Tom Heintjes: Welcome to another episode of the Economy Matters podcast. I'm Tom Heintjes, managing editor of Economy Matters magazine, and today we're visiting with Tao Zha. Tao is executive director of the Atlanta Fed's Center for Quantitative Economic Research, and he recently published some interesting research that I wanted to discuss with him today. His paper is titled "A Theory of Housing Demand Shocks," and his co-authors on the paper are Zheng Liu and Pengfei Wang. Tao, thanks for being with us today to talk about your work.
Tao Zha: Thank you.
Photo: David Fine
Heintjes: Tao, can you briefly describe what you called in the paper the "price-rent puzzle?" What does the price-rent puzzle mean exactly, and why is it a puzzle?
Zha: "Price-rent puzzle" means that the fluctuations of house prices are much bigger than the fluctuations of rent. And you typically think if the house price increases, the rent should increase accordingly—which is true, but the house price actually rises much more so than the rent during the boom, and then during the bust period the house prices actually dropped more than the rent. So the fluctuation of the house price is much, much more than the rent is a puzzle to a lot of economists—and of course, other people as well.
Heintjes: Right. So before we dive into your research more deeply, let me ask you what prompted you to look into this area. Was it the boom and the bust that we went through in recent years? What led you to examine this?
Zha: Well, the reason we were examining it is because it has been a puzzle that everyone wants to solve, or at least explain why there is such a discrepancy between the price fluctuation—which means house price—and the rent fluctuation. And if we can understand that, perhaps we'll be able to inform policymakers what to do, or maybe not to do, with this phenomenon.
Heintjes: Right. Yes, I want to get into the policy implications a bit later, but first let me ask you…in your view, what are we able to infer from how house prices and rents moved together, or as you put it in your paper, "co-move"? Or, in some cases, the way they don't move together? What can we infer from that?
Zha: Well, in most cases the house price and the rent actually move together—they go either up or down. And as I said before, it's just that the house price moves in a much larger magnitude than the rent. But sometimes—sometimes—the rent doesn't move at all, and it perhaps has something to do with rent control, and that aspect we don't examine in the paper. We're more interested in, is it possible for the rent to not even move while the house price will still move, and why that's happening. So if we can explain that kind of phenomenon, we'll be able to explain the divergence of the price fluctuation and the rent fluctuation.
Heintjes: Right. Well, I guess we should note that, from what I gather reading your research, the price-rent ratio is often used to help people decide if it's smarter to buy or rent a home in a given area, and it doesn't really get into the matter of housing affordability, right?
Zha: It has implications, and yes, you're right: it does not address the matter of affordability. But it does have implications about affordability in the sense that if the rent is significantly cheaper, then one could afford the house. And then people will choose to rent a home, which will affect the affordability, and so they're actually somewhat connected. And also, that's why I think this is an interesting phenomenon for policymakers to also think about: why they're moving differently and how this will affect the affordability. But the price-rent puzzle per se does not really address the matter of affordability.
Heintjes: Right. Well, you know, Tao, we rarely—if ever—hear about an unsustainable rent bubble, the way we often hear about house price bubbles. And in a normal housing market, in typical times, house prices and rents move largely in tandem. But when house prices rise and rents don't, what does that typically indicate to you?
Zha: Very good question. That was a phenomenon that I touched on earlier briefly, and there are times that the rent does not move almost at all, while house prices fluctuate quite a lot. So what does that really mean? Well, we need to dig into what factors are driving this phenomenon. In other words, what kind of economic forces are driving house prices, and yet not rent?
Heintjes: Can you describe some of these forces, or just enumerate them briefly?
Zha: In this paper, we are talking about how the factors affect house prices. For example, if you have a home, it's a financial asset. So therefore if you own the home, it's a financial asset. If the house price goes up then your wealth goes up, so you can use that to conduct business by making your house collateral, so that you can actually get more finances for your business. That's actually a very important fact. And then if you rent a place, that means you don't own the place. Then you don't have that aspect of your finance.
Heintjes: Right, the equity.
Zha: You don't have the equity. So I think those factors are very important. But of course, there are other factors: for example, speculation—because it's a financial asset, owning your house, and people may speculate with it like any other financial asset. So that speculation could survive the house prices, which is unrelated to economic fundamentals. And the rent, of course, is determined by economic fundamentals—how much people can afford. If the rent price goes up, that's because generally other prices go up, so that's why the rent has to go up. But if it is purely speculation, then the price can move without corresponding fundamentals' movement.
Heintjes: That's a great point. Now, I know you've looked at this, I'm sure, at the macro level. I assume the price-rent ratio varies widely, and the ratio seen, say, for example, in Detroit is quite different from that that we see in San Francisco. Did you note regional variations or anything like that? How did that go into your research?
Zha: Yes, the regional variations are a very interesting phenomenon, because they do vary quite a lot: Florida, California—the house price fluctuations are a magnitude bigger than house prices in Detroit. So our view is that a lot of those—which is not addressed in the paper—but for the regional ones, cross-region there is a very common pattern that the price-rent puzzle is still there, that price and rent, in normal times, typically co-move, but the price fluctuates more than the rent. So that phenomenon is still there, but when you compare across Detroit and San Francisco—and then of course, in San Francisco both rent and price move much higher, and that has something to do with people's tastes. People like to have a house near the ocean, on the coast, instead of in particular areas. So that can drive the price and rent.
Heintjes: Yes, I was going to ask you about the volatility of the price-rent ratio and why does it exhibit such fluctuation, and I think you've touched on that already with some of the speculation and factors like that, that lead to some volatility. Is there anything else that causes the price-rent ratio to be somewhat volatile?
Zha: Yes, and I should elaborate. As I touched upon previously, that because houses are financial assets—so when the value goes up, it means the house price goes up when you own the house, and then you are able to use that as collateral to borrow money for your own business. So the business uses that to finance their investment projects you can think about as liquidity—premium—that they're looking for, because they want money. Liquidity means money. They want money, and the way they can get money is using the house value as collateral. So you can see, they can get more money, and the economy is booming, and more people are employed. And the demand for the house increases again, and then that will drive up the price again, and then they can use that to borrow more money to invest. So this feedback loop actually will help—or sometimes exacerbate—the fluctuations of the house prices. Because one way you can go up—and they're doing that—but at the same time house prices could go down, and then you cut the financing or borrowing, and the people won't lend you much anymore, because your house is not as valuable as it was before. And then you have to cut your investment, the economy is actually in a downturn. So all of those fluctuations are not something that can be reflected in the rent.
Heintjes: Well, Tao, let me take a little different tack here and ask you about the models you devised to reach your conclusions in your research. Were there existing models that you could leverage? Or, if not, what did existing models lack that you needed to conduct the research you wanted to?
Zha: Well, there are two strands of literature, actually, dealing with this issue. One strand of literature is talking about sentiment, thinking about people's beliefs, or speculation, that house prices will go up from now on, so you had better buy a house now. So by doing that, you actually move up the price much, much faster in the current period, because people do expect the house price will move up more. Or if people expect house prices will continue to decline in the future, then they are going to sell the house right away before the house price actually goes down too far. So all the sentiment can explain this wedge between the price and the rent—so that's one strand of literature. Another strand of literature, as I said before, is called the "collateral channel," and that's how you use a house as a financial asset, as collateral to borrow money for investment or even consumption. So in that case, that mechanism can also propagate the fluctuation of the house prices—which has nothing to do with rent, either. So those existing models are attempting to address this so-called price-rent puzzle.
Heintjes: Right. So in performing the type of analysis you needed to produce the research in the paper we're discussing, what were the primary obstacles? Were there any obstacles you had to overcome to achieve the research you wanted to?
Zha: Yes, because when you have this kind of idea you have to write down the model to verify this idea actually works and when you write down the model, oftentimes you don't know how the model will conclude. And so even though you have some intuition as to how this will work, when you write a specific economic model with mathematical equations, the conclusion could be different. So the obstacle is, under what kinds of conditions will you get results that you think actually explain the data, under other conditions? And then the model may have different predictions than you think. So those are the obstacles we have to overcome. And also, we have a good data set on the price-rent ratio across different regions and metropolitan areas in the U.S., and also internationally. And then they have different patterns, as you noticed, across regions and across countries, besides the discrepancies. And across regions, across countries, you find some common phenomenon. Is your model write-down—which is somewhat abstract—able to explain the data you observed? And so those are the obstacles we have to overcome during the research.
Heintjes: Actually, that leads me very neatly into my next question. I wonder if any of your findings surprised you or went against any of your notions about market behavior or credit or anything like that?
Zha: Yes, we did find that there is a parameter called "risk aversion parameter," which in layman's language you can think about as measuring how much people like to have housing services. In that, I mean some people like to own a house. The reason they want to own the house is because they feel like they get services from owning the house. For example, you can redo the kitchen, you can paint the house the way you want. or having housing services.
Heintjes: The joys of homeownership.
Zha: Yes, the joys of homeownership. And we find that that parameter guides you. How much you like it actually matters. And then it can create a nonlinear relationship, in the sense that a set of the results could be different if that parameter—or how much you like the house—goes from one set of values to another.
Heintjes: In your paper, you look at what happens to house prices, and the price-rent ratio, when there is rapid growth in credit. Historically—at least in the U.S.—what do you find has been the effect of a credit boom on house prices, and are rents similarly affected?
Zha: Across all the countries, and also in particular in the U.S., you'll see that before the 2008 financial crisis, we truly had a credit boom—which is what this paper is also studying, is about how much down payment you have to put down to purchase a house. So during a credit boom, it's easier to get credit. The bank will say, "You may not need to put down 20 percent, 10 percent will be enough, or 5 percent, or sometimes even zero percent." So how easy it is to get credit will have a big effect on the house price fluctuations. The easier the credit is, the house price tends to boom and increase much more dramatically. But if you're tightening credit—for example, requiring you to put a down payment and the amount goes to 20 percent or even 30 percent—then you will see that the house price actually is going to decline. So they do affect, and so that's what's happening in the U.S. before the 2008 financial crisis and after the 2008 financial crisis. And that's precisely because it has something to do with how easy it is you can get credit—"boom," because it's called credit "boom and bust." But rent prices are probably affected in some, but not really nearly that much. So I would say that's why we have this wedge between the house price and rent, because rent does not have to be directly affected by the down payment, because it's irrelevant for the renters. So you can see that rent does not even fluctuate that much, even though the credit could be "boom," and whether there is a credit boom or a bust.
Heintjes: I see. Well, in hindsight the dramatic increase in the price-rent ratio leading up to the housing crash in 2008–09 was a red flag—or maybe it should have been a red flag. Do you think people view it with appropriate seriousness when it spikes in a given area?
Zha: Yes, I think that well implies that. There is a chance that the price will come down dramatically, and any big volatility or fluctuation, either in the house price or in the stock market, is not a good thing because the volatility implies uncertainty, and people don't want to live in a world which has a lot of uncertainty. It makes it harder to make policy and makes it harder to invest. So if you see that spike, I'd always be concerned about the volatilities in the housing market.
Heintjes: You know, in our conversation earlier you mentioned other nations' housing markets, and the price-rent ratio is not a uniquely American metric. In your research, you also looked at situations in some other advanced economies. What sort of observations did you make there? Are the movements similar in other countries, or were there factors that set the U.S. housing market apart in some way?
Zha: Actually, surprisingly, cross-country they have very similar—extremely similar—movement between them in this particular metric, which is the price-rent ratio. And the fluctuation of the house price is always manifold, and higher, and more so—more than the rent fluctuation. And if you look at—this is interesting, this is true for advanced countries, and also for the emerging market economies—like in China, and other countries—and they are quite, quite similar.
So this is really related to your excellent question previously about a credit boom, because why do we have a credit boom? Because we want the economy to grow, and we want financing to be easier for people to invest, and then that will actually generate probably exuberance, in terms of the housing market, and make the house prices probably grow much faster than fundamentals—as you know, the fundamentals push the house prices. It's really common across all the economies, all the countries, which is why it makes this research quite interesting.
Heintjes: Yes, that is very interesting. Okay, Tao, here's the question I always like to ask our economists when I talk to them: What do you think policymakers should take away from your work in this case? How would you like to see it inform their thinking on this topic?
Zha: I think there are two ways this research is useful for us in thinking about policy analysis, and the first is that, if it's truly speculation for people to think about how the house price is going to move—and a lot of speculation has something to do with how the policy is conducted in the past, to lead people to believe the house price will go up. To just give an example—if the government says, "I'm going to impose 10 percent property taxes," in minutes you know the house price will drop. So I'm thinking about the policies are always important to somewhat guide or control people's speculation in the housing market, and so that's why I think it's important.
And the second is, in my research, what we're more talking about is the down payment—which is a direct policy instrument, requiring how much down payment you need to get a loan from the bank to purchase a house. And then that down payment, the ratio is actually very important, and whether you want to reduce that from 20 percent to 10 percent, or none, or increase to 20 percent, or even increase more than 20 percent. Surely it's going to affect the house prices, so therefore affect the price-rent ratio. The government policymakers, knowing that this will generate the price-rent fluctuation, should think about: What is the optimal down payment ratio? Is 20 percent really optimal? And also thinking about, in order to curb some speculation, what is the optimal property tax on the house you own? And all those are relevant for policy analysis, and also give a policymaker a lot to think about.
Heintjes: Yes, I think you have given policymakers a lot to think about here, and you've given us a lot to think about as well. This been a great conversation, Tao, and I appreciate your making time for us today.
Zha: Thank you very much.
Heintjes: And that's it for this episode of the Economy Matters podcast. I should note that we'll have a link to Tao's paper on our website, so you can take a deeper dive into what we've been talking about today. And again, I'm Tom Heintjes, managing editor of the Atlanta Fed's Economy Matters magazine. Thank you for spending time with us today, and I hope you'll come back next month for another episode. Have a great day.