Plato would not be impressed

Lots of blogs (MR, Yglesias, Drudge) have been linking to another bearish article on China.  I find these articles to be very annoying, but not because I disagree with their specific predictions.  Rather I don’t see the point in making these sorts of predictions.  For example:

May 3 (Bloomberg) — Investor Marc Faber said China’s economy will slow and possibly “crash” within a year as declines in stock and commodity prices signal the nation’s property bubble is set to burst.

The Shanghai Composite Index has failed to regain its 2009 high while industrial commodities and shares of Australian resource exporters are acting “heavy,” Faber said. The opening of the World Expo in Shanghai last week is “not a particularly good omen,” he said, citing a property bust and depression that followed the 1873 World Exhibition in Vienna.

“The market is telling you that something is not quite right,” Faber, the publisher of the Gloom, Boom & Doom report, said in a Bloomberg Television interview in Hong Kong today. “The Chinese economy is going to slow down regardless. It is more likely that we will even have a crash sometime in the next nine to 12 months.”

An indextracking Chinese stocks traded in Hong Kong dropped 1.8 percent today, the most in two weeks, after the central bank raised reserve requirements for the third time this year. The Shanghai Composite has slumped 12 percent this year, Asia’s worst performer, as policy makers seek to rein in a lending boom that’s spurred record gains in property prices.

What are we to make of this?  I’m all for inferring predictions from asset prices, so I do concede that the modest fall in Chinese stocks is interesting.  But what are the investment implications of that?  If you are using asset prices to predict economic growth, that’s fine.  But it makes no sense to use asset prices to make predictions that you imply will be useful to investors.

And what’s with the Vienna World’s Fair of 1873?  I presume that’s a joke, but since I don’t recognize humor without smiley faces attached, I’ll proceed as if he was serious.  The US also had a depression in 1873; was that also caused by the Vienna World’s Fair?  And how about the strong economic growth that occurred in the US following the 1933 and 1964 World’s Fairs?

Faber joins hedge fund manager Jim Chanos and Harvard University’sKenneth Rogoff in warning of a crash in China.

China is “on a treadmill to hell” because it’s hooked on property development for driving growth, Chanos said in an interview last month. As much as 60 percent of the country’s gross domestic product relies on construction, he said.

What exactly do they mean by crash?  And what is the theory that links high levels of construction with crashes?  But here’s my biggest problem with the Chanos prediction; it seems to be based on false information.  I wasn’t able to find current construction data for China, but the construction sector was only 7% of Chinese GDP in 2003.  I’d guess it is about 10% today.  Services are currently 40%, farming is 11%, and I’d guess manufacturing, mining, and utilities are close to 40%.  I’d say if you are predicting a crash based on construction being 60% of GDP, and it’s actually only about 10%, that’s a big problem.

Rogoff said in February a debt-fueled bubble in China may trigger a regional recession within a decade.

In its entire 230 year history, the US has never once gone 10 years without a recession.  So I’m willing to concede that China “may” experience a recession within 10 years.  And since the other East Asian economies are heavily dependent on China, and Chinese recession would certainly depress all of East Asia (just as American recessions impact Canada and Mexico.)

Would the Chinese recession have been caused by the debt-fueled bubble?  Well I guarantee that if a recession does occur, everyone will assume it was caused by a debt-fueled bubble.  But how can you tell?  I just read an NBER article that discussed the market for real estate bonds in the interwar period.  These bonds fueled lots of skyscraper construction in NYC and Chicago during the 1920s.  But unlike the authors, I don’t see much evidence that this was a bubble, or that it in any way contributed to the Great Depression, or that lenders were foolish in making those loans.  Instead, I think the Great Depression was caused by NGDP falling from $106.7 billion in 1929:3 to $49.8 billion in 1933:3.  When NGDP falls in half, it becomes very difficult to service debt on brand new skyscrapers built under the assumption that NGDP would not collapse.  I just don’t see irrationality here, nor do I see causation running from debt bubbles to depression.  But I am in the minority regarding the Great Depression; indeed I recall that Rogoff also thinks the Depression was caused by a financial crisis.  If China experiences a recession I will again in be in the tiny minority who think debt has little to do with the problem.

Citigroup Inc. warned in March that in a “worst case scenario,” the non-performing loans of local-government investment vehicles, used to channel money to stimulus projects, could swell to 2.4 trillion yuan by 2011.

I wonder how the average reader interprets that sentence.  My first thought was:  If only the US and Europe could have that sort of “worst case scenario.”  Note that 2.4 trillion yuan is about $350 billion, chump change for the Chinese government.

The government has banned loans for third homes and raised mortgage rates and down-payment requirements for second-home purchases. Prices rose 11.7 percent across 70 cities in March from a year earlier, the most since data began in 2005.

Pretty scary eh?  Except that nominal incomes are growing even faster, so the price of homes relative to Chinese incomes is actually falling.

Housing prices nationwide may fall as much as 20 percent in the second half of the year on government measures to curb speculation, BNP Paribas said April 23.

Obviously this might well happen, but what justifies this prediction?  Earlier we were told that falling stock prices signal economic weakness ahead.  Then we were told housing prices were rising rapidly.  Asset prices should incorporate publicly known information about the fundamentals of the housing market.  Using the logic of Marc Faber, doesn’t this suggest good things lie ahead for the housing markets?   Yes, we know from the US experience that markets are sometimes wrong.  But then why use the stock market to forecast the economy, but ignore the information contained in housing prices?

I believe that Plato said knowledge was justified true belief.  I just don’t see any justification for the predictions made in this article.  If people are making predictions based on the 1873 Vienna World’s Fair, or the assumption that construction is 60% of the Chinese economy, then I don’t particularly care whether the predictions come true.  Here is an analogy.  You are talking to a drunk in a bar.  He suddenly says “The huge Chinese trade deficit will cause the Chinese stock market to crash in 2011.”  Are you going to file that prediction away and check later to see whether it was accurate or not?  Or would you think “That guy doesn’t even know that China has a trade surplus, not a deficit.”

I’m not impressed by predictions that assume the EMH is wrong.  But I am willing to keep an open mind on the EMH.  I am especially unimpressed with predictions that assume the EMH is wrong and that are based on false or misleading economic statistics.  If those predictions turn out to be accurate, it would not make me think more highly of the person who made the prediction, I’d just assume he got lucky.

I understand that predictions are fun, and it’s only human to want to anticipate what will happen next.  I’m all for making conditional predictions based on various public policy options.  And I’m all for drawing inferences regarding the implied predictions embedded in asset prices.  But the sort of unconditional predictions discussed in the Bloomberg article should be placed in the astrology section of the newspaper.  Fun to talk about, but not to be taken seriously.

Most economists obviously disagree with me.  The top economists are constantly making predictions that violate the EMH.  OK, then why not set up a new journal, the Journal of Economic Predictions.  Each month they would publish 25 to 30 short predictions about the economy or asset prices.  After 30 years we’d have 10,000 predictions, a pretty good track record of whether economists actually had the ability to make useful predictions.  Those who think my skepticism is misplaced could set up a mutual fund that invested a little bit of money on the basis of each JEP prediction.  With so many predictions it would be highly diversified.  If economists’ predictions are superior to dart tossing, then the mutual fund would outperform index funds.

PS.  Here is a summary of the Goetzmann and Newman paper I referred to above.  There’s one good thing about the commercial property bubble that preceded the Depression.   The overbuilding in NYC during the 1920s reduced skyscraper construction in the 1950s and 1960s.  With a few exceptions like the Seagram building, the post-war era produced a bunch of boring modernist boxes.  Who hasn’t been delighted by strolling around Manhattan in the late afternoon sun, inspecting all the exotic spires on top of the art deco towers.  The Empire State Building and the Chrysler building are like the uneven towers at Chartres; built for the wrong reasons, but built by people who still had a sense of wonder and awe.  Thank God they were built.

PPS.  Don’t take this post as implying that I am bullish on Chinese markets.  The Chinese markets are highly volatile—I have no opinion on where they are going, except that we will probably continue to see large swings in assets prices.  I am bullish (long term) on the Chinese economy.  I think it will continue to grow rapidly for quite some time, albeit with occasional recessions.


Tags:

 
 
 

38 Responses to “Plato would not be impressed”

  1. Gravatar of Don the libertarian Democrat Don the libertarian Democrat
    5. May 2010 at 09:18

    If you predict a crash, you will be seen as a someone who knows what they’re talking about. That can translate into real money. If you make an erroneous prediction, you can make excuses that explain how the conditions of your particular prediction changed, and so were no longer operative. The key is to make your prediction very clear, and then add some seemingly minor caveats that allow you to weasel out in the event you’re wrong. I don’t think there’s any percentage in keeping a record.

  2. Gravatar of Indy Indy
    5. May 2010 at 10:41

    I like to compare rates of “electricity generation per capita” between developed countries (with stable levels of use) and similar developing countries (with growing rates) as one of my ways to try and understand where things might go in the long-term future in kind of a “eventually converging quality of life” scenario.

    So, let’s put China in context in this regard by ranking some major/notable countries. The convenient unit of measurement is Megawatt-hours per person per year (2008 data):

    1. Canada: 17.6
    2. US: 14.0
    3. Australia: 12.2
    4. Taiwan: 10.3
    5. South Korea: 9.3
    6. Japan: 9.1
    7. France: 8.8
    8. Germany: 7.8
    9. Saudi Arabia: 7.5
    10. Russia: 7.3
    11. UK: 6.3
    12. Greece: 5.5
    13. Ukraine: 4.2
    14. Venezuela: 4.0
    15. Turkey: 2.7
    16. CHINA: 2.6
    17. Brazil/Mexico (tied): 2.4
    18. Egypt: 1.63
    19. India: 0.71
    20. Pakistan: 0.56
    21. Bangladesh: 0.17

    So there we have two orders of magnitudes of different levels of electricity consumption.

    Notice where China is today – between Mexico and Turkey at in the high 2’s. Now notice where the other semi-similar East-Asian industrial countries (Japan, South Korean, Taiwan) are. Between 9 and 10, so between 3.5 and 4 times more intense usage.

    That would be about 15 years of nearly 10% growth to catch up.

    But then again, China uses mostly coal, and there’s only so much coal to mine and burn, what happens if they quadruple use? Also, China has already been the world’s #1 CO2 emitter for 4 years now, what happens if they quadruple emissions?

    Even if they don’t ever rise to such a high level of use, remember, they still have to double even to match Greece!

    Can they do it, or will they hit a wall? Will it be a plateau “smooth landing”, or will it cause breakage when the optimistic expectations of irrational exuberance are suddenly corrected?

    Something to think about.

  3. Gravatar of Doc Merlin Doc Merlin
    5. May 2010 at 13:03

    ‘If you are using asset prices to predict economic growth, that’s fine. But it makes no sense to use asset prices to make predictions that you imply will be useful to investors.’

    So very true…. but alas this is what ARMA models basically attempt to do.

  4. Gravatar of Jimmy Jimmy
    5. May 2010 at 13:48

    Except that real estate prices can be predicted fairly accurately using simple serial correlation. Kind of violates the EMH.

  5. Gravatar of StatsGuy StatsGuy
    5. May 2010 at 14:00

    ssumner:

    “My first thought was: If only the US and Europe could have that sort of “worst case scenario.” Note that 2.4 trillion yuan is about $350 billion, chump change for the Chinese government.”

    This, I think, is the point that seems to fly by everyone… Everyone looks at the US and deems that debt is not the problem. Technically, that’s not true in the immediate sense. Debt is a constraint on future consumption because distribution of wealth matters. You attribute this to wage stickiness. I attribute this to the fact that households become more risk averse as their net worth declines and thus they reduce consumption to rebuild balance sheets to insulate against risk. (My big objection to current policy is the trajectory of social obligations stored as government debt (and debt in general) vs as currency, and of course the structural issues.)

    But China… well, China doesn’t have the ossified democratic process in which a minority (creditors) hollers viciously that they deserve to be repaid in appreciating currency. China will attack the debt directly, either by devaluing or by neutralizing it, or some combination. This creates incentives for people to speculate due to future expectations of debt neutralization (not unlike the Greenspan Put), but China will try to attack this through administrative mechanisms (e.g. higher down payments, etc.).

    China can continue this until it starts running a negative trade balance.

  6. Gravatar of Doc Merlin Doc Merlin
    5. May 2010 at 14:29

    @Jimmy:

    It only seems that way. Its obviously false as the housing crash has shown. This illusion is something that happens a lot when error distributions are power functions otherwise have other fat tails, and you use a method that assumes gaussian error functions (as AR models do).

  7. Gravatar of Doc Merlin Doc Merlin
    5. May 2010 at 14:30

    should read “are power functions or otherwise have fat tails”

  8. Gravatar of Jimmy Jimmy
    5. May 2010 at 15:21

    “It only seems that way. Its obviously false as the housing crash has shown.”

    No it was not. A simple serial correlation predicts both the bubble and the crash in the housing market. Serial correlation is just that, a CORRELATION. It does not mean that if prices go up, they must go up forever.

  9. Gravatar of Tom P Tom P
    5. May 2010 at 15:44

    @Jimmy: are you referring to a particular study or data set? I’m curious as to where this claim is coming from.

  10. Gravatar of Jimmy Jimmy
    5. May 2010 at 19:52

    Case, Karl E., and Robert J. Shiller. 1988. The Efficiency of the Market for Single-Family Homes.” American Economic Review 79(1): 125-37.
    –. 1990. “Forecasting Prices and Excess Returns in the Housing Market.” American Real Estate and Urban Economics Association Journal 18(3): 253-73.

  11. Gravatar of rob rob
    5. May 2010 at 22:29

    i think Plato would have put economics in the same category as politics and thought that its relevance was equivalent to the rules for a lunatic asylum. practical but not profound.

  12. Gravatar of Doc Merlin Doc Merlin
    5. May 2010 at 22:45

    “No it was not. A simple serial correlation predicts both the bubble and the crash in the housing market. Serial correlation is just that, a CORRELATION. It does not mean that if prices go up, they must go up forever.”

    The AR models fail when you reach the crash. Last time I modeled housing starts we had to put in a dummy variable for the crash.

  13. Gravatar of scott sumner scott sumner
    6. May 2010 at 05:23

    Don the Libertarian Democrat, I agree.

    Indy, China is already starting to move beyond coal, although in absolute terms coal consumption continues to rise. In any case China will be able to increase it’s electricity production up the Japanese levels–there is plenty of coal around. The big problem will be the effect on atmospheric CO2.

    Doc Merlin. Thanks.

    Jimmy, In the short run (a few months out) there is momentum in real estate prices. But I don’t think turning points are easy to predict. And it is turning points that are of interest to these pundits.

    Statsguy; You said;

    “Debt is a constraint on future consumption because distribution of wealth matters. You attribute this to wage stickiness.”

    I agree that debt is a constraint on future consumption, I don’t understand the wage stickiness comment. You may be confused by what I said on an unrelated issue. Although debt causes people to consume less, it does not cause them to work less. Unemployment is due to sticky wages, indeed people often wish to work more when they are deep in debt.

    You said;

    “But China… well, China doesn’t have the ossified democratic process in which a minority (creditors) hollers viciously that they deserve to be repaid in appreciating currency.”

    Yes, but China does have an ossified dictatorship where the interests of a minority (urban residents) are favored over the more numerous rural residents.

    Jimmy#2, I am puzzled how serial correlation could predict both the bubble and crash, as nothing like that had happened in all of American history. And if it was so easy to predict, why did so many Wall Street experts fail to predict it?

    rob, I’m afraid I know little about Plato, other than that he believed in philosopher kings.

  14. Gravatar of Jimmy Jimmy
    6. May 2010 at 06:33

    “Jimmy#2, I am puzzled how serial correlation could predict both the bubble and crash, as nothing like that had happened in all of American history.”

    Why should that matter? Once prices start to rise very quickly, a simple serial correlation makes a very strong prediction that they will keep rising at a similar speed. Once they start to fall very quickly, a serial correlation predicts them to keep falling very quickly.

    If you’re looking at monthly or even quarterly data, then it is very easy to gain excess returns in a bubble/bust cycle that lasts 10+ years.

    “And if it was so easy to predict, why did so many Wall Street experts fail to predict it?”

    Questions like these make me want to smack myself in the face. Notice, Case and Shiller wrote those papers identifying serial correlation and developing forecasting models in 1989 and 1990. Long before the bubble began to inflate.

  15. Gravatar of Jimmy Jimmy
    6. May 2010 at 06:35

    “The AR models fail when you reach the crash. Last time I modeled housing starts we had to put in a dummy variable for the crash.”

    Maybe it’s just me, but this post seems incoherent. Who is talking about housing starts here?

  16. Gravatar of athEIst athEIst
    6. May 2010 at 09:08

    Instead, I think the Great Depression was caused by NGDP falling from $106.7 billion in 1929:3 to $49.8 billion in 1933:3.

    NGPD dropped by more than half. That is called the Great Depression. It didn’t cause the Great Depression–it IS the Great Depression.

  17. Gravatar of scott sumner scott sumner
    6. May 2010 at 10:07

    Jimmy, So you are saying that serial correlation can predict the crash after it has started? I can’t disagree with that. But that is consistent with what I said in the post. Housing prices have been rising, so doesn’t this imply they are likely to keep rising? But some pundits want to suggest that if housing prices have been rising, then they are likely to fall in the future.

    The real challenge is to predict turning points, not just extrapolate a few months out into the future.

    athEIst, No. A depression is a big fall in RGDP, which is completely different from NGDP. Not all depressions are caused by falls in NGDP, indeed many depressions in other countries are accompanied by big increases in NGDP.

  18. Gravatar of Doc Merlin Doc Merlin
    6. May 2010 at 10:22

    ‘”The AR models fail when you reach the crash. Last time I modeled housing starts we had to put in a dummy variable for the crash.”

    Maybe it’s just me, but this post seems incoherent. Who is talking about housing starts here?’

    1. ARMA models fit supply or demand far better than they generally fit prices, which usually are far more noisy.

    2. Yes, there is some serial autocorrelation in housing prices, however in order to actually make much (more than say the 3-4% you can get from nearly risk-less assets) money on it in the long run by speculation, statistically, you need to know approximately where the turning points are. Otherwise you will lose money and have huge liquidity problems when you miss a turning point.

    3. Changes in price data tend to follow long tailed distributions, not white noise distributions. All this means is that the turning points are far more important than anything else.

  19. Gravatar of marnues marnues
    6. May 2010 at 12:07

    Jimmy, in your model, how does one predict the crash? Or do prices truly just keep on rising? Have you been living on the moon for 3 years?

  20. Gravatar of Miles Miles
    6. May 2010 at 12:11

    The quote you contest about construction reads:

    “As much as 60 percent of the country’s gross domestic product relies on construction”

    Your retort: “the construction sector was only 7% of Chinese GDP in 2003. I’d guess it is about 10% today.

    Calculating a portion of the GDP that relies on construction and calculating the percentage of GDP that represents the construction sector are not the same. I think that while the 60% number might be a little high, its not that high. Keep in mind that “manufacturing, mining, and utilities are close to 40%.” What percentage of mining products get used in construction? I’d venture to say that its a rather high proportion. Proportions of manufacturing are probably almost as high as mining. The energy usage of construction is certainly not insignificant.

    Furthermore, there are numerous small businesses that would be considered part of the service industry whose primary customers are the construction industry. I’m mainly thinking of mechanics and other repairmen who specialize in cranes, paint sprayers, elevators, HVAC systems, and everything else that goes into a building. Are architects considered part of the construction industry?

    My point is that construction of buildings employs a huge sector of any economy when things are actually getting built. What portion of GDP was due to construction in 2005? What portion of GDP RELIES on active construction for its clients?

  21. Gravatar of Wishing for a chinese crash « Jim’s Blog Wishing for a chinese crash « Jim’s Blog
    6. May 2010 at 12:19

    […] and lots of people are predicting China will crash, though The Money Illusion doubts it. Supposedly, debt fueled construction, decreed by central planners, Soviet style, is sixty percent […]

  22. Gravatar of athEIst athEIst
    6. May 2010 at 12:29

    But YOU said the depression was caused by a drop in NGDP. Perhaps a little clarification about RGDP versus NGDP(specifically with regard to 1929-1933) and some examples of the “big increase in NGDP” depressions.

  23. Gravatar of 8 8
    6. May 2010 at 12:56

    Faber wants to make money. He’s actually quite bullish in the long-run on China, and he and Jim Rogers have pointed to Korea and Russia, for instance, as countries that defaulted and came back quite rapidly.

    Faber is using some socionomics and looking at social mood. Events such as the Olympics and World’s Fair are awarded to a country that has been successful, so they can sometimes mark the peak of a business cycle. It’s just like the skyscraper index. And the panic of 1873 started in Vienna…

    Chanos is different, he’s in the crowd that thinks China is a giant bubble that will pop like the Internet stock bubble. Faber wrote a book saying Asia, including China, is the future.

  24. Gravatar of StatsGuy StatsGuy
    6. May 2010 at 13:25

    OK, let me rephrase your argument …

    Sticky Wages ==> Unemployment ==> Reduced Income & Reduced Income Expectations ==> Reduced Future Consumption

    My argument …

    Net (permanent) reduction in household balance sheets ==> Reduced Wealth Expectations ==> Reduced Future Consumption ==> Unemployment

    Honestly, I’m sure both are true. There’s a cycle here, and an exogenous shock can hit any variable in the model, and affect the rest. The empirical arg is which shocks are most important (which may change over decades).

    But heh, how’s the EMH holdin’ up today?

    Supposedly 60% of the plunge was caused by a trader error (unless you believe in conspiracy theories).

    You realize that a trader with enough liquidity could have bought 10 billion in options (puts) on correlated assets, then trashed an asset (check out the chart for ACCENTURE!!), and made a total killing even if the individual lost the entire billion they used to drop the market.

    I know you would defend the markets, but consider that Fama didn’t really model illiquid markets that well, and when you have massive leverage, even “deep” markets like the Dow seem kind of jittery. Or, think of it this way: Market prices convey information. If one trader secretly causes the plunge, that individual has private information about what is causing the plunge, while everyone else is paralyzed trying to figure out what the heck caused the plunge. That’s not to say that markets are irrational (though I think they are to some degree), merely that the existing models of “market rationality” may employ a very limited form of rationality.

  25. Gravatar of Mari Mari
    6. May 2010 at 13:44

    Scott: What makes a market? People like Chanos and Faber offer their views, and there are people who take the opposite side. I can understand you railing against “arm-chair” economists, but Chanos and Faber are real money guys – not sure about Faber, but Chanos puts his money where his mouth is. If he is wrong, he will lose money…plain and simple.

  26. Gravatar of scott sumner scott sumner
    6. May 2010 at 14:51

    Miles, You said;

    “Furthermore, there are numerous small businesses that would be considered part of the service industry whose primary customers are the construction industry. I’m mainly thinking of mechanics and other repairmen who specialize in cranes, paint sprayers, elevators, HVAC systems, and everything else that goes into a building. Are architects considered part of the construction industry?”

    I would assume that much of this was included in the 7% figure. But I think you are probably right that some inputs get included in with manufacturing. Remember, however, that there is another issue that cuts the other way. When many people hear “construction” they think about new buildings. In the US that’s pretty true, but not so much in China. Much of construction in China is roads, airports, subways, rail, ports, dams, power plants, etc. All that stuff is non-cyclical, and will continue whether housing declines or not. Then there is construction of factories, which may be cyclical, but isn’t directly related to housing. A huge chuck of Chinese manufacturing is for exports. Another big chunk is cars and appliances. So I’m still not convinced the number is anywhere near 60%, although perhaps it is more than 10%. The interesting number would be the value of new housing and offices as a share of GDP.

    In the US housing construction plummeted in the 2 years after mid-2006, and yet real GDP continued growing during that period. I think people exaggerate the impact of housing.

    AthEIst, NGDP is M*V. A fall in nominal spending caused by tight money or falling velocity causes RGDP to fall if wages and prices are sticky. Zimbabwe recently experienced a depression with rising NGDP. Many countries in the developing world saw NGDP rise during depressions (if they had high rates of inflation.)

    8, The Olympics were awarded to China in 2001, so I’m not sure how that relates to the current cycle. The 1873 depression was caused by an increased worldwide demand for gold, along with a falling nominal price of gold in the US.

    I agree that Chanos and Faber are quite different, and that Faber has been an Asia bull. I think Faber knows more about China than Chanos does. The Chinese economy is about as different from the tech bubble as you can imagine.

    Statsguy, You said;

    “Net (permanent) reduction in household balance sheets ==> Reduced Wealth Expectations ==> Reduced Future Consumption ==> Unemployment”

    I don’t see how reduced consumption causes people to work less.

    You said;

    “You realize that a trader with enough liquidity could have bought 10 billion in options (puts) on correlated assets, then trashed an asset (check out the chart for ACCENTURE!!), and made a total killing even if the individual lost the entire billion they used to drop the market.”

    Then why didn’t they? I agree that markets are not perfect. 15 minutes out of every decade there is a glitch. I admit to not knowing much about these issues, as I focus on longer term price trends–like one day.

    Mari, I am railing against journalists who write articles that are completely lacking in intelligible content. Where’s the beef? Where’s the information that should persuade me?

  27. Gravatar of StatsGuy StatsGuy
    7. May 2010 at 11:28

    ssumner:

    “I don’t see how reduced consumption causes people to work less.”

    Because the labor market has a demand side too, and there are other factors involved in clearing the market other than labor price (wage). For example, inventories. If there’s a large inventory backlog (like in the housing market, or consumer durables), then if one imagines just a handful of “real world” dynamics – like cost of inventory storage, loss of inventory value due to damage or technology updating – then you don’t need sticky wages to get unemployment. All you need is transaction costs in the labor market and/or industry specific human capital (e.g. labor is not a purely fungible commodity good; someone with a PhD in electrical engineering will probably spend several months searching for a job with a good fit before settling on a grocery bagger position).

    “Then why didn’t they?”

    Well, sometimes they do. aka, Liar’s Poker.

    http://www.amazon.com/Liars-Poker-Rising-Through-Wreckage/dp/0140143459#reader_0140143459

  28. Gravatar of Bill Stepp Bill Stepp
    8. May 2010 at 04:25

    Scott writes:

    Instead, I think the Great Depression was caused by NGDP falling from $106.7 billion in 1929:3 to $49.8 billion in 1933:3. When NGDP falls in half, it becomes very difficult to service debt on brand new skyscrapers built under the assumption that NGDP would not collapse. I just don’t see irrationality here, nor do I see causation running from debt bubbles to depression. But I am in the minority regarding the Great Depression; indeed I recall that Rogoff also thinks the Depression was caused by a financial crisis. If China experiences a recession I will again in be in the tiny minority who think debt has little to do with the problem.

    1. I think you confuse what the Great Depression was with the cause of it, or should I say lack of a cause. Saying it was caused by a fall in NGDP begs the question of what caused it to fall. As an Austroid, I go with the cause being an expansionary Fed, which forced market interest rates to fall below the natural rate of interest. This caused entrepreneurs (promoters, as they were sometimes called then) to invest in projects (like Florida land schemes) with discount rates that overvalued their expected cash flows, thus causing unsustainable investments, which proved to be malinvestments when the boom ended and reversed. The debts/projects had to be liquidated or restructured during the depression.
    Your focus on–ugh–NGDP overlooks the microeconomic aspects of both booms and busts. See the Austrians, Garrison, Horwitz, Selgin, White, and on the boom in the 20s, Rothbard’s America’s Great Depression.
    I don’t know about Rogoff and China, but I agree that Chanos’s 60% figure is too high.
    I’d advise you to read the Austroids and have a rethink, just as they might read your writing and have a group rethink, or sorts. Don’t know where there might be a middle ground, or if one exists, but that’s what makes intellectual markets, I guess.

  29. Gravatar of scott sumner scott sumner
    8. May 2010 at 06:43

    statsguy, That sounds like Kling’s reallocation theory. But if people want to consume less, that means they want to save and invest more. This should cause labor to be reallocated from consumer goods to capital goods-making industries. You are right that there could be some frictional unemployment during the reallocation, but certainly investment should increase. The problem I have with this sort of view is that generally people apply it to recessions where both investment and consumption are declining. Indeed in this recession investment has declined more rapidly than consumption. So our current unemployment problem is definitely not related to people wanting to consume less. I still think it is sticky wages that are the problem.

    Bill Stepp, Even the Austrians don’t beleive that an expansionary monetary policy caused NGDP to fall in half. Expansionary monetary policy makes NGDP rise. After WWII monetary policy was far more expansionary than during the 1920s, and yet NGDP kept rising almost continuously (until 2009.) I don’t think any serious Austrian would agree with your view. I think most blame the depression on “secondary deflation,” which is an excessively tight monetary policy.

    And regarding NGDP, many Austrians (Hayek, Selgin, etc.) have at one time or another found it to be a useful indicator of the stance of monetary policy. Hayek called for stabilizing NGDP. Had we done so, there would have been no Great Depression.

  30. Gravatar of Bill Stepp Bill Stepp
    8. May 2010 at 08:30

    Scott,

    I didn’t mean to say (or imply) that expansionary monetary policy caused NGDP to fall. It clearly was caused by a contractionary policy, but the boom set up the conditions for the fall, including the malinvestments that were liquidated. You didn’t address the malinvestment question. That’s just completely off your radar screen, as it is with non-Austrian macroeconomists.
    But I hope you don’t think that monetary policy is responsible for economic growth, which is caused by savings, investment, and productivity increases, the rule of law, etc. Free market money and banking institutions (i.e., free banking, no central bank) would promote economic growth, but they certainly would not be the root cause of it.

    The bottom line with monetary policy, is that there is no such thing in a free market, any more than there would be a foreign policy in a anarchistic world, or an education policy in a world of free market schools, etc.
    Down with policy, including monetary policy!

  31. Gravatar of Scott Sumner Scott Sumner
    9. May 2010 at 04:36

    Bill Stepp, i thought you used to read my blog (the name seemed familiar.) If not, no I have never argued that money is responsible for long run growth.

    The malinvestments story of the 1920s is a myth. They only looked like bad investments because we went into the Great Depression. Had NGDP kept growing at 3% year, as in the 1920s, then the investments would have turned out fine. Of course if you create a Great Depression with tight money then people will stop buying cars. To the person who doesn’t understand what is going on it will look like they built too many car factories, houses, offices, etc in the 1920s, but the real problem was too little demand in the 1930s.

  32. Gravatar of Matthew Yglesias » Bullish on China Matthew Yglesias » Bullish on China
    9. May 2010 at 13:29

    […] groupthink, so I linked to this bearish article last week. Today, though, I’d like to link to Scott Sumner’s rebuttal which strikes me as […]

  33. Gravatar of Chris Chris
    9. May 2010 at 19:55

    “But it makes no sense to use asset prices to make predictions that you imply will be useful to investors.”

    I only agree with this statement if “investors” means “those who invest in public securities.”

    If a company is considering making a particular investment (e.g., building a factory), it could use asset prices to infer the value of such an investment. Since the factory is not an investment option for the majority investors, an EMH world provides useful information to make this investment decision.

  34. Gravatar of StatsGuy StatsGuy
    10. May 2010 at 12:23

    ssumner:

    “statsguy, That sounds like Kling’s reallocation theory.”

    Yes, the second part is like Kling’s theory. But Kling believes the exogenous trigger was sudden realization of massive malinvestment, and thus that the increase in “frictional” unemployment was due to reallocation of resources that was in the end functional. [note this relates to your comment about “malinvestment” in the 1920s – a lot of it was only malinvestment because monetary trajectory shifted]

    The argument above states that the increase in “frictional” unemployment is _not_ functional because the trigger event was a decrease in _perceived_ net wealth (that is to say, the sudden realization that we were NOMINALLY poorer, not that we were REALLY poorer).

    But why would an increase in nominal poverty matter?

    Because everyone realized they were exposed to more risk (e.g. had less liquidity than they thought), and realized that everyone else was realizing this, which created a demand for cash, and thus a realization that cash would become more expensive and demand for goods/services would drop as savings increase causing prices to fall.

    But the supposed “malinvestment” isn’t really real. (Or, most of it isn’t really real.) It just looks really real. In other words, we’re not really (that much) poorer, we just think we are, but because we think we’re poorer we make act poorer and this makes ourselves poorer.

    Here’s one big area where Kling and you certainly disagree with me: If one accepts this story (e.g. perceptions of nominal wealth matter due to perceived increase in risk driving demand for money/liquidity), then wealth distribution matters too, because the curvature of the money demand curve is dependent on perceived individual risk (aggregated up), and if that function is nonlinear, then the aggregate of the function does not equal the function of the aggregate.

    Blaslphemy, I know. I mean, an argument like this might even be used to defend the notion that the Middle Class is an important component to stable growth. Egads.

    Anyway, I don’t expect you to agree with the story above, but it isn’t inconsistent with your wage stickiness story (which is how Krugman still thinks of it too). The stories are reinforcing.

  35. Gravatar of StatsGuy StatsGuy
    10. May 2010 at 12:35

    “because we think we’re poorer we make act poorer and this makes ourselves poorer”

    should read

    “because we think we’re poorer, we act poorer and make ourselves poorer”

  36. Gravatar of Matpw Matpw
    12. May 2010 at 12:37

    I was taught investment equals the change in the capital stock. I doubt investment at 60% of GDP is sustainable. If investment falls from 60%, is there any way consumption can lead keep the Chinese economy growing or even flat?

  37. Gravatar of Jeff Jeff
    12. May 2010 at 15:07

    The comment about a “Journal of Economic Predictions” reminds me of this gem of a letter to the editor in the WSJ:

    “HEADS I WIN, TAILS YOU LOSE”
    Appeared in the Wall St. Journal, August 6, 1996

    Your semiannual survey of economists has come and gone (July 1) and I was not included again. This is a shame because I have developed a complex economic model for predicting the yield on the 30-year Treasury bond. Much of the process is proprietary, but I will share some details. I first gather all sorts of economic data, including consumer confidence, nonfarm payroll, the Producers’ Price Index and those all-important commodity prices, with a special emphasis placed on gold, of course. I then contact Eleanor Roosevelt to get her thoughts on the mood of the Federal Reserve. And finally, and this is most important, I reach into my left pocket, pull out a 1993 Canadian penny and flip it.

    Astute readers will no doubt argue that my model gets the direction of long-term rates correct only about half of the time. To these cynics I point out that there have been 29 surveys conducted since the Wall Street Journal began asking economists for their 30-year T-bond yield forecasts. The consensus estimate of these highly-paid economists for the 30-year bond has been in the wrong direction in 20 of these 29 surveys. Furthermore, 43 economists have participated in 10 or more surveys. Only 13, or 30%, have gotten the direction correct more than half of the time, with only one economist breaking the 60% barrier.

    Interest rate forecasts are most important when the interest rate movements are large, and there is where my model shines. The yield on the 30-year Treasury bond has ended up or down more than 100 basis points during the six-month prediction period on ten occasions. The economists’ consensus forecast (and I use this term loosely) was in the right direction only twice. My model has predicted the right direction, believe it or not, in five of 10 periods.

    I will ignore the Journal’s slight and provide my forecast to your readers anyway. The Canadian maple leaf is facing up, suggesting rates will decline. I noticed six economists predicted 30-year rates out to the second decimal place (for example, 7.29% instead of 7.3%). I too have a good sense of humor and therefore predict the yield on the 30-year bond will fall to 6.71%. Readers who desire future forecasts will have to purchase my newsletter.

    Kevin Stephenson
    Assistant Professor of Economics
    Middlebury College
    Middlebury, Vermont

  38. Gravatar of scott sumner scott sumner
    16. May 2010 at 17:29

    Chris, I agree.

    Statsguy, I would emphasize the fact that the Fed did not respond properly to the increase in demand for cash. It is hopeless to try to stabilize the demand for cash, although I agree that income distribution changes could influence cash demand. But I still see the root cause of demand inadequacy as monetary policy, not shocks causing an increase in the demand for liquidity.

    Matpw, You said:

    “I was taught investment equals the change in the capital stock. I doubt investment at 60% of GDP is sustainable. If investment falls from 60%, is there any way consumption can lead keep the Chinese economy growing or even flat?”

    Yes. In America consumption is nearly 70% of GDP. China won’t reach that level soon, but consumption is likely to rise. In any case, growth in the long term is determined by AS, not AD.

    Jeff, That’s a gem. Thanks.

Leave a Reply