Two notes on liquidity traps and decreases in wages

1)  A fall in wages
increases the incentive to hire (call this the substitution effect) but it
decreases the income of people who already have jobs and this in turn
decreases their spending and other people’s income (call this the macro income effect).  In essence, Krugman and others are arguing that the macro
income effect can dominate under certain situations.  See Tyler (here and here) and Scott Sumner (Whom I cannot resist quoting–"no respected macroeconomic theory has ever been so decisively refuted
by the data as the theory that high wage policies can actually help the
economy during a Depression") on this point.  I, however, will ignore this debate and take it as given that this is possible.

Now one reason that people are talking about
a cut in the minimum wage in this context is that it follows exactly the two-part logic given above.  But bear in mind that the real policy choice
we have is to cut the payroll tax and cutting the payroll tax increases the
incentive to hire and increases the income of people who already have jobs.

2)  The idea of the
liquidity trap with its notion of lack of movement, inactivity and firms which "just aren't hiring" is hard to reconcile
with the fact that millions of new jobs are being created every month.  I have said this before but I should have shouted it from the rooftops because this error is very common.


"A fall in wages increases the incentive to hire...."

Not if prices fall in proportion, surely?

...surely prices won't fall in proportion.

The response of prices to a fall in money wages varies from model to model. This discussion might get somewhere if the advocates of wage-cutting would present an explicit model, but they seem remarkably shy about doing so. Krugman and Sethi interpret them as using an (obviously inappropriate) partial equilibrium model. Maybe so, or maybe they have no model at all in mind. Perhaps they are relying on gut instinct.

1. So far you haven't addressed the empirical work by Krueger et al. regarding the minimum wage...that its previous increases did not result in job lossess.

2. So far, you haven't addressed the composition of unemployment and tied it to the composition of low wage earners (those who earn the minimum wage).

3. So far, you have not decomposed the minimum wage from all the costs of adding another employee...wages are not everything an employer pays for adding a new employee, nor are lost training costs when he lays one off;

4. So far, you haven't explained how the sectoral overcapacity in some segments of the economy would have the need for a new employee or would hire an unskilled worker;

5. So far, you haven't tied this to a world economy in which low wages workers in foreign countries will still have lower wages than US workers, so that our only path is to improve the skill levels of US workers, regardless of the minimum wage.

6. So far, you haven't explained how the earned income tax credit, which will kick in when employees income drops, will not just bring us back to the same position, since the government will now be subsidizing a low income employer;

I can barely tolerate monetary policy as monetary policy. Wage policy as monetary policy just sucks.

Bill, if a minimum wage increases during a boom, would you expect employment to drop? And if the minimum wage doesn't do anything, then what's the big deal?

Could someone explain how the cited job turnover numbers disprove the existence of a liquidity trap? It sounds like Alex is saying that because individual firms are hiring, it doesn't make sense to talk about the effects of a liquidity trap in the economy as a whole. Is Alex's logic as flawed as it seems at first blush?

And doesn't Alex's point about job turnover undermine his argument that a change in the minimum wage won't reduce the wages of the currently employed?

Kevin, if you cut the payroll tax AD goes up, not down.

But it doesn't decrease "the income of people who already have jobs", so a cut in the payroll tax can't be what Alex has in mind in his first paragraph, where he is describing the debate so far. And it is in that paragraph that he makes his "substitution effect" claim, which looks wrong to me. As I said in response to Millian, there may be GE models in which there is such an effect. (The fact that there is such an effect in a partial equilibrium model is not in dispute.) It would be helpful to see a model so that we know how the wage-cut party understands its case.

My only concern is that cutting the payroll tax, which funds social security (although we all no it goes into the general fund, which is why we use the SS surplus to fund tax cuts) is that this will "show" in later years that the SS trust fund is lower. If you could divorse it from SS, such as a per employee credit, for example, then it would be viewed as a subsidy to hire or retain. But, then how do you take it off later when the economy recovers, because someone will always say it is too early.

Do not do this a reduction in the SS payroll tax. Call it something else. Optics are everything. And, if you are unwilling to maintain SS and are using this as a device to cut SS, then don't do it at all, because what you will do is create a savings panic (people will increase savings suddenly) which will further reduce aggregate demand.

Bill, nice, but it would be a big deal to that guy that doesn't get hired who apparently stole Krugman's cab.

There may be hysteresis in employment, partly for some of the reasons you mention. I can't imagine too many people firing a worker that has already proven themselves because minimum wage goes up fitty cent. But, I can definitely see them not hiring the several people it takes to keep the one good fit. This is particularly important because the bottom rung of employment is, well, the bottom rung. So, it is an important question, but finding the answer might cost more than the cost to the economy. However, perpetuating the myth of a free lunch is an additional cost, so it is important as a policy kŠan. To say that this is unimportant as a debate is to say there is something more important for these guys to sit around debating.

So, I am reading through this paper's abstract.

And it seems to me their method was to compare New Jersey to Pennsylvania when New Jersey increased their minimum wage (a 'national study'?) but one problem that jumps out is they found that the border franchises increased employment perhaps at Pennsylvania's expense. That and a lot of other complexities with data when looking at a small factor effect highlights the difference between "no difference" and "no statistical difference." I don't know if all their papers are like that but it confirms the difficulty of looking at something like raising the minimum wage an amount that is tolerable to a constituency.

Alex, Bill thinks the minimum wage is the important thing to discuss, so let us wallow in it!

Also I think that elimination of the SS tax is a tax cut that might pay for itself later.

It might get people to think of SS as welfare and so voters may be more accepting of setting a single low payout rate for everyone regardless of how much they paid in and it may also make voters more accepting of raising the retirement age. Thus paying for itself.

Clearly the Great Depression showed that trying to force wages up, often to politically powerful groups, is destructive to growth.

And all workers have a comparative advantage in something if wages are allowed to adjust. So even unskilled workers can find employment providing some service to others at some wage.

Could lower wages make a liquidity crisis worse. Perhaps if falling wages created a general fear in workers about the future and they cut expenditures the situation could worsen.

But it is unclear why seeing large numbers of unemployed workers on the public square would be less negative on the public psyche then people employed at low wages.

Cutting payroll taxes, like social security, has a big advantage over stimulus spending. It would most certainly be temporary, broad based, and free of political distortions. Once the economy recovered politicians of both parties would quickly move to increase the taxes.

I'm not certain how many employers would increase hiring given that the Obama administration is intent on increasing taxes in some many other areas.

A general fall in wage rates *relative to other prices* "increases the incentive to hire (call this the substitution effect) but it decreases the income of people who already have jobs and this in turn decreases their spending and other people’s income (call this the macro income effect)," *ceteris paribus*. But, of course, other things aren't equal: as already noted, the newly hired people have extra income, and (as not already noted) *capitalists* have extra income. So what happens to *total* income? Since some unemployed people are now working, and everybody else is doing whatever he was doing before, the physical product increases. If prices-ex-wages have held steady (only wage rates falling), nominal income increases. (I'm ignoring the adjustment in product mix that might be required if the new distribution of wealth was unexpected.)

(This has nothing to do with a "liquidity trap," *per se*.)

Rick, How does a higher payroll tax promote saving?

As far as the burger flippers go, since it is costly to hire and to fire, presumably they've move from lower valued employment to higher valued employment. Where's the problem?

@Alex, I didn't say lowering payroll taxes would hurt employment. I said, if you reduce it through a SS tax reduction, and later say that SS trust fund is a mess, I would rather you do it through the general fund, and call it a per person credit, unrelated to the SS tax.

Maybe this is overpayment for a new job, however. You give a deduction for the already employed, when you are aiming for the marginal, potential employee to be hired. Now, you could say the marginal employee increase gets a credit to the employer, and then you have the potential for games: employer laysoff someone and a month later rehires the same person to get the credit.

For all the rational expectations economists out there,who would otherwise oppose stimulus spending, I am a little surprised at the rational expectations economists who are proposing this, as they would other wise argue that, if the employer knew that in year 1 after the economy got started that the credit would go away, he would not hire in year 0.

Gabe/Alex, I suppose one could read, "Another way to convert savings into consumption would be to increase the tax on people who have a very high marginal propensity to save," to mean "the _payroll_ tax," but I certainly didn't put the word "payroll" in that sentence, and I'm not sure how one could insert the word "payroll" into that sentence when I also talked about the government using the proceeds from that tax increase on general spending. Is it not generally understood that the proceeds of the _payroll_ tax are already earmarked for something other than general government spending?

Nevertheless, allow me, as I fight desperately to keep my eyes from rolling skyward, to rephrase the first part of that sentence in order to remove any trace of ambiguity whatsoever: Another way to convert savings into consumption would be to increase the non-payroll taxes (e.g. income or capital gains taxes) on people who have a very high marginal propensity to save...

As for burger flippers, first, you are, quite frankly, misrepresenting the data you cite. We might have millions of new job openings, but the data you cite makes no distinction between job openings that occur due to entirely new enterprise vs. job openings that occur because, say, people simply got older and retired. Unless you take a very broad definition of "value," it's difficult to presume that the net effect on aggregate demand is positive (let alone "significant").

Now, you could try, maybe, to make the argument that an increase in net job openings indicates that at least some of those new openings represent new enterprise, but it's really difficult to make that argument when the hires rate (3.0%) was less than the separations rate (3.2%).

Lastly, even if you could get that far, you're still missing the salient feature of a liquidity trap: short-run interest rates are at or near zero. It is the primary feature that distinguishes our current situation from that of a normal recession. How does the raw (as opposed to net) number of jobs created relate to, or alter, this fact or the implications it has for what kinds of policy changes will or will not increase aggregate demand?

Dan C says:Clearly the Great Depression showed that trying to force wages up, often to politically powerful groups, is destructive to growth

In the 1930s the relationship between employment and output was strongly related to wages. When wages fell so did employment and output. when wages rose so did employment and output.

Would you please explain why the data is in direct contradiction to your statement.


Comments by Cole and Ohanian in the WSJ

"Cutting the payroll tax "works" in Krugman's framework, because it effectively converts savings into consumption..."

Ergo, raising the payroll tax converts consumption into savings, raising savings.

I'm using value in the sense of "wealth is created by moving resources from lower valued uses to higher valued uses." Aggregate demand is a nominal quantity, but I'm talking about creating real wealth. As far as a liquidity trap goes, I'm with Philo.

Bill, that's a feature, not a bug.

@Dan C,

I can see your point that "cuts in the payroll tax would be from increased income in the hands of average citizens", but that is not the only person who is getting the cut: the employer or corporation gets it too, and there is no guarantee the money he gets will not be pocketed, hoarded or saved. I would raise the same point again that if it comes from SSec, then what you do is risk that fund, raise the prospect of the SSec fund running lower, and stimulating fearful near retirees to save like mad and defer their retirement because they will think that benefits will be reduced. If it comes out of the general fund, fine, you can even borrow it from the military or ag price supports for all I care.

As to the comment about California--you're kidding right--California has a problem from stupid property tax rules, and we don't use state taxes to stimulate anything; in fact, it usually works in reverse: states have balanced budget requirements which have them working to screw up the economy during times of recession. California argument re stimulus means as much as me saying Miss. or La. or some other state means something. I wouldn't care to live in the last two states, and if California can get over its constitutional problems, get over its stupid property tax fiasco, it will be able to get its house in order. Just takes the crisis to bring the change.

I have read Cole and Ohanian research -- not the WSJ article you reference-- and it does not negate my point that the relationship between wages and output in the 1930s was positive.

It is like all of you complaining about FDR's NRA. Maybe there were a lot of problems with it, but during the time it was in place the US had some of the strongest growth on record and a very sharp drop in the unemployment rate. Something was really working during that time.

The same with the cole argument about wages. They cherry pick some data and use it to show that FDR's policies had big negative impacts. But if you look at all the data-- not just their few months of cherry picked data -- you can not help but come to the opposite conclusion.

I will advise you to look at all the data, not just a select few observations.

I'm still waiting for anyone to show me a single real world example of the theory that lower wages leads to higher employment. Beautiful theory, but there are no historic examples of it working.

@ Alex J

"Ergo, raising the payroll tax converts consumption into savings, raising savings."

Yes. Oh, you want to know _how_ that increases savings. Well, that's what Social Security is. Are you angling toward some kind of discussion about various congressional dalliances with the Social Security Trust Fund? Or would you argue that Social Security is an inefficient means to promote savings? Why? How would either be relevant to any discussion of wages and they relate to the concept liquidity trap?

"Aggregate demand is a nominal quantity...."

That makes no sense. We actually do distinguish between real aggregate demand an nominal aggregate demand.

"As far as a liquidity trap goes, I'm with Philo."

And, Philo's argument assumes that the existing excess productive capacity will be put to use merely by reducing wages. If the demand existed now for that excess productive capacity, wouldn't output increase without a decrease in wages? Philo makes the same mistake that far too many people make: trying to extrapolate from a single firm or industry to the economy as a whole. Rajiv Sethi makes this point rather clear on his blog.


"I was aiming at the idea that cutting payroll taxes was somehow bad for the lower class. It was couched in obtuse reasoning about depleteing the lock box or some such non-sense"

But that was, also, not what I had said. I said, as Floccina accurately quotes, that "the adverse effects fall disproportionately on low-income workers." I don't know how you took that to be a statement of absolute harm. There might well be some other justification for cutting payroll taxes, or even doing away with Social Security entirely, but that has nothing to do with cutting payroll taxes _as a means of increasing aggregate demand_. There are other ways to accomplish the same thing.


"Rick Schaut, how does cutting the payroll tax adversely and disproportionately effect low-income workers"

Well, we did try to reform Social Security into "private accounts" not too long ago. The general public didn't seem to think that was such a hot idea. The whole effort rather went down in flames. So, we can take it as a given that Social Security is generally beneficial, and that reducing it would have a net adverse effect.

Which leaves us with proportionality, and that strikes me as even more obvious than the existence of a net adverse effect. High income people already have a relatively high marginal propensity to save. Reducing the Social Security part of that savings would seem almost negligible in comparison, no?

None of this has much to do with the advisability of using a reduction in payroll taxes as a means of increasing aggregate demand during a recession that features near-zero short run interest rates. Curious, don't you think, that in a discussion thread of a blog post that's ostensibly about a liquidity trap, only one person, who is not the original blog poster, has managed to mention the key condition that defines a liquidity trap.

You can believe in the tooth fairy or FDR for all I care.


People already fear the solvency of social security so I doubt much will change. Perhaps some upper middle class workers will see benefits cut but that will be true if the Democrats stay in power regardless of any temporary cuts in SS taxes.

If employers get a cut in the payroll taxes that they pay they might just save it to pay the other higher taxes that Obama and the Democrats want to impose. Of course if that didn't have that additional revenue they would need to scale back.

Still, if firms pocket or save the money that is wealth to the owners of the firm. That has a wealth affect on the economy. The value of the firm would be reflected in improved cash flows. It doesn't disappear.

California is based on the notion that you can soak the rich. Tax revenue is based on a highly progressive tax code which is more subject to a wide flux on yield as the economy falters. Plus the people who are forced to pay ever higher taxes can vote with their feet. Tax codes that seek to soak the rich have blowback.

But bear in mind that the real policy choice we have is to cut the payroll tax and cutting the payroll tax increases the incentive to hire and increases the income of people who already have jobs.

How is cutting the payroll tax going to work better than the 2001, 2002, 2003, 2004, 2005, 2006, 2008, 2009 tax cuts to increase the incentives to hire?

In those tax cuts were: cuts in both taxes on pump and dump capital gains; cuts in business taxes; cuts in taxes on the rich so they could invest and create jobs; cuts in middle class income taxes so the could spend more wisely than government.

Taxes as a "dead weight loss" have been cut from over 20% of GDP a decade ago to less than 15% of GDP today, so why is employment today the same number as a decade ago even with a larger population? In fact, if it weren't for the Keynesian deficits, I think GDP would be down from a decade ago, for as tax revenues have fallen, the spending has increased as if tax revenues were increasing.

And these tax cuts were never intended to be temporary but merely followed the conservative politicians' deceptive practice of spinning themselves as being the champions of long term fiscal responsibility, basically a conservative free lunch for everyone policy. So, the claim the tax cuts were temporary so no one acted on them to make long term plans is bogus. The economy tanked before 2006, because if the economy, specifically the labor market, had been as good as the 90s, Republicans would have been reelected.

To reduce unemployment, the best policy is to increase taxes.

Taxes reduce real wages and real prices. If I am working and my real wage is reduced, I will need to decide whether the new real wage creates the required incentive to work, or whether I should substitute retirement, vacation, education, working for myself building up my own capital or producing my own consumables. If I am selling goods, the increased taxes on my business, on my inputs, my capital, etc. will force me to chose to hike prices and reduce my sales, or to take a cut in profits or return on invested capital in order to keep as much of my sales volume as possible. And maybe the price cut make me decide I would rather be retired and that removes my excess capacity from the economy.

And as the increased tax rates go to government which is responsible internationally for national savings as reflected in the current accounts balance, hiking taxes increases national savings. The nation becomes less attractive to foreign investors due to the cuts in returns on investment that flow from higher taxes. That will in turn hold down asset price inflation caused by the subsidies on pump and dump capital gains policies.

Remember, capital gains taxes apply only if you buy an asset, pump up its price, then sell. If instead you buy an asset with the intention of holding it for life, living off the dividends, capital gains taxes of 200% would have no impact (other than ensuring people invested very wisely and with capital management focused on sustained long term returns forever).

In any case, the lower real prices after tax hikes for wages, for goods and services, and for capital would reduce the supply which would reduce the unemployment, reduce excess capacity of firms, and reduce the supply of savings chasing capital, so the economy would be more in balance.

Remember unemployment goes up as the economy grows more rapidly as the higher real wages pull more people into the labor force.

However, the worker who is at the lowest end of the wage scale has a far different set of choices than the very high wage earner. A banker making one million might be doing so only because he wants to be able to say he makes a million, but if the wage drop under a million, he no longer is interested so he will retire to his live off his ten million in investments until the economy demands his labor and offers him two million. To make him unemployed, he needs to see the job posting with the possibility of at least a million and probably a lot more. In other words, to increase unemployment in the upper incomes, wages must increase.

The same is probably true for the low end - a wage of five an hour with the costs to get to the job totaling two dollars will all things being equal remove him from the job market and the ranks of the unemployed. Of course, conservatives demand that he be unemployed and looking for a job that doesn't pay enough to justify the effort, so it is the conservative government policies that increase unemployment. To increase the unemployment among the poor, the solution is to hike wages.

I think the place conservatives get confused is by looking at the wages of the young, minorities, school drop outs, and assume they are unemployed because minimum wages are too high and that by cutting the minimum wage, those unemployed will become employed. But the jobs the poor are holding out for are the jobs for a million a year because they have concluded they are as qualified as anyone.

To put it another way, if the pay for a bank CEO were reduced to $10/hour, the logic used to justify cutting the minimum wage would result in BofA and Citi et al to hire ten or a hundred or a thousand CEOs because they are so cheap. As this is absurd, reducing the mimimum wage to cut unemployment is possible only if no wages higher than that are offered anywhere in the economy, and one weren't forced to look for work (and be unemployed) by conservative government policies.

How is cutting the payroll tax going to work better than the 2001, 2002, 2003, 2004, 2005, 2006, 2008, 2009 tax cuts to increase the incentives to hire?

The theory is obvious:

1. A SS tax cut is targeted at lower income people as compared to those other tax cuts.
2. A SS tax cut for business is targeted not only at profitable businesses but at businesses that currently pay no corporate income tax, those that are unprofitable due to the recession.

Are you saying that you think that those reasons are completely insignificant?

Yes! I feel like I’m back in introductory macroeconomics again. AD, AS, long time no see. In all seriousness though Paul, thank you for taking the time to post your opinions for us layman, it is highly appreciated in these troubling times. That you take the effort to do so in terms that people with at least a bit of economics background can understand is also very helpful.

Comments for this post are closed