Japanese fiscal policy in the 1990s

Paul Krugman recommends this chapter by Adam Posen, which gives a good overview of the history.  The piece argues at length that the Japanese didn’t try much expansionary fiscal policy during their downturn of the 1990s.  On p.49 comes the evidence that fiscal policy works, at least as it was tried in 1995:

In the end, the September 1995 stimulus package did add significantly to economic growth in 1996.  Not only was the actual real GDP growth of 3.6 significantly higher than the 0.9 percent recorded in 1995, it was at least 0.9 percent higher than the growth forecasted for 1996 by all of the major international institutions and the financial consensus…This stimulative effect can largely be attributed to the fiscal package, although the decline in the yen also stemmed the decline in net exports (by -1 percent of GDP in 1995 and by -0.4 percent in 1996)…There was actually no other source of positive impetus to the Japanese economy in late 1995 and early 1996 that can be identified except discretionary fiscal policy.

A few observations:

1. This is a piece of evidence in favor of fiscal stimulus and so we should take it seriously.

2. It is, quite literally, only a single data point.

3. In November 1994 there was a big cut in personal income taxes and that may be responsible for some of the increase in economic growth in 1995-6.  (There was also reconstruction from the 1995 Kobe earthquake, as one reader notes in the comments.)

4. Japan was much weaker automatic stabilizers than does the United States.  Some of the fiscal policy boost was to strengthen those economic stabilizers.  The case for doing that is indeed much stronger than the case for initiating new government expenditures in the form of specific projects.

5. The history is fully consistent with an alternative interpretation, as I have discussed in my post on the fetishization of measured gdp.  Namely, the Japanese spent more money putting unemployed resources to work on construction projects.  Measured gdp went up, but the Japanese didn’t get much of value for their money.  (Japanese construction projects from this era are notoriously ugly, wasteful, and unpopular.)  The spending also didn’t set off any kind of lasting recovery.  It was the proverbial ditch digging without much in the way of later-order benefits or multipliers.  In these circumstances a boost in measured, temporary GDP is very different from an economic recovery. 

6. There is a deeper question of why governments so often back away from aggressive fiscal stimulus, if that policy indeed will bring so much recovery and thus bring in so many votes and so much revenue.  Posen in his chaper suggests that ideology is at fault but I am not convinced.  After all Japan is not ruled by Grover Norquist.  The alternative null hypothesis is simply that governments see the fiscal stimulus is not working.

Anyway, that is the evidence we are being asked to spend $600-700 billion on — or $2 trillion for some –so I thought you should see it.

Addendum: Here are very good comments from Greg Mankiw’s blog.


A fair amount of the increase in Japanese government spending in 1995 was for repairs of damage from the Jan 17, 2005 Kobe earthquake. Does that count as aggregate demand stimulus?

Japanese government spent 5.2 trillion yen (= $58 billion) for Kobe earthquake from FY94 to FY99, 1 trillion yen (=$11 billion) among which was spent in FY94 (EOFY=March 31 in Japan). The fiscal stimulus in September 1995 was 14.2 trillion yen (=$158 billion) which was much more than the restoration from the earthquake.

Where is the indication that we have an infrastructure problem at all? Take a look at the flash-animated history of the US Interstate Highway system at this DOT site. Compare 1950 with 2000. Just what is this monstrous, new roads and bridges spending going to buy us? If you ran the government like a business and wanted to make a profit (pay down the debt), would you spend hundreds of billions on roads and bridges?

Next, take a look at our new Secretary of Education and his historical budgets. What has he done with his money? What is the price elasticity of performance for public education? What is the indication that he couldn't simply redirect his existing budget and build new schools with the money he already has? If you ran the government like a business, would you spend billions on this?

In the end, you must run the government like a business or you do what Ecuador just did and default on your loans. You cannot run at a loss every year and not pay the price. Borrowing can make sense if you invest in something that pays off the loan and leaves you a profit. Borrowing on things that cannot possibly pay you back is a sure path to failure.


I don't know Keynesian theory well, so I'm wondering if part of the 'multiplier > 1.0' argument rests on the idea that the spending is "investment" on ugly infrastructure.

I think this multiplier hogwash is built on fuzzy Depression nostalgia and models that work only on some University's mainframe computers. No successful business does mammoth capital investment based on generalized models like this. Can you imagine Intel deciding to simply spend money because a model tells them that spending $X randomly brings in $Y? Or do they instead project revenue streams based on their investments?

We're about to pour barrels of money into a shredder.

1) There are a number of single data points that point to the efficacy of fiscal policy. Individually none are convincing, collectively and with the backing of what seems to be solid theory I think the case for the efficacy of fiscal stimulus is while not exactly strong, adequate.

2) The burden of proof argument doesn't make sense to me. If we demand the level of proof that you require we will never find out whether fiscal stimulus works. Because we can't try until we know it works but we can't know whether it works because it hasn't been tried. I guess the world should have used tariffs as the chief source of government revenue because we didn't have any examples of the welfare benefits of freer international trade until trade was made freer.

3) The projects being proposed have varying levels of merit. In addition there is almost certainly a non-zero economic stimulus effect. In addition costs are going to be lower now than they have been previously. For reasons I don't really understand you seem to characterize these "ands" as an "or" with people vacillating between arguing for projects based on their merits and based on their stimulus effects. Yes the rushing required to make a project effective as stimulus compromises the amount of utility provided by the project itself. The question isn't "is the project effective as stimulus?" or "is the project worth doing on its own merits?" rather it is "is it worth compromising the utility provided by spending to take advantage of stimulus effects, and if so to what extent?"

4) We are not being asked to waste a trillion dollars. We are really being asked to waste some unknown percentage of a trillion dollars in pursuit of making the recession shorter and less severe. In the case of military expansion as fiscal stimulus in history (which seemed to work in at least a couple cases.) There was almost certainly more "waste" than there would be even when compared with infrastructure/energy/telecom projects even if these projects are necessarily suboptimal.

Take it from a lower level bureaucrat operating in the budgetary apparatus of a northeastern state-this is NOT possible.

No matter how close to pure monopoly a business is, it never can force its customers to buy its products and services through compulsion, conscription or the threat of incarceration.


Rothbard couldn't have said it better. Why don't you get a job, I mean a real job, in the real economy a k a the private sector?
All governments are criminal gangs, as Rothbard pointed out.

GOV-ern-ment n. a court 'n cop monopoly which gains its resources by two types of theft, taxation and inflation, and which rewards its supporters with subsidies and punishes its enemies by more theft and sometimes worse.

(Weber forgot the theft part, but Oppenheimer didn't.)

If we should be learning lessons from Japan's history when considering how to deal with our own recession, and if the Kobe earthquake provided a boost to Japan that helped their recovery, then our optimal course out of the recession is clear.

We need to instigate giant monster attacks upon major U.S. cities.

Perhaps Japan could lend us a few? I suggest Godzilla in New York, Mothra in Los Angeles, and Gamera in Las Vegas.

Superheater, I agree with your comment, up to a point. A government is not perfectly analgous to a business. However, it still has to look at revenue and expenses or else it ends up like the USSR. At this point, when the debt is well over $10T, I would suggest that we have indeed tried a massive stimulus package (the aforementioned $10T) and it hasn't worked.

A "data point" is properly speaking a "datum." Anyway, a datum here, a datum there, after awhile they can even add up to real data.

As for all that ugly Japanese infrastructure, heck, I guess Obama needs to hire some highway beautification experts, although I have no idea what their multiplier effects are, Lucas critique and all that.

Oh, and I think private businesses regularly assume greater than one local multipliers all the time in making investments in local communities where other businesses have just made investments that will increase exports from the local community. I guess they need to go read Lucas and get critiqued

"private businesses regularly assume greater than one local multipliers all the time in making investments in local communities where other businesses have just made investments that will increase exports from the local community"

I'm not sure what you are talking about.

While extrapolation of short term data is certainly done by entrepreneurs, there is no reason to believe that their mental models don't include heuristics of the kind: "*This* short term growth is the result of the government spending which must stop at some time, so we should be careful in extrapolating from it."

The fiscal stimulus plan seems to come down to a multi-trillion dollars bet that entrepreneurs don't use such heuristics.

2) The burden of proof argument doesn't make sense to me. If we demand the level of proof that you require we will never find out whether fiscal stimulus works. Because we can't try until we know it works but we can't know whether it works because it hasn't been tried. I guess the world should have used tariffs as the chief source of government revenue because we didn't have any examples of the welfare benefits of freer international trade until trade was made freer.

This is not correct. Even if tried, we can't establish causality so clearly. We can attempt to deduce the effects of stimulus from past experiences and economic models, which is what Tyler has been doing. If you want to aggregate the knowledge of economists like Tyler, perhaps a prediction market would be of some use (and would certainly be less costly than billions of dollars wasted).

1) There are a number of single data points that point to the efficacy of fiscal policy. Individually none are convincing, collectively and with the backing of what seems to be solid theory I think the case for the efficacy of fiscal stimulus is while not exactly strong, adequate.

Considering that today the share of government to GDP is significantly higher - in the range of 30-45% of GDP, the level of indebtness of the nations is highest in the history (which means the economies were stimulated constantly over the last 50 years), don't you think that if stimulus works, there should be no problem with growth now?

As to Tyler's use of GDP as a measure of "effectivity" of fiscal stimulus, isn't it possible that this is a statistical ilusion in the same sense as Robert Higgs describes in his latest econtalk interview?

A government run like a business, one that starts to measure the success of its efforts exclusively or
primarily by its effects on its CAFR (Comprehensive Annual Financial Report) will become as
abusive and as one that ignores the economics of its operations will become stupid.

KT Cat, note my comment above. I think we are in agreement.

We are really now into Keynesian economics.. we have practical men in the federal apparatus.

"Practical men... are usually the slaves of some defunct economist"

I lived in Japan from 1991 to 2000. The Japanese tried 4 separate fiscal stimulus packages as well as monetary policy of near zero percent. Much of the fiscal stimulus went into construction projects (10% of Japan's workforce worked in construction, much like the 10% of America's workforce in in financial services). There was a brief spike of growth in early '96, but it did not last and it did not affect most people.

The result of Japan's fiscal stimulus was lots of new high-rise buildings in Tokyo and other cities.

Having lived there, I can tell you. Japan's fiscal stimulus and monetary policy had absolutely zero effect on the economy. Trust me on this one.

All the fiscal stimulus did was to take Japan's government finances from a budget surplus in 1991 to a deficit of 180% of GDP (comparable to Italy) by 2001, just in time for the mass retirement of much of their workforce.

The keynesians are talking out their arses.

No, this is a matter of businesses locating in, or expanding their capacity in, metro areas where other businesses are making export-base investments, which will generate multiplier effects in the metro area (and most first round multiplier effects are local). They are taking advantage of a growing market, and they take account of multiplier effects in doing so.


There is no need to invoke a "multiplier effect" in making note of this commonsense commonplace, namely that many businesses locate in metro areas and near to export markets. This is how businesses operate.

But the Keynesian argument is that we need government to spend us out of the recession. Ignoring the facts that (1) it was government that caused the recession, and (2) all legitimate (drug laws, organized mass murder campaigns, and knocking off other governments don't count) services performed by government, including "national" defense and provision of roads, can be performed better and more morally by the private sector, the Keynesians overlook the economic discoordinations caused by subsidies, the problem of having to separate winners from losers (and governments tend to side with and subsidize losers and tax winners), and that in fact government expenditures are, as Rothbard put it, waste consumption.

Tyler obliterated the public goods fallacy in, I think, a 1985 article. (Link, Tyler?) Resources are unemployed in a recession/depression because of previous malinvestments and economic discoordination caused by government. Letting markets work, and above all abolishing central banks, is the surest way of returning to full employment--and staying on such a path.
Say's Law does work in a free market, which is one with free banking and no central bank or treasury.
So let's press on to anarchy!

Barkley R.,
I know about regional economics, growth, etc., having studied them, as well as managed the NY operations of one firm with regional, national,and international operations. I'm also an avid investor in similar companies, so I'm guessing I know at least as much about how they operate as you think you do.

You write, insightfully:
Investments in businesses that serve only the markets within the metro areas in question have
much smaller multipliers.

You don't say? You're a bright guy, and I never would have known that. Thanks for the insight, I'm eternally grateful. Will you get a Nobel for this? I mean Krugman's was bestowed for something not too dissimilar.

Then you opine:
I suspect that this is a matter of you not knowing much about urban
or regional economics.

So that's it! Two in a row, you're good!

Secondly, it is not the size of the metro area, but the fact that it is

Sorry to have confused you in my haste, as I was heading into a conference call.
You're right--in my business/investment career, I never encountered the concept of growth.

Now, if you're really bright, you'll figure out someday that SS is not just a Ponzi schemola, but actually much worse, because people are forced to participate, unlike Madoff's operation.
And Thomas Donlan hit the nail on the head in Barron's this week when he pointed out that the whole government is one big Ponzi swindle. Admittedly that intelligence is too refined for a crypto-Keynesian.

Barkley Rosser,

If you admit that we don't know the multiplier, why are people so eager to bet trillions of dollars on such shaky grounds? And I didn't say that Lucas told us what people think -- I said that the fiscal stimulus assumes that people don't use heuristics like: "*This* growth is the result of the governmental stimulus, which must stop at some time, so we should not extrapolate too much from it"

In other words, you (and others) want to bet trillions on the idea that people don't realize the source and nature of the stimulus investment. That isn't anywhere close to the "solid science" epithet used by the earlier poster -- it is just pseudo-science.

I understand that it is easier to engage in a criticism of rational expectations, but such strong assumptions have nothing to do with the discussion.


We may not know exactly what a particular multiplier is for sure a priori,
but I think it is extremely unlikely that at this time in the US, one for
infrastructure investment in the US will be less than one, and heck, even a
multiplier of one implies an increase in GDP as a result of the spending, even
if it does not trigger all kinds of additional expansion. In a situation where
the economy is collapsing, that is not such a bad thing and why there is so much
support for this. We certainly have never seen a multiplier of zero or negative
in this country in the past, and I see no reason to expect it now.

Bill S.,

Well, I am not going to repeat about Ponzi schemes and social security. I appreciate
that if one thinks government is evil, then its biggest program must be evil also,
and can be called all sorts of names with bad reps, even if they do not really fit
if one insists on using the names according to their real meanings. And, yes,
"Ponzi scheme" does have a precise meaning that fits Madoff's scheme.

I note that in your description of what Mises says entrepreneurs do, you list "financial
studies." What financial studies? These do not involve "mathematical relationships"?
Are you telling us that when WalMart or StarBucks decides whether or not to open new
outlets within given communities they are not making forecasts based on assumptions about
future local growth based on recent capital investments by other businesses in those
communities? And, even if multipliers do not have fixed values, if an investment triggers
greater consumption spending, the multiplier effect has happened, and the data suggests that
this has happened many times in many places, even if its amount has not always been constant.
That something does not remain constant does not mean that it does not exist.

BTW, I love it how people put "crypto" in front of various terms. It always makes them
sound even worse than if one simply used the term itself. Which is worse, "crypto-communist"
or just plain old "communist"? How about "crypto-libertarian"? Does that sound bad even to
a libertarian? "Crypto," boy, it just makes whatever it is attached to sound creepy.

And if you want to call me a "crypto-Keynesian," well, on another thread I made fun of a set
of "complexity people" who might have Post Keynesian and behaviorist tendencies (aka, "crypto,"
eeeeek!). I did not say so, but I am probably one of the small set of people who fit that
descrypto, so you can call me whatever names you like, Billy, crypto or decrypto, :-).

Close italics? I hope.

Comments for this post are closed