Luis Garicano on Spain, Europe, and the Future of Work

23:44The politics make it very, very hard. And I told you defense, but I could go through many other examples and show you why, even though you would want to think that the gains from a single market, single procurement, et cetera, would be enormous. The politics make it really, really hard. Right now, what we're doing is harmonizing the shells. You have a big meeting group. People, they say, oh, I want the 55 millimeters to be like this. And somebody says, no, I want it like this, 55 and a half. And they meet for weeks and weeks and weeks and weeks and come up with pages and pages. And people, the citizens are upset.

24:18But the truth of the matter is, there is a political reason for this. And it's the one I mentioned. So in those unanimity areas, so fiscal, Ireland wants to do a separate tax in Luxembourg, separate tax policy that basically attracts multinationals and lowers tax bases and all of those things. The rest would want no arbitration tax bases. But this is by unanimity. So as long as there's one loser, no change takes place. And in all of these policies, there's always one loser. Pareto improvements are the only ones that can happen.

24:50Tyler Cowen:So you have your finger on a button, and you can press the button, and the unanimity rule goes away, and it's replaced by two-thirds majority rule. Do you press the button? You know, everything, everything. Europe becomes, say, not even the U.S. has that, right? Correct. Not even the U.S. So let's say Europe gets the same intercomers clause as the U.S. on the functions. Let me put it like this. On the function that the U.S. government has, Europe loses unanimity? I would say yes. Could a euro bond work at a significant level under current institutions? No, you cannot have the four housemates of the apartment share the credit card bill without sharing anything about how the money is spent and how the money is raised.

25:37Then the incentives are for all the housemates to go around with a common current card bill and penny the wrong. So joint borrowing requires, and I've been a very big proponent, and I was a proponent of the Coventry Construction Fund for Europe. We proposed it together with some revenue functions for Europe that would allow the Hamiltonian compromise. The money gets centralized for some public goods, there are European public goods, and you raise it in some European way. But that just not happened. And we issued those bonds, which we do have, so Eurobonds do exist, without really any joint revenue.

26:18Tyler Cowen:Capital markets seem to be a big problem. What could we do to make EU capital markets stronger and better? It's very interesting because I wrote about this in Silicon Continent in my sub-stack, a piece with Per Stromberg. Sweden has enough money to fund startups and scale those startups up. and Per Stromberg is one of the biggest world experts on private equity, and he happens to be also an expert in Sweden because he's based in Stockholm School. And the basic story that he tells and that I wanted to write in the blog for our readers with him is that it is not really about having the cash. It's about the cash and the knowledge together.

26:58So the money exists. Europe sends the money to the U.S. We invest in all those startups. The talent exists. all those French and Germans in Antropic and all these other places. What is missing is the Spotify chief of marketing that actually has seen how companies run, how it's scaled, etc. and now has a lot of money and starts to fund startups, advise them and help them. So it's possible that it is true our capital markets are a problem, But it is possible that just fixing the other aspects, which is at the national level allowing for the big change that Sweden did, is that you don't pay those taxes on the capital gains as long as they are reinvested in the new startups.

27:47So then you create this ball that starts rolling in which all these startup funders start funding the new startups. And you can do it in a country that was supposed to be one of the most socialist in Europe. In fact, it's very free market thanks to internal reform. So it's true the capital markets are not working. It's also true the capital market changes we're doing are wrong. The main thing I learned about Europe in that respect is what we do efforts are harmonizing. Instead of we put this new law and eliminate the previous one, the way it often happens in the European Parliament, European Commission, European Council, the European institutions, is we add this new law, which is harmonized, but leave all the lower layer laws there.

28:27For example, we are giving right now in the new capital markets package more power to the European Securities and Markets Authority, the ESMA, the SEC at the European level, without removing any power from the SEC at the national level, which means you're not harmonizing anything. to just having more hurdles for people to raise money.

28:49Tyler Cowen:You mentioned your substack Silicon Continent. I'm happy to recommend that strongly as well. Now, in labor markets, should severance protection rise with the tenure of the worker, or why not just let the market handle that? I have written and argued in favor of more flexible markets. Denmark has found a solution that I think achieves both ends in protecting workers by having relatively or pretty high unemployment insurance, but in making it flexible and able to hire and fire. That would be, to me, the best solution to the current rigidity on European labor markets. The problem is that it costs quite a bit of money because you have to insure people quite significantly to make them be happy to move around.

29:31But there are solutions in Northern Europe that allow for this flexicurity, and I would wish that the rest of Europe was adopting them because I do believe that it's not just that rigid labor markets raise the cost for companies. You can then lower wages and neutralize it. The problem is they increase hugely the cost of making mistakes. That's what a new literature is recognizing. People weren't clear about that before. If Facebook makes a mistake, thinks the future is the metaverse, and the metaverse is a completely ridiculous idea, and the future is AI, Facebook can turn on a dime, fire all the engineers in meta bears, and hire all the ones in AI in a week.

30:15If Ericsson or Nokia want to do something similar, they're going to take months and months and months of negotiation with their labor consul and so on, to the point that they might go under in the meantime.

30:26Tyler Cowen:Before we get to your new book, Messy Jobs, are you game for a round of overrated versus underrated? Sure. Go ahead. Okay. Rosalia, overrated or underrated? Rated just fine. She is amazing. Penelope Cruz. Also fine. A very fine actress, doing very well. Javier Marías, the author. I am a big fan. I would say he deserves wider recognition, so I would say underrated. The Spanish painter, Antonio Saura. Maybe. Just fine. Just fine. Let me just say just fine. I think he's underrated outside of Spain, for instance. Yes, he is big, big in Spain, but I wouldn't be able to comment on how internationally he's recognized.

31:04Tyler Cowen:The region of Spain, Extremadura, overrated or underrated? Underrated. It has some of the best history. All the conquistadors, many of the conquistadors came from there. So when they went back, they did a lot of work and making the cities beautiful. The region stopped growing 400 years ago. So it's all frozen in time. And you have some of the most stunning experiences of walking into a portal towards the past in places like Cáceres. Which, by the way, has one of the best restaurants in Spain as well. What's it called? The restaurant in Catherys, I don't remember. I ate there a few months ago. It's a three-star, so people can find it.

31:41The AIs will know, yeah.

31:42Tyler Cowen:Within the world of Beatle fandom, is the album Sgt. Pepper overrated or underrated? I think it's overrated. I'm not as huge a fan of that as of Abbey Road or of Robert Sowell, or for sure, the earlier ones. I don't think it's the biggest piece in the canon. There are some weak songs on it. I think George Harrison's contribution is not so wonderful. When I'm 64 is charming, but it bores me a bit. I think there is some weakness. I would agree with you. Is there an underrated Beatles album? Again, this is within the world of Beatle fandom. People somehow don't like Let It Be, which I think the post-not-quite-last album, but the really last album.

32:21I don't know why. I always adored it. But I mean, it's kind of so relative. I mean, their albums are all hugely appreciated.

32:30Tyler Cowen:What is the special food place in Spain that not everyone knows about? The place where I'm sitting now, I'm in Asturias, recording this from Asturias. It's the region in the north between Galicia and Cantabria in the Basque Country. It's the oldest bit of Spain in many ways because it was the last place to be conquered by the Islamic invasion in 711. And it's where the Reconquista started. So, historically, it has a long history of Romanic churches and all that. But it also has outstanding food everywhere, wherever you stop. It has amazing nature, unspoiled, beautiful beaches, and it's completely climate change proof.

33:12It has amazing weather. How will the Spanish language change? The way it's changing is it's incorporating a lot of English language words and expressions. People, probably I am guilty of that as well. People who are very exposed to English introduce expressions and false friends. You know, we are so careful not to introduce a false friend from Spanish into English, and then somehow they come back from English into Spanish, and you see words that are used. If you are an English speaker, you realize, oh, that's completely the English usage. That's the biggest word. Many words that are ambiguous, instead of shifting towards the Spanish use, they shift towards the English one.

33:47Tyler Cowen:If I'm in Latin America, I notice that the conditional and subjunctive are used much less often than in Spain. Given that Madrid now is, what, 20 % Latin American? Is this going to affect Spanish in Spain? That's a great point. We are getting a very big Latin American influx. I haven't noticed that. I haven't noticed the change in the use of subjunctive. We use, unlike the French, not just the subjunctive present. I don't want to discourage our listeners from learning, but we use a lot the past and the compound past subjunctive. So if you make a future statement that is hypothetical, the hypothetical future is subjunctive, if I were a rich man.

34:27and past subjunctive. And I haven't noticed people using it less, or at least not among people I hear regularly.

34:35Tyler Cowen:And in Mexico, at least non-learned people, they don't use the future tense very much, right? They'll use the present tense and maybe throw in the word mañana, or they'll depend on context. And Spanish in Spain, is it going to go that way as well? I haven't noticed those changes in Spain. I think that the impact of English language is orders of magnitude higher than Latin America. That's my sense, but I would have to consult my favorite LLM before being so sure. Okay, on Messy Jobs, your new and excellent book, you argue very persuasively, in my view, that AI will not lead to anything like mass unemployment, maybe not even to a rise in unemployment, because jobs will become messier and the AIs won't be able to do them.

35:19Tyler Cowen:Is that a fair description of part of your argument? I think that's completely You're right. Yes. Now, I agree with you, but I think my worry is the opposite, that I know a lot of people, they want simple jobs, they're okay with some measure of tedium, and that if most jobs become messy jobs, for them that's quite stressful, and they're upset and disoriented. Do you worry about that? I think that we will have to have a tolerance for human relations. Is that what your friends don't want? We'll have a tolerance for relational work that is complex, that has politics in it, that has a big human component, and that's the part that is going to stay.

35:55So if people just want to be in their desk typing away, you know, I think two good ways to think about it is work from home and outsourcing. So if you think of which works, which jobs were outsourced in a big offshoring, which jobs were offshore, let me say offshore, in the big offshoring wave to India, those are jobs that are not messy. They're clean. The company specifies the jobs. They said, okay, we can specify this job perfectly. Let's ship the job. Those jobs are the same exact ones that are under complete threat of disruption. A lot of the work from home, when it doesn't involve a lot of Zoom meetings, I guess, has the same feature.

36:31Those jobs are clean, single task, and very often verifiable. You can just see how the performance is going and have your RL, your reinforcement learning loop work on those. I think those are gone. The messy component that I want to emphasize, many people will say, oh, stuff, AI, we're getting to the messiness as well. I think that there is some messiness that is contingent and that AI can streamline. But there is a lot of messiness that is both relational and has to do with a deeper aspect of the knowledge problem that doesn't really go away. And that, as AI advances, many aspects of Hayek, Poliani, and all these knowledge problems are still there.

37:14Tyler Cowen:But how much retraining will be required, and how frequent will that retraining have to be? So if I think of me working with agents, I have to retrain myself every month or two. And that's difficult for me. It's not stressful, given my position, but I can imagine it would be stressful. Can we really just put a big chunk of the labor force through that? But Rafael Sadoon and others at HBS have been working on this. I mean, one of the things that is frustrating about retraining programs and not just returning individuals, but returning programs is that they generally don't work. The favor middle of the road, free market economies kind of prescription for technological change is let's retrain people.

37:53And we find that this is really, really hard to do. If you wanted to retrain, you know, a clinician, a doctor who's learned to diagnose illnesses and that suddenly gets very easy. This is not one of our examples, but just to give an example, to become a plumber. I mean, it's not that you couldn't retrain him because, you know, the plumbing skills are so hard, which they probably are, to be honest, but not just that. But because there's a lot of social identity and other aspects that are tied to the kind of job you do. So retraining is a huge barrier. And the way we retrain ourselves, as Kun said, talking about scientists, is by dying and the new generations do new things.

38:30So I share your concern about how easy it will be to get people to learn new tricks. On the other hand, my mate is 84 years old and she talks to chat GPT all the time and she has good long conversations. I've updated it so that she now has the better version that you can talk over it and argue back the very recent improvements. and the LLMs are becoming better at eliminating all the prompting skill and all the skill that went into making agents, etc. It seems like it's easier to use. I think the hardest thing about LLMs, and maybe that's easier to retrain, the hardest skill about LLMs is knowing what they can do.

39:10Most people have no idea what you could do with your AI. And that's the part where a short focus lesson to a relationship manager at a bank, would you tell him, here's the case, you're going to visit your client, here are the six questions you should ask. I'm doing some work with a big bank on this exact topic, a big experiment. Here are the questions that you could ask before you have the client meeting. Maybe that retraining is easier on the use case as opposed to retraining for new jobs, which I think is very hard.

39:41Tyler Cowen:So do you think the natural rate of unemployment will go up? I don't mean cyclical, but just due to search retraining. My sense is that we have several hundred years of automation and we do have moments of blips, but I think the natural rate is going up much more due to other things. Well, if you allow me to group participation in, which has to do with disability and mental illness and people not feeling well and all that, rather than to a permanent increase in retraining needs. You know, we have some degree of recursive self-improvement right now. Arguably, there's more to come. So there's a new model out every month or two.

40:24Tyler Cowen:And for the most part, they're much better than the model that existed a month or two earlier. Just how exactly are people supposed to keep up with that? Like, do the jobs get messier at an equivalently rapid pace? You lose certain bits of the job. Certain things get unbundled. Other things don't get unbundled. other things. I mean, we talk about Gene and Jan Hui and me talk about strong bundles and weak bundles. Weak bundles, the job kind of breaks into pieces and some parts are done by the AI. When the job is a strong bundle, the AI kind of comes into your bundle and helps you, like in a relationship manager, an account manager for a bank who gets to prepare the meetings better, but still needs to have the cognitive task bundled with seeing the clients.

41:06and I think that part doesn't quite require staying as close to the frontier as you and me are staying. I think that part requires knowing that the LLM can help you prepare the client meeting and you ask the client, the LLM, please help me prepare the client meeting. Here are the budgets and the balance sheet statements of the client. Help me see which accounts, which products of the bank I could offer. And I don't think that needs really to know what the actual frontier model is and how does it differ from the previous ones.

41:38Tyler Cowen:Now, as you well know, virtually all new technologies and productivity improvements, they lead to more leisure over time. And in a good way, I don't mean unemployment, actual enjoyed leisure. If I think of the current day, there's some people like the two of us, we choose to work more to learn about AI. And then there's other people, they feel they have to work more to learn about AI, to not lose their jobs. And so far, there's no leisure dividend. People basically are working more. When is the leisure dividend going to arrive? I'm not sure that's the case. Let me give you two data points. So one is a Chinese CEO I had dinner with.

42:16My two co-authors are from Hong Kong University. And we had dinner with this CEO who told us she sees every one of her IT people more productive. She looks at the productivity of the overall IT department. Nothing happens. And quote, she said to us, I think they are smoking the productivity gains in the corridor. The second is from the recent NBER Summer Institute. There was a paper in health which gave doctors scribes and they all adopted. I mean, you know, adoption is slow. People are. Well, the automated scribes went up to 100 percent of adoption really very quickly. You see doctors scaling more meetings and doing more stuff.

42:56And eventually, and there was a recent New York Times article from a doctor exactly arguing this, she was saying, you know, I have more time for my yoga classes. So she was actually using the scribe, which transcribes her meetings and which stops a lot of the bureaucratic work that is nasty for a doctor. My wife is one, so I know how much they hate that. And she was actually going to do more yoga. And this actually shows up in this paper to some extent. So I do think that what we see with this usage, this shadow usage of AI in the job, which is people have their own computer and they actually ask questions that maybe their corporate AI doesn't allow them to ask, has a lot of welfare gains in the form of leisure and lower work stress in the job.

43:45Because you just ask the LLM, you post it there and you just spend some time browsing the web. So I am not as sure that the leisure is not expanding, but definitely not mine.

43:55Tyler Cowen:I've never worked as much in grad school. But say in the short run, when taxes change, almost always the substitution effect overwhelms the income effect, right? Tax cut, people work more. If taxes go up, people work less. Of course, we see this in Western Europe. So aren't these AI productivity improvements a bit like tax changes? You're more productive. And the aggregate, if doctors are more productive, they can earn more. they'll work harder. I'm not complaining about that fact, but maybe we should expect that instead of a leisure dividend. Well, things can happen. I mean, we are seeing Jevons' effects in many of our own utilities and in many of our own decisions to work.

44:34It's definitely the case for you and I, as you said. I was just for the amusement of our readers, I guess. Many of you have a whoop. I was really very happy with my whoop until January this year. I was 53 years old in actual age versus 59, sorry, in physical age. Since January, it's been steadily declining. I'm now 57. I grew four years and six months. And that's because I am not sleeping a lot and I am kind of running things into the night and going to bed thinking, oh, I should have done this old thing. Why don't I try this? So yes, it definitely happens to me. So why wouldn't it happen to those doctors?

45:08Tyler Cowen:Is there a plausible scenario whereby Spain in particular, and I don't mean the EU, I mean Spain, achieves AI sovereignty. Now, you've written about this on your sub-stack, Silicon Continent. What's your view right now? And keep in mind, the whole stack in some way has to be under your control, right? Yeah, I don't think that's a substantive objective, neither for Spain nor for Europe. Europe as a whole, and Spain in particular, are nowhere on hardware except for ASML, which does the capital equipment. I'll go back to ASML, the producer of the machines that make the EUV chips. We have nowhere on the language models, the frontier models, neither in open weights nor on closed models.

45:53We have mistrial, but it's not really at the level of the Chinese ones.

45:57Tyler Cowen:Well, we are not really there on the data centers. So all the elements that are needed are not there except for ASML, which is a Dutch company and is not under European jurisdictions, under a Dutch jurisdiction if you want, and many of the supply chain, which is German. What you can aspire to, I've been defending the second mover strategy for Europe with my colleague, Jesus Arraquejo. What we are defending is being a fast implementer, profiting from the gigantic investment that the Americans are making in many things that will actually, like the railroads, eventually not lead to significant value capture, but will generate enormous dissipated welfare gains for everybody.

46:39and let's be the ones who enjoy those welfare gains by adopting AI very aggressively becoming big clients when you're a big client like China showed with NVIDIA you have bargaining power because the people who sell to you

46:51Tyler Cowen:your vendors are going to lobby the government to keep selling to you so become a second mover try to make sure to have options keep it interoperable make sure the models can easily be switched which at the moment is very very easy you just need to change the API call and API call it becomes seamless. That's the best that we can do. I don't think sovereignty is an objective at this moment. So how much bargaining power does Spain lose in this new world? All countries except the US and China will be dependent on AI, which stayed in control. My view is that I've disagreed with my blog co-author and son, Peter, on this.

47:32This is the one point on which we disagree. He's very convinced that it's all about frontier models, and he's written a couple of posts on that. I think that you need frontier models in research and innovation. You need frontier models in some adversary applications like war, but that in order to have good law, good professional services, good medicine, good accounting, good companies doing proprietary research in their own servers. You don't need the frontier model. You can be three, four months behind the frontier. As long as you have the hardness in-house, data in-house, you're better off with that.

48:13And I think that this letter from very recently that NVIDIA CEO and Microsoft CEO stimulated and everybody else wrote goes in this direction. The open wage model allows you to keep proprietary your own data and your own know-how. And a country like Spain or Europe as a whole could build a strategy on that ability to switch and to keep the market competitive and contestable.

48:40Tyler Cowen:Let's say you were liberated from your AI learning responsibilities and given two or three free weeks in life, all expenses paid, to explore some corner of the world where you've never been. What's your choice and why? I think I'll go to Vietnam. It's one of the places I've been dreaming of going for a long time. I think it has the cuisine, the landscapes, the people, the economic side, the technology to see how they're doing all this transition and nature. I think it would be Vietnam. And final question, what is it you want to learn about next, other than Vietnam? I'm continuously fascinated by economics.

49:17I learned a lot of prize theory in Chicago from Becker and Murphy and Rosen, and those were my advisors and many others. I find that these tools are allowing me to learn much more and to be deeper on many questions and to think harder about what is it that remains messy? What is the knowledge problem once cognition is very cheap? And why is it hard to reproduce with the machine? And all of those questions are infinitely interesting. I spent quite a few years outside of economics because of the politics that tour. Of course, I was in charge of economic dossiers in Parliament, but I was not in the frontier of economics, and I realized how much I like it.

49:58So my desire is to learn more about economics and help the economics discipline and the students make this beautiful transition that we're going to see that suddenly is going to accelerate all sorts of interesting research.

50:14Tyler Cowen:Luis Caracano, thank you very much. All right. Thank you, Tyler. It was my pleasure. Thanks for listening to Conversations with Tyler. You can subscribe to the show on Apple Podcasts, Spotify, or your favorite podcast app. If you like this podcast, please consider giving us a rating and leaving a review. This helps other listeners find the show. On Twitter, I'm at Tyler Cowen, and the show is at Cowen Convos. Until next time, please keep listening and learning.