AI, Financial Markets, and the Economy – Wharton Future of Finance Series

Knowledge at Wharton

Ask about Knowledge at Wharton

Episode Description

Could artificial intelligence reshape the economy on the scale of earlier general-purpose technologies — and what are financial markets telling us about its potential? In this episode, host and Wharton professor Itay Goldstein speaks with Maryam Farboodi and Penn professor Jesús Fernández-Villaverde about the economic implications of AI and how investors are responding to its rise. They explore evidence that financial markets are pricing in the possibility of transformative AI, its potential effects on productivity and inequality, the role of regulation and adoption, and whether soaring investment and valuations could include a bubble component. SERIES OVERVIEW Hosted by Wharton finance professor Itay Goldstein, this limited series takes a deep dive into the cutting-edge insights and pioneering perspectives of experts in finance, academia, and industry. Discover how artificial intelligence is reshaping financial markets and institutions as we explore its impact on investing, financial stability, the economy, and the future of work in finance.

---- The Wharton School of the University of Pennsylvania is committed to sharing its intellectual capital through Knowledge at Wharton, the school’s online business analysis journal.

Launched in 1999, Knowledge at Wharton offers free access to:

- Articles, podcasts, and videos highlighting Wharton faculty research and analysis of current business trends - Interviews with book authors from Wharton School Press - In-depth series of curated content like Wharton Executive Education’s Nano Tools for Leaders - A searchable database of more than 10,000 articles covering all aspects of business - A weekly newsletter that delivers Knowledge at Wharton insights directly to your inbox

Visit the Knowledge at Wharton homepage: https://whr.tn/3qUpSBE Subscribe to the Knowledge at Wharton YouTube channel: https://whr.tn/3DEAgAU

Read the full transcriptHide the full transcript
  1. [music] Uh welcome everyone. Uh this is the third season of the Warton future of
  2. finance podcast. I am Itai Goldstein, a professor in the finance department, currently the chair of the finance
  3. department and in the third season we are talking about AI in finance. One of the most exciting developments we have
  4. in recent times is AI and this is affecting every aspect of finance and the economy. Today we're going to try to
  5. understand uh at a bit of a broader level what is the effect of AI on the economy? How this is reflected in
  6. financial markets? What can we learn from financial markets? and what are some of the risks we might have in
  7. financial markets because of a potential uh bubble. And uh to dive into these issues, we have two perfect guests. Uh
  8. first we have Mariam Faroody. Uh Mariam has been in different universities in the last decade or so. Uh starting in uh
  9. Princeton, then moving to MIT and now is new at uh Cornell. uh where she is joining as an associate professor uh and
  10. and some kind of a joint appointment between the business school and the computer science uh department. Hi
  11. Mariam. >> Hello. Thanks so muchai for the uh very kind introduction. >> Thank you. It's great to have you. And
  12. then we have uh my colleague here at Penn over in the economics uh department uh Jesus Fernandez Vilvard who is the
  13. Howard Marx presidential professor of economics uh and he has been working over the
  14. years on several related topics uh including uh economic history and we can learn a lot from history when thinking
  15. about uh AI and also computational economics and certainly there is a lot uh here that has to do with uh
  16. computation and uh programming. So um we are going to talk about these uh topics
  17. and I'm going to start with you uh Jesus given your uh broad historical perspective. Uh c can you help us
  18. understand uh what is new in the current AI uh wave and how this compares to previous technological revolutions.
  19. Thank you for having me here. Itai I like to think about artificial intelligence as a general purpose
  20. technology. This is a very common term in economic history and it refers to technologies that are a very broad
  21. spectrum that affects everything in life. So think about electricity. Electricity is just not about the light
  22. at your home, but it's also the fact that you can reorganize factories around the electric engine that you can
  23. completely reshape society. And before that, perhaps it was the steam engine. So around every 100 years or so, you
  24. have one of these general purpose technologies coming that completely changed our life in nearly every aspect
  25. of it. And I think that artificial intelligence is one of those. Um, this is really incredibly exciting because
  26. perhaps the internet was also a general purpose technology, but I don't think it
  27. was as much as artificial intelligence and I really think we are in one of these situations that happens once every
  28. 100 years where the whole economy is going to be completely reshaped. Very good. Uh so so Mariam you you
  29. recently wrote uh a paper that is trying to understand based on information in financial markets what the market thinks
  30. about this uh technology and how it's going to affect uh future uh growth or maybe even some catastrophic scenarios
  31. like uh extinction. Uh so uh can can you uh help us uh understand based on on your paper first what motivated you what
  32. you looked at and and what are uh your conclusions from that? >> So it's uh so exciting for me that you
  33. brought up this question because uh in this paper we asked a super simple question. Do markets believe in
  34. transformative AI? We and that includes all the aspects that you said like something some extreme event if you want
  35. extreme good extreme bad or different rates of future growth rates and we approached this question uh by looking
  36. at the government bond market given that uh market prices have long been way be before AI uh used as a tool and a strong
  37. indicator to understand investor beliefs. So to be very honest when we uh entered this we did not expect to find
  38. anything given that we did not think the US government bonds would be impacted by
  39. such a rapidly moving technology which is so young. H but to our extreme surprise and contrary to our initial
  40. intuition we found strong evidence uh that they the investors do believe in transformative AI and in particularly we
  41. find that a long-term treasury yield fell economically and stochastically and and statistically significantly around
  42. major generative AI uh new model announcements since the announcement of CH GBT 3.5 which got attention of all
  43. the market participants to this new wave of technological growth. Then given that
  44. this finding of drop in yields was very and such significant large drop in yields was uh very um counterintuitive
  45. to us because it would go against the observation that investors believe in any extreme events. So what's your
  46. what's your take on it overall if you had to summarize it? >> So this is a this is a very very uh good
  47. question. That's a million dollar question and uh as uh as I said uh we would expect uh so we know kind of from
  48. finance we know that the US government bond yields to the extent that they're safe they reflect how the investors
  49. discounts their future payoffs. Okay. So if the bond yields are falling that means investors um uh have higher
  50. discount rate. So which would for instance uh one way to cause that would be that the investor's expectation about
  51. um um uh about the growth rate of the economy is going to be lower. Okay, which sounds extremely counterintuitive.
  52. Okay, because these were exact these drops were exactly concentrated in events which were positive AI growth. So
  53. um to be perfectly honest we cannot pinpoint a specific uh answer but um there are a couple of mechanism that we
  54. have explored and they're broadly consistent with the drop with good news. uh one is that there might be some
  55. adverse if people believe that gains from AI are very concentrated within a part of the population. Uh this uh
  56. growth in AI would have um adverse wealth distribution away from the average population and so they would be
  57. concerned and that would uh push down the yield. Alternatively, a saving god story where all the rich people are
  58. going to get richer then they want to save a lot and that would push around the bounds. That's another reason. An
  59. alternative reason that some people have um suggested was that is that the progress of AI improves US fiscal
  60. position that brings down the probability of defaults of US bonds. we this is the only uh thing that we can
  61. actually directly test by using at the CDS spread and we do not find any evidence about it. So to us something
  62. about the redist concern about redistribution or saving god seems to be uh the most um
  63. um consistent with what we have seen but a lot more work is needed by people like
  64. you to figure out the actual mechanisms. Yeah, very interesting kind of just looking at uh what is the reaction to
  65. the new release of uh new versions of the product and and also just looking at at rates. So, Jesus, are you are you
  66. worried about some of these things if you think about like uh redistribution and what what that is going to to do
  67. overall uh to inequality and uh distribution in the society. So I tend to be relatively optimistic that
  68. contrary to uh some expectations artificial intelligence can have a positive effect on on the distribution.
  69. If we go to the United States and other advanced economies most of the economic growth from around
  70. 1973 or so on until around 2010 2012 was what we economists call a skill bias. basically economic growth complemented
  71. very well with a skill. So you look at the relative wages of people that go to college versus high school graduates.
  72. People that went to college did comparatively much much better over the last 50 years. The reason why that
  73. happened probably had to do with the type of technologies that were implemented in the 1970s, 1980s and
  74. 1990s, computers, etc. which complemented very well with a skill. For the first time in over half a
  75. century, we have a technology that substitutes high skill but complements at least for the moment the type of
  76. things that are hard for machines to do but that humans have a comparative advantage. So what do
  77. I have in mind? turns out to be the case that right now you can use uh cloth to write a contract. You can use cloth to
  78. deal with low-level legal cases and you don't need a lawyer anymore. I actually got involved just as a personal anecdote
  79. last summer. I got a speeding ticket in Germany. It was completely unfair. So I actually appealed to it and I appeal in
  80. German which I don't speak that well and I certainly cannot write a legal document in German through plot and I
  81. actually won the case you know 20 years ago 10 years ago they would have required me to hire a German lawyer on
  82. the other hand artificial intelligence cannot do anything about the landscaping in my house here in Hford. So that means
  83. that over the next 20 years or so the relative wages of landscapers is going to go up and the relative wages of
  84. lawyers is going to go down. So in that sense yes there is going to be some people who is going to make a lot of
  85. money but by and large my baseline scenario is that in 25 years 30 years we are going to look at artificial
  86. intelligence on the whole as something that has not increase income inequality within society.
  87. What worries me much more that increases in income inequality is that vested interest
  88. may actually prevent the roll over of artificial intelligence in many sectors. And again economic history tells us that
  89. just because we have a new technology doesn't imply we are going to use that technology. And you know my favorite
  90. example always is about the printing press which is the artificial intelligence of the time. Suddenly we
  91. can make a book, we can share information. Well, the printing press was prohibited in many countries. The
  92. printing press was extremely limited. For instance, in the Ottoman Empire for several centuries
  93. and the Ottoman Empire didn't take advantage of the printing press. So what I'm afraid is that we are going to see
  94. in many countries um roadblocks to the implementation of artificial intelligence just because the vested
  95. interests are against it. And that's the place where maybe thinking about redistribution that gets around those
  96. roadblocks may help everyone. >> And you think that those roadblocks are generally bad? you don't see any uh
  97. positive aspect of >> I mean I'm not going to say all of them are bad you know like everything you are
  98. going to find exceptions but I think that nine of each 10 limits that I can see regulatory
  99. agencies or parliaments across the world passing against artificial intelligence
  100. are not going to be a very good idea in the European Union it's extremely difficult to create any education
  101. related artificial intelligence app because there is a regulation in the European Union that says that if you are
  102. going to be dealing with children or with the education service you need to do so much regulatory compliance that it
  103. makes it nearly impossible for any startup to have the personnel to be able to comply with it.
  104. Okay. So you may end up in a situation where China has fantastic artificial education sorry education art artificial
  105. intelligence education apps the US has fantastic apps Europe just doesn't and you know good luck trying to convince
  106. the vested interest in Europe that this is just a silly idea. >> Yeah. So Mariam um how do you think
  107. about those tensions the societal political tensions and how they interact with the economic effects of AI?
  108. So I uh I I don't disagree with uh uh with Jesus that there it is important that regulation is set to allow progress
  109. of um AI and uh uh and the spill over to many many different um um industries u effectively. At the same time, I'm
  110. particularly concerned about education because I agree that uh jobs like um miners or like landscapers would not be
  111. impacted. But there's going to be a lot a broad set of uh not uneducated but lower education peoples who can be
  112. completely replaced. And unless there is and it's possible to have a good education system, an education
  113. regulatory framework to allow these people to train better and learn their skills to um to augment their abilities
  114. with this new technology. Um I am concerned that uh this kind of if you want middle part of the
  115. population which is actually kind of large would be very adversely impacted and um and could not would not benefit
  116. from the growth rate of AI even if the aggregate GDP growth is positive. So I think I am a little especially given
  117. that at least in the US which is at the forefront of this technological growth there does not seem to be a lot of
  118. discussion or investment in how to retrain uh uh individuals to be able to best use
  119. AI. I think that's a concern unlike for instance going back to Jesus is a very very interesting example of him winning
  120. a law case. This is Jesus. So he is a professor at UPEN. Okay. And of course he has the skill set to use this
  121. technology as good as anybody can. But the average population, I am not sure they're benefiting from this um advance
  122. as much as they should and as much as the potential allows. >> Very good. So uh let let me uh move on a
  123. little bit and talk about uh what is going on right now in the stock market and the way that this is reflecting
  124. uh the overall investment and potential of AI. Uh so of course we saw a lot of money uh being uh spent on AI investment
  125. across the board and we see valuations uh at close to peak uh level. There is this constant worry that maybe
  126. we are uh facing a bubble and if the bubble bursts that is going to have very negative implications of course uh for
  127. for everyone and and uh constantly we hear about comparisons to previous episodes. uh most recently the.com
  128. bubble. Uh so Jesus, when you look at all this, what what is your take? I know that you're a little more skeptical.
  129. >> Yes. Uh first of all, I I have a little bit of an aerodox reading of the 2001
  130. episode. Yes, we had a big increase in valuations. The market corrected in 2001, but if you actually look at what
  131. the internet delivered by 2015 2016, it really did deliver what those valuations
  132. in 2001 were suggesting and perhaps even more. Uh it is true that in the short run it was a bumpy ride and we have seen
  133. that with any new technology arriving in the US since the railroad and it is true
  134. that some of the concrete companies that had higher valuation in 2001 are no longer with us but the amount of wealth
  135. and value generated by the Amazons of the world by the Googles of the world by the metas of the world has been just
  136. absolutely incredible. So I think that in retrospective maybe we can go back and ret 2001 a little bit
  137. more positively that what we give it credit for. Now is this time going to be different or not? So first of all I'm
  138. not claiming that we may not have a crash on the market. This very well happen. uh I'm just claiming that when
  139. we go back in 2045 2050 at this moment the preponderance of the evidence is that we are going to say yes artificial
  140. intelligence was >> so so you're more optimistic about the longer term but uh you don't really know
  141. what's going to happen in between >> well if I if I if I knew that I will be trading on it and you know and that
  142. house in Monteito is not going to pay for itself no I'm very optimistic I really think
  143. that we are going to be back in the year 2050 looking at what has happened over the last 10 years and say that was an
  144. important inflection point in the economic history of humanity. But but just just to be clear, I mean I I think
  145. one difference between uh the current uh technology and previous technologies is
  146. the substitutions with uh humans right and uh so so the counterargument to everyone who is optimistic is always
  147. yeah but you know at the end of the day this technology will do everything. So how can humans be uh better off or what
  148. will they do? So how how do you see that uh being resolved? >> So if you actually look again at the
  149. historical record, we have been saying that for 250 years and it never happened. Okay, people were saying oh
  150. you know remember watch Metropolis is by Fritz Lamb and they were saying that you
  151. know everyone was going to be done by robots by now and it still hasn't happened. Now, in the case that
  152. artificial intelligence becomes so incredibly good that we don't really need humans to produce cars or to
  153. produce a lot of other services, I think that we are extremely creative as an species to come up with other ways to
  154. fill our life. Okay? And there's a lot of things I would like to do with my life if I had, you know, someone walks
  155. in and says, "There you have Jesus. $200 million. Do whatever you want with your
  156. life. You don't need to work for the rest of my life." I promise you I will be able to feel every single day of my
  157. life until the day I die. So if this is the case for humanity as a whole, yes, there's going to be an issue about
  158. meaning of life and as a society, we will need to tackle it. But I don't I'm not particularly pessimistic about it.
  159. On the other hand, my baseline scenario is we are just going to find other ways to do things. we are just going to
  160. create new products that are going to be complimentary to what humans do and at the end of the day the result is going
  161. to be okay. >> Okay, that's a very uh good way to think about things. Uh so so Mariam, you you
  162. talked before about the paper that you have on the bond prices and what you're trying to infill from changes in the
  163. rate uh after the release of of new versions of AI. uh but uh over the years you've done research also on uh the
  164. stock market and information in prices um and I'm sure that you are uh looking to some extent on stock prices uh right
  165. now. So what what do you uh make of that? >> So this is very interesting in the sense
  166. that there are two patterns happening at the same time. one in the past two years
  167. everybody is like oh my gosh stock price of AI related firm are like shooting to
  168. infinity so their valuations are extremely high like Nvidia of the world okay and in fact there's an ETF which
  169. follows AI related stock prices okay at the same so that's I think um kind of two year two and a half three and a half
  170. year uh rise in stock prices is undebatable So um at the same time when we use the same methodology
  171. to look at the response of let's say even indivi both like S&P500 and individual country companies like Nvidia
  172. to these new AI uh uh [clears throat] model releases we do not see a large impact okay the statistically that we do
  173. not see a statistically significant uh movement either up or down Okay. So that is very interesting. At the same time,
  174. one thing that our results merging the findings in bonds and stock uh kind of uh uh kind of points to is
  175. that uh to the extent that uh we believe our bond market results and there is a discount rate effect
  176. uh especially for the long uh duration bonds like 20 years, 30 years. This is also supposed to show up in stock
  177. prices. Okay. And particularly AI related firms um because their cash flows are pretty backloaded.
  178. They their valuation would be very much impacted by this. So this kind of discount factor effect would act like a
  179. would have a bubbly effect in them on them. So I'm and in fact we did try this with the ETF that uh uh that I told you
  180. follows AI firms and we do exactly see that that particular ETF after residualizing on all the other market
  181. components okay does uh go up at the time of uh these AI news events. So, and that is completely consistent with just
  182. a bubble discounting fac discount factor without letting us uh disentangle that from an actual cash flow
  183. uh factor which would say that AI would actually make these companies uh more uh
  184. productive. So that is something that we still don't know how to separate from each other and um and that is actually
  185. why we try to uh in the first step look into the bond markets to look at uh that
  186. belief um factor [snorts] before going into the um the actual cash flow increase due to the advent of AI. Um so
  187. from all that uh I could see we cannot rule out uh um a bubbly component in stock prices
  188. and I also think related to the com bubble boom let's say there was a lot of winner takes it all in that boom so
  189. maybe to some extent guided by that investors ers now are throwing a lot of funding into the um the leaders of uh
  190. this new technological revolution to be hope to be part of the winners. So which uh would not be um which would
  191. not be um uh irrational but that can rely on this speculative uh view of the prospects of the AI for instance if AI
  192. if the cost and that's what something that we are seeing as J was uh pointed out that the cost of this technology
  193. seems to be extremely high. So it is unclear if this level of growth is sustainable.
  194. >> Okay. So uh I guess both of you are also uh of course uh teaching in a university
  195. and uh constantly interacting with young students who are looking to go into the
  196. job market and thinking what they should do. Um what do you tell them uh these days? Jesus, let's start with you.
  197. I I wish I knew how what to tell them. [laughter] No. [clears throat] Okay. I actually have been thinking a
  198. little bit about it and if if any of the of the listeners follows my my ex feed,
  199. I try to post on that from time to time. Um I think that there are two fundamental skills that in some sense
  200. were already the skills of classical education and we need to go back to them. The first skill is to know how to
  201. ask questions. So think about all the dialogues of Plato. It's all about how to ask the right question and the Greeks
  202. understood that, the Romans understood that and maybe we have lost a little bit that understanding over the last 200
  203. years. So the universities need to be much more about what are the right questions to ask. And
  204. the second fundamental pillar of a mo education should be how can I tell this answer is good. So you are going to ask
  205. clot tell me about x. Clot is going to give you a very good answer. No it's going to give you something that
  206. looks like a very good answer and you need to tell if this is the right answer or not. So if I were the president of
  207. pen which I will never be or the dean at Cornell which I will never be my my whole point will be how do I reorganize
  208. education around these two points now if you are just a student and you are asking that question for yourself that's
  209. maybe what you can try to get out of your courses how do I learn what are the right questions to ask and how do I
  210. learn how to judge that an answer that I get from the question is the right one and for instance in my own courses at
  211. pen that's what I try to to organize my courses around trying to say look I give
  212. you a paper this is even for undergrads h are you convinced by the answer of that paper or not do you think that the
  213. answer is the right one and if the answer is yes why and if the answer is no why and I I keep pushing them don't
  214. try to pander to my beliefs because I'm not going to tell you what I think about
  215. this paper but I want you to think carefully about how How do you structure an answer and how do you evaluate it?
  216. >> So, uh I I love uh this are about education. A lot of it is about questioning and curiosity and validation
  217. and uh that is what I try to tell my students that it is important for them to learn. They need to under they need
  218. to learn how to augment their skills by AI and not to be able to be replaced by it. So one practical thing that I do is
  219. that um there a lot of MBAs. So I give them um questions which are totally mathematically correct. So they can give
  220. it to AI and give the answer get the answer which is completely right but it doesn't there is something that
  221. economically doesn't make sense. And always I ask them what is it in this question that does not economically make
  222. sense and um how what do you think what part of the input do you think has to be
  223. changed to address that in economic inconsistency and that is something that I'm sure um lll will be perfect
  224. answering this question in six months by the next time that I fe so far I mean at
  225. least in the past um years h they have not been good and that helps students exactly do what uh Jesus was mentioning
  226. hopefully. Okay, excellent. Uh so I think we are coming to the end here. Uh very very
  227. interesting conversation, a lot of insights. Um somewhat positive, somewhat optimistic, but not all uh optimistic.
  228. Certainly a lot to uh think about and a lot of challenges ahead. Uh so thank you
  229. very much Jesus and Mariam for all your uh insights and uh this is the future of
  230. finance uh podcast here at Wharton. This is season 3 uh talking about AI and finance and today we talked about how it
  231. affects the economy at large and financial markets and what we can expect in the next uh years. Uh and I'm Itai
  232. Goldstein, a professor in the finance department and the chair of the finance department. Thank you all.

Brief Summary

AI's impact on the economy and financial markets is discussed, highlighting investor beliefs and concerns about inequality.

Key points

  • AI's Transformative Power. Jesus compares AI to past general purpose technologies, predicting a societal reshaping.
  • Market Beliefs in AI. Mariam's research reveals surprising investor confidence in transformative AI.
  • Optimism About AI and Inequality. Jesus argues that AI will not increase income inequality in the long term.
  • The Importance of Questioning. Education should focus on teaching students how to ask the right questions.
  • Insights on Education. The conversation covers the challenges and insights regarding education and AI.
  • Adapting to AI. Students must learn to enhance their skills with AI, not be replaced by it.

Tension

50/ 100Some tension

Related episodes