2017 Financial Markets Conference-Research Session 2: China's Model of Managing the Financial System

Michael Sockin of the University of Texas at Austin looks at the intervention of the Chinese government in financial markets. He presents a theoretical framework that shows, among other things, how this intervention may exacerbate, rather than improve, the information efficiency of asset prices.

Transcript

Michael Sockin: Thank you, thank the organizers so much for putting our paper on the program.

So one of the benefits of going the second day is, I've learned a lot from the first day from the various policy and research presentations. I'll try to make some links as I see them, to my own work, as I go through my presentation now. I'm going to go through my work. It's a theoretical model, I'll go through it at a fairly high level, but I will try to walk through it because I think it's important to understand the various details and intuition that comes out of it.

China's economy continues to be one of central planning that is still largely mixed with free markets. Its dual-track paradigm is still present in many sectors. The state sector, while much improved, is still large and less efficient than private firms and likely to remain prominent in the future. In this environment, China's central government plays a pivotal role in stewarding the economy by, for instance, setting the policy agenda, determining the allocation of key resources, and providing soft budget constraints to state firms as well as implicit guarantees.

China's government also has a long history of actively intervening in its financial system. For instance, it has frequently changed the reserve ratio requirements for banks, suspended IPO issuances, modified the stamp tax on stock trading, and regulated mortgage rate and first-payment requirements in its housing market. It even directed a "national team" to bail out the stock market in late 2015, after the market crashed earlier that year.

This ubiquitous intervention, however, is not without its consequences, as much of the fluctuations in China's financial markets can be traced to both the intended and unintended ramifications of government policy. The stock market crash in 2015, for instance, had its origins in rumors in late 2014 that the government was going to engineer a stock market boom to alleviate firms with high leverage. These rumors were neither confirmed nor denied by policymakers.

Investors, speculating on this uncertain future government intervention, took highly levered positions and pushed up stock prices in early 2015—well over 200 percent in less than a year, despite the overall weakness of the Chinese economy. The subsequent bust in June forced the government to bail out the stock market later that summer. The expansion of China's shadow banking system, which led to this high leverage among state firms and local governments, can also trace much of its amplification to the unintended consequences of policy—in this case the four trillion RNB stimulus package of 2008 to 2009.

And as we learned from Alex [Bleck] yesterday, the stimulus package also had implications for the real estate market, and helped contribute to a real estate boom across China. This also had a lot of second round effects. For instance, local governments issued municipal bonds backed by this real estate boom—so-called Chengtou bonds—and this can be read in a paper by Andrew Ang. My coauthor, Wei Xiong, also has a paper documenting how firms took advantage of this real estate boom by in some cases abandoning their core businesses to instead invest in real estate for the return—which is a terrible misallocation of resources.

Now the stock market episode I just described highlights that there is tremendous uncertainty surrounding not only the timing but also the scale of China's management of its financial system, and this uncertainty is a source of speculation for investors. This uncertainty I'm going to talk about is a bit different than that from our keynote speaker last night. It's not so much the unknown about the unknown, here it's going to be more of the policy risk he discussed, but since China's financial system is largely still developing, and a lot of these policies that are being undertaken are pretty much unknown and uncharted territory, there is a sense in which this should apply to the unknown about the unknown as well—the things we can only sort of speculate about for the future.

So I just want to take a few slides now to illustrate how active China has been in managing its financial system. So here, I plotted the time series of the required reserve ratio in China—I believe from around 1985 to 2011. We can see there have been about 32 changes between just the eight years of 2003 to 2011. This is China's management of the IPO [initial public offering] issuances, the arrows being the points of intervention, and finally, just the historical rate of the stamp tax and stock market transactions.

So, why is China so active in intervening in its financial markets? Well, China has a large population of inexperienced retail investors. These investors hold about 50 percent of all tradable shares and contribute to about 90 percent of trading volume. In part a result of their participation, China's stock market is characterized by both high price volatility and high turnover. In fact, asset prices often deviate from their fundamental values, as documented in the literature. The CSRC [China Securities Regulatory Commission] has a paternalistic philosophy toward protecting these retail investors in addition to stabilizing financial markets.

Intense and uncertain intervention by the Chinese government, however, entails unavoidable policy risks. So, since a lot of these interventions—such as the installation of circuit breakers in early 2016—are largely new and untested by policymakers, since they have such a profound impact on financial markets, these interventions attract speculation by investors, which can reinforce and even trigger policy errors. In this environment, intense intervention by the Chinese government makes any sort of noise in the policymaking process a factor that drives asset prices, and this noise can attract and be amplified by speculation.

This has implications for real efficiency since by distorting asset prices, this impacts cost of capital. Instead of reflecting asset fundamentals, prices may instead reflect market expectations of future government intervention—which may be useful for the government to know, but is less so for directing capital flows. And as we saw from Nikolay's [Gospodinov] presentation yesterday, a lot of asset correlations rose in the post-crisis environment.

To the extent that government policy may be a contributing factor, we saw it across a very wide array of asset classes; this might be something to think about more broadly, even though it is true correlations do tend to rise, for instance, just in recessions.

So this rich interaction between the Chinese government and investors gives rise to several conceptual questions that are highly relevant for China's financial system: first, how does government intervention impact market dynamics? Second, how do market participants react to this intervention, do they trade along with, or against, the government? Third, what is the right objective of government intervention, is it to reduce price volatility, or to improve informational efficiency? Are these two motives the same?

For the rest of this talk, I will try to answer these questions in the context of a conceptual framework. I'm first going to introduce a perfect-information benchmark setting, to justify the need for government intervention. I'm then going to introduce an extended setting, with realistic informational frictions, to show that intense intervention not only makes noise in policymaking a factor that drives asset prices, but this noise can also distract market participants from analyzing economic fundamentals.

Finally, through several numerical examples, I plan to just highlight a potential tension between these two aforementioned objectives of reducing price volatility and improving informational efficiency. Key to this result is that market participants can choose to learn about future government intervention instead of analyzing economic fundamentals—which can reduce overall price volatility but decreases price informativeness.

So let's just start with the perfect-information benchmark setting—again, I'm going to stay fairly high level, but walk through the basic ideas of the model. So assume there's just a single risky asset, which pays a stream of dividends. They're driven by a time-varying but predictable fundamental. For now, we're going to assume that this fundamental is known to all market participants. We will later make it unobservable when we introduce realistic informational frictions in the sequel.

Let's assume that a subset of these investors have to trade every day for noninformational reasons, and submit random market orders to the market. These random market orders are both persistent and price insensitive, and are meant to capture unstable market forces. In the context of China's financial system, these noise traders represent those inexperienced retail investors I mentioned earlier, and they are, as we know, a very sizable part of the market.

In equilibrium, prices are going to reflect two things: they're going to reflect the asset fundamental, but they're also going to reflect the aggregate position of these so-called noise traders. Importantly, it is possible that the intensity of noise trading is so severe tomorrow that investors today are unwilling to accommodate the market orders of these noise traders at any price. This can lead to an explosion of volatility, and market breakdown for sufficiently strong intensity. In work by Larry Summers, who seems to have left, but in some of his seminal work he points out that the investors can be ineffective in insulating prices from noise trading behavior, because they're either risk averse or they have short-term investment horizons.

The intuition for this is quite straightforward: noise trading today represents a mispricing that's an arbitrage opportunity, but noise trading tomorrow represents a source of risk that prices will move against investors tomorrow—and this risk may be so severe that investors are unwilling to participate in markets today. This potential for market dysfunction introduces a role for the government intervention.

So now let's just add a government, and assume the government can now participate in financial markets and can trade against these noise traders. Importantly, when the government trades, it's also going to introduce some noise and surprises as well through its trading. This trading motive of leaning against the noise traders is consistent with the paternalistic philosophy of the CSRC to protect these retail investors, in addition to stabilizing financial markets. We're going to assume that the government has two objectives: to minimize price volatility as well as to minimize the price's deviation from its fundamental value, which is inversely related to informativeness.

These two motives are often treated as equivalent in policy discussions, with the former being easier to implement in practice. As I mentioned before, China's stock market is characterized by high price volatility and deviations of prices from their fundamental values, so these two motives seem somewhat sensible objectives.

Finally, the government is going to fully internalize that it impacts the market failure by taking a sufficiently large position to mitigate the region of market breakdown. So as one may expect, the government is able to slow down the volatility explosion and stave off market failure by adding risk-bearing capacity to the market, so the plot here that I have is of noise trader intensity—I'm measuring it by the variance in noise trading against price volatility, both without intervention—this black line—and with intervention—this dashed line. As one can see, price volatility is higher for any level of intensity of noise trading, without intervention, and breakdown occurs sooner. One can imagine times of high volatility for this noise trading as corresponding to periods of high market distress.

So how does the government impact asset prices in this setting? Well, to understand this impact I would just like to decompose holding period returns into two pieces: the dividend yield and the capital gain—which is roughly related to the expected cash flows from the asset, and discount rates that are related to time-varying risk premium. Importantly here, intervention is strictly about intervening in financial markets, it has nothing to do with affecting the operations or the cash flows of the underlying assets. In this sense it is similar to the different types of interventions of the Chinese government I used to motivate my talk. It has nothing to do with perhaps taxes or bailout policies, which are often the focus of the literature.

This also distinguishes the channel I'm discussing from the works of, for instance, [Lubos] Pastor and [Pietro] Veronesi (2012) and [Philip] Bond and [Itay] Goldstein (2015), which focus on interventions that affect cash flows. In fact, much of literature is focused on interventions that affect cash flows. Here we're thinking about purely effects on the discount rates—through changing interest rates or through changing transaction costs—more broadly in practice.

So now that I've built the intuition of the perfect-information benchmark mull, I want to add those realistic informational frictions I talked about earlier. So assume now that the asset fundamental is unobservable to everyone. There's going to be no government—we're going to start with that setting again, just to highlight intuition. Investors, however, now acquire private information about this asset fundamental that may be eventually reflected in the price. Similar to the perfect-information benchmark setting, prices are going to reflect two components: the fundamental given market expectations, which was just the fundamental itself before, and the noise traders' price impact. Importantly with informational frictions, there's now a third component that's related to the aggregation of the information that investors bring with them when they trade the assets.

Asymmetric information actually makes things worse, however, as we can see from a plot of price volatility. Price volatility is always higher with asymmetric information—which is the solid line—and market breakdown occurs even sooner, suggesting with informational frictions there's even more of a motive to want to have the government intervene.

The right plot plots the two objectives I had mentioned earlier: the price volatility and the deviation of the price from its fundamental value. As one can see in this simple setting without the government, trading against these noise traders by reducing the volatility of noise trading (by having the government participate) appears to accomplish both objectives. As one can see if one is able to reduce this volatility or accommodate these noise traders, price volatility and the price deviation both fall, suggesting that these two motives are one and the same. We're going to show that this intuition breaks down, however, when we actually introduce the government trading along with these investors. But keep this plot in mind, because I might return to it a bit later.

We're going to also assume that the government is now going to trade, the government is going to have no private information (for simplicity), and it's going to, again, trade based on these two objectives of reducing price volatility and improving price informativeness. Importantly now, we're allowing investors to now choose what private information they want to acquire. Before, everyone acquired information about the fundamental—now they have a choice between learning about the asset fundamental or instead learning about these future government policy errors when the government trades.

There are three possible equilibria that can obtain in this sort of setting. One is the fundamental-centric equilibrium, in which all investors choose to acquire information about the fundamental—this corresponds to the case without the government. There is also now a government-centric equilibrium, however, in which all investors instead choose to acquire information about future government policy areas. Finally, there's also a mixed equilibrium, in which some investors focus on the fundamental and some instead focus on government policy errors.

Interestingly, in this government-centric equilibrium, price volatility can be lower because now market participants are trading along with the government—in a sense, they're front-running them, because they are collecting information about future government policy—but price informativeness can now be worse, because now no one is collecting any information about the fundamental.

We can illustrate this intuition graphically by focusing on three different types of governments: the first government is only going to care about price informativeness or informational efficiency—that's going to be this solid black line right here—and we expect there to be a fundamental-centric equilibrium in this outcome. Second, we're going to have a government that only cares about price volatility. One would expect that only government-centric equilibrium would obtain when we have this sort of government—that's represented by this dashed line here.

Finally, there's just laissez-faire—we're going to compare it to the benchmark with no government intervention at all—that's going to be this dashed line up there. What I want you to take away from these two pictures is, one, price volatility appears to be lowest in this government-centric equilibrium where everyone's acquiring information about government policy, but actually, price informativeness—or the deviation in the price from its fundamental value—can be worse than with no intervention at all. So even though we're reducing price volatility, we're destroying this informational efficiency in the process.

So when do we expect this government-centric equilibrium to obtain? Well, the intuition we have here is plotted by this sort of phase diagram. We expect to see a government-centric equilibrium, where volatility is lower but informational efficiency is worse in two different types of situations. One, the higher the intensity of noise trading, because when the intensity of noise trading is higher we expect the government to be more active and taking a larger role in financial markets, which would attract speculation from investors, and second, when the government cares mostly about reducing volatility. So if the government has a motive of stabilizing prices, we expect to see this government-centric equilibrium, but bear in mind this does not mean that prices are more informative or reflect their fundamental values.

So, just to summarize this part of the talk: government intervention can help to stabilize financial markets. The perfect-information benchmark setting kind of motivated why you might want to have the government intervene. However, this can have some adverse effects: intense intervention by the government makes noise in the policymaking process a factor that drives asset prices, and this factor can in fact attract speculation by investors.

And finally, I've highlighted a tension between these two objectives—that are often treated as equivalent in policy discussions—of reducing price volatility and improving informational efficiency. While price volatility can be lower with intervention, informational efficiency can in fact be worse.

I just want to think a little bit back to Marc's [Saidenberg] discussion yesterday, when talking about financial regulation, that there is a time-consistency problem that regulators in the government do face—and so far I've assumed the government could commit ex-ante to any sort of trading strategy. So without commitment, however, it's well known that there's a time-consistency problem for the government, because any time it can set expectations ex-ante it has an incentive to violate them ex-post and do what's ex-post optimal, and being a sort of discretionary government.

So without commitment, and investors are still able to choose what information to acquire, what the government would want to do is kind of intuitive: first, it would want to pretend not to trade too aggressively. It'll say, "I don't want to intervene in financial markets." This will induce the investors to acquire information about the fundamental, but once that information has been acquired, it's a sunk cost. And I showed you in that plot with just asymmetric information: once the fundamental information has been collected, reducing price volatility and improving informational efficiency go in the same direction.

So once that information has been acquired, the government would love to trade as aggressively as possible, since it improves both objectives at the same time. But because investors realize this, and they have rational expectations of what the government will do, this generates the well-known time-consistency problem—for instance, going back to classical works of [Finn E.] Kydland and [Edward C.] Prescott (1977) and [Robert J.] Barro and [David B.] Gordon (1983). We expect this issue to be more severe for emerging-market governments, because they have less reputational capital than the governments of their more developed counterparts.

And I'm just going to use this time also to just talk about a related companion paper we have in the AER [American Economic Review] that thinks about a different type of policy that the Chinese government has, which is known as "crossing the river by touching the stone." So what this means is, typically the Chinese government will implement a series of policy changes and examine how the economy reacts to try to figure out what the right policy adjustments will be. However, in the presence of financial markets, this sort of approach may not work. Why? Because financial agents have an incentive to front-run government policy and cause the adjustment to happen too quickly for the government to really learn anything from the policy to make the right adjustments. And we expect this problem to be much worse when the government cannot commit to a policy ex-ante.

So, just to conclude, there's a lot of commonly concerned risks people have about the Chinese economy right now. There's these noise traders, inexperienced retail investors in the stock market, rising leverage across the nation, its shadow banking system, a potentially overheating housing market, and surging capital outflow. However, the government is very active in these markets, it also has a lot of power in being able to resolve a lot of these issues internally. The savings rate is high in China right now, and all of the debt is internal and could be potentially renegotiated through government intervention. What we want to highlight as perhaps a more important policy risk is that policy areas may be magnified by market speculation, and any intervention today may have long-term consequences for the future by distorting asset fundamentals, where you have a misallocation of capital.

So the stock market turmoil in summer of 2015, the exchange rate crash in August, and the breakdown of the circuit breakers in January 2016 might represent some of this speculation about government policy that we're actually seeing in real time. So it's true government intervention can stabilize, but it introduces a new risk factor to prices, and it could potentially shift investor information acquisition away from learning about fundamentals, which ultimately determine long-run growth and the correct and efficient use of capital across the economy.

And finally, we expect these problems to be magnified if there's a time-inconsistency problem, and the Chinese government cannot credibly commit to its policies. Thank you.

Paula Tkac: OK, now we've got Neil Pearson from the University of Illinois, with some of his comments on China.

Neil Pearson: It's a stylized model. It's a discrete time model, with a single asset, in an infinite horizon—so you might wonder to what extent such a stylized model could possibly be relevant. I'm just going to have a few slides about just how stylized the model is.

The model is in the Chinese context, the discussion is about the Chinese context, the asset is referred to as a stock, but it could potentially be any risky asset—a mortgage-backed security, Greek sovereign debt acquired by the ECB [European Central Bank]—so the model doesn't necessarily apply just to the Chinese context.

Make all the convenient assumptions: all random variables are normal, investors have nice utility function so that they end up caring only about mean and variance, everything's very nice. Important components of the model are that there are noise traders. What's a noise trader? In the paper, a noise trader is interpreted as a retail investor; but they're just exogenously imposed in the model. But the way to think about them, they're people who are trading on information, but aren't; who think they're trading on information, but are not; people who make mistakes; people who trade because they have cognitive biases. For whatever reason, they introduce some noise into the system.

And then importantly, there's a government—you could think of it as a central bank, but a government—that trades against the noise traders. X is the government demand, and the first term the government acts to offset the demand of noise traders, and then, importantly, the government is almost perfect...but not quite perfect. What I mean by "almost perfect" is the government has a reasonable objective function, it tries to minimize deviations from fundamental value, and it also tries to minimize price variance.

The two γs here are the weights, the government has a reasonable objective function. The government is pretty good, knows Bayes' rule so it updates its beliefs optimally, so the government is almost perfect. But it's not quite perfect, it introduces a little bit of noise, because even the Federal Reserve occasionally makes a mistake, OK?

But almost perfect, right? And then the other component is that investors are able to acquire information. Rational investors are able to acquire two sorts of information: they're able to learn about fundamentals, or they're able to learn about government actions—and in particular, the mistakes the government makes (that is, the noise introduced by the government actions), right? If the government is an important player in the financial market, then learning about government actions could be important even if government actions don't affect fundamentals.

Sort of a special feature of the model, introduced for convenience, is that investors have to pick. They can't learn about both fundamentals and government actions, they have to pick which one they learn about, but that seems to be just convenience. And in the usual way, of course, markets clear and there's an equilibrium. Now, the sort of interesting feature of the model is that there are two equilibria: one that I call a good equilibrium—in which investors learn, investors choose to acquire information about fundamentals. And then a second equilibrium, which I'll call a bad equilibrium—it's not the words the paper uses, but it's obvious that they intend that one is good and one is bad. Bad equilibrium is that in which investors don't learn about fundamentals but choose to learn about what the government is going to do, or choose to acquire information about what the government is going to do, OK?

So in the bad equilibrium, prices turn out to be less informative. That's the one sense in which the bad equilibrium is bad, so even though there's not production in the model, less informative prices—if we step outside of the model a little bit—would presumably affect investment decisions and asset allocation decisions. And then even stepping outside the model...because in the model, investors have to choose what to learn about. In the model, learning doesn't require real resources, but even stepping outside the model a little bit, the bad equilibrium is particularly bad because perhaps tens of thousands of very smart people in the economy stop trying to learn about fundamentals, but start trying to learn about what actions the government will take.

So it's a very stylized model, and the question that comes up is, what should you take away from such a stylized model? So the first thing to take away is that it's possible for a bunch of smart guys—it's possible to construct a sensible, internally consistent model in which intervention by an extremely sophisticated regulator or government, maximizing a reasonable injective function, behaving optimally, subject to occasionally making some mistakes, knows Bayes' rule—actually makes things worse. In the mechanism that if the government is a sufficiently important player in the market, it's no longer optimal for you to try to learn about fundamentals, it becomes optimal for you to just try to learn about what mistakes the government is going to make, OK?

So the first thing to take away is that it's possible—even with almost the best case—for a regulator or a government or a central bank...when by the best case, the regulator, the government, the central bank, is completely rational, it's, even for the best case, the intervention can make things worse, OK? And it seems like it's very costly to have people diverted from acquiring information about fundamentals to acquiring information about what the government is going to do, OK?

So the first message, if you had been acting under the assumption that intervention by a benevolent regulator—an extremely sophisticated benevolent regulator, a regulator who doesn't make mistakes—well, makes some small mistakes, but anyway. But even if you're a pretty good regulator, a regulator is not subject to agency problems, not subject to political pressure. A good regulator can make things...so if your assumption had been that such a regulator would make things better, one message from the paper is that you should rethink.

So then another message that comes up is, well, another question you might ask is, "OK, smart guys can construct such a model. Does that actually apply to the Chinese financial market, or even could it potentially apply to a Western financial market?" OK? So in thinking about, "Well, does that apply to the Chinese financial market?" we'll try to take a more or less scientific perspective and ask whether the model is consistent with the data.

Well, so here we fail. I don't mean that the model is not consistent with the data, but we actually can't take the model to the data. It's a single asset with a stationary equilibrium...oh, it's a single asset, so there are no cross-sectional implications of the model. Oh, it's a stationary equilibrium, so there are no time-series implications of the model. It's very difficult to think about taking the model to the data. One could imagine thinking about the comparative statics of the model—but the comparative statics of the model are actually varying parameters—corresponds to looking at a sequence of different models, it's actually stepping outside the models.

The short answer to all this is, I can't figure out what the implications of the model for the data are. So it's very hard to think about how relevant the model is, at least in terms of taking it to the data. So a question for the authors—which I don't expect Michael to answer right now—is, does the model have any actually testable implications? And then, I will indulge myself and take a cheap shot: if the model doesn't have any testable implications, maybe it's not science?

Anyway...but that's a cheap shot, because I actually think the model is interesting and relevant, right?

So we can't take the model to the data—that really is not going to help us think about whether the model is relevant—so let's try to think about the reasonableness of the model. One feature of the model is that if the regulator—if the government, the central bank, whatever—cares only about the deviation from fundamentals, you don't get the bad equilibrium.

Well, this is what happens in the Chinese market, in China, when the market goes down. This is the headquarters of the China Securities Regulatory Commission, and this is in the summer of 2015 after a rapid decline of stock prices. And the people in the front of the picture are taking photographs of the protesters, but the protesters are preventing the regulators from leaving their office. Probably most of the regulators in the room hope that this is not what would happen in the U.S. if the stock market goes down a little bit.

So the other message here is that, of course, Chinese investors think they're in an equilibrium, where the government determines stock prices (or the government has a lot of influence on stock prices), and that what the government does is an important determinant of stock prices, OK?

In thinking about whether the model is possibly relevant: again, the model is not—the bad equilibrium—is not relevant if the government only cares about fundamentals. Does the Chinese government care much about fundamentals? Well, short sales are restricted—severely restricted—and actually, historically, they were forbidden. An easy way to get prices closer to fundamentals would be to permit short sales, OK? And other regulatory practices also suggest a lack of concern with fundamentals, right? IPOs are rationed, there are limits on the prices at which shares can be offered—this actually creates various pernicious side effects, there is a queue of several hundred companies—there's a queue of several years, at current rate—to clear all the companies that want to IPO.

It has various pernicious side effects. Listed companies with no assets have nontrivial values, because their demand for the listing, to do a reverse merger, because it's one of the devices to go to public. Anyway, it seems obvious that the Chinese regulatory system is not geared toward making prices be consistent with fundamental value, but that the real focus seems to be on maintaining social stability, or social and political stability, that is, on reducing volatility.

Well, is that important enough to put us in the bad equilibrium again? I really don't know; it would be nice to be able to take the model to the data, and we can't. But casual empiricism strongly suggests that China is in the bad equilibrium. It appears that a lot of Chinese investors think the government is committed to having stock prices trade in a range.

So the other thing I was thinking about, the paper is set in the Chinese context, and the risky asset in the paper is called a stock, but the risky asset doesn't have to be a stock, it could be a mortgage-backed security. It could even be a 30-year Treasury bond, since even though there's no default risk, those have price risk. So could the model potentially apply to the U.S.? You might think U.S. regulators would be more focused on fundamentals and more competent, but of course, Chinese investors and Chinese regulators are convinced that they're more competent than U.S. regulators. Who's right? I don't know.

But another thing I was thinking about, I was struck over one and a half days of this conference, the amount—of course, it's a Fed conference—so I was struck by the amount of concern, discussion about when is the Fed's balance sheet going to be unwound, how aggressively will it be unwound, will it be fully unwound. Now, of course, the Fed affects fundamentals, so some of this concern is about fundamentals, but I hypothesize that people would be very concerned about the Fed's actions even if they thought the Fed didn't affect fundamentals.

So I look around here and I listen in to the discussion, and I wonder if we are already in the bad equilibrium where a tremendous amount of resources are devoted not to trying to learn about fundamentals, but rather to trying to learn about, trying to understand, what mistakes the central bank might make, and what the central bank might do. If we think about whether the haircut on Greek sovereign bonds is going to be X, or X plus ε, or X minus ε, a tremendous amount of resources went into thinking about that, and trying to figure out what's going to happen, and it's not going to affect productivity and growth. Greece is dead, European banks are going to have to be bailed out, but tremendous amounts of resources get devoted to things like, is the haircut going to be X, or is the haircut going to be X plus ε? And I think of this as, that's an example we're in a bad equilibrium, where tens of thousands of very smart people are trying—because governments and regulators are such large players in the financial market—tens of thousands of the smartest and most highly educated people in the country are thinking about what the regulators are going to do, and not thinking about productivity and growth.

So I just offer the possibility—and I actually suspect most of the regulators in the room will disagree, but...that maybe the model is relevant because we're actually already in the bad equilibrium here. And you say, "Well, we're at a Fed conference, of course, at a Fed conference people talk about what the Fed will do, what else are they going to talk about at a Fed conference?" But the very fact that we're at a Fed conference, and we're not at a conference on productivity and growth, tells us that maybe we're not focused on fundamentals, right? Because you actually think about, as [Harvard Professor] Bob Barro [the quote is actually attributed to Robert Lucas Jr.] said, "Once you start thinking about growth, it's hard to think about anything else." Because, I think about this, the only thing that matters for my children is whether growth over the next 20 years averages 1 percent or averages 3 percent.

Obviously, there have been discussions that I haven't heard, but I have heard no one say a single word about productivity and growth. And there's a lot more interest about how aggressively the Fed will unwind its balance sheet, so I think we should think about, we can't directly take the model to the data, but we might want to think about whether we're already in the bad equilibrium, where tremendous amounts of resources are focused on figuring out what the Fed is going to do and not focused on learning about fundamentals. Thank you very much.

Tkac: All right, thank you, Neil. Before I turn it over to Michael to see if he wants to rejoin the conversation from your side on some of Neil's comments, let me say to all the Fed watchers in the audience, we'll be done here in 15 minutes, and you can have it, Neil, all you want [laughter] about your activity. I do think—and I'll use my moderator prerogative—I do think there's another interpretation of Fed watching—and again, those of you out there, you feel free to weigh in afterwards. It's not about mistakes, it's about being a really large player in markets, and trading over a variety of horizons, and taking actions that are going to impact—not just the term structure, but out through credit markets and other things. And of course, our dual mandate is price stability and maximum employment.

So behind the scenes, that is certainly part of the objective function, even if we're not focusing on it here. It's the most important thing.

OK, so Michael, would you like to comment on any of Neil's insights?

Sockin: I want to thank—do I have this back? OK. I want to thank you Neil for his insightful discussion. In terms of trying to take things to the data, we have at least been trying to think about that, but before I was going on the job market two years ago I went to my adviser, Wei Xiong, and I asked him, "Maybe I should do an empirical paper instead, that might help me get a better job." He goes, "No, Michael, don't do that, that's not your comparative advantage." [laughter]

I have not touched data ever since, but I will certainly try to think more about the empirical implications. Thank you.

Tkac: OK, so Neil, you kind of got to one of the most popular questions, in U.S. markets, investors lamenting the amount of time they spend thinking about global central banking, so I think we've kind of covered that one. But here's a question that I think is interesting, especially as it does relate to some broadening of the perspective outside of China: how does the government know about fundamental values, and how do you think about—either in the context of your model, or more broadly—the role of promoting price efficiency and informational efficiency?

Because, to your point, I think it supports the macroeconomic objectives that sit behind certainly the Fed's mandate. How do you think about implementing that in a policy sense? You've got in your model some very stylized ways to think about that, but I think there are various alternative mechanisms by which a government might try to promote informational efficiency, and getting prices to fundamental values—perhaps not even knowing what those fundamental values are.

Sockin: So, at least one would imagine maybe like a market-based solution, if in the absence of people focusing on learning about future government intervention people would be focusing on analyzing the fundamentals. And I think to some extent that's maybe what the Chinese government expected in the background when it was doing all of these interventions. With the expansion of credit with the stimulus package, it basically gave a carte blanche to lending, but as we're seeing with a lot of research, for instance, by [Wei] Xiong and [Chong-En] Bai, and Zhiguo He, that we're seeing a lot of this allocation of resources that went to less efficient firms, went to well-favored firms of these state banks.

You might expect that the market might be able to allocate resource effectively, so perhaps trying to step back a bit, and pull the reins in, kind of encourages one of the roles of what we think financial markets are for in the first place. Maybe the regulators don't know, and they shouldn't be the ones making the loans and doing the credit rationing themselves, but perhaps they can try to at least encourage or facilitate that among the players that we do think are best at doing that, such as banks, intermediaries, and market participants.

Pearson: Can I offer an answer to that question?

Tkac: Please.

Pearson: I have a simple answer to that question. They don't know what fundamentals of values are, they have no idea, they make the same mistakes everyone else does. I think in the context of the model, it's just as easy to have the government minimize a convex combination of price variance and deviations from fundamental values as it is to have the government minimize price variance. So they do it, but I think the interesting case is the more pernicious case where the government—even if the government thinks it's minimizing deviations from fundamental values, it's not.

So that in practice—and I think this is characteristic of the Chinese market—in practice the Chinese government, the Chinese regulatory system, is focused on volatility. And the reason is that even if they think they can estimate fundamental values, they can't, and when they think they're doing it, if they think they can, they're just making a mistake.

Now that's a particularly strong view, but my answer is they really can't, because they make the same mistakes that everyone else makes.

Tkac: OK, so let's turn to another potential solution that China could move to. The question here is, the U.S. also has millions of uninformed retail investors, but those of us that study markets know that there are a variety of intermediaries that help those folks manage their money—and we heard about many of them yesterday—wouldn't another policy be for the Chinese government to encourage the development of that sector to achieve these objectives?

Pearson: I'd like to answer that question, because...I think that would be a great policy. In the context of the model—so noise traders come in exogenously, but also rational investors have a very short-term horizon, in the context of the model. And that's just taken exogenously. In Michael's model, Chinese investors have a very short horizon, but I wonder if that's created by the policy? If the policy is more or less that stock prices are going to trade in a range—and we don't want prices to go down, definitely—and we don't want prices to get high, because if they get high then there's a risk they're going to collapse. So then, it doesn't make sense to be a long-term investor because prices are not going to go up at 10 percent per year.

And it's actually pretty safe to trade aggressively over short horizons, because you think the government's not going to let the market collapse. So I wonder if the equilibrium consequence of the Chinese policy is actually to create the market, or to encourage the market, in which people have a very short-term horizon.

Sockin: I would agree with Neil. I also think that the financial markets in China are still developing. It's a growing pains process for how the Chinese government wants to regulate and intervene, and decide best how to decentralize this role among investors. I have some colleagues who've gone over to China and talked with some hedge funds there, and it seems like they are a bit focused on policy as well, and so even some of the delegated asset managers have this sort of view. And perhaps, at least in the context of the model, if it's rational, if you care about your returns, if you care more potentially about the capital gains than the long stream of dividends, it's perfectly sensible to have this sort of approach to trading and to running strategies.

So this may actually be part of what is going on. How to shift away from it, or to develop better practices—again, perhaps the government is shaping how investors trade and how they think about their horizons.

Pearson: And in the picture I showed, what I think is really going on is people thought the government would prevent the stock market collapse, so it was very safe to trade aggressively, and then they feel like the regulators reneged on them and allowed a stock market collapse. And that's why they're angry, because they feel like there was reneging on the implicit promise to make sure the stock prices don't go down.

Tkac: OK. So I'm going to ask you to put your policymaker hat on—and I'm guessing I may know Neil's answer to this question. But I don't know, he might surprise me.

So, is it always the case that the government should think about reducing volatility? Is it a reasonable objective for a policymaker to care about volatility, or should volatility be looked through?

Pearson: I'm just a discussant; it's your paper. [laughter]

Sockin: That's fine, but I thought she was directing it...it seemed like a natural segue...

Tkac: Well, I'm not so sure of your answer, I'm a little more certain of what I'm guessing...

Sockin: In some cases, it's hard to tell. I mean, volatility can be driven by fundamental volatility as well—things becoming more volatile may just be a reflection that people are acquiring better information, and so prices are more reactive to the news about the fundamental. So it's not necessarily clear in all cases, volatility is equal—or should be treated equally—between the two.

I've heard some criticism whenever I've had my paper discussed at other places: "Well, what China is really trying to do is prop up asset prices, at the end of the day." Well, you might think by reducing volatility it's reducing the price of risk in the background—that, as a level effect, will be raising prices over time—so some of this intervention may be kind of contributing. We see, even for the U.S., that the VIX is at an all-time low, really since 1993, and this might be contributing just to an overall higher level of prices and more compressed risk premium—which introduces a lot of issues that are outside the context of the model, but it's true, you do get a lot more out of it, reducing volatility has a lot more effects than those I just described with my framework, for sure.

Tkac: Thoughts?

Pearson: I agree.

Tkac: [laughter] You agree. OK, we're almost out of time so I'm going to head to the sort of bottom line here, and the question literally asks, what is the bottom line? Chinese intervention: good or bad?

So, you listed a lot of examples of events that we've all seen over the past several years, of various kinds of interventions or actions by the Chinese government. Ultimately, how do you see this fitting into the framework of your model? You've got the good equilibrium, the bad equilibrium, and then we'll say there was the hybrid equilibrium—right?—where clearly the government is a player, but maybe not necessarily driven all the way to crowding out fundamental information acquisition.

So, if you had to say—since you're sitting up here, for another three minutes—what would you say?

Pearson: So, it's difficult to think about the counterfactual, right? So, I think there are a lot of pernicious effects from intervention. Prices seem to be not well connected to fundamentals, and we actually already talked about one thing where I think that the focus on reducing volatility actually encourages aggressive short-term speculative trading, because it makes it safer.

But I think I can identify lots of pernicious effects of the Chinese regulatory system. I think the Chinese regulators would say, "Well, maybe so, but it would be even worse absent us." So we really don't know what the counterfactual is, but I think the regulatory system—it was actually related to one of the earlier questions—inhibits the development of a healthy financial market, so that even if right now things would have been worse without the current interventions, I think the current interventions inhibit the development of a healthy financial market because they discourage thinking about fundamentals.

Sockin: I like to think of these more in terms of trade-offs rather than just unambiguously good or bad. In terms of a trade-off, what are the benefits? Obviously, they are engaging in these interventions for some purpose, some objective in mind: protecting the investors in financial markets, trying to prevent defaults and civil unrest, for instance.

What are the costs? The costs, I think, are harder to try to quantify and over a longer time horizon. It's true, intervention is going to stabilize things in the short term, and perhaps rationally put people to thinking of things on short-term horizons, but the long-term consequences are a bit more subtle, and they've been borne out—at least in the data we're seeing so far—in terms of misallocation of resources in the cross-section. The four trillion RNB stimulus package, which was applauded by the world as helping to stave off a deeper financial crisis, well, what happened to it? It was lent out mostly to state firms that are less efficient than private firms, it was lent out to real estate—and then when the government tried to rein that in, it ended up developing a shadow banking system through these entrusted loans and municipal bond issuances.

So the government tried to engage in something that seemed at the time the right thing to do, and had a lot of benefits in terms of boosting investment, but we're now seeing that this investment doesn't seem to have gone to the most productive uses of capital, and people are estimating that's going to have a permanent impact on long-term growth.

So there are benefits. I think the benefits here are more short term, but they're definitely there. The costs are much harder to quantify, and probably are going to be borne out over a much longer time horizon, that's kind of how I would think about.

Tkac: OK. So, on that note, we will end the session, take a break, and come back for our last discussion. Thank you.