Matt Yglesias recently made this remark in a post criticizing Tyler Cowen’s assertion that the GSEs significantly worsened the US housing/banking crisis:

The part where unwise public policies to subsidize homeownership would seem to come into this is step (1), but we in fact see this happening in many markets (Spain, commercial real estate) where Fannie and Freddie weren’t players.

I’m not sure what Matt means by “many markets.”  According to the interactive housing price graph constructed by The Economist, only the US and Ireland experienced major housing bubbles.  Spain had a smaller one (although I suspect more accurate figures would show a more serious price decline in Spain.)  Japan and German also saw price declines after 2006, but they had had zero price run-up before 2006–so no bubbles there.  The vast majority of industrial countries saw big increases in real housing prices between 2001 and 2006, just like the US.  But unlike the US, prices tended to move sideways after 2006, even in real terms.  (Of course you can find some brief price dips in the severe 2008 recession, but nothing like the long collapse in the US, which pre-dated the recession.)

But let’s take a look at Spain.  It’s true that Spain does not have a single institution called “Fannie Mae.”  But do they have a similar problem of governments deeply involved in promoting real estate lending?  I’m no expert (and I welcome critiques) but my initial impression is that the answer is yes.  Here’s The Economist:

Another duality lies in the banking system. Observers fret that the Spanish state may have to pump a lot more money into the banks than the roughly €25 billion ($36 billion), or 2.5% of GDP, it currently reckons will be the total bill from Spain’s epic housing boom and bust (Ireland’s bank bail-out bill is over 40% of GDP).

The problem is concentrated in the cajas, local savings banks that make up around half the domestic banking system, rather than big Spanish banks like BBVA and Santander, which are protected by big international businesses. Not before time the government is overhauling the cajas. Their ranks are being slimmed””the number has fallen from 45 to 18″”and they are being reorganised as joint-stock companies that can raise equity capital. José Maria Roldan, director-general of banking regulation at the Bank of Spain, says that the reform is “a huge step forward, replacing the caja model with a standard banking template that is more secure and comprehensible to international investors.”

So I decided to further investigate the cajas, and this is what I found (also from The Economist):

IN THE end CajaSur trusted in God. On May 22nd the small savings bank (or caja), which is controlled by the Roman Catholic church in Córdoba, was seized by the Bank of Spain after failing to agree on a merger with Unicaja, a larger Andalusian rival. The move shocked investors and prompted other savings banks to hasten consolidation. But mergers between wobblier cajas, which are unlisted and make up close to half of Spain’s financial system by assets, are merely a first step in a longer process.

CajaSur is tiny, holding just 0.6% of Spain’s financial assets. Yet its seizure unsettled investors for two reasons. First, it was a reminder that politics often trumps rationality.

.   .   .

The politicians who control the cajas like this virtual structure because it allows the banks to keep their own brands, governing bodies and local retail operations while combining treasury and risk-management functions.

.   .   .

Encouragingly, both the government and the opposition have agreed to reform the law governing savings banks.Attracting private capital requires other changes, too, such as a reduction in the influence of politicians, something caja managers would relish. Greater openness about banks’ balance-sheets would also help: on May 26th the Bank of Spain moved in this direction by announcing plans for tougher provisioning rules.

“The politicians who control the cajas”?  I thought banks were supposed to be controlled by businessmen, not politicians.  I’m still no expert on Spanish banking.  But I’d wager that further investigation would turn up the same incestuous relationship between politicians and the cajas that we saw between politicians and the GSEs.

So we observed clear-cut housing bubbles and busts in just two countries with more than 5 million people; the US and Spain.  And both suffered from the same problem—politicized institutions that will require massive transfers from the public.  Both also had large private banks that made mistakes, but at least they didn’t impose huge burdens on the taxpayers.

Matt also mentioned commercial real estate in the US, but I don’t think that proves what he wants it to prove.  As you know, the profession has not accepted my argument that lack of AD, not the banking crisis is responsible for our macroeconomic problems.  But the one weakness in my argument is that the subprime bubble blew up well before the recession.  This leads Matt to draw a connection between the financial crisis and the recession.

Banking activities need to be regulated or else asset price fluctuations will lead to macroeconomic instability.

That’s why the stakes in this debate between progressives and libertarians are seen as being so high.  If it was just the TARP bailout, I’m not sure people would care that much.  The big banks are repaying the loans.  Indeed I’ve seem progressives praise the auto bailout, even though GM may never pay back all the money.  No, the reason this is so important is that it’s seen as a crisis that led to 9.2% unemployment three years later.  Fair enough.

But in that case Matt can’t use commercial real estate, as that was clearly a symptom of the recession.  Indeed commercial RE was still booming in mid-2008, and only turned south toward the end of the year.  Now in fairness to Yglesias, the fall in commercial RE did bring down lots of smaller regional banks, and this resulted in costly FDIC bailouts of depositors.  If we insist on having FDIC (a big mistake in my mind), and insist on not reforming it by placing $25,000 caps on insured deposits (and even bigger mistake), then yes, we need to regulate banking.  We tried to re-regulate after the 1990 S&L debacle, and it didn’t work.  We tried again with Dodd-Frank, and again failed to deliver an effective set of regulations (like, umm, banning subprime loans, for instance.)  But I would agree that in the presence of FDIC, a completely unregulated banking system will take far too many risks.  I don’t consider that the failure of “laissez-faire”, I view it as the failure of a banking system where much of the liabilities are essentially nationalized.

PS.  Interested readers may want to play around with the Economist’s interactive graph at this link:

It’s a bit hard to read the graph below, so a few comments.  The steady drop (orange line) is Japan.  The only other two countries down in real terms from 2001 are the US (grey) and Ireland (green, of course.)  Spain is a rather dark line that rises steadily to almost 180, then slips back to 140.  As I see the graph, most other countries had substantial real price run-ups, like the US, but then real prices trended sideways.  This shows that not everything that goes up must come down.  I see lots of commenters patting themselves on the back about how they predicted the bubble would burst.  Count yourself lucky that you live in America; otherwise you probably would have been wrong.  Germany is not listed because of incomplete data, but it had no bubble in the first place.

Also note that the default option is nominal prices, which is actually more supportive of my “very few bubbles” claim.  I converted to real for the graph below, which puts more of a downward bias on prices after 2007.

When I read the newspapers from the 1930s I’d occasionally see hopeful articles about how consumer spending would have to pick up soon, because of all the “pent-up” demand.  If people hadn’t been buying cars for a while then presumably their cars would wear out, and this would trigger new demand for replacements.  Of course I knew that there actually was no light at the end of the tunnel, which made these articles seem slightly pathetic—as if they were grasping for straws.  They overlooked the fact that the depression made America much poorer, and that low consumer demand reflected that poverty.  For similar reasons, there isn’t much “pent up demand” for cars in Somalia, despite low sales in recent years.

Sometimes I see this argument applied to the housing slump.  Housing construction is down 70%, to levels far lower than at any time in post-war history (relative to population.)  And this slump has been going on for a number of years.  Surely we’ll soon need to build more houses, to meet our growing population.  If only that were true.  Unfortunately, as sharply as housing construction has fallen, household formation has fallen even faster.

Jim Glass sent me some very interesting data on household formation, which casts a very different light on the recent housing crash.

By my simple measure recent housing starts peaked with an inventory of 40% of an average year’s worth of starts above the trend line in 2007. That’s a cyclical high but a typical one. About the same or a little higher was reached in three other cycles since 1960.

But the plunge in starts since 2007 is unprecedented “” by the end of 2010 cumulative starts were 72% of an average year’s worth of starts below trend. The next-lowest figure was 46% below trend back in 1970. If things were “normal” this would predict a huge boom in housing starts soon.

But housing starts are *following* household formation, which is plunging even faster, like an ICBM heading straight to its target.

In 2007 household formation was 1627k (average 1998-2007: 1499k) and housing starts were 1355k (average 1998-2007: 1716k). In 2010 household formation was all of 357k, down 78% from 2007 and down 76% from the prior ten year average. Housing starts were 587k, down 57% from 2007 and down 66% from the prior ten years. That’s a big fall, but it is still *well behind* the fall in household formation.

If I still had my blog I’d post the graphs “” the line for household formation is heading straight down like to the bottom of the sea, it’s three times the fastest-deepest decline of the last 40 years. The line for housing starts looks like it is just striving to not fall too far behind.

I hate to be the bearer of bad news, but that light at the end of the tunnel is an onrushing train called falling household formation.  It’s caused by three factors:

1.  Less immigration due to the post-2006 crackdown.

2.  Less immigration due to the severe recession and high unemployment

3.  20-somethings who can’t get jobs are living with their parents.

The problem is not that we built too many houses and need to work off the excess.  Yes, we did, but we worked off that excess long ago.  No the current problem is crashing demand for homes due to an unprecedented plunge in household formation.  Call it a demographic depression.  And the root cause?  I know I’m going to sound like a broken record, but the biggest cause in low NGDP (although obviously other factors are also at work here–including immigration crackdown, minimum wage increase, extended unemployment insurance, etc.)

PS.  I can’t wait for some smart alec commenter to write in and tell me the minimum wage increase can’t possible affect household formation, as no one can afford to live by themselves on the minimum wage.  I already look forward to slapping you down.  So go ahead and make my day.

PPS.  This website shows that the Census had forecast household growth of about 1.1% per year.  In this website you can find the data Jim used; only a 0.3% growth in household formation last year.  And it’s possible the Census is a bit behind the curve, as for instance they didn’t catch up to the 1990s immigration surge until the 2000 census.  Thus it might be even worse than Jim’s figures show.

There’s a lot of discussion now about the role of the GSEs in “the crisis.”  Unfortunately, not everyone is talking about the same crisis.  Some are talking about the housing bubble/crash, some are talking about the late 2008 financial crisis, and I believe both groups have the 2011 unemployment crisis in the backs of their minds (otherwise why is the debate seen as being so important?)  After all, there is no similarly high-charged debate over the auto crisis/bailout/sales slump.

Let’s start with the housing crisis.  A major theme of the Austrians is that too many houses were built in the mid-2000s, and the resulting slump has led to high unemployment.  Here are US housing starts per capita going back to 1960:

As you can see, housing starts over the last decade have been far below the level of previous decades.  We certainly don’t have a weak housing sector in 2011 because an extraordinarily large number of homes were built in the past decade.  Rather it seems the recession has caused many families to double up.  BTW, I will concede that we built too many homes in the mid-decade period, so I don’t completely deny the Austrian story.  I just don’t think too many houses is the huge “crisis” most people are talking about.

Instead, it seems to me that both sides of the GSE debate tacitly accept that lax lending standards due to either:

1.  deregulation and moral hazard causing banks to take excessive risks, or

2.  the GSEs and other federal housing rules, regulations, tax breaks, etc.,

caused a housing price bubble in the mid-2000s.  When this bubble collapsed, it created a severe banking crisis, which then led to a severe recession.

I believe this is mostly wrong.  I’ll concede that part of the housing bubble was due to the factors mentioned above (both banks and the GSEs played a big role.)  But the link between the housing bubble and the severe financial panic is much weaker than people realize.  And the link between the severe financial panic and high unemployment in 2011 is almost nonexistent.

The mistake both sides make is to look for monocausal explanations.  Here’s what the facts show:

1.  The economics profession almost entirely disagrees with me.  Yet in mid-2008 the consensus view of the economics profession was that we were NOT going to have a severe financial crisis and we were not going to have a severe recession.  Indeed growth was forecast for 2009, along with moderate unemployment.  And yet the scope of the subprime crisis was almost completely understood by that time.

After things blew up, Bernanke was mocked for early statements suggesting that likely subprime losses, even in the worst case, were not large enough to bring down the US banking system.  But of course he was right.  Here’s what actually happened:

1.  Between June and December 2008 both NGDP and RGDP fell sharply.

2.  The cause of the fall in RGDP was the fall in NGDP

3.  NGDP fell at the fastest rate since 1938 because the economy was already sluggish due to the housing slump, which reduced the Wicksellian equilibrium rate of interest.  Ditto for oil and auto sales.  Normally the Fed would cut rates enough to prevent a recession.  But this problem coincided with a severe commodity price shock, which drove up headline inflation and frightened the Fed.  They did not cut rates once between April and October, by which time the great NGDP and RGDP crash was nearly two thirds over.

4.  The big NGDP crash dramatically reduced almost all asset values in the second half of 2008.  About half way through this crash, the severe phase of the banking crisis started.  This doesn’t mean the earlier subprime fiasco played no role.  It did greatly weaken the system in 2007 and early 2008.  So the GDP crash was imposed on an already weakened banking system.  A cold turned into pneumonia.  GDP fell even faster.

5.  Estimated losses to the entire US banking system soared during this crash, and peaked in early 2009 at roughly $2.7 trillion, only a modest fraction of which were subprime mortgages.  Then expected growth rates recovered somewhat, asset values partially recovered, and estimated losses to US banking fell back under a trillion.  So the proximate cause of the financial crash was tight money which drove NGDP expectations much lower, although the earlier subprime fiasco certainly created an environment with a low Wicksellian equilibrium rate, making monetary errors much more likely.  And of course when rates hit the zero bound (which by the way didn’t occur until the great GDP crash had ended in December!!) the Fed had an even more difficult time steering the economy.

6.  To summarize, the severe financial crisis could not have caused the great GDP collapse, because monthly GDP estimates show it was half over before the post-Lehman crisis even began.  But even if this view is wrong, there is not a shred of theoretical or empirical evidence linking the current 9.2% unemployment with the 2008 financial crisis.  Theory suggests that if a central bank inflation targets, it drives NGDP.  The Fed says it has the economy where it wants it (in nominal terms), and doesn’t think we need more inflation.  When it did think we needed more inflation mid-2010 (when the core rate had fallen to 0.6%) it did QE2, which raised core inflation back up to roughly where the Fed wanted it.  Of course (just as in mid-2008) commodity price inflation is distorting Fed policy, but that’s a problem attributable to the Fed, not the financial crisis.

This graph shows how IMF estimates of total US banking losses are inversely correlated with expected total inflation and RGDP growth in 2009 and 2010.

I see three separate crises.  A “misallocation of resources into housing crisis,” a “federal bailout of banks crisis,” and a high unemployment crisis.  Who’s to blame for each?

1.  The private banking system and the GSEs both played a major role in causing too much housing to be built in the mid-2000s.  The errors of the private banking system were due to both misjudgment (they did lose money after all) and bad incentives (moral hazard due to various government backstops.)  Pretty much the same is true of the GSEs, although their role has always been a bit more politicized, and Congress must accept some blame for pushing them to boost the housing market.  But this was a modest problem, as the first graph shows.  It’s not the “real” issue that the left and right is debating so vigorously.

2.  The GSEs are far more to blame than the banks for the bailout problem.  And the banks most to blame are often smaller banks that made loans to developers, not the more famous subprime mortgages.  Last time I looked the estimated losses to the Treasury from the GSEs was a couple hundred billion, from the smaller banks (i.e. FDIC–which is financed by taxes, BTW) was over a hundred billion, and the big banks was near zero (depending on how much they lose on Bear Stearns.)  That’s all you need to know about where to apportion blame for the bailout crisis.

3.  As far as the high unemployment crisis, the proximate cause is low NGDP, which means the Fed is to blame.  Then we can apportion some blame to Obama for not putting more of his people on the Fed, and not doing it sooner.  But ultimately we macroeconomists are to blame, as both the Fed and Obama take their lead from us.  We were mostly silent on the need for vigorous monetary stimulus in the last half of 2008, and many have remained silent ever since.

The hero is the EMH, as markets warned the Fed that money was way too tight in September 2008.

In the history books it says the 1929 stock market crash triggered the Depression.  After an nearly identical crash in 1987 had zero effect on GDP, we learned that was false.  But it’s hard to blame historians for connecting a high profile financial collapse, with an economic collapse that was barely underway, and suddenly got much worse.   Economists should know better.

Here’s the GDP data I referred to (from Macroeconomics Advisers):

I recently had lunch with a guy who is very bright, and pretty well-informed about economics.  He had seen Inside Job and immediately spotted the flaws.  He started lunch by asking me what I thought about the crisis.  I said; “do you want to talk about the financial crisis that took place three years ago, or the current problem of recession and 9.2% unemployment?”  He responded that most people see the two as related, that one led to the other.

I asked why a financial crisis in 2008 would cause high unemployment in 2011.  He talked about the loss of wealth, that people were cutting back on their spending—basically an AD-side explanation of the recession.

I pointed out that he was confusing “spending” (i.e. consumption), with aggregate demand, which involves all types of output.  It’s not immediately clear that less consumption means less AD.  After all, savings equals investment, and investment is part of AD.  Less consumption is more saving, which means more investment.  At this point the Keynesian readers will be throwing bricks at the computer screen, we will go through a long fruitless discussion of planned and unplanned, and end up with the conclusion that more saving depresses interest rates, which lowers velocity, which lowers AD.  Fair enough.

But that would be equally true if the collapse of the Soviet Union led Russians to hoard $100 billion in US base money.  Yet I doubt you’d find a single economist arguing that that action would cause a recession.  The reason is that if the Fed were targeting interest rates, rather than the money supply, the Russian hoarding would be smoothly accommodated by the Fed, and NGDP would remain unchanged.  In contrast, the financial crisis led to velocity falling at a given interest rate, and thus the Fed might have needed to drive real interest rates much lower in order to prevent the rush for savings from lowering AD and NGDP.

Even so, they did have the power to prevent NGDP from falling 10% below trend in 2011, Ben Bernanke just reiterated that point yesterday.  So why does the banking crisis in 2008 cause low NGDP in 2011?  And then it hit me.  The intelligent but uninformed analysis of my non-economist friend had become the OFFICIAL DOGMA of the economics profession.  To deny that the banking crisis of 2008 causes the low NGDP of 2011 makes you a heterodox economist.  Out on the fringes.

And here’s what’s even more interesting.  I’m simply doing mainstream economics, right out of all the textbooks.  The rest of the profession is making it up as they go along.  Go back and look at your economics textbooks, and see where is says a banking crisis will cause NGDP three years later to be far below the level the Fed wants it to be at.  You won’t find it.  You’ll find 100s of models explaining how the Fed determines the path of NGDP.  But nothing about how banking crises make it impossible for the Fed to determine NGDP three years later.

You might be thinking; “But didn’t Bernanke argue the banking crises made the Great Depression worse?”  Yes he did, but he was also very critical of the Fed.  He argued that once the US cut the tie to gold, they had essentially unlimited ability to raise NGDP.  And of course in this crisis Bernanke has never once argued that the banking crisis prevents the Fed from raising NGDP three years after the crisis—just the opposite.

But that’s where we are.  The common sense view of AD, that when America suffers great losses, it is important for millions of Americans to go on 99 week “vacations” in order to “rebalance” the economy, has become official dogma.

You might argue that the recession is structural.  Yes, that’s possible, but nevertheless the official dogma is that demand is low because Americans are broke.  That’s not structural, that’s AD.  We have way too many houses, with no one to move into them.  No one except the millions of 20-somethings that are doubling up in their parent’s house.  And why are they doing that?  Because they are broke, they have no jobs.  And why do they have no jobs?  Because there’s no AD, because Americans are broke.  And why are Americans broke?  Because we built too many houses, so their value fell.  No one to live in those houses.  No one except all the young people who can’t afford houses because they have no jobs because there’s not enough AD.  And we all know the official dogma says there’s nothing the Fed can do about that lack of AD.  That’s how the banking crisis of 2008 mysteriously causes low NGDP in 2011.

In the future we should just let non-economists write our textbooks; at least if we plan on adopting seat-of-the-pants, common sense views as the OFFICIAL DOGMA of macroeconomics.

BTW, during the decade of 2001-11 housing construction in America was well below normal.  Too many houses or too little income?

PS.  Hang in there for just a couple more days and I promise my onslaught of posts will end, I’ll return to my normal one-a-day.

In several recent posts Tyler Cowen has tried to draw a distinction between how much wealth we believed we had, and how much wealth we really had.  I was somewhat skeptical of his argument, but also thought it had some merit.  Indeed in an earlier post (not posted yet!) I tried to distinguish between wealth we correctly thought we had, which was later lost due to bad policy (1929), and wealth we thought we had, that we never really had (2006.)  Now I have doubts about my argument, indeed I think we might both be wrong.  But I’m not sure.

Consider the following 5 scenarios:

1.  The bank makes a typo, which leads you to believe you have more money than you actually have.  The typo is eventually corrected.

2.  Your family believes it owns 10 1933 $20 gold pieces, worth $80 million.  Later you find out the government has a different view.

3.  The public believes it has lots of housing wealth in 2006, but there was never any prospect that these values could be maintained.

4.  The public believes it has lots of housing wealth in 2006, but later an immigration crackdown followed by tight money reduces housing prices.

5.  The public believes it is very wealthy in 1929, but later the Fed cuts NGDP in half and caused mass unemployment.

A few days ago I thought there was a clear distinction between case 1 and case 5.  Now I don’t know where to draw the line.  Indeed I don’t know if there is a line to be drawn.

I’d like to say the public really was wealthy in 1929, and that later decisions by the Fed destroyed that wealth.  But is that a scientific way of looking at things?  At levels about subatomic particles, we tend to assume that things follow deterministic laws.  Why couldn’t someone argue:  “That national wealth in 1929 was never real, because the Fed is a part of our economy.  It was a dysfunctional institution in 1929, so it was only a matter of time before they screwed up.  We just didn’t know it yet.”

Rorty argued that when people say “Most people think X is true, but I believe Y is true,” they actually mean “most people think X is true, but I predict that in the future people will come to believe Y is true.”  Rorty saw no distinction between what is true, and what we believe is true.

Rorty also believed; “That which has no practical implications, has no philosophical implications.”  So what are the practical implications of the distinction between believing one is wealthy, and actually being wealthy?  Obviously society acts on the basis of beliefs.  So for most people it is a distinction without any significance.  Like the difference between saying I believe X, and I believe X is true.   On the other hand the skeptic who believes the wealth is phony (i.e. predicts it will later be seen as phony), would obviously see practical implications for his belief.  Indeed policy implications.

Imagine I’m debating Tyler Cowen on the question of whether the 2006 wealth was real in 2006.  What’s at stake?  I might argue that the wealth was 100% real, but later policies like immigration crackdown and tight money reduced the wealth later on.  The practical implication is that we might want to reconsider those policies.  Or, one could argue that the extra wealth was only 40% real and the other 60% was irrational exuberance.  In that case the policy implication might shift slightly.  It doesn’t mean easier money couldn’t have helped a bit, but you’d also want to put in place banking regulations robust enough to prevent housing bubbles from damaging the banking system.  Indeed you might also want to do that if the problem was 100% the Fed’s fault, but the necessity would be greater if optimal monetary policy couldn’t solve the problem.

Are we rich if we believe we are rich?  I can’t answer that question.  Rorty would say beliefs are all we have.  Yet he also allows for dissenting voices.  Just because most people get swept up in the housing bubble, and believe ranch houses in San Bernardino are worth $500,000, doesn’t mean Shiller, Krugman, Baker and Roubini have to believe that.  One the other hand, current market values have a very practical implication, they’re what we can sell things for.  In that sense they are real.

Each day that goes by we find out that the previous day’s value of the S&P 500 was wrong, as new information comes in.  Or maybe it was right; maybe it was “true,” based on what we knew at the time.  What’s the TRUE value of the S&P 500?  God only knows.

PS.  I just noticed an interesting shift in wording between the first and second posts that I linked to above.  First post:

We were not as wealthy as we thought we were.

And second post:

We are not as wealthy as we thought we were

See the difference?  If applied to someone in 1933 talking about 1929, I’d have once said the first was false, and the second was true.  Now I don’t know what to think.  I wish Rorty was still alive to help me out.  Now I feel alone in the universe, with no one to provide true answers.

PS.  The second Cowen post that I linked to above didn’t make sense to me.  I left a comment over there—maybe someone can explain the connection to AD.