit on one thing you said.

54:21And I seem to have this ongoing debate with economists who work in a larger corporate framework. We're here in Bloomberg, giant company, big operation. my day job is a much smaller company under 100 employees and I have noticed just over the course of the past decade how our productivity has skyrocketed and it's a services business finance is a services business and it just feels like the things that used to take so long to do 15 and 20 years ago are now automated and it's not that we're hiring fewer people and it's not that we're working shorter hours, but the same size team can just accomplish so much more than they were capable of.

55:11Like I recall the days of quarterly reporting and having to literally run a model, create a printout for every client, print it out, stick it into the right. And like, It was like a week-long process that all hands on deck every quarter. And now it's updated 24-7 tick by tick. It's automated. No one cares about quarterly reports because you could get it. And the joke is you have 24-7 access to your daily, weekly, monthly, year-to-date, five-year, 10-year performance reports. just try not to check it second by second. But the way, and that's just one example, being able to communicate with clients, to record and embed an interactive video with charts and everything else, that was like a massive undertaking.

56:12And now it's like child's play. Even though you're doing the same thing, you're just doing it faster, better, cheaper, easier, are we somehow underestimating the productivity gains? Or are these just specific to, you know, that one area? Yeah. So I have some pushback on that. I think the bar for productivity growth is a little higher and it's very specific. It's less inputs per output. So do things get more comfortable? Are they moving faster? Are they qualitatively perhaps better? Yes. But are we using less inputs to generate the same value? Or are we using the same level of inputs to generate more value?

56:54That is what we need to achieve to speak of productivity growth. And let me give you an example that we use in the book. You know, I took an Uber from my apartment to come here into the studio today. And Uber is often upheld as the epitome of progress in tech. And it is fascinating. It's a great app. I love to use it. It's nice. But look, if you want to improve the productivity growth in taxi transportation, we have to talk about inputs and outputs, right? And the inputs are on the capital side, a car, and you're not getting rid of that car. And on the labor side, it's the driver. And the Uber car still has that driver.

57:30Not in Waymo and parts of the West Coast. Yes, and this is why I said it takes time. Incrementally, that will happen and that will unfold. But do you think you're going to have driverless taxis in New York in 2028 or 2030? I don't. Well, we'll have it in 2050, probably in 2040. I can't tell you what exact year it'll happen, but it's coming. I agree with you. And the sooner we embed those RFID devices in vehicles and on street corners, like doing it visually in LIDAR is very 20th century. Right. Yeah. And that's why I said it takes time. Over time, this will be substantial lift to economic output, but it doesn't happen overnight.

58:13It's actually, it takes time, right? And there's an additional important point about productivity growth that can also be shown in this taxi example. When technology is truly productivity enhancing, you see that in falling prices. Technology is deflationary, right? As technology does away with input costs, firms will compete with lower prices to gain market share. So across history, wherever you look, as technology is becoming a credible force in production, prices will fall. Now look at Uber. Uber prices in New York tend to be higher than a yellow cab. Why? Because despite this expensive technology, you're not able to produce this ride more cheaply.

58:55You're not. In fact, you kind of have to monetize the technological expense. The app is expensive. All of this is expensive. So generally, you're paying a premium for the smoothness of the app and all that. Over time, that may change. But watch prices. You want to see productivity growth, whether it's happening or not, you've got to look at prices. And that's one of the arguments we make in the book. So let's hedonically adjust. We'll stay with Uber. Let's hedonically adjust that. In New York City, if you want a taxi during rush hour, hey, sorry, you're out of luck because the monopoly that was imbued by the Taxi and Limousine Commission and a handful of big medallion chain owners decided in their infinite wisdom that we don't need to move people around rush hour.

59:42We're going to change shifts then, which, by the way, is my pet theory for how Uber penetrated. And so, A, you could get an Uber during rush hour that you can't during cab rides. You could get an Uber when it's raining. Good luck hailing a cab in New York City rain. And you have the ability to schedule an Uber. You have the ability to get a higher quality car. You could get an electric car if you choose, a larger car. like I'm not a huge fan of traditional hedonic adjustment because it was a way of kind of tamping down on the cost of living adjustments always felt sort of disingenuous but I don't think you could get anybody to say that Uber is not only better and I'm not a big Uber fan but as a user Uber is certainly better than a cab and in many ways orders of magnitude better more choices, more options, and just a higher quality experience, plus just the idea of having, hey, is this a work thing or I'm going to use that card on the app?

1:00:50Well, no, this is personal. I'll use that card. So maybe taxis aren't the best example, but let's talk about economists. Again, I want to stay with this because I love the topic. Think about the quantity of research you push, you push out, the ability to integrate charts and data. And like, I am been in this business long enough that I can remember, first of all, when I started, the guys in the technical group, they were doing charts with pencil and graph paper. I'm not exaggerating. Maybe that's just a function of my age, but think about how, and the cheat was, you get a different feel when you're doing it point by point than when you're just generating it.

1:01:37Whether that's true or not, at least that was the, when computers came along, people continued to do that. But think about the access you have to the just endless array of data, the ability to do that. I haven't even mentioned your fortune column. Think about how much time and effort goes into putting out a column and you go back 25 years and it was just a horrific grind. Like at this point, everybody seems to use some version of Grammarly or some other editing software. The ability to put out, and I'm not talking about asking ChatGPT to generate a garbage article for you. You writing something, cleaning it up, embedding a lot of data and images, It just feels like, to quote Hemingway, gradually and then all at once, it just feels like it's so much easier to put out a much higher quality product with either the same or less effort than 25 years ago.

1:02:44Maybe I'm just hyper-focused on the junk I do. But what's your experience been like? Absolutely. Incrementally, there's progress. But again, the bar we need to meet is value. Are we generating more value with the same inputs or are we generating the same value with less inputs? That's the definition of productivity growth. So if you can make all these charts faster and you save one economist on the team, well, that's productivity growth. Or you keep the economist and you double your number of reports and you also manage to monetize them and earn revenue for it. Well, that's productivity growth.

1:03:18If the charts get prettier, faster, fancier with the same number of economists and the same number of revenues, well, from an economic sense perspective, that's not productivity growth. So it's got to be a change in the relationship of inputs to outputs if we're comfortably talking about productivity growth. And back to the Uber example, you're right. You can get different cars to ride in. You can get the car, the Uber car, when it's raining, but you're paying for that. So it's not produced more productively, right? You're paying a surcharge. You're paying the surge pricing. I think they call it an Uber.

1:03:54So, you know, yeah, you can get it when it rains, but you'll pay twice as much. So it wasn't done more productively, right? Really interesting. The gap between the increased quantity and quality of output, if we're not monetizing it, or as a consumer, if you're not seeing price declines, then it doesn't really count as productivity. No, it's got to be a change in the ratio of inputs to outputs on either side. Either we keep all the staff and we earn more revenue with it, that's productivity growth. Or we keep the revenue constant and we do it with less inputs, that's more productivity growth. But, you know, again, I'm not saying there isn't productivity.

1:04:32There is, and there will be more, and AI will have impact. It just needs to show up in value in that relationship between inputs and outputs. I see it qualitatively, but I completely get what you're saying. quantitatively. Are you still doing the Fortune column on a regular? Yeah. We publish in Fortune relatively regularly. Whenever we see a cyclical or a thematic topic that we feel is pressing, we publish with Fortune. Yeah. Really, really interesting. All right. I only have you for a limited amount of time. I know you're catching a flight today. Let me jump to our favorite questions that we ask all of our guests, Starting with, what are you streaming these days?

1:05:15What's keeping you entertained, either Netflix or podcasts or whatever? Yeah, I'm not very big on shows or Hollywood. I mean, to give an idea, I think I'm on the second season of Slow Horses. I think there are four seasons of it. I'm kind of slowly making my way through the second one. It's very entertaining. I love Gary Oldman. So interesting. He's great in it. Yeah, it was sort of taking down the genre of spy movies in a very entertaining way. So I'm doing that. But also, I tend to watch late in the day when I'm tired, so it's entirely possible I fall asleep and I take like two, three evenings to get through one episode.

1:05:49Yeah, so I'm not all that big on that front. Tell us about your mentors who helped to shape your career. So many people, right? Because a lot of it is teamwork and you don't progress without mentors and role models. I would say in my current role, I would probably call out two people. Rich Lesser, our longtime CEO and our chairman, he had the vision for a macro product, as did Martin Reeves, who runs our research institute, the Henderson Institute. And they're really the two people who brought me into this role and coached me. So they stand out. Outside of BCG, Kathleen Stephenson, she had many, many different roles on Wall Street and economist roles.

1:06:31She's been a great help navigating my career the last many years. And further back, in academia, thesis advisors and many others, it's always teamwork in a way. So you have many role models and mentors. Let's talk about books. What are some of your favorites? What are you reading right now? Right now, I'm almost done with Making Sense of Chaos by Doin Farmer. Came out last year. Doin Farmer is a very interesting character. He's a complexity scientist at the Santa Fe Institute and I think at Oxford University as well. And his book is interesting to me. I bumped into him at one or two conferences, but it's interesting to me particularly because he kind of argues the opposite of what we argue in our book.

1:07:15So he thinks, he agrees that economics is poor if you just take standard models and theory, but he believes you can crack the complexity of it. So he thinks with complexity science and better data and better models, you'll essentially be able to make those forecasts. I read it because it's always important to see what others are arguing. I don't read stuff that reconfirms what I think. I want to see what other people are saying about the same topic from different angles. So that book's been very useful and also well written. That's what I'm currently reading. I think of other books that I've read over the years.

1:07:52I mean, there's so many, many great ones, of course. I think one that early on made an impression on me was Seeing Like a State by James Scott. It's at least 25 years old. I read it as a grad student. And what he does, he looks at the ability of governments to do top-down policy to improve the lives of large amounts of people. And he shows all the pitfalls in a sort of Hayekian way. It's tough to have the local knowledge. It's tough to do the top-down improvements. Things have to grow bottom up. And that book kind of stood out for being very, very eclectic, very multidisciplinary, and still, I think, an excellent book to how to think laterally and not in a sort of strict model-based way.

1:08:36Really interesting. Our final two questions. What sort of advice would you give a recent college grad interested in a career in economics, investment, finance, anything along those lines? Yeah, I think a career as an economist is challenging in some ways. There are so many economists out there. Often when I hire, you see the flood of CVs and often very good CVs. And I think there's been an overproduction of economists. So I think doing something adjacent to economics, work in finance, work on the buy side, work on the sell side, Unless your heart truly beats for economics, I think you can use economic skills in many adjacent disciplines.

1:09:23And careers, I think, are plentiful in those adjacent disciplines. If economics graduates really feel strongly about economics, it's fascinating, but your heart has to be in it. And there aren't all that many seats as economists. So one has to build that over the long term. And our final question, what do you know about the world of economics today you wish you knew 25, 30 years ago when you were first getting started? Yeah, well, I mean, that's really what I wrote down in the book. You know, the book is the 20, 25 year journey through the maze of the economics profession and discipline. The themes we touched on, the master model mentality, the pitfalls of treating economics like a physical science, the doom mongering, which we have to simply ignore most of the time.

1:10:15and then the eclectic approach to economics. I call it economic eclecticism, drawing on a broader range of disciplines. Those are the things that I learned through that path the last 20 years. I wrote them up in the book. You know, it would have been interesting for me to read that 20 years ago, but I wrote it now, and so I'm happy with that. Really, really intriguing. Philip, thank you for being so generous with your time. We have been speaking with Philip Carlson-Slezak. He's global chief economist for the Boston Consulting Group. His new book, Shocks, Crises, and False Alarms, How to Assess True Macroeconomic Risk, co-authored with Paul Schwartz, is an absolutely fascinating read.

1:10:58If you enjoy this conversation, well, check out any of the past 500 we've done over the previous 10 years. You can find those at iTunes, Spotify, YouTube, wherever you find your favorite podcasts. And be sure to check out my new book, How Not to Invest, The Bad Ideas, Numbers and Behavior That Destroys Wealth, coming out March 18th, 2025. I would be remiss if I did not thank the crack team that helps us put these conversations together each week. My audio engineer is Andrew Gavin. My producer is Anna Luke. Sage Bauman is the head of podcasts at Bloomberg. Sean Russo is my researcher. I'm Barry Ritholtz.

1:11:43You've been listening to Masters in Business on Bloomberg Radio.

1:12:11Bloomberg Intelligence to bring you deep dives into the companies that are moving markets from publicly traded companies like Apple to those that are privately owned but known by everyone on earth like OpenAI. Now, I helped to build Bloomberg Intelligence to what it is today, Scarlett. And now our analysts are the best in the world, covering more than 2 ,000 global companies. That is your legacy, Paul. And we speak to those in-house experts every day. They are Bloomberg's go-to authorities on sectors, companies, and legal processes. And we do it all live each weekday, then bring you the best conversations in our daily podcast.

1:12:42So be sure to search for Bloomberg Intelligence on YouTube, Apple, Spotify, or anywhere else you listen. Listen in the afternoons on your way home from work to catch up on the market news you miss during the business day. That is the Bloomberg Intelligence Podcast. I'm Scarlett Fu. And I'm Paul Sweeney. Subscribe today wherever you get your podcasts.