A simple recalculation story about the stimulus

Let's say a real estate agent is laid off and, at some point, needs to start working elsewhere in the economy.

One scenario is that the former agent searches for twelve months and finds a job in the health care sector.  The economy loses twelve months' worth of output from that individual.  (These numbers are chosen for illustrative simplicity and they are not estimates of actual variables.)

A second scenario is that the former agent is reemployed immediately, improving the energy efficiency of school buildings, and he is paid by stimulus funds.

Two years later, that stimulus money ends.  The former real estate agent then searches for twelve months and finds a job in the health care sector.

The net effect is to sub in two years of insulating work for two more years working in the medical sector, or wherever that individual ends up in the later years of his career.

Both scenarios involve the cost of twelve months unemployment and the associated foregone outputs. 

If you measure the progress of the stimulus early on, it will appear to yield higher employment and gdp growth prospects.  Those benefits are to some extent an illusion because you are not picking up the possibility that labor market search is postponed but not avoided.

A plausible intermediate scenario is that an economic downturn is a mix of real and weak AD factors.  So, after the fiscal stimulus, and after the insulation work is over, the former real estate agent can reemploy himself in six months rather than twelve.  By this point in time AD is higher (perhaps) and labor market adjustment is easier.  Still, the short-run measure of stimulus benefits will be about twice as high as the actual net benefit, all long-run adjustments considered.  It will look, in the short run, as if twelve months unemployment have been avoided when in fact the savings are a net of only six months unemployment avoided (toss in discount adjustments plus consider the costs of taxation and debt).

Many people argue that "we're not yet out of the water" or that we are seeing a "jobless recovery."  I agree on both counts.  I would stress that those arguments do not unambiguously favor the case for more stimulus.  On one hand, troubled times may suggest that we can't let the economic recalculation happen all at once.  On the other hand, if the labor market is still sluggish when the stimulus ends, we are just postponing search and unemployment, not much reducing it.


Beyond improving the economy enough that it takes less time for a worker to find a new job, might it not also provide enough of a bridge to keep people at work until the situation improves as the private market takes over again? The example above in of a public worker employed in a WPA-type position. How about a worker at the cement plant that provides material for repairing the school who stays on because of this contract and then is able to continue in his job when housing picks up again?

On the other side of the coin, two years from now, our school inspector is going to face strong incentives to lobby to maintain her 'job' rather than spend 12 months retraining and searching. Multiply that by hundreds of thousands (millions?) and square for rapacious public sector unions desperate not to lose all of their new members, and you've got a perfect storm for an indefinite extension of this unsustainable spending.

Does it complicate matters at all that six months of unemployment is not necessarily only "half as bad" as 12 months? On the macro level, maybe. But for individual families, the first few months might be OK as people get by on savings. Then you start putting stuff on a credit card. Then more stuff. Then you start slipping on your mortgage. They you get forcelose. Then Junior drops out of college, you sell your car, etc.

Later in this scenario, it will start to take years and years to recover from the accrued debt. You need to buy another car and start making payments, while the old one was paid for. Your new mortgage is at a higher rate because of your credit. Junior is in jail, etc.

So month one does not seem equal to month 12.

Keeping our individual employed and therefore spending his income in the short run clearly also prevents ripple effects that maintain other individuals in jobs. It was the threat of a widening cycle of layoffs that the stimulus was designed to prevent.

It's clear that the economy needs a different mix of jobs than the one that prevailed at the peak of the latest expansion. It's also clear (at least to me) that we will move to the new and better mix of jobs faster because the government took steps to prevent systemic collapse of the banking system and steps to bolster aggregate demand.

Let's not edge back toward Treasury Secretary Mellon's point of view on stimulus :)

Good post, Tyler. Reminds me of a point Sumner made on his blog, but yours is a much more concise and intuitive description.

The scenario assumes that some real, underlying non-government economy never recovers., i.e., that once the stimulus ends, the economy reverts to the exact state it was in before the stimulus started.

OTOH, in the real world, the length of time unemployed is correlated with the unemployment rate.

Once the assumption is made explicit, the scenario looks a bit silly.

Oh, come on. When the person who was hired to insulate a building goes back on the market a year later, the economy has changed. And, probably his prospects for using his acquired skills (from both jobs) increases. You also make the assumption that the work under the stimulus program had no value (just digging a hole to fill it) whereas insulating a building has future savings effect. But, also, don't forget that the insulated building now operates at less costs, and that the insulation manufacturing, when the economy picks up, will not have to bear the costs of restarting its business (as it would have had to do if there had not been a stimulus) and can continue operating.

Now, let's try what Bush tried in 2008: give a tax cut to stimulate the economy. I took my $1500 and bought a real nice Korean LCD television. Why is it that no one discusses the impact of the Bush 2008 tax cut? I frankly think the comparision should be made between that tax cut and the current stimulus program, don't you think?

This story suggests that stimulus actually prevents recovery.

OK, let's talk about the other stimulus program: supporting the lifestyles of the rich and famous.

The US just stepped in and saved the bacon of the bank bondholders and investment bankers. You would think that the wage rates of investment bankers would change, but of course it didn't because the government stepped in and saved the investment banks (Merrill, et al), banks holding mortgage backed securities, etc. In a dynamic economy where the risks were born by the institutions, that the workers at the institutions would adjust their earnings if their institution took a loss. Of course, the institutions did not take a loss that they would have taken had there not been government support. Consequently, the investment bankers wage rate did not decline.

That to me is a particularly galling stimulus program. Tyler bit.hes about some guy and his family earning less than $50k getting some short term support whiile the economy tanks (and by the way, unemployment insurance isn't free), whereas the guy who has his institution supported goes on his merry way. Oh, and maybe he's even luckier: someone writes a column that says the cause of his institutions failure wasn't caused by the investment banker's pay and incentitive structure and that the big bad government that was footing the bill shouldn't comment on pay practices of investment bankers because it would disturb the "market".

Stimulus comes in many forms: tax cuts (wow that really worked), support for institutions (wow that really worked for people making making millions of dollars in comp), and short run job training and interim employment measures for the lower middle class. Of the three, which do you think works and what is your preference. Or, you could say, none of the above, but then I would challenge you to take back the subsidies you already provided to the beneificiaries who didn't need it.

I also think your hypothetical ignores the important point that the job finding rate is very procyclical so that the duration of unemployment should be less than 12 months in the second scenario. Of course, it depends on whether you think the post-2001 "jobless recovery" experience is the right reference point.

Economists are not as dynamic of thinkers as you would assume.
Tyler's problem posits that a worker searches for one year, and gets a job, or he gets a stimulus job and searches one year after and gets the same job he would have gotten with the one year wait.

I posit that the world is more complex, and job search is more like a dynamic queing problem with backup inefficiency. Imagine an airport with a steady flow of takeoffs and landings. Suddenly, some of the planes don't take off. There is a backup. Planes can't land either because of the backup. Soon the dynamic system goes off the rails (sorry for switching metaphors) due to small initial pertubations that have big dynamic effects.

Job search and job switching have those dynamic system aspects to them. It is a STRONG assumption to make that you can move a person from period one to period two without disrupting the system and creating a bigger problem.

Not only that, but there are circumstances where waiting or search for the same job is inefficient and the government stimulus assisted job is part of the solution, and not part of the problem. So, for example, assume the worker is a construction worker, and our level of housing starts will not return to the levels of 2004-2007. It is inefficient to wait, but it is efficient to have on the job retraining with some government assistance. This also avoids the dynmaic queing problem (as the person is gainfully employed while being trained), and the problem of frustrated search which arises when, two years later, the construction worker has still not found a job and has exhausted unemployment benefits.

Don't you also have to factor in the notion that the person is less likely to be as productive in a 'stimulus' job as in the job he was trained for, and that providing a stimulus job makes it less likely that the person will go back to work in his area of expertise in a timely manner?

Assume I'm an unemployed financial expert. With no stimulus, times will be very hard, but I'll also be working my butt off to get another job in the financial sector where my training makes me valuable. But if I get a two-year stimulus job, I'm probably not going to try as hard to get back into my own field - and by the time two years are up, I may not be as employable, or employable at all, in my old field.

I'm sure this isn't a dominating factor, but it's just one more illustration that the real world effects are probably far more complicated than any simple models would be able to capture.

Amen Bill. Sorry I didn't see your post before I wrote mine. Great minds think alike though, no?

Well, if stimulus work is paid by magic pixie dust, then there may be no downside considerations whatsoever to stimulus.

Consider something- there are businesses, even in downturns, that might have the ability to expand, but can't due to the fact that labor and other inputs are just expensive enough to make the proposition a losing one. As the downturn happens, labor and other inputs start dropping in price and these businesses start looking at the possibility of expanding their operations to take advantage of the newly available inputs, but then along comes the government and it bids up the prices of these inputs. Those businesses on the margin of expansion then can't expand their operations, and not only that, they have to help fund, along with every other actor in the economy, the consumption of these stimulus workers and their work.

On one hand, I see what you're saying. On the other hand, there's just one thing about your story I can't get over. Real estate agents are almost always contractors, so they don't get laid off. They just quietly starve to death, as the 9.5% of Americans who are unemployed can't afford to buy homes.

Now, to un-derail. I don't think there is such a thing as a jobless recovery. I think any "recovery" that doesn't create jobs is a statistical anomaly and not actually a recovery at all!

Further, many stimulus jobs will have a positive economic effect many years after the money is spent. In your green school project example, the money the school district saves on energy costs can hire more teachers or buy more textbooks (which have to be written, proofread (I hope!) printed, packed, and shipped by people who have jobs). Build a new road today? Not only will it make it faster and easier for goods and services to get around town tomorrow, but somebody will build gas stations and stores and restaurants on that road and put people to work.

Think about it. Which would you rather have? A space program or really cool pyramids in the desert? I can't afford to visit the moon, but I can afford to go to New Mexico, if I save up. We could spend generations building them. Did a worker on the Great Wall of China worry about what they were going to do next?

We're not building pyramids, but we are building space shuttles and missile defense.

We are also not addressing reality.

The hypothetical should really be about a construction worker or an auto parts worker. Two years from now they may still be looking for a job unless they are retrained and the economy turns around.

In 1982, unemployment reached 10.2%. 2.7% of the workforce then was unemployed for more than 27 weeks. Today, 3.5% have been unemployed for more than 27 weeks and unemployment is still climbing.

We should be thinking about what we are going to do about structural unemployment, not moving a worker from one period to another and asking what the stimulus accomplishes as a placeholder. It's not a placeholder we need, and I don't think that is what they are aiming to get with the current stimulus program. What they are not addressing, and this hypothetical does not address, is structural unemployment, which I believe is going to be the subject of more discussion as this economy does turn around, in part because of the stimulus, but not enough to address structural unemployment. The time to start is now.

A nice story except that you assume that the current unemployment is caused by a wave of industrial restructuring which somehow causes unemployment. At least some past waves of industrial restructuring have - say the dot com boom - have caused low unemployment. How do you tell which is which?

Also we should remember that most of the industrial workers who were unemployed during the Great Depression did not go to work in new sectors after the depression. They just got jobs filling the increased demand that was stoked by wartime production. The recovery was not some vast restructuring of industry. The old industry got bigger.

And of course this increased demand - created by the evil and wasteful government - absorbed the agricultural labor that was also thrown out of work by the Depression.

By your model this all should have run out at the end of WWII when the economy should have returned to the painful process of creative destruction that the stimulus only postponed.

And if we artificially employ people, aren't we discouraging them from creating the new businesses needed to hire the idle workers?

The problem with the scenario is the assumption that hiring is done from the pool of unemployed on a first in, first out basis. In reality, hiring is more properly described by last in, first out.

In this market, someone out of work for a year is likely a professional or other high wage skilled worker. Potential employers are leery of hiring them for a lower level job or for something out of their speciality, because once the job market picks up, they are expected to leave. So, in a market like this one, they will remain unemployed while the rate of layoffs slows, and then remain unemployed as those of similar skills more recently laid off are hired first.

Of course, while Tyler is arguing that the market should determine where resources should be allocated, he then goes with the government forecasts and policy recommendations of politicians like John McCain who famously told the auto workers their manufacturing jobs were gone forever and they needed to go into the biggest growth sector in the US, health care.

Now, let's consider health care as a future job for the US. The assumption is that the US will continue to become even less efficient and competitive in the world, with the overhead cost to society of health care increasing, burdening exports with 20% going to 25% health care overhead costs.

Also implicit in Tyler's politically driven model is that the government will continue to subsidize oil imports and burning with policies that promote polluting without cost, and foreign policies that promote cheap oil by supporting dictators willing to expropriate the oil of the people they dominate and sell it for personal profit and weapons.

Tyler has rejected the policy change that more than half the people voted for, toward a reduction in the overhead of health care in the US economy, and a responsible energy policy that imposes on oil burning the cost of pollution and eliminates the subsidies to the authoritarian regimes who exploit the the people and environment of the territory they control for personal profit.

Training someone for a sustainable energy economy is far more productive use of tax dollars than retraining for future higher inefficiency of the US economy by increase health care spending, or for keeping them warehoused on the unemployment roles until they are forced to find some way to live off the dole by becoming disabled or able to retire early.

Note that the number of early retirements has increased in the past year, along with the expected rates of early retirement, which is shifting the Social Security flat tax from a general fund revenue source to a general fund liability.

But in any case, the most important flaw in Tyler's reasoning is to believe that someone out of work for a year at this point is somehow not going to be out of work for much longer given the current economy because lots of people with the same skill are being hired constantly as firms poach workers from their competitors. No way does the guy out of work for a year get a job until the employers are so desperate for workers they hire people based on a mirror under their nose fogging up.

Stimulus can be considered just another form of income support similar to unemployment compensation. At best it can be spent on something worthwhile, at worst it buys time for recovery. The real importance is not how it is spent though, but that it is spent. Being spent it is the income of others and even if not well spent by government, it will be respent well by those receiving it. The only way harm can occur is through crowding out, but under any credible assumptions, it is far, far, more likely to lead to crowding in.

In Ireland (where I am from), the government has just approved NAMA - the scheme to buy bad assets from the banks. Each Irish citizen has just seen their portion of national debt go from €33,000 to €46,000. One little seen statistic is that 90% of the NAMA funds go to cover the bad debts of the banks to a list of 200 individuals or companies. I am less interested in the potential perceived "unfairness" of this deal and more in the underlying theory for why it is a better thing to load us all up with increased future tax in order to "save" the system. Are we really saving the system or just delaying the pain?

OK what if it takes 18 months to find the new 'recalculated' job in health care and the stimulus is 12 months of work in the 'insulating' sector. So we get 12 months of additional output and we avoid deteriation of human capital (long term unemployment breeds depression, detachment from the labor market, 'giving up' etc.)

This doesn't really make any sense. If you're suggesting this simply postpones unemployment - and that unemployment is caused by a shift from one industry to the next - then why wasn't there mass unemployment during the housing boom as a shift from one sector to the other. It should have caused the unemployment rate to rise as people were looking to become construction workers instead of working etc...?

You also forget to include the cost of the timing of cash-flows to an individual, how one budgets that cash flow (i.e. one can plan to spend more than proposed income, but if someone has no job are less likely to spend beyond necessities), as well as positive feedback loops from knowing that one has a job (that indicators seem to be improving - thus inducing higher investment and spending in the short run due to better prospects for the long-run). A better economy means the ability to find a job quicker and perhaps almost immediately.

I think that one point that needs to be answered by this model is the drop in economic activity accross the whole economy, seemily disproportionate to the size of the construction sector. I know that you have brushed on this with stress on multiplier effects but I think a more complete explaination on this point would go along way to making the model more convincing.

A possible explaination could be disruption to the credit lines of non-construction firms after panic and uncertainty that was apparently widespread in the financial sector as a whole. Perhaps some sort of variation of the Minsky Moment, although I am no expert on that area.

Comments for this post are closed