Nobel Prize-Winning Economist Sees Era Of "Permacrisis" Ahead | Michael Spence (w/ Adam Taggart)

AlphaFold (DeepMind/Alphabet) solved the protein structure prediction problem by taking the amino acid sequence of proteins and predicting their three-dimensional structure with near-perfect accuracy, then released predictions for 200 million known proteins as open-source public goods—representing a seismic productivity advance for biomedical science.

“deep mind which is now was bought by alphabet Google uh has done a lot of stunning things the one most people you know may have heard of was we kind of winning the game of Go in two rounds on the second round by playing by the AI playing against themselves um but the the one that I think is very striking is is called Alpha fold and basically they they set themselves the task of predicting the three-dimensional structure of proteins from the amino acid sequence that defines it sounds technical but the three-dimensional structure is absolutely crucial because that determines what you know what molecules bind to so it's for Pharmaceutical you know vaccines Etc it's just crucial and it and it can be done in a lab but it takes you know somewhere between months and years um to figure it out and so they said well maybe AI can help and we'll see if we can predict this the three-dimensional structure of these things and basically it worked I mean the predictions aren't perfect but they're pretty good and and once they realized they' done it then they took the 200 million noan proteins in the world predicted the three-dimensional structure and published it as open source F anybody in the world can go look it up”

Implementation and soft skills (how structural change and business model transformation are managed) will be just as important as the technology itself in determining whether technological change is successful and doesn't create excessive social disruption.

“implementation is going to be critical you know you can't stand up and say geez we found this wonderful technology we're going to introduce it and have people s sitting there thinking you know they're coming for my job um so the soft skills the implementation of major structural change or business model change is going to be just as important as the technology itself”

Loss of trust in institutions has been a major consequence of distributional failures in globalization; because governance and policy changes require collective action and buy-in, widespread loss of trust makes it harder to implement solutions even when they're rationally beneficial to most people.

“part in part because we've come to understand and this is really political economy you can't Implement stuff um in terms of governance and policies that drive change you know if everybody's decided that you know to get off the train uh and that in some ways is a pretty good description of kind of where we are there's this you know a massive loss of trust in institutions of various kinds in various groups of people exactly as you just said Adam and so um and that then makes it harder to kind of move forward and do do something that you know is designed to kind of help everybody and their children and grandchildren”

Generative AI is likely to have its dominant impact through human-machine collaboration and augmentation rather than pure automation of jobs; the question is how to define the task—if you define it narrowly (write first draft of report) it looks like automation, but if you define it properly (produce the needed report), it's augmentation of human capability.

“so that that probably won't happen too many more times once people get the hang of it um you know and and so yeah I mean I'll give another example these an AI is going to write the first draft of a report that a doctor writes doctors spend an incredible amount of time writing up reports because they have to right we need that information in the system the AI will write the first draft that probably reduces you know 80% of the doctor's time right so if you define the task as the first draft that's automation but no doctor in their right mind is going to hand in the first draft right they're going to check it that's the other 20% and if you define the task as producing the report that's needed it's human machine collaboration right it's augmentation you have a powerful digital assistant that does something that uses up a lot of your time uh that we didn't have before”

Pandemic fiscal stimulus had dual consequences: it prevented balance sheet collapse and supported consumption (positive), but it also increased velocity of money and drove inflation (negative consequence that was underestimated).

“and I think you know that's partly a result of some pretty smart policies I mean they have consequences you know sovereign debt that 100% of GDP is different from sovereign debt at 60% of GDP in a rising interest rate environment but you know and so we got to deal with that right and inflation was a was was a a consequence as well right I mean sending direct stimulus to households increased the velocity of money and therefore we got inflation”

Labor market behavior has shifted significantly with shortages across major US labor markets (government, healthcare, traditional retail, hospitality, construction) as people have changed behavior and don't want to work roles that are inflexible, dangerous, stressful, or low-paid.

“we have shortages in all the major labor you know I mean Big labor markets in the United States pick anyone government Health Care you know traditional retail uh Hospitality construction right right I mean people have changed their behavior they don't want to work whether it's inflexible or dangerous or stressful or low paid”

Global supply chains are being reconstructed on the basis of resilience and security rather than cost minimization for the first time, which is an expensive process, as evidenced by Europe's rapid diversification away from Russian fuel dependence.

“in the global Supply chains are being constructed on a completely different bases you know in pursuit of resilience um and uh uh and security there's a cost being an important part of that right going from low low cost to higher cost in most cases yeah I mean here in Europe we're diversifying it very at light speed away from dependence on Russian f fil fuels it's a very expensive process”

The Internet and digital access have transformed the informational structure of markets by making large amounts of information available at negligible cost, which democratizes information and reduces asymmetries; however, this raises new questions about what information is credible and trustworthy.

“the presence of the internet and the fact that you can access very large amounts of information at essentially negligible cost does transform the informational structure of markets and does democratize it in many ways I mean there's some you know issues having to do with well what's believable and what's not and so people have to sort of sort that out and maybe there's a role for signaling in in that context but but but the overall impact is just just enormous”

Information asymmetries between buyers and sellers are far more prevalent than typically taught in introductory microeconomics, and they cause markets to function poorly when gaps are large; reducing these asymmetries through signaling, disclosure, and transparency is crucial for market function.

“there is an issue pace so that part's right right you can do some stuff so fast that the that kind of structural um adaptation that's essentially inherently Limited in any economy you know people don't change their behaviour they don't change their skills organizations don't change their business models you know overnight um so there's natural things that slow things down but you can imagine some serious imbalances you know that and that are disruptive of people's lives on the way through true so that's point one point two is everybody thinks that digital is when it comes to jobs is automation just automatically if a senior hospital administrator starts talking about using AI there's panic and the Panic is they're coming for our jobs right right and it and it and it's not necessarily an accurate description of what's going to happen”

Signaling—actions or credentials that are costly to obtain and negatively correlated with underlying quality characteristics that buyers want to know—helps close information gaps; examples include graduating from prestigious institutions (MIT, Stanford), executive dividend cuts, and management stock purchases.

“there are activities that people can undertake that have differential costs with respect to with you know that are negatively correlated with the underlying quality uh characteristics that buyers want to know and those signals turn out to be pretty important not just in economics but you know more broadly in society”

The financial impact of higher education is a combination of two effects: actual human capital acquisition (skills, knowledge) and signaling (screening, credentials); Spence's work focused on the signaling component without denying that human capital acquisition also occurs.

“I mean I got when I after after I wrote my thesis and published a couple of papers I was called out to the University of Chicago where they were quite unhappy with me um because they basically said you know you're saying the return to higher education is just the signaling effect and human capital doesn't matter which was a kind of ridiculous extreme form and I said no no no wait a minute I I wasn't saying there's no human capital acquisition in higher education I was saying that in addition at least for the uh for the top institutions there's a pretty powerful signaling effect as well and the return to education is a combination of those two things”

Imported goods that declined in nominal terms over the 20 years leading to the 2008 crisis represented powerful deflationary forces, and the fading of emerging market growth means that deflationary force is gone, so inflation would have been higher even without pandemic demand surges.

“if you look at the 20 years coming up to the crisis at at the components of the Consumer Price Index everything that was imported you know went declined in nominal terms... I mean no that it's just dramatic I mean if you took that away we would have had an inflationary environment”

There are two dimensions to AI policy: (1) preventing downside risks from inappropriate AI use (intellectual property, bias, misuse), and (2) accelerating positive adoption across sectors by providing support to small and medium-sized businesses that lack the resources of large firms like JP Morgan.

“the policy thing has has two sides to it um one is there's a lot of potential damage and you know downside risk associated with the inappropriate use of these Technologies so you know policy you know has to deal with that um and that's multi-dimensional I mean you know is the AI read your book and integrated it into the way it thinks about the world is there some intellectual property property issue there probably yes and we don't know how to handle it yet the the other part is positive right I mean you know we could have this you know Tech and finance going full speed ahead with AI and a whole bunch of sectors lagging behind we've done this before in earlier rounds of digital transformation um and we're not going to get the productivity Surge and we'll get you know highly unequal results across sectors you know if we get that kind of Divergence again and government has a role in you know accelerating and increasing the footprint of the adoption and diffusion options for small and mediumsized businesses you know JP Morgan can spend a ton of money you know experimenting with AI in the way they do business you know a smaller business doesn't have the resources”

The world is transitioning from a 40-year period of disinflation and declining interest rates (starting from the early 1980s) into a structurally different macroeconomic regime characterized by persistent inflation, higher interest rates, higher costs of capital, and lower valuations, driven by non-transitory supply-side constraints.

“we're going into a world in which INF as long as these Supply constraints are as binding as they are that inflation will be a persistent threat and by that I mean any demand surge is going to produce inflationary pressure it's going to be a world with higher interest rates higher costs of capital lower valuations and a difficult transition from a relatively highly indebted world including sovereign debt”

10,000 baby boomers are essentially taking themselves out of the workforce every day through retirement, which combined with low birth rates creates a structural demographic constraint on labor supply that cannot be easily reversed.

“we just have 10,000 Boomers basically taking themselves out of the workforce every day right”

Three major scientific and technological revolutions are currently underway with enormous potential: digital transformation (multi-decade, accelerating), biomedical and life sciences revolution, and energy transition technology—and these technologies are not only delivering breakthroughs but are becoming widely accessible and cost-declining.

“we're living in a world in which there's three enormously uh powerful scientific and technological revolutions underway now one is the digital transformation it's multi-decade and it looks to be getting more and more powerful as we go along we were writing this book when the you know generative AI kind of Hit the public awareness and it's stunning um we'll come back to that but there's a revolution in biomed science and Life Sciences going on um and then there's this fair amount of Technology that's pretty impressive in the energy transition that's part that is the essence of the um sustainability agenda”

Central banks failed to recognize non-transitory supply-side constraints and bet instead that inflation would be transitory, leading them to start interest rate increases late, which then had to be done faster and more aggressively than might otherwise have been necessary.

“and the BET and remember the bet the central banks made and the reason why the interest rates went up so fast is because they got started late was that the supply side constraints were transitory would be transitory yeah we were transitory and so all of a sudden this demand surge would be accommodated by at you know adjustments and some of them were transitory I mean let's be honest right ocean shipping costs aren't $110,000 per container anymore right um across the Pacific they're back down to 2,000 China wasn't going to stay in zero coid mode uh although the bounceback has been less dramatic than most people thought”

The U.S. government has engaged in a fiscal-monetary policy mismatch: the Federal Reserve is raising rates and doing quantitative tightening to fight inflation, while the administration and Congress are running historically high government spending as a percentage of GDP (outside wartime), putting stimulus and brakes on the economy simultaneously.

“right now the in the US at least uh the the Federal Reserve is largely pumping the brakes right it's it's been raising interest rates it's been doing QT it's trying to get inflation down you know back under its 2% Target here right but on the fiscal side government spending is been at one of the highest certainly percentages of GDP we've ever seen and never seen this High outside of like a wartime situation so in many ways the the administration and Congress are kind of hitting the gas at the same time that the the Federal Reserve the central banks are trying to hit the brakes”

Young people have a disproportionate advantage in creating and implementing breakthrough ideas because they are not yet 'trapped by conventional wisdom' and have less accumulated cognitive baggage, though there are also advantages to age and experience.

“I think when you're young you're at your most creative right for sure uh there's just too many examples of this you know because you're not trapped by conventional wisdom you know a whole bunch of conceptual apparatus that you can't kind of get rid of I mean there's advantages to being a little older and having experience but but I think unbalance in in most lines of you know human endeavor uh there's a big there's a big advantage to Youth and so there it's the time to take risks”

The economics profession has shifted in the past 10-15 years toward much more serious attention to distributional issues and inequality, moving away from the prior position that globalization's aggregate benefits justified losses for some groups (because losers could theoretically be compensated, even though in practice they weren't).

“I think one of the more you know attractive Trends in the economics profession in the last 10 to 15 years is much much more serious attention and I think it is serious not just you know sort of words um to distributional issues and they're incredibly important um part in part because we've come to understand and this is really political economy you can't Implement stuff um in terms of governance and policies that drive change you know if everybody's decided that you know to get off the train”

Macroeconomic underperformance (slower growth, higher debt service costs, investment challenges) will likely result from the regime change, but this is different from existential risk; the key risk is that policymakers, conditioned on 40 years of low rates, may make errors of judgment that amplify damage.

“many of the other things we talk about you know it's it's the question is basically underperformance right... if you're talking about to ordinary citizens in the wide range of middle income countri R and the and the developed countries this is just you know underperformance and same thing goes for macroeconomic policy you know if you if your mindset is conditioned on the previous three decades with very little inflation and so on um then there's a lot of accidents and mistakes there's an awful lot of debt out there sovereign debt Etc you know we're living in a rising interest rate environment unless you think after the inflation fight all the interest rates are going to comp plummeting back down to where we were before close to zero um you know there there's a kind of difficult um few years of transitions here to deal with”

AI is being used to increase transparency in global supply chains, which are too complicated for any human to fully understand due to their extreme complexity and data fragmentation.

“they're starting to use AIs to try to increase the transparency of these incredibly complex things called Global Supply chains they're completely opaque no human being can kind of figure out all because they're too complicated”

Housing markets will experience lag effects as transactions dry up first (sellers don't believe prices have fallen), followed by months of adjustment before sellers accept new market valuations, and people's balance sheets will experience damage from higher mortgage rates that we haven't yet seen.

“in general I mean you know we've seen this before in in when you have a big change in housing prices there's a lag I mean what happens first is the transactions dry up right why because the sellers don't believe it uh and so after it takes a few months before they decided actually the value of the thing fundamentally changed I know I saw this here in Italy um you know when trouble hit uh you know in the sovereign debt crisis”

Debt service costs on rising sovereign debt burdens (from levels like 100% of GDP in the U.S.) will become increasingly constraining in a rising interest rate environment, reducing fiscal space for other priorities.

“do does our current economic model um have sort of a you know I don't know a a reckoning moment kind of built into it on the debt side of things if we keep amassing debts at the way that we we currently are at least in the major Western economies right now um does that get to a point where the system can't operate as intended and as a you know a good example of that which we seem to be approaching at a faster and faster rate right now is when The Debt Service cost on your national debt starts eating up all of your tax revenues that begins to become a really big problem”

Adverse selection is a market failure where inability to distinguish between different product qualities causes high-quality products and participants to exit the market because the equilibrium price reflects average quality; this can cascade and destroy entire markets.

“George talked about the they kind of illuminated a phenomenon called adverse selection which is what happens in markets when you basically can't distinguish between different products that have different levels of quality in them and in this extreme case you know that that kind of structure can lead to the complete destruction of a market that's that's the phenomenon of adverse election that sort of high quality people facing that price that reflects average quality essentially pull out of the market um and then that Cascades down uh and eventually you can destroy the whole market”

The American economy has proven 'stunningly resilient' because: (1) fundamental dynamism remains in place, and (2) the pandemic-era government transfer of balance sheet damage from households and corporations to the sovereign balance sheet left critical sectors with stronger balance sheets compared to the Great Financial Crisis, enabling continued consumption and reducing near-term default risk.

“I mean the the economy to I think everybody's surprised the American economy has turned out to be stunningly resilient I I attribute that to a couple of things one one is uh the fundamental kind of dynamism is still there um and secondly you know in the pandemic the government took a huge amount of balance sheet damage in the in the in the corporate sector including small businesses and the household sector and transferred it to The Sovereign balance sheet which left the balance sheets you know as compared with the GFC in in critical sectors I mean in the household sector if you wreck the balance sheets it won't take long before you've wrecked consumption consumption is holding up fine”