If you want any proof that Koo's message-not just his-is hitting its mark, you have even Sumner admitting that the debt burden is a real thing, not some 'mere epiphenomenon-which is how he used to dismiss it.
Last week, we saw where Mark Sadowski while quibbling around the edges had to admit that the credit impulse has a strong empirical basis behind it.
http://diaryofarepublicanhater.blogspot.com/2014/03/mark-sadowski-reveals-marko-made.html
Now even where Sumner is insisting that Abenomics is working-particularly the monetarist aspect of it-he's promising that it will help with the debt burden.
"Tyler Cowen has a post entitled “Is Abenomics Working?” The first line of the post says:
Last week, we saw where Mark Sadowski while quibbling around the edges had to admit that the credit impulse has a strong empirical basis behind it.
http://diaryofarepublicanhater.blogspot.com/2014/03/mark-sadowski-reveals-marko-made.html
Now even where Sumner is insisting that Abenomics is working-particularly the monetarist aspect of it-he's promising that it will help with the debt burden.
"Tyler Cowen has a post entitled “Is Abenomics Working?” The first line of the post says:
Japan’s economy grew by 0.7% in 2013, down from an initial estimate of 1%.
"That reflects the year over year figures, which makes the “in 2013″ phrasing slightly misleading. The rate is quite low because Japanese RGDP was falling in 2012, before Abenomics was enacted. A better test would be 4th quarter over 4th quarter figures, which show about 2.5% RGDP growth during calendar year 2013. That’s actually a decent number for a country with a falling population, and a working age population that is falling even faster than its overall population. Unemployment has fallen to 3.7%. Only once in the past 16 years was a lower unemployment rate reported."
"Abenomics will not produce RGDP growth miracles due the the bad supply-side characteristics of the Japanese economy and the fact that unemployment is already fairly modest (even accounting for the bias in their data.) But it will boost growth modestly, and has the potential (if pushed more aggressively), to help reduce their debt burden."
http://www.themoneyillusion.com/?p=26342
So while he's always claiming Keynesianism is dead, this is hardly the case if he admits that Koo's right. Actually the whole point of the credit impulse idea is that it's not a liquidity problem in which case the solution isn't monetary. Koo's whole point is that because it's not a liquidity problem but a problem of borrowers paying down debt, a monetary solution won't work.
However that may be, it's nice to see that Sumner acknowledges the reality of this. It shows that Koo's basic point is already very widely accepted to the point that Sumner wouldn't dare to try to deny it-as this would make him look silly in front of his 'peer group'-his fellow economists- as it were that he's still trying to influence for Market Monetarism.
Actually, Koo has some good things to say about Abenomics at least but just not so much the monetary part.
“This is the first government in 20-some years that is trying to do the things Japan needs very desperately,” Mr. Koo said in an interview with The Wall Street Journal."
http://blogs.wsj.com/japanrealtime/2013/12/05/leading-japan-economist-says-abe-on-right-track-but-boj-a-risk/
Yet, Koo is not being wholly rigid about monetary policy though in the past he's been accused of being such by Krugman among others.
http://krugman.blogs.nytimes.com/2010/08/17/notes-on-koo-wonkish/?_php=true&_type=blogs&_r=0
Here he concedes that Abe's monetary policy has had some effect.
“Monetary easing, even though I don’t like it, I have to admit it did produce results” including an 80% rise in stocks to their peak in May, and a 20% drop in the yen against the dollar, the economist said.
Though “what happened is great,” Mr. Koo said it was the result of foreigners buying Japanese stocks, not renewed optimism by Japanese investors.
Krugman had said before that he doesn't get Koo's anathema for monetary easing. Yet my problem with the MMers was that they were always dead set against any fiscal stimulus-with the notable exception of David Glasner who said that while FS was not his preference why not try it along with anything else during a recession like this?
Meanwhile Bernanke has never been against fiscal stimulus at least not categorically. He suggested back in 2003 that Japan have a big tax cut monetized by the BOJ. During our Great Recession he suggested among other things deep loan modifications to heal the housing industry. This would seem to be a great policy in that it could kill more than one bird with its stone. You could not only help people that needed help an repair the housing and mortgage industry but also cut into the amount of debt that needs to be paid off. If you believe that the balance sheet recession is real-and it seems that even Sumner now knows it is-then there is hardly a better policy we can imagine than some kind of debt jubilee-where at least a large sliver of debt would be pardoned.
"notable exception of David Glasner"
ReplyDeleteBeckworth too has advocated a king of "helicopter drop." And Benjamin Cole (guest poster on Marcus Nunes and Lars Christensen) has advocated something along those lines too... a lottery maybe. Well shoot, Ben thinks we'd be better off with counterfeiters if I recall correctly.
Also, Cullen has a couple pieces up on Cullen. The 2nd links to the 1st:
http://pragcap.com/keynesian-myths-and-misunderstandings/comment-page-1#comment-169596
Beckworth's plan was more of an emergency backup idea. I don't think Sumner approved of it.
DeleteAnd even Morgan Warstler had some kind of Fed lottery idea I think: designed to net lose money. Same for Ben's scheme (it's pretty fuzzy now, so I may be getting people mixed up here).
shoot, I didn't proofread well. Should read:
Delete"kind of 'helicopter drop'"
"Also, Cullen has a couple of pieces up on Keynes"
Well Ben is a little different. He's not one of the big shot economists. That doesn't mean I don't like reading him but I mean the econ MMers-Sumner. Lars, Nick Rowe, Bechworth, etc.
ReplyDeleteYes, that's my impression of Ben and Morgan too. But there is a real disagreement between Sumner and Beckworth over this variant of the helicopter drop proposed by Beckworth.
Delete"Last week, we saw where Mark Sadowski while quibbling around the edges had to admit that the credit impulse has a strong empirical basis behind it."
ReplyDeleteAs usual, this is a lie.
The link takes you to a post discussing my comments in which I pointed out that the research suggested that there is a threshold beyond which low monetary velocity (i.e. tight monetary policy) has an adverse effect on RGDP growth.
As far as the credit impulse is concerned, I reiterate that there are serious econometric problems with the concept owing to the use of overlapping observations when estimating the equation. The estimation output produces an extremely low Durbin-Watson statistic suggesting severe autocorrelation.
In fact I have reestimated Biggs, Mayer and Pick's equation using quarter on quarter observations in order to overcome to problem with autocorrelation. The Durbin Watson statistic is fine, but the credit impulse ceases being statistically significant even when using Newey-West standard errors. More importantly the slope coefficient isn't even the right sign anymore.
Mark you haven't shown me to usually lie and neither did you here. I find this kind of tactic tiresome and unnecessary-I know Sumner is a fan of it and you mimic and ape him on everything even his body language.
DeleteI think I may have put up the wrong link in one of the posts I wrote last week or mixed a quote up. However, there's just no question about the credit impulse. Sumner himself is admitting it in the above quote. Last week you admitted there is a strong correlation.
" "Now, as a matter of record, despite the strong correlation, I am somewhat skeptical of the credit impulse concept. I think its theoretical justification is wanting, and I'm not entirely sure it is not committing some econometric sin, since the dependent and independent variables correlate strongly with lagged terms of themselves owing to the fact that they are year on year differences. In fact although Biggs, Mayer and Pick first wrote about their results in 2009, to my knowledge the credit impulse still has yet to appear in any peer reviewed research journal."
http://diaryofarepublicanhater.blogspot.com/2014/03/on-merits-koo-its-mark-sadowski-vs-marko.html
True you still left open a question of whether or not there is real causation. Still despite your reference to its theoretical justification being wanting you never showed this. What you did do was quibble with some of the researchers that gave us the empirical results, not based on this empirical work on the credit impulse but rather based on their personal views.
Scott Sumner,
ReplyDelete"Tyler Cowen has a post entitled “Is Abenomics Working?” The first line of the post says:
"Japan’s economy grew by 0.7% in 2013, down from an initial estimate of 1%."
That reflects the year over year figures, which makes the “in 2013″ phrasing slightly misleading."
It turns out that the BBC, Cowen and Sumner are all wrong. The 0.7% is the rate of increase in RGDP in 2013Q4. RGDP in 2013 as a whole was up by 1.5% over 2012 (Table 2-1.1):
http://www.esri.cao.go.jp/jp/sna/data/data_list/sokuhou/gaiyou/pdf/main_1.pdf
Sumner's main point that year on year RGDP growth was 2.5% in 2013Q4 is correct however.
If you pay a visit to the Wikipedia aggregate demand page you’ll notice a section on debt where a version of the following equation is used:
ReplyDeleteAggregate Spending = Aggregate Income + Change in Debt
To derive something called the “credit impulse”, which is essentially the idea that the change in aggregate demand is related to the change in the change in debt, or the acceleration of debt. This concept has been shown graphically in many financial blogs, and was adopted by Steve Keen and renamed the “credit accelerator”. But note the obvious system of national accounts problem that aggregate expenditures is not equal to aggregate income.
Michael Biggs, economist at Deutsche Bank, introduced this idea in November 2008 as the “the change in new credit issued as a % of GDP”. I’ve replicated Biggs, Mayer and Pick's econometric estimation of credit impulse for the US. On first glance the correlations appear to be statistically significant, but there’s much more to the story, which I’ll get to in a moment.
Back in 2006, Calvo, Izquierdo and Talvi published a paper in which they described a very similar pattern in emerging market economies following financial crises in which output recovered with virtually no recovery in either domestic or foreign credit, a phenomenon that they termed a Phoenix Miracle. Calvo also showed that the Great Depression could be classified as a Phoenix Miracle.
Biggs, Mayer and Pick developed the concept of the credit impulse specifically to explain the phenomenon of credit-less recoveries:
http://www.banccentraldecatalunya.ch/wordpress/wp-content/uploads/2013/Biblioteca_ERC/2013Oktober/11102013/BankOfNetherlands-Working%20paper%20218_tcm46-220409.pdf
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1595980
The second paper more useful since the sources for the data are described in detail.
The credit impulse is essentially the rate of change of the rate of change in credit market debt as a percent of GDP, or the acceleration of credit. So even though the credit impulse may be positively correlated with growth in real consumption and investment, it does not imply that the stock of debt must increase in order for economic growth to take place. On the contrary the credit impulse was developed specifically to demonstrate the importance of credit in credit-less recoveries.
The credit impulse’s theoretical justification is obviously wanting, since the above equation is a clear violation of the system of national accounts. In addition, since the dependent and independent variables correlate strongly with lagged terms of themselves, owing to the fact that they are year on year differences, there are serious econometric problems with its estimates. In fact although Biggs, Mayer and Pick first wrote about their results in 2009, estimates of the Biggs’ version of the credit impulse, with its overlapping observations, still has yet to appear in any peer reviewed research journal.
When I repeat the very same analysis of US data over the period 1954Q1 through 2008Q4 as Biggs, Mayer and Pick, I find that the Durbin-Watson statistic is far too low (0.48) indicating there's a serious problem with autocorrelation. This should not be too surprising because of the overlapping observations (i.e. everything is year on year). Biggs, Mayer and Pick don't mention this problem at all in either of their papers. The only indication that they tried to deal with the problem is they report using Newey-West HAC standard errors.
When I re-estimate Biggs, Mayer and Pick's equation using quarter on quarter observations, instead of year on year observations, in order to overcome to problem with autocorrelation, the Durbin-Watson statistic is much better (1.36), although still too low to be acceptable. But the credit impulse ceases being statistically significant. More importantly, the slope coefficient isn't even the expected sign anymore.
To give you a feel for how spurious the credit impulse correlations are, here’s a graph of quarterly changes of nominal wages and real productivity at an annual rate from 1965-2013:
Deletehttp://research.stlouisfed.org/fred2/graph/?graph_id=165127&category_id=0
When one regresses wages on productivity the estimation is nowhere near being statistically significant, and this is not surprising from a theoretical standpoint. Despite the fact there are no overlapping observations, the Durbin-Watson statistic (0.42) still indicates there is a serious problem with autocorrelation. First differencing each variable totally eliminates the problem and the results remain statistically insignificant.
Here is the graph of the same variables but this time in year on year form:
http://research.stlouisfed.org/fred2/graph/?graph_id=165126&category_id=0
Now the variables are negatively correlated at the 5% significance level. The Durbin-Watson statistic is 0.05. Simply by making the observations overlap so that autocorrelation becomes a very serious problem, we can get two variables which should be completely unrelated in theory to be highly correlated.
This is the problem with the credit impulse in a nutshell.
The credit impulse is directly analogous to the fiscal impulse , and there have probably been thousands of papers investigating the effect of the fiscal impulse on growth. Various definitions of fiscal impulse are used , various choices are made as to data frequency and lags , the statistical boxes are mechanically checked , and then comes a robust debate on the effectiveness , or lack thereof , of fiscal policy. The conclusion most authors come to on the fiscal impulse depends on their ideology , not the empirics , and certainly not on the DW or any other of the many statistical methods used. The same will be true of the credit impulse studies , as their numbers grow.
ReplyDeleteMonetarists can be expected to come down on the side of "lack of effectiveness" as regards the fiscal impulse or the credit impulse. Endogenous , debt-based money driving economic growth is simply not their bag.
This is the problem with the credit impulse in a nutshell.
Marko
Marko
An interesting new paper from the Bank of England :
ReplyDelete"Money creation in the modern economy"
http://www.bankofengland.co.uk/publications/Documents/quarterlybulletin/2014/qb14q1prereleasemoneycreation.pdf
I wonder if Mark Carney signed off on it. It seems a little blunt to me :
"This article explains how the majority of money in the modern economy is created by commercial banks making loans."
Under the ZLB :
"QE is intended to boost the amount of money in the economy directly by purchasing assets of the rich.........raising the price of those assets and stimulating spending by the rich in the economy."
Marko
( OK , I sneaked in the bits about "the rich" , but we all know that's what they're talking about. )