Not Another Investment Podcast
Understand investing beyond the headlines with Edward Finley, sometime Professor of Finance at the University of Virginia and veteran Wall Street investor.
Not Another Investment Podcast
Supply, Demand & Securities: What Your Economics Professor Never Told You with Aditya Chaudhry (S2, E7)
Take everything you thought you knew about how markets work and prepare for a fundamental rethinking. When NVIDIA's stock price skyrockets after negative news, or the Federal Reserve successfully manipulates interest rates through bond purchases, classical economic theories struggle to explain what we're witnessing in real time.
Professor Aditya Chaudhry from Ohio State University joins us to explore the revolutionary concept of "demand-based asset pricing" - a framework that's challenging decades of financial orthodoxy. Traditional models suggest that when investors irrationally pile into a stock, deep-pocketed arbitrageurs should quickly step in to correct any mispricing. But what if there simply isn't enough arbitrage capital available to enforce rational pricing, especially at the macro level?
We dive into groundbreaking research showing that shifts in investor demand might impact security prices up to 100 times more than classical theories predict. This "inelastic markets hypothesis" helps explain everything from market booms and busts to why quantitative easing actually works. The implications extend from central bank policy to government debt issuance, potentially reshaping how we interpret market movements.
The conversation bridges complex financial theory with practical examples, making sophisticated concepts accessible without sacrificing depth. Whether you're trying to understand the current bull market, puzzling over Federal Reserve decisions, or simply seeking to make smarter investment choices, this episode provides crucial context for navigating today's financial landscape.
While uncertainty remains a fundamental feature of markets, the one clear takeaway reinforces timeless wisdom: for most individual investors, passive products remain the most reliable strategy rather than attempting to outguess the market's complex demand dynamics.
Show Notes:
Gaigax & Koijen, Inelastic Markets Hypothesis (2023)
Chaudhry & Li, Endogenous Elasticities (forthcoming)
Thanks for listening! Please be sure to review the podcast or send your comments to me by email at info@not-another-investment-podcast.com. And tell your friends!
Hi, I'm Edward Finley, a sometime professor at the University of Virginia and a veteran Wall Street investor, and you're listening to Not Another Investment Podcast. Here we explore topics in markets and investing that every educated person should understand to be a good citizen. Welcome to the podcast. I'm Edward Finley.
Speaker 1:When you open the newspapers today, what are the headlines you read? You read about a soaring stock market, two years of 20 plus percent per year returns. You read about AI frenzy everything connected with AI. You read about the federal budget deficit, the federal debt being enormous. You read about central bank under pressure from the president to reduce interest rates. There's a lot going on.
Speaker 1:The question that it seems to me the headlines beg is are securities priced by supply and demand, what people want and don't want, or are they priced some other way and what are the implications of that? With me to discuss. That is Aditya Chaudhry. Aditya is an assistant professor of finance at Fisher College of Business at the Ohio State University. His primary interests are asset pricing and macrofinance. He earned his PhD in finance at the University of Chicago Booth School of Business and he earned a joint degree in PhD in finance at the University of Chicago Booth School of Business and he earned a joint degree in finance and mathematics at the University of Virginia where he was my student. Adithya, welcome to the podcast. Hi, edward, great to be here, glad to have you. So I set up in the introduction a little bit about what we're talking about. Let's do just a little bit of nerdy, dorky econ professor stuff for a second. Classic asset pricing theory tells us what about supply and demand in the price of securities.
Speaker 2:Right. So if we take the most standard, straightforward, classic asset pricing theories, they're going to tell us that the price of any security let's call it a stock should be equal to what people expect the future cash flows of that stock to be, under some notion of rationality, discounted back to the present at the appropriate discount rate that reflects the risk of the stock.
Speaker 1:So does that mean that if more investors want a stock, we shouldn't, under classical asset pricing theory, we shouldn't expect to see a change in price?
Speaker 2:Basically so. The idea is that if you have a bunch of what we would call noise traders who come into the market, they start buying up, let's say, nvidia, for no good reason. There's no rational reason to want to buy more Nvidia. We don't really think the cash flows have gone up. We don't really think it's any more or less risky than it was before. Just like, completely randomly, a bunch of people come in and want to buy more Nvidia. In a classic asset pricing theory, the idea is, this should have very little impact on price because in theory, you have essentially what we would call deep-pocketed arbitrageurs who are willing to step in and make a trade against whatever these Norse traders are doing.
Speaker 1:I see. So the idea isn't that the classic asset pricing theory assumes that demand doesn't have any effect. It instead assumes that, yeah, there might be lots more people who want to buy NVIDIA, but let's just use that as an illustration Lots more people who want to buy NVIDIA, which would otherwise, if there's no change in supply, have the effect of bidding up price. But if the price gets bid up beyond what our pricing theory tells us it should be, there's a big pool of arbitrageurs out there who are going to short the stock Precisely Exactly.
Speaker 2:And now they build up a big short position. Maybe they say, look, I have a lot of risk in my books and maybe I want to be compensated for this, and so maybe there's a small price impact. But you know, in standard models is it going to be very small. Basically, OK.
Speaker 1:So in standard economic models, asset pricing models, we would expect there to be little impact from changes in demand. Same for changes in supply. Precisely Exactly so. If a company is issuing new shares in a secondary offering, let's say, or if a company is buying back its shares, it should be of no effect in the end, ok, a company is buying back its shares, it should be of no effect in the end, okay. There are, though, some effects, like some material effects on price, for some changes that one could interpret as changes to demand. So, for example, changes in risk aversion, or changes in inter temporal choices, choices about do I want to consume my money today or do I want to save it for a future day. The classic asset pricing theory does account for those changes, right?
Speaker 2:Yeah, absolutely so. These two particular ones you've mentioned would enter prices, what we call the discount rate. Basically, in both cases, you're essentially going to view the future cash flows that NVIDIA is going to generate as either more or less risky, and because of that, you're going to want to discount its future cash flows back to the present at a higher or lower rate, and so the quote unquote rational price ought to adjust to reflect these. I gotcha.
Speaker 1:And so that change in demand is really a function. So it is a change in demand right, because if I'm using a different discount rate it means I want more or less, depending of that stock. But you're saying then, in classic theory, that can be a response to these choices, but at the same time that's going to reflect itself in price as materially as, say, the changes in demand that we described about a minute ago. Or is one larger or smaller?
Speaker 2:A change in discount rate in a lot of theories is going to have a very large impact on asset prices. If you, for example, double the risk aversion probably a lot of models that's going to end up having what we would call the efficient or the rational price.
Speaker 1:Yeah, no, and so I think this is crucially important, right? Because when we're reading headlines and we're seeing NVIDIA's price go massively up, it's not like we can just look at that chart and say, oh, that's more demand, because, for example, if I'm understanding you the right way, it could also be a real decline in risk aversion.
Speaker 2:Exactly. It could also be a real increase in the expected future cash flows of NVIDIA, and maybe we are really entering an AI golden age, and for very rational reasons. You know, everyone's going to need a lot more GPUs. Nvidia is basically the only one making them, and so this is going to need a lot more GPUs. Nvidia is basically the only one making them, and so this is going to create a lot of future profits for NVIDIA.
Speaker 1:And so listeners who've listened to season one will remember discussions in earlier episodes in which we talk about efficient markets hypothesis and we talk about the reason that we have markets and fundamentals about how the equity market works. For those listeners who haven't, or maybe you've forgotten, I encourage you to go back to season one and take a listen to some of the episodes on equity markets, just to sort of bone up on what we're saying. Cool, ok, so that's classic asset pricing theory. And so again, staying with our NVIDIA example. So in the Financial Times just earlier this week it was reported that on Monday and Tuesday there were retail inflows that were quite substantial. You know millions of dollars of retail inflows and when you look at the chart that the FT provides, you see that flows have been on a pretty steady positive clip for a while, until around May of last year, when there's a pretty big increase. But the Monday and Tuesday increase in flows even topples that one. So if we're thinking about that article from a classic asset pricing theory point of view, what's the interpretation?
Speaker 2:So it's either one of two interpretations. So either it's the case that people think all of a sudden, on Monday and Tuesday we had some new news that would raise investors' expectations of the future profits, the future cash flows NVIDIA is going to generate. Hence they're going to buy more and in a classic asset pricing model, this ought to increase the price of NVIDIA. Or these are coming back to what we called it before noise traders, who are buying for no good reason, in which case a classic asset pricing model would tell us this ought to have very little impact on prices.
Speaker 1:In sort of the medium term, like it might have an immediate impact, but then, sure, sure, it should reverse itself Exactly. Yeah, so let's take up that point in its two iterations. That kind of drove this big jump was the release last week of the new AI model, deepseek, which sent people into a frenzy because it appeared that they could deliver similar not the same success rates as something like ChatGPT, but at a far lower cost of development and at a far lower cost of running. So that news points towards which of your two explanations.
Speaker 2:So this news would actually suggest that people should lower their expectations of the future cash flows of NVIDIA. So, if anything, people ought to pull out of NVIDIA, when that would be consistent with the big declines in NVIDIA that we saw on Monday. But the FT article you're mentioning is about retail investors pouring into NVIDIA and not pulling out of NVIDIA, and so this seems to be inconsistent with a story of. We have deep seek comes out. This is really bad news for the future profits of NVIDIA and people take that into account, and so this seems to imply that we might be in the second category, where you have a lot of investors, these so-called retail investors noise trades, if you want to call them that who are pouring into NVIDIA regardless of the cash flow news that has come out.
Speaker 1:Got it, so you would not expect to see inflows.
Speaker 2:You'd expect to see inflows? Yeah, exactly.
Speaker 1:Though I would add that the chart I haven't seen any of the data, I don't know if you've seen them, but the chart in the FT shows net flows and so, for all we know, there are outflows of Nvidia, right, it's just that the inflows are bigger. Nevertheless, it suggests that when we're looking at retail really signals is that there's going to be quite a lot of competition in the AI space, and if there's going to be quite a lot of competition, then that might imply much higher cash flows in the future, which, if I'm doing my math the right way, points to your first explanation.
Speaker 2:Precisely Exactly so. The Jevons Paradox just to take a quick tangent is the idea that if the cost of a resource becomes a lot cheaper, this would actually be good for the company or the person providing the resource, if this enables a lot of downstream customers applications to use a lot more of this resource. So if DeepSeek has made some large technological advancement that all of a sudden lets everyone use a lot more AI, large language models, et cetera, jevons Paradox would argue that this could actually be good for NVIDIA. In the long term, this actually could be a positive cash flow shock.
Speaker 1:And so it may be that there are some retail traders who are taking the view of this second opinion Exactly, and they're actually coming into the stock, which listeners will recall that when we're talking about markets and how they work, this is the whole reason why we have them in the first place, because there is uncertainty what the right answer is and that's exactly what markets are really designed to do is have a collective way for actors to express views in price and to find out what the right answer is. We just have to wait to see the end of the story. So I don't want listeners to take away from what I just did there as saying, ah, so there's no answer. I want the listeners to take away the idea that there will be an answer. I just don't know that. We know it at the moment. Ok, cool, so let's shift gears a bit. Now. Let's talk a little bit about classic microeconomic theory, because when we're talking about classic microeconomic theory and we're talking about pick your commodity, what would you like to talk about?
Speaker 1:It's called apples, apples, right. When we talk about apples, we don't have really any of the conversation that we just had a moment ago. No one is imagined to be thinking about the future cash flows of apples and discounting it back to our present state, etc. So what does classical microeconomic theory tell us about supply, demand and price?
Speaker 2:Exactly so. If we go back to our first microeconomics class in undergraduate, it tells us that prices are set by supply and demand. You have some people that produce apples, that supply apples, and you have people that consume apples. They demand apples, and so the people supplying the apples want to be paid for supplying apples. If the price goes up, they're going to supply more apples paid for supplying apples. If the price goes up, they're going to supply more apples. The people who are consuming or demanding apples would like to pay as little as possible for those apples. If the price goes up, they'll demand fewer apples, and so we see that in equilibrium, the price that gets set in the market is the price that equates supply and demand.
Speaker 1:Got it so for listeners who might never have encountered supply and demand curves or understand it, and in sort of very broad brush, we can imagine, you know, an X-axis and a Y-axis, where and I'll put this picture in the show notes for anybody who wants to have a look On the X-axis we'll put quantity, on the x-axis we'll put quantity and on the y-axis we'll put price. The demand curve describes how much people in the market would demand apples at different prices, and we all understand that, right, intuitively. If apples cost $15 a piece, right, there are still some people buying apples, but probably many fewer than if apples cost 15 cents a piece. And so the demand curve on that chart I'm describing would start in the upper left-hand corner and slope down and to the right. The supply curve describes how many apples apple producers are willing to provide. And likewise, if the price of apples is very high, suppliers want to produce and sell lots of them, and if the price is very low, they don't want to sell many of them at all.
Speaker 1:Again, very intuitive. And that line is going to start in the lower left corner and slope up to the right. And so what Aditya is meaning when he says in equilibrium is. We imagine a market where the place where the demand curve and the supply curve intersect will describe to us where the market will settle in terms of how many apples get produced and sold and bought and at what price. Okay, why isn't that applicable to securities markets? Why do we have a separate asset pricing theory? What about that? Or what about securities, makes us feel like no, no, no, that's a different bird altogether.
Speaker 2:Right, exactly. So the idea in securities markets is that basically, cash is fungible, that if you have $1 dividends paid by Apple, this is the same as $1 dividends paid by Google. By Apple, this is the same as $1 of dividends paid by Google. And so, classically, we have this notion that, for example, different stocks are going to be very close substitutes for one another. That if you have two companies that are doing basically the same thing I don't know, let's say, for illustrative purposes, call it Apple and back in the day, when they're both making computers, they're both having basically the same business risk, the same product, et cetera. Their cash, the risk of their cash flows, is going to be very similar. And so, in a classic theory, we would treat these two stocks as very close substitutes, meaning that if the price of Apple goes up by $1, people are going to say investors are going to say look, apple looks overpriced. Right now, microsoft is not looking overpriced. Microsoft is basically the same as Apple. Let me move my money from Apple to Microsoft.
Speaker 2:So what this means is that this demand curve you're talking about, edward, is, in classic theories, very flat. We say that the price elasticity of demand is very elastic. In classic theories of financial assets, the idea being that, look, there is a rational price that should prevail for Apple, and if the actual price of Apple on the market deviates at all from that rational efficient price, I'm going to demand way less. Okay, that's what it means for the demand curve to be flat, and because of this, a lot of classic asset pricing theory is not thinking about the demand curve or the supply curve for assets. It's thinking about what is the efficient price for this asset? And once we've known that, we know where the demand curve is going to be and it doesn't really matter how the demand curve shifts, how the supply curve shifts, you're just moving along a kind of flat horizontal line, along whatever the efficient price ought to be.
Speaker 1:That's what our classic models would say, with the ability to discern apples from Apple, which is a different thing. Ok, let me unpack that a little bit and make sure I understand you. So, substitution we say that securities, we think that securities, like stocks, are close substitutes of each other and that makes the demand curve very, very flat, meaning that it's really just going to be demanded at one price and if it's not offered at that price I'm out Cool. And we imagine, I suppose implicitly. We didn't say it explicitly but if we're talking about the supply and demand curve for apples, we're thinking about the demand for apples relative to pears or peaches, which are not close substitutes. They're substitutable, but they're not close substitutes. Close substitutes, they're substitutable, but they're not close substitutes. Does that logic change if I'm talking about the supply and demand of Macintosh apples versus golden, whatever delicious apples? Because there I would imagine those are much more substitutable. And do we see in microeconomic theory something more similar to asset pricing theory when we're talking about very close substitutes?
Speaker 2:Yes, precisely so. We in economic jargon would say that the more substitutes this particular good Macintosh apples. Macintosh, I realize, is also maybe not the best thing to call it, because that's also the name of the computer.
Speaker 1:We're deep in this hole and I suggest we keep digging.
Speaker 2:We have Macintosh apples which have a very good substitute, which might be Fuji, golden Delicious apples, etc. If the price of Macintosh apples goes up by $1, in classic microeconomics people ought to demand a lot less of Macintosh apples and demand a lot more of Golden Delicious. This is exactly the same idea as for stocks, now that if you have, if the good in question has very close substitutes, then shifts in asset demand or shifts in supply are not going to have much price impact, because people will just substitute to other goods unless this one particular price is what prevails in the market.
Speaker 1:Perfect. So let's take one little step back. Then We've got two different theories at work. We've got one theory for goods that don't have many perfect substitutes, and we expect supply and demand really to be the main drivers not the only drivers, but the main drivers in what the market price will be. And then we have financial assets securities, stocks, bonds, etc. Where we imagine that these things are themselves close substitutes of each other. Do we think that stocks and bonds are close substitutes in asset pricing theory?
Speaker 2:So no exactly, and that comes down to kind of the risk-return properties of stocks and bonds are very different. Buying a treasury is a very different proposition than buying a share in an index fund that invests in the S&P 500. Invest in the S&P 500. And so even in classic asset pricing theories, investors shifting money from bonds to stocks is going to have a much larger impact than investors randomly shifting money from Apple to Google. I would note that in our classic asset pricing theories, even shifting from bonds to stocks is not going to have as big of an impact on price as what we observe empirically, but in relative terms, the classic theories do suggest that because bonds and stocks are less good substitutes than Apple and Google are, we ought to see a bigger price impact when investors move money from bonds to stocks versus from Apple to Google.
Speaker 1:That's great, thank you. That's a great explanation and it triggers in my mind a follow-up question. So all of these theories, economists are developing these theories and proving them out mathematically. But then of course in the end one has to sort of then observe the world and ask whether empirically sort of measuring what really happens in the world do we observe that in the classic asset pricing model? Do we observe empirically that it works?
Speaker 2:So, on many dimensions, yes, on this particular dimension of shifts from Apple to Google having very low price impact, that appears to not be borne out in the data based on our growing literature. That's really kicked off in the past, let's say five years or so.
Speaker 1:And how about between stocks and other things? So, for example, I'm thinking about booms and busts. I mean, it would seem that whether we're talking about microeconomic theory stocks versus bonds, versus cash, let's say or whether we're talking about asset prices theory just what's the right price for stocks generally those theories seem inconsistent with the idea of booms and busts.
Speaker 2:So it depends. So, to answer your first question exactly, we also see in the data that investors moving money from bonds to stocks has a much larger impact on the prices of stocks than our classic asset pricing theories would predict.
Speaker 1:So is that the consequence of substitution per se, or is it the consequence of changes in risk aversion, or is that the same thing?
Speaker 2:No, no, so, exactly. So when we're talking about the price impact of shifts in demand, either from Apple to Google or from bonds to stocks, we're really talking about hold everything else fixed. Just move some money from Apple to Google, just move some money from stocks to bonds and our classic theories both of these should have quite little impact on prices versus what we see in the data. Now there's a bigger question about in real life. When we do see big swings in stock prices either individual stocks like Apple and Google or for the stock market as a whole where's this coming from? Is this coming from people just randomly moving money from bonds to stocks or is this coming from things like change in risk aversion?
Speaker 2:That's an open question because, as we talked about before, even in classic models, if you have a big economy-wide increase in risk aversion, people are all of a sudden basically a lot more fearful. The move money from stocks and into bonds, the move money towards safer assets and this ought to have, in classic models, a large impact on price because, as we said before, risk aversion has changed. Discount rates have changed. The question is really more of the flavor of if risk aversion has not changed, if the rational expectations for aggregate cash flows in the market have not changed and all of a sudden people just pour money into the stock market. Can they push the price up a lot more beyond what we would otherwise think the efficient price ought to be?
Speaker 1:And we're not going to observe that in real life. I mean, the bottom line is, as we mentioned in the first part of the conversation when we were talking about NVIDIA and flows, we're not really going to be able to observe it concretely. Right, we have to sort of and listeners should understand that this doesn't make economic theory bad or wrong. It's very, very constructive in order to understand what's going on in the real world. But it means you have to make some assumptions, you have to sort of take a view, and so if we're reading the paper and we see two years of consecutive 20-plus percent equity returns, should we be nervous or not? The answer to that question is partly. Do you think that the future cash flows of American corporations that are trading publicly have greater prospects? Do you think that people's risk aversion has gone down, or do you think that this is just noise? You have to make a judgment about that and we'll only know when we know. We won't know in advance.
Speaker 1:Ok, and so I think I understand that we've got these two theories. They're not competing. I think that's the wrong way to think of it. They're very, very consistent with each other. They just understand securities prices to be different in how they operate than, say, goods. And so now the real crux of the thing, which is there's a new-ish, I suppose new, one could say new literature on demand-based asset pricing and, in particular, a paper, a 2023 paper by Gbex and Coyen, which shakes that up a bit, doesn't it Sort of reorders our thinking. Walk the listener through what that paper tries to explain.
Speaker 2:Right, so exactly. So. The point of this paper by Kipix and Coy and it's called the inelastic markets hypothesis is essentially that the price impact of in their setting and moving money from bonds to stocks essentially is much larger Think like a hundred times larger than what our classic asset pricing models would predict. You know the types of models we've been talking about so far, and so the idea here is that, in terms of a sliding scale between, on one hand, we have apples, the good that you eat. On the other end of the spectrum, you have the classic understanding of the stock of Apple as something that's very substitutable with something like the stock of Google. The point of this paper is essentially saying that we are much close, that we ought to think about securities prices as much closer towards Apple's the good than perhaps the profession has previously thought.
Speaker 1:In their paper, particularly or just more broadly, when we think about changes in demand. I think the paper contemplates permanent changes in demand. Just unpack that idea a little in lay language. What should we understand that to mean?
Speaker 2:So the paper thinks about a spectrum of changes in demand, from purely just there's more demand today and it goes away tomorrow to the other extreme, which is permanent there's more demand today and it remains there forever. It's probably easiest to think about the case of a permanent demand shock. So, for example, let's consider a situation in which, over the past 20, 25 years, it's become much easier to make passive investments. We've had the rise of going back. Earlier we had the rise of passive mutual funds and then more recently we've had the rise of passive ETFs. Let's assume that this financial innovation has attracted a lot more money that people were previously holding in bank accounts, in bonds, into the stock market. And once these people have shifted their money into the stock market, it's just kind of going to remain there in perpetuity. They're not going to withdraw it in mass quantities going forward. That's what we think about as a permanent demand shock.
Speaker 1:So I suppose not changes in risk aversion. Changes in risk aversion maybe fall somewhere in the middle of the spectrum, like it's not, you know, noise today and tomorrow, right, but by the same token we don't think of it as permanent, right.
Speaker 2:So we usually think about risk aversion as going up and down but kind of in the long run reverting to some mean. So like in 2008, risk aversion spikes up but you know you look at like 2017, risk aversion was quite low, one could imagine.
Speaker 1:Okay, and so the idea. When we talk about these changes in demand, the paper talks about the whole range, but it's just easier conceptually to understand the permanent change. And so when you talk about the change in demand that affects change in price far more dramatically than what classic theory would tell us, you're talking about, or the paper is talking about permanent changes in demand.
Speaker 2:So both will have much larger, both permanent and, like one period, quote-unquote transitory shocks will have larger impacts on price than predicted by classic models. But yes, the permanent demand shocks are going to have far larger impacts. That's exactly right.
Speaker 1:And if I'm a listener and I'm not an economist, and so I'm thinking demand and I'm not an economist, and so I'm thinking demand, what can I grab onto that helps me understand demand? Is trading volume an example? A good way to measure demand?
Speaker 2:So I would say not quite trading volume per se. I think a better measure of like a concrete measure of demand is going to be inflows and outflows into particular stocks or into the market as a whole. So basically you should think not so much about volume but more about what is the like how many total shares have people initiated buy orders for in a given time period? That's more of the notion of flow that we're talking about here.
Speaker 1:Yeah, and I mean I guess that's intuitive, because if I'm looking at trading volume, you and I might be trading, buying, selling a certain share, as we're doing price exploration. That's going to give rise to volume, but that doesn't tell us anything about how much demand there is to hold the stock. So it's a slippery subject, but I hear you. So flows makes more sense. Going back to our NVIDIA example. So the flows that happen on Monday and Tuesday, if we measure them as a percentage of the average daily trading volume Monday's was almost 2% of average daily trading volume in dollars and Tuesdays was 1% and so, measured as a percentage of trading volume, that sounds pretty substantial. But if we talk about the flows on Monday and Tuesday put together as a percentage of NVIDIA's market cap, it's three basis points. So is that a large inflow?
Speaker 2:No Three basis points.
Speaker 1:three one hundredths of a percent is pretty small of the research in this paper and the topic that we're talking about. I suppose one could say maybe the NVIDIA illustration is not the best illustration because it looks like right, the headline is screaming massive inflows, but the flows if measured. Would you agree that market cap is a better way to understand it than trading volume? Yeah, so if we look at it as a percentage of market cap, it's a really tiny. It's bigger than it had been, but it's still pretty tiny, and so it's not clear whether we're thinking about this in terms of, you know, a permanent change in demand. Okay, cool. So the next question that I have is you mentioned, when you explained to listeners the what the rational investor should think it is. There stand ready willing and able arbitrageurs who are going to short the stock and bring the price back in. But I found it very interesting. Gobex and Coyen had some interesting data and an interesting response to that theoretical problem. Describe to listeners what they said.
Speaker 2:So essentially the amount of quote unquote total arbitrage capital that is sitting on the sidelines that can actually jump in and short a stock when the price gets too high or buy a stock when the price gets too low, is kind of small. Is the point, and this is in stark contrast to what our classic asset pricing theories would suggest, which is that you have basically an effective infinite amount of wealth sitting on the sidelines that can come in and enforce that price equals what it ought to equal under some notion of rationality and efficiency. And so in the real world, arbitrage capital is kind of limited and arbitrageurs are going to face constraints, they can't take infinitely large positions, et cetera, and so these frictions are going to limit the ability of these people who, in theory, can enforce that prices must equal their efficient prices.
Speaker 1:These frictions are going to limit their ability to do this in reality tells us should be a rational price, but probably not because arbitrageurs step in and reset the price by shorting, but instead probably because as information gets incorporated into price whatever exuberance, there was sort of fades and the price sort of settles back down to its place. I think that's a super interesting point of taking theory in real life and kind of putting them together.
Speaker 2:So one quick note on that it absolutely could be that. It could also be that, look, if NVIDIA ends up being, for example, 25% overpriced, there's limited capital. That can come in short that today, if that mispricing let's call it remains there for a very long time, over time, you could have a reallocation of capital. These arbitrageurs might attract more capital when you have this very big profit opportunity If it's sustained.
Speaker 1:If it's sustained. I see, yeah, and this goes to the point when we said earlier, the news last week was about DeepSeek, which could suggest on the one hand, oops, the price is too high, it's irrationally high. But you could also read the deep seek news as oh, this is Jevin's paradox, this is going to be great for NVIDIA. We just don't know. And so I think listeners should be really very sensible about reading headlines and understanding the talk that people are giving that these are not changes that can be affected instantaneously. These are changes that might take a lot of time, all the while the market is trying to figure out what the right price is. This brings me to a follow-up question, which is we know like there's lots and lots of research to suggest that equity markets, for example, are rather efficient at the micro level. Isn't it just logical, then? Like to say well, if they're rather efficient at the micro level, then at the macro level they should be equally efficient, and therefore demand shouldn't affect price, affect price.
Speaker 2:The idea is the following that look, if Apple becomes overpriced relative to Google, you do have a long short equity hedge fund industry that is not enormous but that is able to make this trade of short the expensive stock Apple, long the cheap stock, google. Now we have to think about, let's say, stocks as a whole, like the S&P 500 as a whole becomes overpriced, who are the actors that can actually come in and take a short position out against the market? Global macro hedge funds is a much smaller set of institutions is able to actually make this trade, and so this is the notion this is actually a very old idea going back to, like Fisher Black in the 1970s and 1980s, that stock markets can be micro-efficient but macro-inefficient. Just because, going back to this notion of the set of arbitrage capital is going to be much smaller when we're talking about the aggregate market and the opportunity set's a lot smaller, like I, I mean, it's not the size of the arbitrage market period, isn't it?
Speaker 1:It's the size of the arbitrage market relative to the pricing kernels, and so if it's just Apple and Google, there's probably enough arbitrage capital to make that efficient. If you're talking about all US equities versus other securities, I don't think there's enough capital versus other securities?
Speaker 2:I don't think there's enough capital.
Speaker 1:Exactly Our philosophers would say that's a fallacy of composition. We can't imagine that just because we observe something at a micro scale that it applies at a macro scale. Does the size of the change in demand or, I suppose, the change in supply matter?
Speaker 2:So this comes to some very new and recent work that I'm working on in which it seems to be the case. So throughout this whole conversation we were talking about, you know, let's buy 1% of the shares of Apple, let's buy 1% of the shares of the S&P 500. What's going to be the impact on price? The question is imagine you bought 2% of shares of Apple instead of 1%, or 2% of shares in the market instead of 1%. Would this have twice as much price impact? Would this have two and a half times like more than two times as much price impact? Would this have less than two times as much impact? Based on some recent work I'm actually doing, we find that the larger the demand shock this is joint work with Jay Lee from University of Utah we find that the larger the demand shock, the smaller the per unit price impact. So, roughly I mean, think about, if you double the size of the demand shock, you don't really double the price impact. Maybe you increase the price impact in totality by one and a half times.
Speaker 1:And why do we think that is?
Speaker 2:There could be many possible reasons for this, and I think one of the plausible explanations comes back to. We just talked about this notion that when profit opportunities become very large, you might have capital reallocate itself in the market to take advantage. Essentially, imagine that you have like a 0.1% increase in demand that raises price by like 0.1%. Based on this whole conversation we've been having, you know there's a lot of uncertainty in markets and whatnot. It's not even trivial to detect whether or not the price has diverged from its efficient price, and so investors these hypothetical arbitrage are sitting on the sidelines, might think, look, maybe the price is overvalued right now.
Speaker 2:It's not at all obvious that it is overvalued right now. We have a lot of uncertainty about this. Maybe we don't want to step in Contrast that with a situation, let's say, you have a 10% increase in demand that, in the absence of arbitrage, forces would raise price by 10%. Now the arbitrage are sitting on the sideline and thinking, look, we think this might be overvalued. If this actually is overvalued, there's a lot of money to be made by shorting this stock. Let's do things, let's acquire some information or attract some more capital to figure out whether or not this overvaluation is actually a true overvaluation or is this basically just noise, a lack of our understanding?
Speaker 1:So, bringing it back to our NVIDIA illustration, we said that it probably isn't a terribly big shock in demand. It's modest, it's three basis points of market cap. But what you're saying is, if we understand your paper, if we understand Gabex and Coyen's paper, we would anticipate that to have a fairly substantial effect on price. We would anticipate that to have a fairly substantial effect on price and we wouldn't yet expect to see arbitrageurs stepping in and so on, because perhaps that is the new efficient price a la our original conversation, or perhaps it's not, but there isn't yet an opportunity for there isn't enough profit to be made yet for investors to take the other side of the trade.
Speaker 2:Precisely. It's not worth our time, it's not worth our capital to get into a situation where price has not been distorted that much in totality If the mispricing becomes even larger. Now, all of a sudden, we as arbitrageurs think look, maybe it's worth taking a look at this and figuring out if this mispricing is there Because essentially big. If true, if the mispricing is actually there, then we're going to make a lot of money potentially by trading against this Got it?
Speaker 1:Are there explanations, or that those are probably the most plausible explanations for-.
Speaker 2:Yeah. So the most plausible explanations are going to fall into this set of things. Basically, you could think about arbitrage. Lawyers have some fixed costs of taking out a position and so imagine that for every position we initiate we have to pay. Let's just say $5, make it simple. If the potential profits from opening this position are less than $4, we're not going to do it. Information acquisition is another potential explanation where, in general, we have a lot of uncertainty. It's only worth it for us to go out, do research, acquire information to reduce our uncertainty If the mispricing looks really big, if the potential profits for making this trade are very big. We have this thing in financial economics called slow moving capital, which is the notion that that over relatively short horizons maybe these arbitrage errors are constrained, but over longer horizons, if the profit opportunities are really there, they can go out, raise more capital, try to kind of relax on the constraints they face to trade aggressively against these potentially large mispricings. This is the class of mechanisms we have in mind.
Speaker 1:And so moving away from the micro NVIDIA example and moving toward the macro S&P 500 example, so two years of 20 plus percent returns in US equity markets. We said that markets are macro inefficient. That may be what leads to booms and busts, and so our theory that we're developing here with your paper and the paper from Gavex and Coyen tell us that, in fact, we might see a correction in markets, that this might be the result of an increase in demand, not just a change in risk aversion, and that that increase in demand may or may not be permanent. How could we know?
Speaker 2:In real time it's hard to know. In real time, it's actually even deeper than that. The fact that in real time it is hard to know is one of the reasons that people don't trade against these increases in demand. To begin with, the fact that you know you see equity prices go up by 20 percent. You don't know. First of all, is this because of new cash flow news? People are more bullish on the economic prospects of US companies. You don't know if this is just you know. Rationally, people think the market is less risky now and so the price ought to be higher. Versus, is this just a noise demand shock, reallocation of capital, irregardless of what's happened to fundamentals? And even if it is a noise shock, you don't know how long it's going to last. And the fact that you don't know exactly why the price has moved makes you unwilling to trade against the price movement. And so, yes, to answer your question, you don't know.
Speaker 1:And the fact that you don't know is why we see these large price movements to begin with, let's just like skate around a little some implications of the demand-based asset pricing theory, if it holds, if it turns out to be right. So let me talk a little bit then about central banks. So central banks, during the financial crisis, engaged in quantitative easing, which meant that they, in order to inject even more money into the system, they went out and they bought bonds. And then, more recently, prior to COVID, the central bank began quantitative tightening. They want to remove some money from the system. They went out and sold bonds to the market. Apply this thinking to things like central bank action and what we think that might mean for it.
Speaker 2:No, precisely so. The entire idea of quantitative easing is that, look, the Federal Reserve usually controls short-term interest rates, but in times where we need extraordinary measures, such as in the aftermath of 2008, the Federal Reserve wants to lower also long-term interest rates, and their idea to do this is to go out and buy a bunch of long-term US government bonds. When you buy the bonds, according to this whole notion of quote unquote, inelastic demand, the notion that shifts in demand can have large price impact when the Federal Reserve goes out and buys a lot of long-term US government bonds, that's going to push the price of these long-term US government bonds up and hence lower the interest rates on these long-term US government bonds, conversely-. But it worked and it worked. Yes, it worked, I mean it held.
Speaker 1:It held empirically.
Speaker 2:Yes, Empirically, the estimates of the price impact of QE purchases are that they did matter for sure.
Speaker 1:An asset pricing theory would have told us it shouldn't have mattered.
Speaker 2:Classic asset pricing theories would have told us it should have mattered far less than what we observed empirically Exactly.
Speaker 1:Could the central bank decide that it wants to do something similar with equities?
Speaker 2:Precisely so. The US government, the US Federal Reserve, does not do this. Other central banks around the world have extended their asset purchases beyond just bonds to include other asset classes, including equities. So, for example, the Bank of Japan, for a while now, has been buying Japanese equities as a more extreme form of to make it very simple injecting money into the economy.
Speaker 1:Right and so like these actions. When we read about quantitative easing and quantitative tightening, whether they're just the traditional US central bank notion of government bonds or whether they're the more modern Bank of Japan aspect of equities, we can understand that it works in the way in which it works because of this notion of demand-based asset pricing. That's what makes it work. How about fiscal policy? Right? How much debt can the US have before it starts to affect interest rates? What does demand-based asset pricing tell us should be the answer here.
Speaker 2:No, this is a great question. So, quantitatively, we don't have an answer to this yet and there's a lot of ongoing work on this. Exactly, what is the like quote-unquote fiscal capacity of the US government? What is the like quote unquote fiscal capacity of the US government? But to our discussion here. The notion that demand is more inelastic than our classic theories recognize, the notion that shifts in demand and shifts in supply can have large impacts on prices, suggests that if the US government does issue a lot more debt, this can actually lower the price of US government bonds and raise the yield to the interest rates on. Us government does issue a lot more debt, this can actually lower the price of US government bonds and raise the yield to the interest rates on US government bonds by a lot, essentially more than our classic theories would predict.
Speaker 1:Which is rather interesting to consider because there are two different organs of the national government engaging in these behaviors. So the central bank might be engaging in quantitative easing or quantitative yeah, let's say quantitative easing for a moment but the Congress and the Treasury might be engaged in racking up the debt and selling massively more bonds in the market, and they're working at cross-purposes.
Speaker 2:Exactly precisely, and so in COVID we did have. Simultaneously, the Federal Reserve bought a lot of outstanding US Treasury debt, while at the same time the Congress issued a lot more US Treasury debt.
Speaker 1:And how is this idea? Because in this context we've many people have heard about this notion of crowding out, but this seems different than crowding out. Just explain the distinction here.
Speaker 2:So the notion, the classic notion of crowding out is if the government issues a lot of government debt, there's only so much total demand of the economy to make loans. And so let's say we have $100 of capital that is available to make loans either to the government or to private businesses. The crowding out theory says look, if the US government issues $80 worth of debt, that's only going to leave $20 left for private sector loans to be made. That's the notion of crowding out classically.
Speaker 1:And that would have the effect of increasing the rate of interest on the private debt, because it's scarcer.
Speaker 2:Precisely Exactly. The amount of capital that can be used to make private loans in this example is going to be very small. Hence, the supply of private loans is less. Price of private loans is higher. The interest rates on private loans are going to be higher.
Speaker 1:But then is that contrastable with demand-based asset pricing theory or is it really echoes of the same thing? It just wasn't expressed the same way. And then you might imagine that my follow-up question is can't you say the same about equities? So if you've got private equity and buyout, increasingly acquiring companies and holding them private for longer, isn't it the case that what you might be doing is crowding out? You know, is crowding out really a similar concept to what we're describing here in demand-based asset pricing theory, or is it contrastable? Is it something different?
Speaker 2:Yeah, so exactly. That's a really great point, edward. Conceptually these things are definitely similar, and so the notion is that when the government issues a lot more debt, this reduces the supply of capital available to make loans to the private sector, raises the price of loans made to the private sector. This is conceptually similar to what we're talking about on demand-based asset pricing, except it's a change in supply. It's a change in supply as opposed to a change in demand, precisely Exactly. And now you bring a very good point for private equity the rise of private capital. Broadly, we can imagine that there is some supply of capital that can either go to public companies or private companies, and over time we've had a shift of capital towards private companies and away from public companies. This is basically a negative. We could think of it potentially as a negative supply shock for public markets, which is going to raise the expected return.
Speaker 1:Exactly the expected returns you would earn in private markets?
Speaker 2:Exactly, yeah, actually, it has the consequence of lowering the prices of public equities, because if we raise the expected returns, we're raising the discount rates on public equities, which ought to lower the prices of public equities Got it, and so that would not explain.
Speaker 1:That wouldn't explain the world we see today.
Speaker 2:Exactly Right right.
Speaker 1:Thank you, this has been tremendously interesting. I want to end our episode with something that I do with every guest, and it's a little bit of a curveball. I'd like you to offer our listeners a pearl of wisdom. It can be personal, it can be professional, it can be professional, it can be academic, it can be anything you want it to be.
Speaker 2:Yeah, so that's a great question. I think the pearl of wisdom that really stands out coming off the rest of our conversation is that I mean, you might be listening to this podcast and thinking, wow, there seems to be not only a lot of uncertainty in markets, but also a lot of uncertainty about how academics think about financial markets. And it's probably true. I think the actionable takeaway from all of this is something that we've known for a long time now, which is that one ought not try to beat the market unless they have some very clearly defined edge, and so, for a very long time work that Eugene Fama, Nobel Prize winner, has been doing for 50 years now we think that, for most people, investing in passive products, passive mutual funds, passive ETFs is probably the way to go, as opposed to trying to time the market, as opposed to trying to beat the market, and for most individual households, for most individual, normal people, this seems to be a much better strategy than hopping on Robinhood and day trading a lot you know every second of every day.
Speaker 1:Got it Great. Aditya Chaudhry, thanks for joining us. This was great.
Speaker 2:Thanks so much. Edward had a great time.
Speaker 1:You've been listening to Not Another Investment Podcast hosted by me, edward Finlay. You can find research links and charts at notanotherinvestmentpodcastcom. And don't forget to follow us on your favorite platform and leave comments. Thanks for listening.
.jpeg)
