Mark Sadowski on fiscal stimulus

Mark Sadowski is the only guy I know who actually understands the fiscal data put out by governments.  Here is a great comment he recently left, discussing how fiscal stimulus failed in Japan, whereas fiscal austerity succeeded:

Off Topic.

Izabella Kaminska discusses the Japanification of the Euro Area:

May 16, 2014

Charting Europe’s Japanification
By Izabella Kaminska

“…Meanwhile, there’s also an eye-opening comparison “” which should be of interest to anti-austerity campaigners such as Paul Krugman “” setting out real GDP in terms of investment, GDP and consumption data for the two countries:


As the analysts note, a powerful fiscal stimulus in Japan helped to counter the demand shortfall. That caused personal consumption to continue to grow until 1997 and investment to rebound almost to its previous peak in just six years “” something which isn’t slated for Europe any time soon…”

This caught my eye because the charts, which come by way of Credit Suisse, fail to take into account the idiosyncrasies of the System of National Accounts (SNA), which both Japan and the Euro Area use. Under SNA, investment spending includes both private and government investment spending. The item labeled “government expenditures” in the graphs is actually just government consumption spending, and doesn’t include government investment spending.

Why does this matter? Because in the 1990s the Japanese government engaged in a massive old school Keynesian fiscal stimulus, heavily tilted towards government investment spending (i.e. “infrastructure”). The Credit Suisse graphs totally miss that, and imply that the decline in Japanese private investment spending was much smaller than it actually was.

In fact, David Andolfatto did a post on Japanese government spending, and after I alerted him to how SNA treats government investment he did a followup post:

In particular, I want to draw attention to the following graph:

Japanese real government investment spending increased by over 50% in between 1992Q1 (the peak in RGDP before the recession) and 1996Q2. This was not a trivial amount because unlike in the US and the Euro Area, where government investment spending is only 3.4% and 2.1% of GDP respectively, Japanese government investment spending reached a staggering 9.9% of GDP in 1996Q2.

Now, one can argue that things would have been much worse in the absence of this massive infrastructural spending, but as Kaminska goes on to note, Japan didn’t lose monetary policy traction until much later. In fact the BOJ’s call rate didn’t really hit the zero lower bound until March 1999.

One other thing that I think is worth calling attention to is the fact that the failure to disaggregate Japanese investment spending leads to the failure to notice that private investment spending soared during the Koizumi Boom in the 2002-2008. This is because while private investment spending increased, public investment spending decreased, obscuring a boom in private investment within the aggregate investment statistic. Thus I encourage people to compare the Credit Suisse graph with David Andolfatto’s.

And of course we all remember when Japan did its first ryōteki kin’yÅ« kanwa (QE). That was from March 2001 through March 2006. Coincidence?

Incidentally, in the case of the Euro Area, government consumption and investment spending is the only major component of GDP that is higher in real terms than it was in 2008Q1 (the peak in RGDP before the recession), over six years ago. In contrast, in the case of the US, the only major component of GDP that is lower than it was in in real terms in 2007Q4 (the peak in RGDP before the recession) is government consumption and investment spending.

The difference is obviously attributable to the fact that the US has done QE, and that the Euro Area is only now considering its possibility.

BTW, Izabella Kaminska ends her post by arguing that Mr. Draghi needs to step up to the plate if the eurozone is to avoid Japanese-style stagnation.  I completely agree.

PS.  My only quibble with Mark’s comment is his characterization of Kaminska’s view on the monetary transmission mechanism.  She doesn’t really say when it broke down, only that it was clearly operative at least until the mid-1990s.



18 Responses to “Mark Sadowski on fiscal stimulus”

  1. Gravatar of Major_Freedom Major_Freedom
    18. May 2014 at 07:03

    So I guess the favorability of Sadowski’s point about the importance of disaggregating government spending from private investment would imply that NGDPLT is not the actual standard, but rather the private component of total spending.

    I welcome the day when MMs cease treating spending on imperialist war and boondoggle bridges to nowhere as effectually equivalent to private investment spending on goods and services individuals choose to buy. I won’t hold my breath, because once we cease putting importance on aggregates, it ‘s only a matter of time before it is fully realized that it is relative spending and relative prices that matter for economic well-being, not aggregate spending and price levels, which will then lead to the realization that only a free market in money can accomplish this. That would take the ground away from MM, and expose it as intellectual malinvestment.

  2. Gravatar of benjamin cole benjamin cole
    18. May 2014 at 07:57

    Excellent work by Sadowski.
    Independent central bankers are a menace to prosperity. Egads, the Europeans have agony ahead if they cannot assert democratic control over the ECB. But then maybe the same can be said for the Fed. Ironically, the best major central of the last 20 years?
    The growth-oriented PBoC. But sadly, the PBoC may becoming central bankerfied. It is some sort of professional affliction, a squeamish aversion to prosperity and a peevish fixation on inflation.

  3. Gravatar of Tom Brown Tom Brown
    18. May 2014 at 16:11

  4. Gravatar of Tom Brown Tom Brown
    18. May 2014 at 16:36

    … I should have mentioned that the above link refers to this post.

  5. Gravatar of Peter N Peter N
    18. May 2014 at 21:51

    You have to consider the state of the banks.

    We’ve seen 3 different recent policies concerning zombie banks:

    The US S&L bailout

    The 2008 bailout

    The Japanese post 1992 workout

    I’d give them grades of A-, C+ and D- respectively. Europe seems to be closest to the European model. I don’t think Europe’s Japan-like result is a coincidence.

    And, of course, Major_Freedom has a point. Fiscal stimulus should be used for things people need and want. If it can’t be used for that expeditiously, it shouldn’t be used.

    However a recession is usually the least expensive time to build infrastructure, so if you have stuff you were going to have to do anyway…

  6. Gravatar of ssumner ssumner
    19. May 2014 at 05:13

    Tom, I left two comments over there.

  7. Gravatar of Tom Brown Tom Brown
    19. May 2014 at 06:48

    Scott thanks, I saw them. Mark commented too.

  8. Gravatar of Mark A. Sadowski Mark A. Sadowski
    19. May 2014 at 07:04

    David Andolfatto has responded:

  9. Gravatar of Mark A. Sadowski Mark A. Sadowski
    19. May 2014 at 07:45

    Peter N,
    Zombie economies beget zombie banks. Japan’s banks only recovered during the Koizumi Boom, *after* the economy recovered. Similarly, Europe’s bank bailouts are enormous and are still ongoing. They completely dwarf anything that was done in the US, and the US bank bailouts have long since been paid back (with interest):

  10. Gravatar of Dtoh Dtoh
    19. May 2014 at 08:24

    My rough recollection is that the Japanese boomlet was somewhat real estate driven. Better supply demand balance. Better credit availability after banks worked off bad debt. All led to a pop in real estate prices and construction.

  11. Gravatar of Mark A. Sadowski Mark A. Sadowski
    19. May 2014 at 09:17

    Interesting, because that’s the exact opposite of what actually happened.

    Japan’s Urban Land Price Index (ULPI) fell 26.9% from 2002 to 2008. This compares to the 27.5% decline in the previous six years:

    In fact the ULPI has fallen every single year since 1991.

    And between 2002Q1 and 2008Q1 real residential investment fell by 21.0%.

    In fact when you strip out residential investment, the increase in private investment between 2002Q1 and 2008Q1 rises from 25.1% to 40.0%, or an average annual rate of increase of 5.8%. That’s the fastest rate of increase in private non-residential investment over any six year period in Japan since 1974Q4, nearly 34 years previously.

    When they say “Koizumi Boom”, they mean *Boom*.

  12. Gravatar of Mark A. Sadowski Mark A. Sadowski
    19. May 2014 at 09:24


    “That’s the fastest rate of increase in private non-residential investment over any six year period in Japan since 1974Q4, nearly 34 years previously.”

    should be changed to

    “That’s the fastest rate of increase in private non-residential investment over any six year period in Japan since 1993Q1, 15 years previously.”

    Much less impressive of course, but still reminiscent of the old go-go days.

  13. Gravatar of David Beckworth David Beckworth
    19. May 2014 at 09:35

    Mark A. Sadowski,

    You are killing it! You recently responded to a post of mine about regarding Sufi and Mian’s view of household debt and the Great Recession. It can be found here:

    I was hoping you could send me the individual Euro area data you used to make the weighted Euro area. Thanks!

  14. Gravatar of Mark A. Sadowski Mark A. Sadowski
    19. May 2014 at 11:41

    I sent you an email with the data.

  15. Gravatar of Peter N Peter N
    19. May 2014 at 13:21

    Mark A. Sadowski,

    “Zombie economies beget zombie banks.”

    I’m not sure what you mean here. In all 4 cases the banks got into trouble during the boom preceding the crash. They made bad loans, bought high risk securities and engaged in wholesale criminal activity (NOT nickel and dime stuff). It’s when the economy started to tank that they were found to be overextended.

    In all 3 cases accounting quirks were partly to blame (Unrealized capital gains and regulatory tolerance for nonperforming loans in Japan, Euro zone sovereign bond risk treatment in Europe, off balance sheet entities and abuse of derivatives in the US in 2008).

    The question was how to deal with the insolvent banks. The longer you let the situation drag on, the worse it gets and the more it costs. Both in Japan and in Europe political opposition prevented any resolution. In Japan it was the strong negative public reaction to the (relatively tiny – 680 billion Yen) Jusen bailout. We all know about Europe, by now.

    Which brings us to debt. How much does it matter. Is the accounting rule identity that one persons debt is another persons asset the end of the matter?

    Compare 3 cases:

    1) A borrows $10,000 from B in 2000 and defaults on the debt in 2002.

    2) B gives A $10000 in 2000.

    3) A embezzles $10000 in 2000 and the fraud is not discovered until 2002.

    The cash flow in all 3 cases is basically the same. The differences are that accounts closed in 2000 in case 2, and A receives a compensating accounting gain for the forgiven debt in case 3.

    In all cases the money was spent in 2000. Recognition of the loss by B in cases 1 and 3 doesn’t cause A to have any additional money.

    What did happen was that in cases 1 and 3 A could to some extent (depending on the nature of the debt or nature of the embezzled account) act as if the asset was good with a corresponding gain in credit and liquidity. With loss or default this gain disappears for A, while B usually gains at most additional freedom of action.

    Note that with the loan, the assumption was that A would reduce consumption to pay it off, which would balance B’s additional income, and the only difference would be in propensities to consume and preferences.

    Instead we find that there was more effective spending in the past than we thought and there will be less in the future.

    When you consider that almost all of the over $1 trillion in consumer debt reduction came from default, not repayment, this isn’t such a small thing.

  16. Gravatar of TallDave TallDave
    20. May 2014 at 07:43

    This was not a trivial amount because unlike in the US and the Euro Area, where government investment spending is only 3.4% and 2.1% of GDP respectively, Japanese government investment spending reached a staggering 9.9% of GDP in 1996Q2.

    Heh. This reminds me of something I read recently — apparently in the late 1980s there was a knockdown, drag-out fight between VP George Bush and Don Rumsfeld over whether Soviet defense spending was between 9-11% of GDP or 7-9%. Later they found out it was something insane like 38%, which was comparable to the United States at the height of WW II.

    I’ve heard the Japanese have a bridge built during that investment splurge which cost $1.2B, has a $12 toll, and is used by three dozen commuters. The depreciation and maintenance of all that infrastructure must be truly awesome right now.

  17. Gravatar of dtoh dtoh
    20. May 2014 at 07:47

    That doesn’t sound right. You might want to take a closer look. I’m pretty sure ULPI is nationwide, land only, and not weighted for value. I’m pretty sure if you at land prices for Tokyo or the 6 big cities, you will see an uptick starting in 2004 or 2005. If you look at building prices (average house or condominium prices), it will probably show a more pronounced bump and maybe starting a little earlier. My memory is not perfect, but that was my recollection of real estate prices during that period.

  18. Gravatar of dtoh dtoh
    20. May 2014 at 08:17

    Also, I think you need to look at both commercial and residential real estate investment. My recollection is that Japan came out of a period of over-supply of commercial property and a lot of investment in new construction was in commercial rather than residential. Another point to keep in mind is that in Japan a lot of commercial lending is directly or indirectly tied to the value of real estate held by the borrower so the run up in prices would have significantly increased the availability of credit.

Leave a Reply