Arnold Kling  

Morning Commentary

Ayn Rand Discovers Public Choi... Preview of Book 2...

Bill Easterly is not of fan of military nation-building in Afghanistan. He writes,

critics of top-down state plans for economic development are also not fans of top-down state plans for military development. If the Left likes the first, and the Right likes the second, that just shows you how incoherent Left and Right are.

Winterspeak writes,

to keep responsible lending, we put private capital in front of public capital and ask that private capital take the first loss on loans it makes which turn out to be bad. Ultimately, taxpayer money is there as backup, but it should not be directing investment. We call this institutional arrangement a "bank".

This simple sensible construct is utterly lost on policy makers and the commentariat alike. For banking to do the job it is meant to do (ie. make loans that will be paid back), a bank should be required to keep all loans it makes on its books until maturity.

Robert Pozen gave a talk where he offered ideas to make securitization work better. I asked why we should want securitization. He said that when a bank securitizes a loan, it can turn around and lend the money again, increasing lending. I asked why this is a feature, rather than a bug. Pozen did not answer. He is against excess leverage and in favor of securitization, yet seems to be unaware of the tension between those two positions.

Scientific American writes,

If there were a massive conspiracy to defraud the world on climate (and to what end?), surely the thousands of e-mails and other files stolen from the University of East Anglia's Climatic Research Unit and distributed by hackers on November 20 would bear proof of it. So far, however, none has emerged. Most of the few statements that critics claim as evidence of malfeasance seem to have more innocent explanations that make sense in the context of scientists conversing privately and informally. It is deplorable if any of the scientists involved did prove to manipulate data dishonestly or thwart Freedom of Information requests; however, it is currently unclear whether that ultimately happened. What is missing is any clear indication of a widespread attempt to falsify and coordinate findings on a scale that could hold together a global cabal or significantly distort the record on climate change.

Pointer from Mark Thoma. The headline for the article is "Seven Answers to Climate Contrarian Nonsense." As the tone of the headline suggests, the article preaches to the converted and attacks the weakest arguments of the skeptics. [UPDATE: As Julian Sanchez once wrote, there is such a thing as the "weak man" argument.] For example, in the paragraph above it says that skeptics see a climate-change conspiracy, and such a conspiracy is not possible. In fact, what skeptics see is a scientific discipline that is self-defined by the position it takes on global warming. This self-defined science requires no conspiracy. I have seen it happen in economics. Macroeconomic modeling as of 1970 was self-defined as Keynesian orthodoxy, with no conspiracy taking place. The fact that the CRU leaks uncover some conspiratorial emails and computer code is what leads me to say that climate science is at least one standard deviation worse than what would be normal in economics.

On the other hand, if your goal is to persuade people with whom you disagree, you have to focus on your own weakest arguments and show why, in spite of the problems in your position, you still think that it is reasonable. My problem with the climate believers is that I continue to see a reliance on models that are "tuned" to fit data. I want to see those models make interesting predictions that pass empirical tests.

Moritz Schularick and Alan M. Taylor write,

money and credit began a long postwar recovery, trending up rapidly and eventually surpassing their pre-1940 levels compared to GDP by the 1970s. Credit itself then started to decouple from broad money and grew rapidly, via a combination of increased leverage and augmented funding via nonmonetary liabilities of banks. In recent decades we have been living in a different world, where financial innovation and regulatory ease has permitted the credit system to increasingly delink from monetary aggregates. By 2007, the typical level of broad money had risen to about 70 per cent of GDP, but bank loans now exceeded 100 per cent, and bank assets over 200 per cent.

Again, thanks to Mark Thoma for the pointer. A longstanding debate in economics is whether money matters or credit matters. The MV=PYers tend to focus on money, and the rest of us are willing to look at credit. Read the whole Schularick-Taylor article.

COMMENTS (12 to date)
fundamentalist writes:

Schularick and Taylor: “Past credit growth spectacularly improves the forecasting power of an early warning banking crisis model in our data. Using long-run historical data, the growth rate of lending emerges as the single best predictor of future financial instability, a result which is robust to the inclusion of various other nominal and real variables.”

Well, with enormous effort mainstream econ is beginning to catch up to Minsky and Kindleberger 40 years afterwards. In another decade or two, they might catch up to Hayek who wrote the same thing in the 1930’s in Monetary Theory and the Trade Cycle. Ain’t progress wonderful?

Ryan Vann writes:

Why is there a debate about whether credit or money matters when, in practice, they are essentially the same thing.

fundamentalist writes:

Ryan, I agree, as does Austrian econ. I never understood why they separate the two.

steve writes:

It seems that the same people who can't imagine the decentralized co-ordination of markets can't imagine the decentralized co-ordination of conspirators. Perhaps this is no surprise.

SydB writes:

"attacks the weakest arguments of the skeptics"

I increasingly think that your currency is slogans and pop terminology (suits/geeks, etc etc) rather than actual substantive analysis. Your latest is to--in an almost postmodernist sense--rewrite any oppositional writing in terms of your world view, one in which the opposition is now attacking the "weakest arguments." I see this more as name calling than anything else.

"one standard deviation worse than what would be normal in economics."

Here you use statistical/scientific language, a "standard deviation." But your use--unless I'm mistaken--is entirely rhetorical. What standard deviation are you talking about? In what terms? What criteria are you using?

Or is this all just anecdotes and rhetoric?

wm13 writes:

"He said that when a bank securitizes a loan, it can turn around and lend the money again, increasing lending. I asked why this is a feature, rather than a bug. Pozen did not answer."

The answer is, because origination and investment are two different skills. So first, there is no reason to believe that an organization that is good at the first will be good at the second. And second, even if an organization is good at both, there is no reason to believe that the portfolio it originates is the portfolio it wants to invest in.

Ryan Vann writes:


Perhaps you should re-read the paragraph from which the phrase, "attacks the weakest arguments of the skeptics" comes from. I thougt Mr. Kling did enough substantiating for a blog post; he even posed an argument that he believes is stronger.

I'm not sure how using the concept of standard deviation is anecdotal, but it does seem a cutesie interjection of statistical vernacular. I'm not sure exactly what was meant either. What do climate scientist deviate from exactly (the mean doesn't make much sense to me)? Truth?

mark writes:

"For banking to do the job it is meant to do (ie. make loans that will be paid back), a bank should be required to keep all loans it makes on its books until maturity."

This statement is very simplistic and bad policy and also philosophically at odds with the blog's libertarian preferences.

Two examples of why it is overly simplistic. First, placed as it is after a passage on securitization I wonder if the poster and the blogger understand that transferability and securitization are two different things (indeed, credit, transferability and securitization are three different things). One can transfer to an unleveraged buyer - what would be wrong as a matter of policy in that? How would restricting that transfer enhance "banking"? Suppose on day one, a bank invests all the deposits it takes in in a 30 day loan to a government at a 1% interest rate. If on day two, a commercial customer comes in that is willing to pay a 2% interest rate for 30 days, why should the bank have to wait for enough other deposits to come in to make that loan, as opposed to selling the first one off?

Second, it ensures duration mismatch and liquidity problems except to the extent it forces banks not to make loans of more than a few days'maturity. A banking system with demand deposits cannot make many long term loans and hold them to maturity without a liquidity source. If you were in finance in the Volcker era you know this.
As one of the commenters above noted, banks have an infrastructure for originating loans. Other financial intermediaries with different liability profiles such as insurance companies and pension funds have an appetite for longer maturity paper but no infrastructure for originating it. Why should the latter be made to duplicate the banks' origination infrastucture?

Simon K writes:

I don't see a fundamental problem with securitization. It seems to be a reasonable way to transform big, unique, illiquid, long term loans into small, liquid bonds, without the originator needing to carry interest rate risk, as they would if they simply financed their lending by selling their own short term debt. The problems with the latter model were demonstrated in the S&L crisis, which is why we ended up moving to securitization. It should spread risk.

Of course, it didn't work out that way, but the problem was specific to the particular model of securitization and the reasons it was adopted, not the general idea. The model that was in use (sadly still is for prime loans) encouraged originators to ignore default risk, because they didn't have to carry any of it. In fact they could make the balance sheets look better by securitizing their loans and buying the resulting securities. That should have tipped people off that something was wrong ...

In spite of the duration matching advantages of securitization, though, I doubt it would have become nearly as popular if it weren't for regulatory capital arbitrage advanatges it gave banks.

Pete writes:

I think this is the second time you have mentioned code manipulation with the stolen emails. You should probably read: or this:

Ryan Vann writes:

I'm not buying the origination vs investment story; if it were the case, securitized instruments wouldn't have been nearly as pervasive as they were. In other words, there wouldn't be so many companies with so much securitized debt instruments. Likewise, you wouldn't have instruments like life insurance securities if you really thought some firms weren't able to originate the necessary underlying assets.

If anything, life insurance companies are far more capable of producing securities because they don't have demand deposits, and people tend to keep insurance over a long time horizon.

Securitized debt basically performs two functions; it allows for risk pooling. This helps the credit rating of the underlying debt. The latter can lead to a higher risk appetite, which is partially what led to much of the current recession. Secondly, it is another way to raise capital. These two characteristics explain why so many firms deal in them.

George X writes:

Scientific American wrote:

If there were a massive conspiracy to defraud the world on climate (and to what end?), surely the thousands of e-mails and other files stolen from the University of East Anglia's Climatic Research Unit and distributed by hackers on November 20 would bear proof of it.

In other words: where there's smoke, there's no fire.

Comments for this entry have been closed
Return to top