Contact Us



NYT Roubini Article, IMF Transcript

The August 16 issue of the New York Times magazine included an article about Nouriel Roubini written by Stephen Mihm, who went on to co-author a book with Roubini. The magazine article keyed off of a September 2006 presentation that Roubini made at the IMF. Below is the transcript of that IMF talk, and the response by Anirvan Banerji, who the IMF had invited specifically to critique Roubini's presentation.

Contrary to the claim made in Mihm's article, the transcript shows Banerji did not characterize Roubini as either a “naysayer" or a “permabear."

However, the transcript does include specifics of Roubini's forecast that Banerji questioned: for example, that in September 2006 the U.S. economy was on the cusp of a new recession analogous to where the economy had been in early 2001. Roubini went on to say that "all commodity prices – including oil – should be falling," contradicting Mihm’s claim that Roubini had “warned” of “an oil shock.”


Washington, D.C., Thursday, September 7, 2006, Introduction by Charles Collyns

COLLYNS: Welcome here this afternoon. We certainly have an impressively large crowd, but it does not surprise me at all since I think we have a very hot topic and two very distinguished speakers to listen to this afternoon. This is a topic which is of great professional interest to us all, of course, but I think it is also of interest to us at a personal level since many of us at least have a substantial part of our personal portfolios exposed to risk in the Washington real estate market.


So we will be listening very closely to what our speakers have to say.

Our main speaker is Nouriel Roubini, who is I think very well known to the Fund's staff. I think his website and his blog are probably the most widely read economics websites among the Fund's staff, and probably outside the Fund's staff as well.

But Nouriel wears many hats. He is a professor at the New York University's Stern School of Business. He received his Ph.D. in economics from Harvard in 1988, and he has done important work on the effects of fiscal rules and on political cycles and the macroeconomy. But he has also been very active in policy circles. During the Clinton Administration, he worked at the Council of Economic Advisors and the U.S. Treasury. He wrote a book which many of us will have read at the time with Brad Setser called "Bailouts or Bail-Ins: Responding To The National Crises in Emerging Markets." Nouriel is also an entrepreneur. He has recently set up the Roubini Global Economics Group.

And most recently he has become notorious as perhaps the first and most prominent promulgator of the view that the U.S. economy is heading I think inevitably -- that may be putting his view too strongly, but maybe not -- towards recession. So I think we are going to listen very carefully to what Nouriel has to say.

We have an excellent commentator and discussant in Anirvan Banerji. Anirvan is someone who is an expert in possibly the most difficult task that economists face, calling turning points and business cycles. I think we are all quite good at projecting trends, but projecting a turn in the cycle is much more difficult. Anirvan has received a lot of attention for his feat of predicting the last two recessions, so I think he has a great deal of credibility in this area.

Anirvan has written a book, "Beating the Business Cycle: Can Turning Points in the Economy Be Predicted?" which was featured in an IMF Book Forum last year. He helped establish the Economic Cycle Research Institute in 1996, after a decade working at Columbia University with Geoffery Moore on the construction of leading indicators. He is a member of the OECD Expert Group on Leading Indicators, and he is the Forecast Chair of the Forecasters Club of New York. He was educated at the Indian Institute of Technology, the Indian Institute of Management, and Columbia University.


Talk by Nouriel Roubini

ROUBINI: Thanks very much for the kind introduction and thanks for also to Prakash Loungani for inviting me to this event.

Actually, I sent a PowerPoint presentation that was all text, not even a picture. It was nicely reformatted by a staff member here, [Dominique Raelison-Rajaobelina], who added this little chart on the top and side. They say a picture is worth a thousand words. This is just the perfect picture because it is essentially the way I see the United States.


So maybe I should not even give the talk, just the chart will be just enough.

The other comment I should say at the beginning is that I have been quite outspoken about the fact that I believe we're going to have a recession, but that in some sense clashes with scholarly evidence. Prakash Loungani has done some of the leading work on it, showing that forecasters are usually pretty bad, or lousy actually, in forecasting a recession. He wrote this article in 2001 that reached the following conclusion: "The record of failure to predict a recession is virtually unblemished."

So given that, I would say that I'm not a professional forecaster, so maybe I could get it right. That might be the way I justify it. The conjecture in Prakash's paper was that usually forecasters because of a series of professional biases tend to essentially be too optimistic and, therefore, when there's a big turning point of the business cycle they tend to miss it. So maybe not being a professional forecaster might allow me to end up saying something correct.

Summary of main points of the talk

I'll just summarize my main points here right at the beginning and then flesh them out in a little bit in more detail. These are essentially the five points I've been making for a while now.

One, there is a risk that the U.S. growth slowdown that we are observing right now this year is going to end up in recession by early next year. So we're going to have a hard landing rather than the soft landing that most people believe. The consensus is still for a soft landing.

Two, I will argue that the Fed is going to ease interest rates at some point this fall or winter as there are more signals about the recession, but this easing is not going to prevent the recession. The recession is going to occur regardless of the Fed easing.

Three, there is a view that even if the U.S. goes into a mild slowdown or a severe one, the rest of the world could decouple, with growth in Asia and Europe picking up while the U.S. may be slowing down. I believe for a number of reasons that the world is not going to decouple from a U.S. hard landing. As they say, when the U.S. sneezes, the rest of the world gets a cold. I think that the world still gets the cold when the U.S. slows down.

Four, my analysis has implications also for markets. Risky assets will underperform in the U.S. and the rest of the world, and not just the equity markets but also commodities and other ones.

Five, the final point is about the unsustainability of the U.S. current account deficit, the risk that essentially at some point foreigners become less willing to finance the U.S. The story I'm going to tell you today is one where shocks to the economy lead to consumer burnout, but they may also trigger foreigner flight from dollar assets.

So those are the five main points. I would say that the first one is the crucial one because if you believe the first one, then the other ones more or less logically derive from it.

Three bearish forces and consumer burnout in the U.S.

What are the main three bearish forces in the U.S. economy in my view that are going to trigger this recession -- first a severe slowdown and then a recession?

The first one is the housing sector bust. Calling it a slump or a soft landing at this point is too mild. When you look at all the indicators of housing -- I'm not going to list them all -- they are all sharply down very severely. Now there is also evidence that house rices are falling for the first time since the 1930s, and the futures prices also reflect now an expectation of falling prices.

The second point is that we have had in the last few years an energy price shock. As we know, this has potentially stagflationary effects on the economy. Oil price increases are a source of potential inflation, and being a tax on income have also negative effects on growth.

The third bearish factor is the delayed effects of the Fed's interest rate increases. As you know, the fed funds rate was brought from 1 percent all the way to 5.25 percent in the last 3 years, and even long rates have been going up in the last year. They have fallen a little bit in the last few weeks with the slowdown expectation, but they are significantly higher than they were. Last year the debate was about the bond conundrum or what drives it. Today now rates have gone up by about 100 basis points even on the long end of the yield curve.

So you have these three bearish forces, and I will discuss that similar forces were present also in previous recessions.

These bearish forces essentially in my view are bringing about consumer burnout. Not only are U.S. consumers hit by housing, oil, and interest rates shocks, but they has been also facing flat or falling real wages, and relatively mediocre employment growth in the last few months. U.S. consumer have had for a few years now zero savings on aggregate, so they are consuming more than their income. They have relatively high debt ratios, consumer credit, credit cards, home mortgages and so on. And now they also have high and rising debt servicing ratios as interest rates have been going up and there is repricing of their ARMs and so on. So it not surprising that the latest indicators of consumer confidence are sharply down. The consumer is in a relatively lousy or foul mood because these shocks are occurring in a situation where there are underlying vulnerabilities.

Comparison with past recessions

Now let's look at the past because I think it is interesting comparing the past with the present to guess what might happen next.

Start with the recession in 2001: what were the causes of it? First of all, the most important, you had the bursting of the tech stock bubble that had led to overinvestment in tech goods. There was a price bubble and it led to overinvestment and increased capacity, and then the investment bust followed.

Second we also had a mini oil shock. Oil was rising from low-teens to the high-teens. In 2000 was the beginning of the second Intifada between Israelis and Palestinians and other tensions in the Middle East. So there was something of an oil shock.

And then, third, the Fed tightened interest rates between 1999 and 2000 by 175 basis points with the fed funds rate going all the way to 6-1/2.

A crucial point here is that the peak of the business cycle we know now was in March 2001, 6 months before 9/11, so 9/11 certainly did not cause the recession. It might have worsened it for a little while, but actually the bottom of the recession was 2 months after 9/11, so 9/11 had a marginal effect and a temporary one.

Let's look at the recession prior to that, in 1990-91. There again you had had essentially a real estate bubble -- commercial real estate rather than housing -- and then a bust, leading then to the S&L banking crisis. That led to a credit crunch in the banking system.

You also had also monetary tightening by the Fed to counter the inflationary pressure that you had in the late-1980s. And you had also an oil shock that followed the Iraqi invasion of Kuwait.

Again, the peak of the business cycle was in June 1990 before the invasion, which occurred in August. But certainly that shock exacerbated the economic downturn.

Severity of the coming recession

I have been arguing not only that there is going to be a recession next year, but this recession is going to be more severe than the one we had in 2001, which was relatively mild -- only two quarters. But it was not mild in the impact on employment: for 2 years after the recession there were job losses. So it was a job loss recovery and then for a while it remained a jobless recovery; only after 2005 was the growth of employment being meaningful.

Why do I think that the coming recession is going to be more severe?

First, because I think the effects of the housing bust will be larger than those of the tech bust for reasons that I will tell you in a minute.

Second, oil is close to 70 [dollars a barrel] rather than 20. I will say why this is important.

And the Fed has been tightening up by more than 175 basis points it did prior to the last recession. Of course, the main caveat here is that what matters is not just the change, but also the level both in nominal and real terms. At that time, the peak was 6.5 percent, today for now it's 5.25 percent, and in real terms you could even argue that today's real rates are maybe lower than they were in 2000 given that inflation is slightly higher. So those are relevant caveats, but there has been a meaningful amount of monetary tightening. We do not have the liquidity that we had up to 2 years ago.

Let me say why oil matters today at 60 or 70, when it did not matter when it was going higher and higher, for instance in 2004-2005. We've seen oil go from 10 to 20, then to 30, then 60. People were worrying about a slowdown, and it did not happen, so why should we care about oil prices ? Maybe we can live with oil at 70 or even higher.

I think that the reasons why you did not have a slowdown then despite increasing oil prices were two-fold. One was that oil prices were high because of strong demand from global economic growth in India, China, U.S., and others. And, two, while there was an oil shock, monetary conditions were very easy. After all, up to 2 years ago interest rates were 1 percent on the short end, and housing was very bubbly.

Today, instead, oil prices are high in part because of supply kind of constraints; there is not much excess capacity. And there are these geopolitical shocks -- one day it is Iraq, the next day is Iran, the next day is Venezuela, the next day is Russia or Nigeria. So the situation looks pretty ugly in terms of supply.

So right now we have a triple whammy. In the past oil was high, but then you had low rates and a bubbly housing market. Today you have the housing bust, you have high oil prices, and you have rising interest rates all the way to 5.25 percent. That is why I think this time around, the oil shock matters more than it would in a different condition.

Housing bust vs. tech bust

Why do I believe that a housing bust is going to be more severe in terms of its macro implications than the tech bust?

The direct effects of a fall in residential investment I argue are going to be similar or larger than the fall in tech goods investment in 2000-2001. At the rates we are observing right now, you could have residential investment falling between this year and next year by about 2 percent of GDP. That is similar to the direct fall in investment in tech goods during the 2 years before the previous recession.

Of course, people say housing is only 6 percent of GDP while consumption is 70 percent. So a recession call has to be based not just on the direct effect from the housing bust but also the indirect effects. Let me mention three.

First, the indirect effects I think are important because the wealth effects of housing are larger because, unlike tech stocks, housing is a significant fraction of a household's wealth. In 2000 you had the big bubble in tech stocks. These stocks were held by a bunch of people in Silicon Valley, day traders and some households, but they were not widely held by the average household. But, as we know, about 50 percent of household wealth today is housing.

Second, the other point that is relevant here is that, savings have been negative, and the only way you can consume more than your income is that either you run down your assets or you increase your liabilities. There are not that many liquid assets to run down for U.S. households because half of wealth is locked into long-term savings plans, and the other half is essentially illiquid in housing. Therefore, the only way you can liquify your wealth is by using your home as your ATM machine, and that is exactly what has happened in the last few years. Last year we had almost $800 billion of home equity withdrawal or extraction. Of that, almost $200 billion went to consumption, another $100 billion went into home improvements, and the rest essentially to essentially financial balances -- reducing credit card debt and so on. It has been crucial to have home equity withdrawal as a way to sustain this excessive spending and consumption over income. If this effect is going to disappear because house prices are going to now flatten and then fall, then you are going to have a significant problem in sustaining consumption.

The third reason I think the housing bust is going to be more important than the tech bust is that the tech sector was not very labor-intensive. You had high-skilled, high value-added kinds of jobs in the tech sector, while employment growth in housing has been much more important. I have estimated that about 30 to 40 percent of the increase in employment since 2001 has been due to housing either directly or indirectly. So the fall of housing is going to have a much more meaningful effect on employment.

Why Fed easing cannot forestall a recession

Now even if all of my analysis so far is right, the question is whether we can actually predict when the recession will start. As Prakash Loungani showed, it is very hard, and lots of other evidence also suggests that getting the turns of the business cycle right is not easy. It is not easy to do it early on especially because we know that downturns are often very sudden and rapid.

Take 2000. The economy was growing like crazy, it was overheating up to the second quarter, 5 percent growth and the Fed was worried about inflation. And by Q4, the growth rate was down to zero, and by Q1 of 2001 we had the recession. So essentially the downturn was very rapid, unexpected, and it was led by the tech bust, together with the other two shocks I mentioned.

The interesting point here is, again, that the Fed missed completely both in predicting and reacting quickly to the recession because they did not expect the recession to occur. Even if you go back to FOMC minutes of September-November of 2000, essentially the Fed was still worried more about the inflation rather than about the slowdown in growth, and they still had in November a tightening bias. Of course, after that the Christmas sales came in at the end of 2000 and were a total disaster. The the markets opened after Christmas on January 2nd, the Nasdaq collapsed, the Fed went into a panic, and that same day -- in between FOMC meetings for the first time in years -- they decided to cut rates. So that is what happened. It took 6 months for the turnaround and the Fed completely blew it.

In the present situation, even if the Fed were to start easing to avoid a recession, in my view it is not going to prevent the recession. Why? Again, look at 2000. The Fed essentially stopped tightening 6 months before the recession started in June 2000, and then they very aggressively eased as soon as there was the signal of the start of the recession.

But why then was the Fed not able to prevent the recession? Because essentially you had a glut of capital goods. The Fed slashed rates from 6-1/2 all the way to 1 percent between 2001 and 2004, and real investment fell by 4 percent of GDP between 2000 and 2004. You had a huge glut of capital goods, and if you have that glut, you have to work it out -- interest rates effectively do not matter.

What is happening today instead is that instead of a glut of tech goods, we have a glut of housing stock, and also a glut of consumer durables. Why consumer durables? Because a lot of durables are essentially things like home appliances and furniture, which are highly correlated with housing. And auto sales that have been stimulated through tons of incentives, so there are way too many cars relative to need. So we have a glut both in the housing stock and of consumer durables. Even if the Fed were to ease -- and as I said, I think the Fed is going to ease in the fall or winter once there is a stronger signal of recession -- it is not going to prevent the recession for exactly the reason that Fed easing did not work in 2001. Whenever you have a glut of either capital goods, housing, or durables, demand for these goods becomes interest insensitive until the glut has been worked out. Essentially, in that situation, easing is like pushing on a string, it does not have any effect.

That is why I think at this point it is kind of too late, and what the Fed does is not going to make much of a difference. It might ease a little bit the slowdown or the recession, but it is not going to prevent it. If we are going to go to recession, it is going to happen regardless of what the Fed can do from this point on.

The same mistakes were made in the past. When the Fed started easing in 2001, for a while everybody started saying that we were going to have a soft landing, that we were going to prevent a recession. Even in March 2001, 95 percent of economists were still predicting no recession in 2001. Guess what? The recession did start in March 2001. So the hope that the Fed is going to come to the rescue did not work then, and I do not think it is going to work this time around.

Can other components of aggregate demand pick up the slack?

If you look at the components of aggregate demand, I see that the signals are exactly the ones that I consider worrisome. We have the Q2 GDP numbers so we know what is happening to aggregate demand.

Residential investment is falling at a 9.8 annualized rate, and I think that everybody now believes that it is going to fall at a much higher rate in the next few quarters. My estimate is it could fall for the next two or three few quarters at an annualized rate of 20 percent or more. That is the first observation. So you have a bust in housing. That is pretty clear at this point. I think that those who were arguing that there was a soft landing in housing have been proven wrong.

Second, when you look at the components of consumption, you already see that durable consumption is not growing at all. Essentially it is flat. That again is correlated with the housing bust. It is in autos, in furniture, in home appliances. That is the direct link.

Third, inventories are up because sales growth has been slowing and production has not slowed yet. You have the typical inventory build-up. And actually the revision of the GDP number from 2.5 to 2.9 was exactly due in part to this inventory adjustment. If you look at sales growth as opposed to GDP growth, sales growth is now down to 2.3 percent, not 2.9 percent. So I think that again is a bearish thing. Once the adjustment of inventories starts to occur, then you have to cut production, and we are going to see it the second half of this year.

Fourth, net exports are not anymore a drag on growth as they have been for the last few years, but they are not much of a stimulus either. Nobody really believes, given what has happened to the dollar, that you are going to have a radical improvement for now in the trade balance. So it is not going to be a drag and it is not going to be much of a stimulus. And my view is actually that the improvement of the real trade balance may be a signal again of an economic slowdown. We know usually when the economy booms, the current account tends to worsen, while in recession it tends to improve. The fact that now imports are sluggish, close to zero growth, is a signal that the slowdown is coming. So for me it is a signal that actually things are worse.

Government consumption is weak or almost flat. There is not much action from that. You cannot expect much demand drive from there.

Will nonresidential investment pick up the slack as housing consumption slows? That is the argument many have made. But if you look at the numbers on Q-2, real investment in equipment and software fell. It is true that firms today are highly profitable. The profit share has been unprecedentedly high, and they are flush with cash. But there is something of an investment strike because essentially firms have tons of profits, tons of cash, but they are looking at demand ahead, from construction and consumption and so on, and they see the glut. They see that essentially there is not much demand growth. If anything, there is a slowdown, so why increase capacity in a situation you expect a slowdown of the economy? Of course, if there is no reason to invest more, what is happening is that instead of investing this money, they are giving back to their shareholders. You have had in the last year or so the biggest share buy-back bonanza in U.S. history. We are speaking about hundreds of billions of dollars. If firms have any opportunity of doing productive investment rather than returning money to shareholders, they are going to use that cash for something productive. That is not happening. I think it is a signal of the concerns of the corporate sector when they look ahead.

Informal and formal indicators of recession risk

My analysis has been based on circumstantial kind of observations. I am not a professional forecaster and I do not use a big global macro model. Even then I said the probability of a recession is "70 percent." If you ask me where I got that number: just out of my nose, I will be very honest about that. I think if you had said "50 percent" you look like a wimp, it means you are not sure. So if you have the guts of believing there is going to be a recession, you should say something higher than that, and that is where the "70 percent" comes from.

So my model is like a "smell test" or a "duck test," if it looks like a recession and walks like a recession and quacks like a recession, it should be a recession. Or we can think of it as being the famous "obscenity test," I'm referring to the Supreme Court Justice who said "I cannot define obscenity or pornography, but I know it when I see it." So I see a recession that is based on this analysis and based on data and historical evidence.

But of course there are more formal indicators of a recession risk. Some people have been looking recently at the yield curve inversion, Jonathan Wright at the Fed for instance. If you plug the numbers into the yield curve equation, you get a 44 percent probability of a recession. There are a number of leading indicators that are showing weakening, even if they do not yet show a red flag but only a yellow flag.

Then there are more sophisticated leading indexes such as those that Anirvan and his colleagues have developed at ECRI, and they are also showing some weaknesses, but he is going to say more about what those models say. So certainly the formal indicators are showing an increasing probability of a slowdown, not yet a full high probability of a recession.

Let me make a related point here about what guesses we can make from what the Fed has been doing recently. They have paused, and their argument has been that while inflation is a little above their comfort target range of 1 to 2 percent and it is going to increase for a while, it is then going to fall to the middle of the range. But if you look at the macro models that have been used both by the Fed and other people, say Larry Meyer at MacroAdvisers, based on an augmented sophisticated Phillips Curve approach or whatever, there is no way you can get such a sharp fall in inflation down to the middle of the range unless unemployment the rate goes sharply up, closer to 6 percent, and that means essentially a recession. So my reading of it is that either the Fed does not believe in its forecast that inflation will fall -- but that would be reckless, and I don't believe Bernanke or the rest of the FOMC is an inflation dove -- or that they truly believe in their inflation forecast but that they might be assuming a bigger slowdown in the economy than they are officially saying in public.

Essentially, any kind of basic macro model will tell you you cannot have it both ways. Either you believe that inflation is going to fall because you have a recession or a hard landing call, or you have to bump up your inflation forecast. It does not make sense to believe that you are going to have a soft landing with growth closer to 3 percent and core inflation going back to 1-1/2 percent. So I think there either is a bit of a contradiction there or the Fed is more worried about the extent of the slowdown than it is saying in public. Now because I do believe we are going to have a nasty recession, I am today less worried than other people, maybe including the Fund, about the inflation risk for the United States. I believe that during a recession commodity prices are going to collapse, you are going to have a huge slack in the labor market, and slowdown in wages -- all the standard reasons why in spite of high oil prices, we are not going to have higher inflation, barring, of course, a further oil shock for geopolitical reasons. So barring that, all commodity prices -- including oil -- should be falling during a slowdown should be falling and, therefore, I worry less about inflation than others.

Soft landing

Of course, there are a whole series of arguments in favor of the soft landing that you hear every day. The market consensus is still that there will be a soft landing.

The first argument is that everybody who has been batting against the U.S. consumer has been proven wrong over and over again. The consumer has always been proven to be much more resilient. When push comes to shove, they borrow more. If they have to do something, they do not cut back on consumption, and they can keep on piling up credit card debt.

Some people say oil prices now are falling below 70 dollars a barrel, and that is good news. Some people say that the wealth effect of the housing slump will be modest. They say there are still several trillion dollars of untapped home equity that people can borrow against. Some people suggest that income generation is still quite robust both through jobs and through real wages.

Another argument that has been made is that monetary and credit conditions are not that tight, and long rates actually in the last few weeks have been falling. So there is some evidence of something of an easing.

And, finally, people have said that history shows that you need a credit crunch to get a recession, and right now we have monetary tightening, but we do not have a credit crunch.

I would like to elaborate on this last argument. The other argument we may go back to them later, but I think that one is important.

The question I want to ask is: Do you need a credit or monetary crunch to get a recession? My answer is: Not necessarily. The first observation I make here is some tightening has already occurred, 4.25 percent on the short hand, and 100 basis points on the long end. So we are not anymore in that slush of liquidity. Secondly, it is true when you look at surveys, credit supply conditions have not tightened. You do not see any signal of a real tightening of standards and lending by the banks for housing or whatever. But we see certainly that credit demand, especially in housing, is starting to fall, so it is kind of like a demand-driven fall in credit, not a credit crunch.

I think the crucial point here is that you do not need really a credit tightening in order to get a bust of the economy. In 2000, the bust of the tech sector occurred without a credit crunch. Right? It was not because the Fed hiked the rate to 6-1/2, it is just that there was a bubble and at some point people realized their pet dot-coms were worth nothing. And when they decided that it was a bubble, it collapsed; the return on that investment turned out to be much lower than previously expected and so we got an investment bust. So once the bubble burst, there was a period of underinvestment to work out the excess capacity. This can occur completely without any credit crunch or monetary tightening. Of course, some monetary tightening might be the tipping point essentially to prick the bubble, but you do not really need a severe credit crunch to get this kind of cycle. Whenever you have a bust that comes from a bubble that goes wrong, you can have a significant fall in real investment regardless of interest rates.

Today I think the same thing is essentially happening with housing. Once the bursting of the housing bubble occurs, you are going to have a fall in prices, a fall in demand for credit. Of course, at that point -- after the crunch has occurred in terms of real economic activity in the housing sector falling -- you are going to see some credit tightening because you are going to see increasing amounts of delinquencies, defaults, and foreclosures. The same thing happened with the S&L crisis. First you have the bubble, then the bust, all these properties went belly-up, and at that point then you got the credit crunch because the Fed decided to crack down and said "enough of this." I would argue that actually the regulators in the U.S. have been relative lax for the last few years. They have allowed all these reckless things like zero interest rate mortgages or negative amortization -- things that have been crazy, and they have done very little or nothing about that. So it's the same story as in the 1980s.

So essentially the point here is that the bursting of the commercial real estate bubble of the 1980s led to a bust, and then to eventually a credit crunch and the S&L crisis that contributed to the 1990 recession. And my concern today is that the bursting of the housing bubble -- we have not seen it yet -- is going to lead to broader systemic banking problems. It is going to start with the subprime lenders -- they are already in trouble because of increases in delinquencies and foreclosures -- and then it is going to be transmitted to other banks and financial institutions all over the country.

Financial implications of the housing bust

That leads me to the question: what are the systemic financial and banking risks of the housing bust? This is something that people only now are starting to think about. And one of the important issues here is, where is the housing mortgage risk concentrated given there is a huge amount of mortgage risk right now, given the trillions of dollar of mortgage debt?

I think that in part it is still the banking system that is directly or indirectly holding this risk; directly because a good chunk of the mortgages are still on their books, and indirectly because they dumped some of that mortgage risk and they got in an exchange mortgage-backed securities (MBSs) and effectively through the MBSs they are holding part of the risk. Part of it of course was distributed as a risk to asset managers and to hedge funds that are holding tons of these mortgage-backed securities. So those guys could get in trouble, but if those guys get in trouble, then the counterparties of those hedge funds are highly leveraged institutions, meaning prime brokers or investment banks -- they could also get in trouble. You could not rule out some systemic effects if one of these big highly leveraged institutions goes belly-up.

And of course a good chunk of that credit risk is now in the hands of the government-sponsored enterprises (GSEs), Fannie Mae and Freddie Mac, that are holding really trillions of dollars of mortgage credit risk. In addition to that of course they are also holding the market risk of changes in interest rates. Some of this agency debt over the last few years, a lot of it has been bought also by foreign central banks, so some of the shocks could be transmitted to the rest of the world, but most of it is still domestically held, so it is somewhere in the system, and I do not think it is very diffused.

So there is a meaningful risk now that if you have a real housing bust that is not just a sectoral shock but leads to a economy-wide recession, then that housing slump could lead then to a systemic problem for the financial system. Again, in the 1980s people said the S&L was just a regional shock. Right? It was just Texas and so on. But it became then one of the factors that led to the nationwide recession and it led to a nasty S&L crisis that cost the U.S. taxpayers $200 billion.

The premium that the GSEs impose for providing the guarantees are based on historical experience in which you had always regional shocks. One was in Texas, the other in California, another time it was in Boston or New York. A 20 basis point premium is okay based on default rates that are based on regional shocks. This time around I think we are going to have a national rather than just a sectoral or regional kind of housing slump. It is not just the two coasts, prices are falling even in the Midwest. So the risks actually from a systemic point of view I think are just much larger than those that you would have in the case of the 1980s, of course conditioned on having a recession.

The other point I think that is crucial here is the ability of macro policies to prevent a recession this time around are going to be much more limited than in 2001. Remember what happened in 2001: you had very aggressive Fed easing and also by other central banks. The Fed got aggressive, cutting rates from 6-1/2 to 1 percent, and you got aggressive -- and I would even say reckless -- fiscal easing from a 2-1/2 percent GDP surplus in 2000 to a 3-1/2 deficit in 2004. And of course the dollar sharply fell from 2002 to 2004 and that also helped. At that time, you remember that the Fund's chief economist Rogoff said that what the U.S. had in terms of macro policy was the best recovery money can buy -- he said it sarcastically, and I think he was absolutely correct.

The unfortunate thing is that today you cannot do the same trick. Monetary easing is going to be limited by the inflationary pressures that are still there. I do not think they can cut interest rates all the way down to 1 given where we are in terms of inflation -- maybe 200 basis points is all the Fed can cut and that is not going to make much of a difference. Fiscal easing is limited by the fact that we have a huge now structural budget deficit. It is improving on a cyclical basis, but it is going to worsen once we go to a recession, and having an even bigger deficit is going to be even more reckless, so you cannot do what you did in 2001.

And the fall of the dollar I think is limited because the dollar has already fallen sharply relative to the floater currencies -- the pound, the euro, Canadian dollar. If China does not move its exchange rate for whatever reason, then the rest of Asia is not going to move and, therefore, there are limits to how much the dollar can fall under those conditions.

Global outlook: can the world decouple from the U.S.?

Let me end by why I think that the U.S. recession is going to lead to a severe slowdown the rest of the world. I do not expect a global recession, I think it is going to be a slowdown. Again, the consensus view today is that of a "decoupling" of world growth from U.S. growth. People say there is enough momentum in domestic demand in Asia -- in Japan, Korea, China, and India -- that even if the U.S. goes into a slowdown that is severe, those countries are going to decouple. People also point to Europe -- where right now growth is better expected -- and say its resilience is based not just on net exports but on domestic demand. Therefore, the world is going to decouple from the U.S.

I am very skeptical about this for the following reasons:

First of all, the weakening dollar is going to impart deflationary effects to the rest of the world. Remember what happened in 2004-2005, the euro went from 120 to 136, and growth in 2005 in the euro area slumped to only 1.2 percent, so you had a significant effect.

Second observation, trade links with the U.S. are still very important, both the direct ones and indirect ones. Some countries are directly relying on trade with the U.S. -- China, Mexico, Canada, and so on -- and some of them indirectly. For instance, some people say Latin America is going to do fine in spite of a U.S. slowdown because they do not export too many goods to the U.S. directly but mostly export commodities to China. But if you have a severe U.S. recession, for sure China is going to be negatively affected. And if there is a slowdown that is severe in China, their demand for commodities is going to fall and, therefore, Latin America is going to be hurt. So those are important indirect trade challenges that are also meaningful. Or take East Asia, which now maybe does not sell as much directly to the U.S. but instead mostly exports components to China -- again, it is going to be indirectly affected by a U.S. and Chinese slowdown. When you have the consumer of first and last resort and the producer of first and last resort having a severe slowdown, meaning U.S. and China, you are going to have trouble for the rest of the world.

Third observation, the oil shock is as much of a shock to E.U., Japan and Asia as is it to the U.S. because they depend on oil imports as much if not more than the U.S.

Another observation: Monetary policies are being tightened everywhere. ECB and BOJ have just barely started, and certainly, after the turmoil in May and June, even emerging markets now are tightening. So that slush of liquidity that you had a few years ago is shrinking significantly. And, as I said, there is a limited role of macro policies that counter the slowdown. In 2001, it was not just the Fed slashed, but BOJ went all the way to zero and even the ECB went, although too slow and too late, to 2 percent. And everybody had the fiscal easing -- massive in the U.S., deficit of 10 percent in Japan, and even in Europe everybody broke their growth and stability limits. This time around, monetary policy has to be tightened and everybody is tightening because of inflation. And given the fiscal mess in all of the G-7, with maybe the exception of Canada and the U.K., you cannot have much fiscal easing. And exchange rate changes are a zero-sum game -- the dollar could fall a lot relative to the floaters, but that is negative for the rest of the world.

Another couple of points. One, I think that while the housing bubble is perhaps more severe in the U.S. than in other parts of the world, we know from work that many have done at the Fed and the IMF and elsewhere that there are housing bubbles in Europe and other parts of the world. You can be in a situation in which bursting of housing bubbles lead to a slowdown in other parts of the world. Two, a number of measures of business and consumer confidence in other countries are also heading south. I know that the signals coming from Europe and Japan are mixed, you can see optimistic ones, but I think that you can make a very good argument actually that the legs of the recovery in these two regions are relatively fragile, and that essentially if you had a severe U.S. slowdown these countries are not as strong in terms of domestic demand as people make them out to be.

Finally, emerging markets. Of course they did lots of reform, they cleaned up their act -- macro, structural, and financial -- after the crisis of the last decade, we know it all. But they also got very lucky, let's admit it. Right? In a world economy where you had high global growth -- 5 percent expected this year -- high commodity prices, and very low interest rates until 2 years ago, it is hardly been for lousy credit to go belly-up. I mean, take a country like Ecuador, even with oil prices at 70, they have changed presidents seven times out of eight in the last 8 years. So what would happen if oil was at 40 or 30 as opposed to 70? So a number of these countries I think have really benefited from this global condition being as ideal as possible, and in a world in which growth is slowing down and commodity prices are falling, and now interest is going up, is going to get rough, and you all know better than I which are the most vulnerable among them. Of course, those that have weaker marco and external vulnerabilities are going to be more under severe pressure of financial distress if not outright crisis than those that have better fundamentals.


So to conclude: my view is that the risk of a hard landing is very high for the U.S. economy. I see essentially a recession coming by next year. I give it a very high likelihood. I argue that housing today, like the tech bust in 2000-2001 will have a macro effect; it is not going to be just a sectoral effect. I argue that U.S. consumers are now close to a "tipping over" point given all the vulnerabilities I have discussed. I argue that the Fed easing will occur, so the next move is going to be a cut, but it is not going to prevent a recession. And, finally, I argue that the rest of the world is not going to be able to decouple from the U.S. even if it is not going to experience an outright recession like the United States. So on that cheerful note, I will stop. Thanks.


COLLYNS: Thanks very much, Nouriel. I think perhaps we will need a stiff drink after that,


but maybe Anirvan you can cheer us up a little bit.

BANERJI: First of all, it is a privilege to be asked to be a discussant for this talk by Professor Roubini. I can tell you that you do not win popularity contests by forecasting recessions, so if you take the plunge, you better have the courage of your convictions.

Over the past couple of months, Nouriel Roubini has been a clear voice predicting a recession, one that he believes will be nastier, deeper, and more protracted than the 2001 recession. And he has performed a signal role in questioning a complacent consensus, and I think that is very good that he has. He certainly has the courage of his convictions. The question is: is he right?

We know that economists worldwide have a rather dismal record of predicting recessions. Nouriel follows an approach based not on a specific model, but on more of a neglected one, which is to draw analogies between today's situation and those in previous recessionary episodes.

I actually agree with many of Nouriel's fundamental premises, but I have four concerns.

First, back in July, I believe Nouriel saw a 50-percent probability for recession based on the lagged effect of interest rate hikes, high oil prices, and the housing downturn. This was partly because he expected the Fed to raise rates all the way to 6 percent before they were done. A month ago, following soft GDP and employment numbers just before the Fed announced a pause in the rate hike cycle, Nouriel raised his recession odds to 70 percent. It is interesting that he still sees the recession probability remaining high even though, at the margin, oil prices have dropped and the Fed is probably done, contrary to his early expectations. In fact, with crude oil inventories of a 20-year high, oil prices have dropped about $11, about 15 percent from the high. Logically then, at the margin, the recent housing numbers must be so much worse than his early expectations that they overwhelm the effects of the pull-backs in oil prices and interest rate expectations. It would be interesting to know what it would take for him to change his recession call. That is my first question.

Now for my second issue. Nouriel's basic premise is that in 2000 the recession was triggered by the combination of an oil shock, a rate hike cycle, and the tech bust, and this time around it looks similar, except that instead of a tech bust, we have a housing bust whose impact would be arguable worse. In fact, this is the analogy used by those who see the similarities to 2001. But many of those who expect to avoid a recession see the soft landing of the mid-1990s as the better comparison. So how does one decide which is the better analogy? That is not immediately obvious because we can catalog all the similarities and the differences, and the bulls and the bears would still differ about the weight to be assigned to each factor. Forecasting by analogy is tempting, but it is hard to choose the right analogy.

In my experience, forecasters often assume that because the present resembles earlier periods in some way, it does so in other ways, and then predict that the current period will exhibit similar patterns. If the chosen analogy does not work, they look for other analogies, other periods this might better resemble, and if none of the analogies work, they proclaim a new paradigm, as in the late-1990s. The danger of such a subjective approach is that instead of letting the objective facts shape your views, you may be tempted to selectively emphasize the facts that support your views. The only remedy is to have an objective, stable framework rather than what Nouriel has referred to in earlier conversations with me as a nonrigorous framework.

In any case, there are many factors that are different in every single episode. Just to pick on one, 10-year Treasury yields, which have a more direct impact on the overall economy than the Fed funds rate, are quite a bit lower today, around 4-3/4 percent, than they were in the lead-up to the 2001 recession, and that has been true pretty much for a while. Does that offset the much worse housing picture than in 2000-2001? It is not clear.

My third point has to do with Nouriel's own track record of forecasting turning points which has to do with the difficulty of picking out in advance which shocks are recessionary and which are not. In 2005 we had already had a series of rate hikes and an oil shock that was already much larger than in 2000. A number of mainstream economists predicted a recession especially after Hurricane Katrina, but what followed was a revival in growth. Nouriel had predicted a slowdown in 2005, and on September 2nd last year he was quoted by Reuters as saying, and I quote, "This is a very delicate moment. The economy is already very imbalanced. On top of that, we have had a massive oil shock and now we have a natural disaster that might be something of a tipping point." He was hardly alone.

The chief economist of the world's largest forecasting company also said that Katrina could be a tipping point, and the markets agreed, pushing 10-year Treasury yields down to 3.98 percent in the expectation that the Fed would not hike rates in September 2005 given the supposed recession threat.

I remember writing that same day that the recovery was likely to remain resilient. Contrary to popular wisdom, the decisive issue is not the level of oil prices, but of the resilience of the economy, as revealed by our leading indexes which are nowhere, still nowhere near recession territory, so the economy is not likely to fall apart.

Nouriel has explained that the economy did not take the predicted hit partly because the oil shock was demand-driven based on the global expansion, but that could also be said of the 2000 oil shock which did help trigger a recession. Actually, 2000 was another period of synchronous global expansion. I do not quite agree that it was the Intifada that raised the oil prices that much. And incidentally, in 2000, the oil prices did not go to the high teens, they went to $37 a barrel, Nouriel. So it was quite an oil shock, from about $10 or so in early 1998, to $37 in 2000.

Economists have sounded other false alarms historically. Back in the mid-1990s, several prominent economists predicted a recession. Sure, we had the Mexican peso crisis and the bankruptcy of Orange County, California, linked to what became at the time the worst year in history of the bond market, but we still avoided a recession. In the fall of 1998 in the wake of the Russian default and the LTCM crisis, many predicted a recession. As Nouriel might recall, President Clinton himself called it the worst financial crisis in 50 years, and "Time" magazine had Alan Greenspan, Robert Rubin, and Larry Summers on its cover as the "Committee to Save the World". But, once again, a recession was averted. Using our retrospectoscopes we can explain why there was or was not a recession in each case, why some shocks were that potent and not others, but that is not so easy before the fact. How do you pick out in advance which shocks are recessionary? That is my third point.

Over the last three generations that our group has been studying business cycles, there has been a wide variation in the combination of factors that has helped trigger each recession, from the defense spending cutback after the Korean War that helped set off the 1953-1954 recession, to the Penn Central bankruptcy and the 2-month General Motors strike that contributed to the 1969-1970 recession. Over the decades we have also established recession dates for 20 countries using an approach comparable to that used to determine the U.S. recession dates.

Our conclusion from all this research is that there is a wide variation in factors that trigger recessions, which are all different in their specific details, making it critical to monitor a wide variety of indicators. Still, it is hard to predict a recession well in advance without a fair amount of luck.

While we at ECRI were lucky enough to forecast the 1990 and 2001 recessions 5 to 6 months in advance, I always recall the words of my late mentor, Geoffrey H. Moore, who told me that if you could predict a recession just as it was beginning, you were doing very well as a forecaster.

To be clear, I actually agree with many of Nouriel's key premises. For example, I think he is right, and many others are in denial, about the severity of the housing bust. Parenthetically though, if there is a presumption here that home prices will come all the way down to where they were before this boom began, that may be not quite right, because not this entire rise is cyclical, some of it is actually structural because of zoning restrictions that have proliferated over the last many years, and we are not going to go back to the levels we saw 10 or 15 years ago. So as far as I can see, Nouriel is right, the prices are going to keep going down, but it may not be as far as some people think.

I think Nouriel is also right that investment spending is not going to come to the economy's rescue. And I think he is right, though most do not yet realize it, that this is not just a U.S. slowdown that we are facing, but a global one, and that there is unlikely to be a decoupling between the U.S. and foreign economies as some have forecast. The question is whether these factors and the others Nouriel has mentioned will be enough to trigger a recession.

In 2005, we at ECRI asserted that the economy would be resilient even to shocks as big as Katrina. In recent months, we have stopped saying that because there has been a material increase in the risks to the economy. At the beginning of 2006, we started predicting a U.S. slowdown, but regardless of our subjective feelings, our objective leading indexes are just not ready to predict a recession the way they did in 1990 and 2001.

Which brings me to my fourth and final question: what if Nouriel turns out to be right? Given his eclectic approach, how would he be able to repeat his performance the next time around? And if he is wrong, if there is a 30-percent probability of no recession that pans out, what lessons do we take away from this exercise in recession forecasting?

Let's not forget that the housing bubble grew so large partly because of the Fed's risk management approach which led them to cut rates to record lows in 2003. Recall, if you will, then Governor Bernanke's warning of deflation risks in mid-2003, and as a top Fed official explained to me at the time, they thought there was a small but significant risk of a double-dipped recession. We did not agree. In fact, right after the June 2003 rate cut, GDP growth in the third quarter of 2003 soared to a two-decade high, surely not the lagged effect of that rate cut.

My point is that it is risky to adopt a risk management approach for decision-making if your assessment of risk is not accurate. You may end up just kicking the can down the road, or worse, create a bigger problem for the future, as we seem to have now.

So the question is, what is a policy maker to do with a 50-percent, or even a 70-percent recession probability? How do they know these numbers are accurate? If they are, should they stimulate the economy aggressively, taking a 50-percent or a 30-percent change of being wrong?

In conclusion, the economy is such a complex system that relatively small influences can shift the balance between a slowdown and a recession. Often times, while the most salient factors take the blame for a recession, it is the mix of more subtle shifts that create the decisive difference, making the economy either vulnerable or resilient to a shock. Nouriel mentioned how nonlinear relationships -- he was talking about how sharply in 2000 we saw growth tumble and how nonlinear relationships in the economy actually relate to that kind of plunge in growth -- can make downturns sudden and unexpected as in 2000-2001. But such nonlinearities also imply that at times, subtle undulations in the economic landscape can make the difference between hurtling towards a recessionary abyss and veering away from it. That is why an apparently plausible scenario based on the most salient factors may not always be probable. That distinction can spell the difference between a prescient and an erroneous recession forecast. You see, highly nonlinear systems can be highly sensitive to small changes in obscure indicators that can set the conditions for a cascading collapse. The trick is to figure out when such a shift appears on the radar screen. It is only when an array of leading indexes indicates that this threshold, this tipping point, will be passed that we find it possible to make a reliable recession forecast without a significant danger of a false alarm. According to the leading indexes we monitor, we are not there yet.

To sum up, I agree with many of Nouriel's premises, and he is right to predict slowing growth. But first, it is not clear to me what data would make him change his recession forecast. Second, I am concerned that his approach amounts to subjective forecasting by selective analogy to past episodes that favor his bearish views. Third, his recent track record of turning-point forecasting does not inspire confidence, and I am not persuaded that part of his explanation for the error relating to the demand-driven oil shock in 2005 is okay because that was just what helped trigger the 2001 recession. Fourth, recession probability guesstimates like 50 percent or 70 percent are neither falsifiable nor actionable. In fact, action could be counterproductive. But my fundamental problem with his approach is that as we have seen repeatedly in the highly complex system that is the U.S. economy, piling up a list of factors to make up a seemingly plausible recession scenario may not provide much insight into the actual likelihood of recession. Thank you.


COLLYNS: Nouriel and Anirvan, thank you both very, very much, that was an excellent discussion. I think anyone who came into the room with any cobwebs of complacency would have had those cobwebs blasted away. It was good for us to hear you, and certainly I think it was a lot for us to think about. Thank you very much indeed.

* * * * *