Arnold Kling  

Why I Think the Multiplier is Large

PRINT
Cold Spouses... John Taylor: The Fed Done It...

Mark Zandi says


A payroll tax holiday and a permanent payroll tax credit would be effective tax cuts, particularly if designed to help harder-pressed lower- and middle-income households and smaller businesses.

Tyler Cowen also favors payroll tax cuts. I like the idea because there is minimal time lag, it lowers the price of labor, and it keeps the government out of picking winners and losers in terms of industries.

I have zero faith in econometric estimates of the Keynesian multiplier. I think that today's asset-deflation economy is too different from most of what we have observed over the past fifty years for econometrics to be of any help.

Logic suggests that the multiplier today is large. It more likely to be closer to 3 (or above) than to 1.

1. Interest rates on Treasuries are so low that printing Treasuries is tantamount to printing money.

2. Unless the fiscal stimulus is really stupidly targeted, there should be few supply bottlenecks. The demand increase will go mostly to output and hardly at all to prices.

3. Investment is more likely to be "crowded in" by stronger demand than crowded out by higher interest rates.

4. We are likely to see a strong "Minsky effect." That is, higher income should increase wealth and improve balance sheets. This will stimulate spending by both businesses and consumers.*

Having said all that, it is still a bit strange to think of a larger government deficit as a stimulus. If I wrote myself a check for $5000, obviously that would not be a stimulus. If I wrote my daughter a check for $5000 and said that it was coming out of her inheritance, we are saying that would be a stimulus. At a national level, what we are doing is writing checks to our children and taking it out of their inheritance. And, yes, I do think that under present circumstances that will be a net stimulus.

*It is the Minsky effect that is the real kicker here.

Let E be expenditure and let Y be income or GDP. Suppose that without the Minsky effect, we have:

E = $100 + 0.5Y
E = Y

The multiplier is 2, and Y = $100/(1-0.5) = $200.

Now, introduce a Minsky effect, where wealth is denoted by W.

E = Y
W = 5Y
E = $100 + 0.5Y + 0.08W

Now, E = $100 + 0.5Y + 0.08(5Y)

= $100 + 0.9Y

and the multiplier is 10, so that Y = $1000


Comments and Sharing


CATEGORIES: Macroeconomics



COMMENTS (24 to date)
C.M. writes:

Wow. The comments in that NYTimes post are very depressing. *sigh*

James writes:

I've got a better idea. Let's call my income J. Let's call national income minus my income K. Empirically, J = 1/300,000,000 * national income. By accounting, Y = J + K. Hence the multiplier for a subsidy to me is 300,000,000. If our politicians would just do the right thing and give me the trillion dollar stimulus our economy needs, that would increase output by $3 x 10^20.

Of course no sane person would believe that. GDP is not a linear function of my income, even if the ratio looks like a constant. The same applies to the coefficients in Arnold's example. Just because you can construct a ratio between two numbers doesn't mean that the ratio is a constant.

Andrew Garland writes:

You wrote: "At a national level, what we are doing is writing checks to our children and taking it out of their inheritance."

And "I have zero faith in econometric estimates of the Keynesian multiplier. "Yes, I do think that under present circumstances that will be a net stimulus."


It is not going to "my children" or even "our children". It is going to whoever is most politically connected at this time.

Why do "economists" feel that they can judge and alter the time preferences of millions of others? Why is maintaining a constant, upward statistic of "national production" a goal that justifies forcibly changing the plans of millions?

You recommend taking from those who are creating the resources, to make that value part of a statistic today. Do your equations measure the side effect of altered production by those who don't want their future incomes commandeered by the government?

So, it is coming out of our children's inheritance. Just lazy old money sitting around doing nothing, better used (by force) to produce more TV's today. This inheritance is funding the business structure of today. Why should it be better to decrease long-term investment for immediate consumption? Not even my consumption, but someone else's consumption based on my future earnings?

Where is the empirical proof that handing out resources now to the masses produces prosperity later? You say, not econometrics. Just your feelings? Just a cloud of assumptions embodied in 10 lines of x and y equations?

It is easy to provide prosperity today, for some lucky individuals, by emptying out the banks and giving them gifts. But, what do we do for an encore?

KipEsquire writes:

How does a FICA tax holiday help the unemployed who, news flash, are not currently paying FICA taxes?

winterspeak writes:

Arnold:

You really struggle with fiat currency don't you? The government does not use tax money to spend. The government does not need tax money to spend, because it can print its own money. Taxes reduce aggregate demand, they do not fund spending.

Therefore, running up a deficit bears no relation to writing an IOU to the next generation. All it means is that the Government will be printing new dollars.

Where do these new dollars come from? They dilute the value of existing dollars. Since the value of existing dollars going up is called "deflation", and since the Government wants to avoid "deflation" then printing new dollars, and reducing the value of existing dollars, is what they will do. The people paying for this are current dollar holders, not your daughter. And this is how we will avoid the world of the ten cent hamburger.

The US Govt should increase the deficit by suspending FICA taxes, and only reinstate those taxes once CPI starts to tick up.

KipEsquire: FICA tax holiday does not directly help the unemployed, but extended unemployment benefits will. It indirectly helps them by making them cheaper for businesses to hire, though, so hopefully they will not be unemployed for long.

Jeffrey writes:

Thank you, I appreciate this insight, particularly government picking winners and losers.

I do have one question. Why should consumer spending increase? Would not additional personal income, due to payroll tax holiday, just as likely go toward savings/investments? I don't mean to imply S/I is unimportant; however, which is preferable at this time consumption or saving? I seem to remember that the stimulus package this past spring went mainly to savings.

Mankiw seems to find this recession different than the last in terms of investment http://gregmankiw.blogspot.com/2008/12/investment-then-and-now.html

I would like to see personal investment increase as I would consumption. Do you have and idea of the desirable mix between the two?

Thank you,
Jeffrey

Eric Rall writes:

Lowing the employer share of the payroll tax will help the unemployed by reducing the cost to employers of offering them jobs.

Employers lay people off in order to reduce payroll costs. A payroll tax holiday that includes the employer share of the payroll tax would temporarily cut every employer's payroll by about 7% without them needing to lay anyone off.

winterspeak writes:

Jeffrey:

Consumer spending will increase -- but only after they've repaired their household's balance sheet by paying down some debt. That's why it's important to maintain the payroll tax holiday until we start to see some serious spending again, enough spending that we see CPI tick up.

Offering over indebted consumers even more debt (albeit at cheap rates) does not work because they don't want more debt until they pay down some of their old debt first. A FICA tax holiday will give them a chance to do that, and then eventually they will start spending again

pgl writes:

A transitional (Keynesian) effect on Y increases W (permanent income) by a factor of 5? Me thinks your model is mis-specified

Babinich writes:

winterspeak writes (December 14, 2008 12:53 AM):

"Consumer spending will increase -- but only after they've repaired their household's balance sheet by paying down some debt. That's why it's important to maintain the payroll tax holiday until we start to see some serious spending again, enough spending that we see CPI tick up."

Until when? So a threshold on consumer spending is the trigger for ending the holiday?

Doesn't the lack of permanence mean that some consumers will measure their spending and save for the day when the holiday has ended?

Wouldn't a permanent change to payroll taxes, personal/corporate taxes help people and businesses far more than a "tax holiday" which is designed to be rescinded after a set (defined by bureaucrats) period of time?

fundamentalist writes:

I apologize for the long post, but I couldn't find any better response to the Keynesian multiplier than Hayek's response. In essense, the multiplier exists, but it works in the opposite direction that mainstream econ expects. Here is Hayek:

John Stuart Mill’s celebrated proposition that " demand for commodities is not demand for labour " is to the present day one of the most disputed theories of economies…. it was almost immediately assailed, and has ever since been the butt of attack and even ridicule by a long list of eminent economists from Jevons to E. Cannan and J. M. Keynes. It has, however, always had its defenders, including Marshall and particularly Wicksell, and Leslie Stephen even described it, as Mr. Keynes has recently reminded us, as " the doctrine so rarely understood, that its complete apprehension is, perhaps, the best test of an economist… "

“Before we proceed further, however, it will be advisable to re-state Mill's proposition in a form which leaves no doubt about its exact meaning. In the first instance it is probably clear from that use to which the doctrine has been generally put that we are entitled, as we have already done, to substitute consumers' goods for " commodities… "

“If, on the other hand, we decide to measure demand in real terms, as we clearly ought to do so long as we treat the proposition as one of pure theory, it will quickly be seen that the opposite proposition becomes almost a pure tautology. And increase in the demand for consumers' goods in real terms can only mean an increase in terms of things other than consumers' goods ; either more capital goods or more pure input or both must be offered in exchange for consumers' goods, and their price must consequently rise in terms of these other things ; and similarly a change in the demand for labour (i.e. pure input) in real terms must mean a change of demand either in terms of consumers' goods or in terms of capital goods or both,
and the price of labour expressed in these terms will rise. But since it is probably clear without further explanation that if the demand for capital goods in terms of consumers' goods falls, the demand for labour in terms of consumers' goods must also fall (and vice versa), and that if the demand for labour in terms of capital goods rises (or falls) it must also rise (or fall) in terms of consumers' goods, we can leave out the
capital goods for our purpose and conclude that an increase in the real demand for consumers' goods can only mean a fall in the price of labour in terms of consumers' goods, or that, since an increase in the demand for consumers' goods in real terms must be an increase in terms of labour, it just means a decrease in the demand for labour in terms of consumers' goods.”

“We see, therefore, that if we treat the problem in real terms and in its simplest forms, an increase in the demand for consumers' goods not only does not increase but actually decreases the demand for labour… The doctrine still retains its validity, in so far as the effect on the real demand for labour is concerned, if we merely introduce money into the picture” Hayek “Pure Theory of Capital” Appendix III Derived Demand p. 433-437

johnleemk writes:

3. Investment is more likely to be "crowded in" by stronger demand than crowded out by higher interest rates.

I am not well-versed in macroeconomics, but this sounds to me as if it implies we are almost in a liquidity trap? If the government is borrowing money to spend, the higher demand for loans should imply a higher real interest rate. If the interest rate in fact does not go up despite greater demand for loans, it suggests that bankers are holding far too much money which ought to be loaned out...and that we are therefore in a liquidity trap. Or am I looking at this all wrong?

Mick writes:

You guys are still using Keynesian models to try to get us out of this?

lol?

Nothing is clearer from this crises than the utter failure of the Keynesian model, the question is not what new and more hardcore form of fail the government need indulge in, but rather what other models actually reflect reality.

Check that, instead of more and gauranteed failure, we might have an economic model that, **cough**, can be observed working in the real world.

winterspeak writes:

Babinich:

"So a threshold on consumer spending is the trigger for ending the holiday?"

Not sure what you mean -- the trigger for ending the holiday is that CPI starts to increase.

"Doesn't the lack of permanence mean that some consumers will measure their spending and save for the day when the holiday has ended?"

Some people will continue to save, which is why the holiday does not end until enough are spending such that there is enough extra money to trigger inflation. The commitment would not be "fiscal stimulus for two years". The commitment would be "fiscal stimulus until we get inflation". In this way, fiscal policy is run in a similar manner as monetary policy.

"Wouldn't a permanent change to payroll taxes, personal/corporate taxes help people and businesses far more than a "tax holiday" which is designed to be rescinded after a set (defined by bureaucrats) period of time?"

Excessive fiscal stimulus, like excessive monetary stimulus, will eventually create (major) inflation. I'm no fan of our current tax regime, but that's a separate topic.

Demand Side writes:

Winterspeak:

What do you say to the fact that a person with a job who gets an increase in income via the payroll tax cut has a different spending profile than a person who gets a job by way of direct government spending, and so the multiplier will be different?

In the first case, one might suppose that savings and consumer discretionaries would be the object of spending. In the second, the range of purchases from housing to daycare to whatever would be the object.

In the first case, savings produce no follow-on effect, and consumer discretionaries presumably in a downturn may keep somebody in retail from losing their job, but produce no new employment. In the second, already there is one newly employed person.

In the first case an occasional sales tax or gas tax may be paid. In the second, the range of taxes from income to property may be affected.

Another question. In the context of a downturn or depression fear, the mechanics of getting the stimulus into the private economy are quicker with tax cuts or checks in the mail, but the actual economic effect may not be any quicker than simply doing something useful like building protection against the poaching of the planet (no reference to pirates). That is it may be true that the money gets to the consumer quickly, but it may also be that the higher the level of fear, the slower the consumer is to spend. Apparently the rebate checks in the spring took some time to actually enter the economy.

Thanks

Anthony writes:

I'm all for eliminating the FICA tax. But let's eliminate the earned income credit while we're at it. And lets get rid of all refundable credits.

People who need welfare should sign up for welfare, not rely on the tax system. Sure, there might be some extra costs in administrating these welfare programs, but on the other hand we won't have to worry about companies like H&R Block taking away huge portions of people's welfare checks. Hopefully refund anticipation loans would go the way of the dodo bird. Low-income workers would certainly have a simple choice - don't have any taxes withheld, and don't receive any refund. Those who want to save up could still have a portion of their paycheck direct-deposited into a savings account.

Refundable credits are by far the largest factor contributing to the refund anticipation loan. Without them I'd guess we never would have experienced the phenomenon. Additionally, if refundable credits were eliminated, tax preparation fees of low-income taxpayers would probably go down by 30% or more. Fraud would drop too.

winterspeak writes:

Demand Side:

I'm actually with Arnold on your first point -- I think previous estimates of the multiplier may not apply in today's economic situation.

I support a payroll holiday over targeted government spending because i) it's faster to turn on, ii) it's faster to turn off, and iii) it's more likely to flow into real goods and services production than Congressional allotment, and so you will get more monetary and real expansion for your $.

Think of it this way: bridges to somewhere that needs a bridge is better than bridges to nowhere, but you can run up the same deficit to build either. Which do you think is more valuable? Which is more likely via a payroll holiday? Which is more likely via Congressional allotment. Which helps rebuild household balance sheets across the economy? Which helps enrich lobbyists?

The rebate checks last year were totally asinine. I have no further comment on them.

The consumer is certainly slow to spend due to fear, and part of the reason they are fearful is because their household balance sheet has a huge hole in it! This hole needs to be plugged before people will take on new debt -- so we should start plugging it.

Fiscal spending is slow, does nothing to repair household balance sheets, increases money supply, but does not necc. increase the quantity of real goods and services in the economy, and so has limited impact on jobs (outside of politically favored firms). It is also impossible to turn off, so will result in greater inflation later.

A payroll tax holiday is fast, repairs household balance sheets, increases money supply, increases the quantity of real goods and services in the economy, and benefits everyone. It is also easy to turn off.

MattYoung writes:

A bit sardonic (I looked it up), but borrowing from future entitlement grants is about the most direct method to borrow from future generations.

How Keynesian.


winterspeak writes:

MattYoung:

I used to think the same way, but for the Federal Government, deficits are not a mechanism for borrowing from future generations. All the transfer happens today, and it comes from diluting current dollar holders. Life is very different for currency issuers than currency users.

Things are pretty "simple" if we run a model of Economics for Joe the plumber

Barkley Rosser writes:

Dani Rodrik's back of the envelope calculation of about 1.8 still looks pretty good, and has long been about the rough number that urban planners use for a proper export base multiplier for when a new business comes to town that sells out of town.

Demand Side writes:

Winterspeak,

The consumer is not going to lead us out of this recession. "Repairing balance sheets" means retaining stimulus checks in the savings category, does it not? Trillions in lost asset values will not be repaired by a payroll tax holiday, is my guess.

The stimulus value of temporary tax cuts has always been suspect. And repairing balance sheets so consumers get comfortable enough to take on new debt for consumer durables is a long way to the stimulus bang. Who will lend them money anyway with a broken financial system?

The Treasury's borrowing rate is very cheap. They don't need banks to work. They don't need to rebuild their confidence. They can start cutting checks in three weeks. There are ways to get going on bridges to beyond global warming that do not take very long. Simple retrofitting projects, for example.

Appreciate your angle, though.

winterspeak writes:

Demand Side:

You seem to believe that Congress building bridges to nowhere (once they pass an environmental review) is a faster way to a better economy than giving every working person in the United States more money tomorrow, and committing to giving them that money until we can see an effect. You are 110% welcome to that belief.

Keynes, if he was alive today, would not share your view. Infrastructure is to economists what billiard balls are to physicists. You don't see many economists digging ditches, just like you don't see many physicists in pool halls, but certain things make it into text books, and then get repeated zombie-like throughout the ages by academics reading from those same textbooks.

Keynes wrote GT at a time when Federal taxes were very low, the Govt was a small fraction of the economy, labor was fungible, and the US had no roads. Today, Federal taxes are high, Govt is a large fraction of the economy, labor is specialized, and the US has a fair amount of infrastructure. But the cry goes out "dig ditches" as if nothing has changed since the 1930s.

If the last 12 months have proven anything, it's that academic macroeconomists have no clue what they are doing. All the output from that profession has been, essentially, phrenology. Now these same witch doctors parrot what was written 80 years ago, and the press and pundits parrot the professors. It certainly beats trying to understand what is actually happening.

Don writes:

I'm with you Mick. I was reading the comments, feeling drowsy, and then I think I fell asleep. In my dream I was in USSR reading the newspaper.

But wait, I'm awake and reading a libertarian website. Weird.

Comments for this entry have been closed
Return to top