Finally, someone noticed

by Russ Roberts on December 19, 2008

in Government Intervention

There are four factors that helped drive up the price of real estate in the United States and create the housing bubble: The GSEs (Fannie and Freddie), the Community Reinvestment Act, expansionary monetary policy starting in 2001, and the 1997 Taxpayer Relief Act that for the first time let people avoid capital gains on home price appreciation without having to rollover the gains into a bigger house.

All of these factors pushed up the demand for real estate. But by how much? Over the last few weeks I have been focusing on the capital gains change because the run-up in housing prices began in either 1997 or 1995 depending on which data you use.

This New York Times article from today is the first one I’ve seen that focuses on the role of the 1997 tax change in the mortgage mess:

By itself, the change in the tax law did not cause the housing
bubble, economists say. Several other factors — a relaxation of lending
standards, a failure by regulators to intervene, a sharp decline in
interest rates and a collective belief that house prices could never
fall — probably played larger roles.

But many economists say that
the law had a noticeable impact, allowing home sales to become tax-free
windfalls. A recent study of the provision by an economist at the
Federal Reserve suggests that the number of homes sold was almost 17
percent higher over the last decade than it would have been without the

Vernon L. Smith, a Nobel laureate and economics professor at George Mason University, has said the tax law change was responsible for “fueling the mother of all housing bubbles.”

favoring real estate, the tax code pushed many Americans to begin
thinking of their houses more as an investment than as a place to live.
It helped change the national conversation about housing. Not only did
real estate look like a can’t-miss investment for much of the last
decade, it was also a tax-free one.

The authors do a nice job looking at the politics and some of the economics. But they miss one key point. They did not look at prices, and focused instead on sales. But it is prices where the impact is going to start and it is the increase in prices that made all of the other mistakes possible.


It is an unavoidable fact that when you tax-exempt an asset, it’s going to appreciate. The question is how much and what evidence we might look at to confirm the importance of the 1997 tax change. There has been an explosion in ownership in second homes but it’s hard to quantify. The National Association of Realtors has data that show a big increase after 2003, when they started collecting the data. But that could be the result of the price appreciation not the cause of it. I can’t find reliable data before 2003.

Just look at that picture. If it’s accurate, something dramatic happened in 1997 that fueled an explosion in housing prices. I don’t think it was irrational exuberance. I think it was a change in a tax policy that suddenly made houses dramatically more attractive as an investment vehicle. The other factors mattered in pushing up demand. And lots of ugliness had to happen along the way. And maybe the timing is just a coincidence. But I doubt it. You would expect making something tax exempt would increase its price. And the price kept going up when Wall St found new ways to let new homeowners borrow in anticipation of higher prices.

Be Sociable, Share!



64 comments    Share Share    Print    Email


jwilliams December 19, 2008 at 1:47 pm

What conclusions do you draw from this interpretation?

That the government should not have reduced taxes on home capital gains? That the taxes should have been decreased more gradually? That al capital gains taxes should be equal to avoid favoring one investment over the other?

Marcus December 19, 2008 at 2:17 pm

I think jwilliams asks some good questions.

I suspect we libertarians as a group strongly supported the tax break. I have no data to support that, just an impression I have from the fact that we basically support all reasons for cutting taxes.

So what's the lesson from that? Selective tax cuts are not a good idea?

It is certainly more social engineering.

Russ Roberts December 19, 2008 at 2:32 pm

jwilliams and marcus,

Good points and questions. See this new post:

maximus December 19, 2008 at 2:37 pm

"So what's the lesson from that? Selective tax cuts are not a good idea?"

Exactly. Government interference in the market process, picking winners and losers.
The action favors one investment type over another, diverting investment funds to the favored industry, creates a malicious bubble that goes boom. Now you got a big ol' mess on your hands.

Sam Grove December 19, 2008 at 2:37 pm

If you want my anecdotal experience, it wasn't the tax change, it was the modest down payment and relatively reasonable interest rate that provoked us into buying when we did.
Later, it was the appreciation on our property that made an upgrade possible.

Ironman December 19, 2008 at 2:42 pm

The change in capital gains taxes for homes in 1997 certainly contributed to booming housing prices, but is not the only factor.

Remember that the change in capital gains taxation applied to all real estate markets in the U.S., but not all real estate markets saw anywhere near the same kind of housing price appreciation. Some markets and states saw astronomical gains. Others, basically treaded water.

You need to combine the effects of the tax incentive change with local and state regulations affecting each market to account for the differences between them. The biggest appreciations, by and large, occurred in the areas with the greatest restrictions on new housing.

Let's not also forget the mortgage interest tax deduction, which supports the homeowners while they own their homes (see this post by James Cullen for an interesting idea for undoing the effects of that incentive). That also made ever larger mortgages more desirable.

And then, there's all those other factors Russ identified in his main post above.

The effects of each of these things are not statistically independent from each other – you can't apply standard econometric analysis to isolate a single factor that launched the housing market into space.

Ironman December 19, 2008 at 2:46 pm

One last thing – here's why serious people will have a tough time pointing their fingers at the change in the capital gains tax treatment for housing as being a direct causal factor.

Dave December 19, 2008 at 2:55 pm

This is the best explanation of the housing boom. It lays out a clear cause-and-effect relationship between the Fed's policy and the decline in lending standards.

Of the four factors you list, I believe that the GSEs had a significant but smaller role, and the CRA played a relatively minor role (why did some countries in Europe experience housing bubbles of similar or even worse proportions with no CRA in place?). The capital gains tax change got the ball rolling, but the Fed dropping its rates accelerated the bubble and had the biggest impact as my first link explains.

Marcus December 19, 2008 at 3:00 pm

"You need to combine the effects of the tax incentive change with local and state regulations affecting each market to account for the differences between them."
– Posted by: Ironman | Dec 19, 2008 2:42:54 PM

Good point.

And one of those factors I imagine are the states where when people get under water on their mortgage can stick the lender with the loss.

Combine that with little or no down payment.

It's like a tax-free margin account with no margin and where you can stick the broker with the losses.

Don N December 19, 2008 at 3:20 pm

That is a good list, but you can add Enron/Worldcom to that list. Enron's blowup (and Worldcom and Tyco…) soured many individual investors on the stock market just when yields on fixed income investments and other savings vehicles were diving to very low levels (monetary policy). So with these two major asset classes losing appeal, real estate looked quite good by comparison.

rog December 19, 2008 at 3:22 pm

Here in Australia there is no capital gains tax on your own home however CGT is applicable to homes that are not or never have been your principal place of residence.

For other homes all costs including interest are treated as a tax deduction which can be offset against income (eg rent)

So I dont think CGT is a cause more the change in tax policy became a factor.

rog December 19, 2008 at 3:24 pm

As Ironman pointed out, changes to CGT were not a factor.

James Kibler December 19, 2008 at 3:35 pm

I think it is an astute insight that the 4 factors listed contributed to the housing bubble. I would add as a cause (not as a result, as implied in the follow-up post) the massive purchases of Treasuries by the Chinese government.

I'm not sure I understand the mechanism that would lead from increasing house prices to China's buying treasuries. I would definitely be interested in hearing how that causal chain worked.

I can see how the claim might be that higher house prices led to more borrowing (and consumption) and hence more need for lending (from China), but I think the root of this effect is in the artificially low cost of capital created by the massive Treasury purchases as China maintained their exchange rate in the face of a huge current account surplus.

A lower price to borrowing seems to be a much more direct causal link to more borrowing than an increase in collateral.

Is there more to the argument than that?

James Kibler December 19, 2008 at 3:44 pm

…nevermind–I see that I misread the original post. The point was not that housing prices necessarily increased investment by China, but that funds coming from China and elsewhere were misallocated due to bad incentives caused by the tax policy… Good point.

Dave December 19, 2008 at 4:05 pm

I'd just like to add to my original comment that another piece of the puzzle is the role of regulatory arbitrage in raising the demand for mortgage securitization. The ritholtz link I provided is good explanation for the increased supply of risky mortgages. Kling helps fill in the blanks on the demand side.

This was enabled by ratings agencies, which are protected by the government as "Nationally Recognized Statistical Rating Organizations" helped to encourage demand by lowering their criteria for AAA ratings in order to gain business of securities issuers. Also, other lazy investors who relied the ratings (b/c of the trust granted them by the government) would've also led to increased demand.

Charlie December 19, 2008 at 4:23 pm


I always find your thinking so fuzzy when you try to explain the rise in housing prices. Suppose we can attribute the rise in housing prices to fundamentals, like a tax change, then we still need a fundamental explanation for why housing prices dramatically fell. Right? We define a bubble ex post, because it pops. So unless you have an argument that the collapse was also driven by fundamentals, you still need a theory of bubbles.

Now maybe our theory of bubbles should have the attribute that fundamental changes can spark bubbles. The tulip bubble was precipitated by pretty cool tulips. Beanie babies were, in fact, cool for stuffed animals, just not two hundred dollars cool. And yes, the internet and technology is a big part of the future, of course that doesn't mean etoys should have a larger market cap than But we still need a better theory of bubbles. I can't believe I just really found out that bubbles exist in lab experiments . How did all the GMU bloggers I read not tell me about that?


Oil Shock December 19, 2008 at 4:41 pm
Kent Lyon December 20, 2008 at 11:55 am

One wonders how other factors still impacted the price of housing. After all of the factors that Dr. Russell has identified, local and state factors on restrictions on new housing, etc., what impact did the bursting of the bubble have on the flow of money in to real estate, and how did that affect prices? Has anyone looked at that specifically, or is it possible to look at? How about the stock market crash after 9/11? Did that encourage more money flowing into real estate investments, and add to the bubble? Some markets, like Florida, perhaps benefitted from state tax policies that encouraged New Yorkers and others from the Northeast to purchase homes in Florida, and declare residency there, to avoid state income taxes in the Northeast (Rush Limbaugh among them). There is also a strong cultural inclination of New Yorkers toward retiring or otherwise moving to Florida, even making it into the Seinfeld series. Texas real estate prices did not sky-rocket, but Texas has the highest property taxes in the nation, when expressed as the median property tax bill divided by the median home price. Perhaps that has something to do with why the Texas real estate market did not take off like the Florida real estate market, where property taxes were quite low and fixed (Florida had a 3% cap on appraisals for calculation of real estate taxes, since the mid 1990's, whereas Texas did not, and real estate taxes rose inexhorably, regardless of increase in home values, as local tax appraisers used property taxes as local piggy banks to increase the income of both local and state governments). And then the weakening of the dollar, relative to the Euro, as a policy of the Bush administration to boost exports, made real estate costs in places like Florida much less for Europeans, who purchased property in Florida, much like the Saudis did in Beverly Hills in the 1970's during the oil crisis of the Carter era (at one time, home prices in ritzy areas of Los Angeles were increasing 10% a month in those days, driven at least in part by Arab dollars flowing into Beverly Hills real estate).

indiana jim December 20, 2008 at 12:52 pm


Oil Shock is right, there is a clear explanation for the housing bubble; indeed its pop was predicted by the Austrian theory that he posts a link to.

Conventional AD-AS textbook macro (with a natural rate built in) predicts only that a positive credit shock would push us above the natural rate in the short-run and then falling back to the natural rate in the long-run (as prices adjust upward). But the Austrians (correctly, as clearly evidenced by recent experience) say no, conventional macro is too optimistic because it ignores the misallocations across "stages of production". The artificial credit expansion by government, in the Austrian model, creates incentive for "malinvestment" and "overconsumption", misallocating resources that have alternative uses away from the middle stages of production. This dooms many new projects (begun only due to credit subsidies) to subsequent failure. A bust is predicted by the Austrian model that causes aggregate output to fall below previously sustainable levels.

John Seater December 20, 2008 at 5:46 pm

Both Russell Roberts and some of those posting comments argue that expansionary monetary policy starting in 2001 was one of the causes of the housing bubble. The only evidence offered seems to be that nominal interest rates were low. I disagree. Start with the quantity equation: MV=PY. In its growth rate version, we have m+v=p+y, where lower case letters represent growth rates of corresponding upper case letters. Rearranging, we get p=m+v-y. Although money is not superneutral, it is nearly so for rates of inflation below 20% (actually perhaps much more). The "long run" shows up in 18 months or less. Consequently, if monetary policy was unduly expansionary for 6 or 7 years, we would have seen substantial inflation for at least the last 4 or 5 years. That is, if m was inappropriately high for so long, we should have seen a corresponding increase in p. We didn't. In fact, the only substantial inflation arose from the price of petroleum. Now, I admit that I do not understand all the forces that drove the spike in petroleum prices, but excessively expansionary monetary policy does not seem to have been one of them. For one thing, if it was monetary policy, then petroleum prices should not have fallen precipitously over the last few months, when monetary policy has been expansionary in the extreme. For another thing, how could monetary policy cause such a gigantic (temporary) spike in petroleum prices but nothing else? The evidence seems consistent with some sort of special effect in the energy market, not an economy-wide inflation caused by excessive money creation. The frequently heard argument that Roberts and some commentors repeat, that the housing bubble was caused in part by excessively stimulative monetary policy, does not seem consistent with the full set of data. Indeed, if one believes as I do that the Fed's main job is (or at least should be) maintaining price stability, it has been doing that job with flying colors.

In sum, then, the evidence suggests to me that monetary policy was not excessively expansionary over the period in question and so does not share the blame for the housing bubble. Does anyone have another interpretation consistent with the evidence or perhaps evidence I have overlooked?

Marcus December 20, 2008 at 7:03 pm


A couple of thoughts.

1) We did see substantial inflation: in house prices. I think we're trying to understand why so much money was focused into that one market.

2) We also saw the dollar weaken against other currencies over the coarse of several years. This may have been due more to government debt, I don't know.

Sam Grove December 20, 2008 at 9:38 pm

Devaluation of the dollar can be obscured by increases in productivity.

Also, price changes can be resisted by reducing produce value, production costs, or service.

Hence the success of Walmart, etc.

Dave December 21, 2008 at 1:17 am


The link I presented above isn't quite the standard Austrian argument, but still implicates the Fed, by showing the low rates, in my view, as an enabler of the principal-agent problem.

Here's the relevant quote:

After the Greenspan Fed took rates down to ultra-low levels, home prices began to levitate. More and more mortgages were being securitized — purchased by Wall Street, and repackaged into other forms of bond-like paper. The low rates spurred demand for this higher yielding, triple AAA rated, asset-backed paper.

In this ultra-low rate environment, where prices were appreciating, and most mortgages were being securitized, all that mattered to the mortgage originator was that a BORROWER NOT DEFAULT FOR 90 DAYS (some contracts were 6 Months). The contracts between the firms that originated mortgages and the Wall Street firms that securitized them had explicit warranties. The mortgage seller guaranteed to the mortgage bundle buyer (underwriter) that payments were current, the mortgage holders were valid, and that the loan would not default for 90 or 180 days.

So long as the mortgage did not default in that period of time, it could not be "put back" to the originator. A salesman or mortgage business would only lose their fee if the borrower defaulted within that 3 or 6 month contractually specified period. Indeed, a default gave the buyer the right to return the mortgage and charge back the lender the full purchase price.

What do rational, profit-maximizers do? They put people in houses that would not default in 90 days — and the easiest way to do that were the 2/28 ARM mortgages. Cheap teaser rates for 24 months, then the big reset. Once the reset occurred 24 months later, it was long off the books of the mortgage originators — by then, it was Wall Street’s problem.

Also see his explanation for the earlier part of the bubble in his reply to a comment here. The relevant quote:

[...]you had the Taxpayer Relief Act of 1997 — that dropped cap gains to 20% from 28%, and exempt the first $500,000 for married couples selling house (allowable once every two years).

Around that time, guess what else was going on? The stock market boom and tech dot com bubble put ALOT of money in people’s hands in 1997, 98, 99 and even 2000. I had many discussions with clients, real estate agents and traders about rotating money from equities into Houses back then.

John Seater December 21, 2008 at 9:03 am

(1) In response to Marcus: The argument that the inflation was concentrated in housing does not fit the whole time series. Average housing prices have fallen quite a lot over the last year. That should have led to a "rebalancing" if you will of the inflation by substantial increases in other prices, which has not happened.

(2) In response to Dave: My point is that the Fed did not cause the low interest rates at all. The Fed can lower nominal and real interest rates only temporarily, which it does by increasing the money growth rate above the level consistent with an unchanging inflation rate. After a time, however, inflation increases in response to the higher money growth rate (see the growth rate version of the quantity equation in my original post), the real interest rate goes back to where it was, and nominal interest rates end up higher than they were. We did not see any such increase in nominal rates over the past 7 years. Interest rates were low, but it could not have been monetary policy that put them there because money growth cannot hold nominal rates down for such an extended period. Low rates may have led to the principal-agent problem, but the low rates were not the result of excessively loose monetary policy.

Lorenzo December 21, 2008 at 9:15 am

Any discussion of housing prices in the US which does not grapple with the huge differences between housing markets is not worth bothering with.

Check the data at

See also Krugman's discussion of the difference between "Flatland" and "the Zoned Zone" at

Marcus December 21, 2008 at 10:34 am

"In response to Marcus: The argument that the inflation was concentrated in housing does not fit the whole time series. Average housing prices have fallen quite a lot over the last year. That should have led to a "rebalancing" if you will of the inflation by substantial increases in other prices, which has not happened."
– Posted by: John Seater | Dec 21, 2008 9:03:13 AM

There has been substantial increase in price in at least one other asset: treasuries. An increase unprecedented in history.

Henri Hein December 21, 2008 at 10:40 am

Everybody seems to accept the Shiller graph. I don't. For one, it contradicts the Census Bureau data. As a secondary problem, the run-up in prices from 1997-2006 seems unlikely absent some drastic factor. Given the problems a collection of experts are having in identifying that factor, I think there are some problems with the underlying theory.

The premise behind this line of analysis is that the recent housing bubble was particularly bad. The only evidence I've seen for this is the Shiller graph. Given that the Shiller graph contradicts the Census data, I don't think we can take this as established.

Dave December 22, 2008 at 1:41 am

The Fed funds rate has been used to control nominal short term interest rates quite effectively in recent history. The Fed has less influence on long term rates. The ARMs that inflated the bubble had low teaser rates enabled by the Fed's easy money policies. Greenspan's defense has been that the Fed doesn't control long term rates. But the use of ARM teaser rates were dependent on short term rates.

I'm not sure why you are skeptical that there was a severe housing bubble. Even in the graph you provide, there is a clear break from the historical trend. Try establishing the trend from 1940 through 1990 or 2000 and then comparing it to the growth in prices from 2000 to 2005/2006 (also try using inflation adjusted home prices). Even in your linked graph, you can see that prices rose from about $120K to $180 in 7 years (a 50% increase). The previous 50% increase started between 1970 (~62K) and 1980 (~97K). For the sake of argument, we'll say that the price hit $80K around 1975. That means that there was a 50% increase from about 1975 to 2000 (a span of ~25 years), and then a subsequent 50% increase from 2000 to 2007 (a span of 7 years). That seems like quite an acceleration during a time of relatively mild inflation.

Henri Hein December 22, 2008 at 5:48 am


I'm not saying there wasn't any bubble, just that we don't know if it was any more severe than previous bubbles. Housing prices rises cyclically and bubbles build every 10-20 years. I contest that it has been established this particular bubble is worse than others. Maybe it was and maybe it wasn't, but I haven't seen a convincing case either way.

The Census data is inflation adjusted.

The Shiller hockey-stick graph, in particular, seems designed to alarm, like other hockey-stick graphs. It may represent accurately same-house prices, and if so, is useful towards identifying specific events in the housing markets. It is still not useful at the macro-scale.

BillS December 22, 2008 at 10:57 am

That capital gains tax decrease inflated housing prices is completely spurious because some 95% of housing sales prior to the change in the law paid no cap gains because of the rollover provisions in the old law.

More complete analysis here:

spencer December 30, 2008 at 2:52 pm

I realize that I'm a week late in commenting but there are significant differences between the Shiller and the Census price indices shown by Henri. but essentially across the board the Shiller index is superior to the Census data.

The census data is an average price while the Shilller data is an index of same home sales. The census data simply collects the data on the average homes sold during a period and does not adjust for differences in the quality of the homes. The Shiller index is a different methodology designed to address this major problem with the index based on the Realtor data used by census. The Shiller index calculates the price change for the same home when it is resold and only when it is resold. This index is clearly a very superior approach to measuring the changes in home prices because it avoids many of the fundamental problems that frequently distort the census data.

It was developed by Case-Shiller because essentially every housing expert realized the existing home price indices were inadequate.

spencer December 30, 2008 at 2:59 pm

Russ, to accept your thesis that the run-up in home prices was only due to the four government actions I will require an explanation of why these four factors only generated soaring home prices in a limited number of cities. There were many urban areas that did not see an extreme run-up in home prices that were also influenced by these 4 government actions.

So why did government policies that impacted all urban regions only cause prices to rise in selected areas?

Previous post:

Next post: