How much did cutting unemployment benefits help the labor market?

by on January 26, 2015 at 11:03 am in Current Affairs, Data Source, Economics | Permalink

Quite a bit.  There is a new NBER Working Paper on this topic by Hagedorn, Manovskii, and Mitman, showing (once again) that most supply curves slope upward, here is one key part from the abstract:

In levels, 1.8 million additional jobs were created in 2014 due to the benefit cut. Almost 1 million of these jobs were filled by workers from out of the labor force who would not have participated in the labor market had benefit extensions been reauthorized.

There is an ungated copy here (pdf).  Like the sequester, this is another area where the Keynesian analysts simply have not proven a good guide to understanding recent macroeconomic events.

1 Hojat January 26, 2015 at 11:25 am

I remember a paper by Robert Barro claiming that most youth unemployment after 2008 crisis was due to the raise in minimum wage and not the crisis itself. This paper is similar to that one. I always take such papers with a grain of salt, and a lot of doubt.

2 Ray Lopez January 26, 2015 at 12:15 pm

@Hojat- yes, I am no Keynesian but this paper is not Bayesian. What it finds is this: states that had high unemployment benefits had a faster return to work when unemployment benefits were abruptly ended in Dec 2013 than states that had poorer unemployment benefits. But from a Bayesian analysis, the authors fail to account for these facts:

1) could it be that by Dec 2013 the crisis in state having severe unemployment, and thus higher unemployment benefits, was over, so they recovered faster? By contrast, low unemployment benefit states never had a “big problem” so they showed slower rebound. Stylized example: low unemployment in the Midwest, since lots of jobs, even during Great Recession, nearly everybody had a job. By contrast, New Jersey, high unemployment, so no jobs during Great Recession but by Dec 2013 finally the market thawed out, and when Congress took away the benefits there was more room to rebound in those states.

2) Necessity is the mother of invention, but are people better off working? It could be that people in Jersey took some miserable menial job just to keep alive, or maybe even turned to illegal crime to make ends met, so arguably, from a social welfare point of view, it would have been better to let these Zero Marginal Productivity workers stay on the dole. What is the marginal utility for society in letting people work? Maybe it’s better if they did not work? OK so this point is weak.

3) There’s no three. It just looks better if there’s a third point here.

3 Dale January 26, 2015 at 12:24 pm

Ray
What, no reference to your wealth or romantic life (that could easily be put into point 3). You should try reading the paper, however. My first reaction was to worry about endogeneity – that states that ended the extended benefits did so because they had recovered more quickly. But the paper actually compares counties sharing a state border and find the same thing. It is a more robust finding than you think.

4 Bryce Dayson January 26, 2015 at 2:41 pm

Who is Ray Lopez? Why do so many people in the MR comments talk about his wealth or romantic life?

5 Al January 26, 2015 at 3:08 pm

If you browse far enough back in the comments on this blog over the last couple of years, you will find that one “Ray Lopez”:

* Lives in the Philippines

* Has a girl friend half his age

* Enjoys cut rate dental work

* Knows a great deal about chess

* Has identified the next master race

* Is a first-class world traveler

* Personally knows all the important people in the US technology sector

* Can beat-up Chuck Norris with both hands tied behind his girlfriend’s back….

If you go back and read his comments, I am certain you will find that Ray Lopez is not a man at all — but something far greater. He is a legend, a state of mind, an enigma, and, at the same time, the final answer to the deepest riddles of our universe.

6 Bryce Dayson January 26, 2015 at 3:36 pm

1) And where is the proof that these claims are true?

2) Does he actually say anything that advance the substance of the conversation, or does he just use MR to blow his own horn about his wealth and love life?

7 msgkings January 26, 2015 at 3:56 pm

@ Bryce: both actually, he’s a pretty good troll. I like having him around to tease.

8 JWatts January 27, 2015 at 1:12 pm

“@ Bryce: both actually, he’s a pretty good troll. I like having him around to tease.”

I second this. Also, Ray does make many good points.

9 Ray Lopez January 26, 2015 at 11:34 pm

@Dale–either you or I did not speed read the paper properly. “My first reaction was to worry about endogeneity – that states that ended the extended benefits did so because they had recovered more quickly” – no, not true. The benefits were ended across the board by Congress for all states. The paper measures what happened after the benefits were ended by Congress for all states.

10 Doug M January 26, 2015 at 1:33 pm

regarding 2 call me a puritan, but even the most menial of jobs has a greater social benefit than no job at all. That is, there is social value in merely showing up to do work.

11 Matt January 26, 2015 at 2:15 pm

I think I largely agree with the underlying sentiment here insofar as that has to do with human flourishing and all that, but I feel like this kind of thinking is tied up in the non-economic definition of ‘leisure.’

That is, I’m not sure we know what people were doing with their time before they returned to work. It’s possible, certainly, that these hours were just filled with tv, but if people used that time to get more exercise than they otherwise would’ve, or cook meals at home, or read a book, well that might have a larger positive impact on social welfare than showing up to flip burgers or some such.

12 Kurt Mitman January 26, 2015 at 3:36 pm

Ray,
In response to point 1), we do several things. First, like someone pointed out we are looking at the evolution of border counties, not just the states. Second, we also allow for a flexible factor model to account for the fact that the county pairs could have had different trends before 2014. Finally, we also run a placebo test to exactly overcome the potential endogeneity that you suggest.

In response to point 2), we make no statements about welfare, we simply make statements about the number of employed.

Kurt

13 Ray Lopez January 27, 2015 at 12:16 am

@Kurt Miltman: Thanks, but I’m busy and so are you. By answering me you force me to respond, and likewise you have to likely respond to my response. Like a chess match, each of us has to anticipate the other’s move. This potentially will never end… sigh. You may think a ‘troll’ enjoys a response but it’s not always true. Good trolling is hard work. Good trolling–factual stuff with spin–is hard to do (because *everything* I say is factual, but with spin; when I say I am worth over $7M-10M it’s true, but, it includes my inheritance money, not just what I worked for and saved, which is also close to a million dollars).

Anyways, here are my thoughts:

Oh great, a model on a model. Indirection at its finest: “However, the effect of the benefit cut is estimated along with a flexible specification of the difference in trends between border counties in each pair using an interactive effects model developed in Bai (2009)”

Your attempt to avoid Holmes (1998), could have been simply rebutted by observing any cross-border differences are largely small these days and letting your critics respond, pace the below (most borders are exactly where the ‘good stuff’ is for resources, so states at these borders compete with each other to attract people and, like a hard fought chess match between two grandmasters, usually will result in a draw), but instead your team introduces another layer of indirection, the ‘interactive’ Bai model: “Thus, there are numerous aggregate shocks that potentially induce different trends across border county pairs. The interactive effects estimator accounts for these trends by identifying the important unobserved aggregate shocks and measuring their heterogeneous impacts across counties”.

“We find that changes in unemployment benefits have a large and statistically significant effect on employment: a 1 percent drop in benefit duration increases employment by 0.0161 log points. While large, this estimate is smaller than that implied by the simple experiments described above.” – again, don’t use Bai’s model and you would not have this problem.

What to make of this paper? It’s much ado about nothing. Here’s a simple model based on the DC tri-state area: northern Virginia’s Fairfax county has smaller UE (unemployment benefits) than say Maryland’s Montgomery County. This is because higher-productivity people cluster in Fairfax county, and more benefits are found there in terms of good jobs (it’s also more expensive). I know because I own lots of real estate there, it supports by jet setting lifestyle (I own lots of real estate near Metro stops, I am still kicking myself over forgoing expansion in Tysons’ Corner around 2007). So naturally, once UE stops in Montgomery county, people will go back to work “from a lower base” meaning there’s more of them that need to find a job (or starve). Naturally the numbers will go up for Montgomery county, but this does not prove much. Simple thought experiment: suppose welfare was abolished. Where would you see the greatest increase in welfare people getting jobs? A: in Gross Pointe, Detroit area, Michigan, where the rich whites live and almost nobody is on welfare, or, B: Highland Park, Detroit, Michigan, where the black slums are. If you answered “B” you win a prize. And you don’t need fancy math nor Bai’s model to prove it. In any region the laws of entropy say you will get a power law “cluster” distribution. So naturally in any region you’ll have ‘slums’ and ‘tony places’ next to each other, and given this disparity you’ll get the effects noted by your paper.

However, as a hard money advocate I do see a place for your paper: it gives ammunition to rich folk like myself to say: ‘see, UE makes people lazy, we should cut it’. And that’s a good thing, since I am concerned about preserving my cash wealth give the out-of-control spending by government (my family has about 2M in cash at the moment, and it’s a hassle just trying to think of different name combinations to qualify for FDIC insurance, believe me, we’ve even put in some relatives as nominal owners).

14 msgkings January 27, 2015 at 11:57 am

And cash portfolio management to the list of things about which Ray Lopez doesn’t know shit

15 Veritas January 26, 2015 at 11:26 am

how much did this decrease the wages these workers were willing to accept (didn’t read the paper, maybe it didn’t)…as someone who may soon be on unemployment, I am not going to get a retail job when UE pays more. UE will allow me to (hopefully) find a job that pays closer to my old wage (whether or not that was inflated…maybe).

16 Veritas January 26, 2015 at 11:28 am

also I think the benefits of (limited) UE outweigh the disincentive to work.

17 Vivian Darkbloom January 26, 2015 at 1:38 pm

Please note that by “cutting unemployment benefits” one is not talking about cutting the “monthly benefit amount”. What happened during the recession is that unemployment benefits were extended from 26 weeks to a maximum of 99 weeks and the extension of the duration of benefits was then subsequently ratcheted down more or less coinciding with the recovery.

Is it your view that someone who has received unemployment benefits for almost 2 years should not “consider getting a retail job when UE pays more” (or, even less)? If so, this may explain the effect the authors of that paper claim to have found. It may, in part, also explain why, statistically, average wages have been more or less stagnant since the recovery and the reduction of the duration of UI benefits. Many people may have been forced to accept lower wages due to the lack of UI benefits. Could this explain the faster recovery and lower unemployment rate? Did it help, in part, to “un-stick” average wages?

18 Veritas January 26, 2015 at 2:40 pm

I think that the cohort that I know (lol sample size) who received UE benefits is not really representative of many who get UE benefits…Upper-middle class,no-debt, has a rainy-day fund (not counting UE benefits)…I think (and I don’t have a study to back this up) that much of the benefit was allowing people who over levered their personal balance sheets during the boom to hopefully find replacement jobs at a level that allowed them to service their debt and not just default on their loans. Much of that money went back into the economy helping consumption, etc. I just looked, benes in DC are capped at 359 a week (taxable) so while that may stop me from being sales associate or working for my friends construction company (laborer), those who do not have a rainy-day fund would quickly feel the fiscal pinch, IMHO

19 Jason Smith January 26, 2015 at 11:47 am

It is pretty amazing how a cut in unemployment benefits leads to a spike in job openings …

http://research.stlouisfed.org/fred2/series/JTSJOL

What are the incentives there?

20 JaredN January 26, 2015 at 11:50 am

Yes, Tyler, do tell.

21 8 January 26, 2015 at 12:17 pm

If I’m an employer and I know there’s a large pool of workers who are going to be looking for work, I want to get first crack at the highest quality labor available.

22 John January 26, 2015 at 1:15 pm

Not sure I follow that. The same people who were getting the benefits had to be willing to work in order to qualify for the benefits so it’s not like suddenly there were a new group of people to offer jobs to. I’ll have to go look at the link but the only bit I see for a case where benefits are cut and a simultanious increase in open jobs occurs is either coincident or some form of job hoarding by employers waiting for a signal reservation wages will be lower.

23 Jason Smith January 26, 2015 at 1:38 pm

So you don’t offer jobs until you know the labor supply is desperate. That was pretty much the conclusion I tried to imply by my question: what are the incentives there?

That is to say, if UI benefits sometimes go away, employers will wait until they do. That is analogous to companies keeping their profits overseas until a repatriation holiday (since they’ve been offered in the past).

Extended UI doesn’t offer an incentive to not work (it’s not ‘laziness’), but rather an incentive not to hire (it’s ‘exploitation’).

24 Careless January 26, 2015 at 10:15 pm

How is not hiring ZMP workers “exploitation”? And unemployment benefits will always go away at some point.

25 JonFraz January 26, 2015 at 3:10 pm

Having had some experience with unemployment as recently as last summer– and with a very generous chunk of severance pay to live off of– I can report that there is very little incentive to sit around and not job hunt, or reject employment at pay less than one has been making (I was back to work at 8K less than I had been making, albeit the benefit costs were less and the new job has been less stressful than the old one).

26 Kevin Erdmann January 26, 2015 at 12:18 pm

Narrative explanations are dangerous. But, I can imagine that an employer would be less likely to commit to potential new employees if she knows that for each new hire, there are one or 2 better qualified workers who are temporarily out of the labor force.

27 Greg January 26, 2015 at 12:53 pm

Doesn’t this explanation require that the “better” employees are the ones who are sitting around doing nothing and collecting (relatively paltry) UI benefits, while the worse employees are the ones hustling for a job? That would seem to be an odd definition of “better.”

28 Kevin Erdmann January 26, 2015 at 2:02 pm

That’s a great point, Greg. I think you’re right that it mitigates my point.

On the other hand, faster re-employment does tend to correlate with youth, lack of experience, and lack of financial security. So, at the high levels of EUI we had, some older, experienced, and formerly higher wage employees might have tended to be the potential employees who were on the sidelines.

I think there might be a duality here. There are about 1 million workers still listed as unemployed who have been unemployed for 2+ years and appear to still be having trouble. On the other hand, workers with shorter durations of unemployment who were eligible for EUI saw their re-employment rates improve right as EUI was canceled.

There might be a category of worker who was able to tactically use EUI, and another who was spit out the back end with a red mark on their employment record. Remember, in either case we are talking about a very small sliver of the total labor force.

29 Kurt Mitman January 26, 2015 at 3:42 pm

We have another paper (blogged about here in October 2013) where we explore the entire mechanism. http://www.nber.org/papers/w19499

The basic idea in the context of the Diamond-Mortensen-Pissarides model, increasing UI durations puts an upward pressure on the wages that firms have to pay. All things equal, this decreases job creation by firms, since profits are lowed from creating new jobs. This explains the effect on vacancies (written about here http://libertystreeteconomics.newyorkfed.org/2014/09/do-unemployment-benefits-expirations-help-explain-the-surge-in-job-openings.html)

30 Jon January 27, 2015 at 10:34 am

This previous paper depends on what is almost a form of reverse causality. Rather than unemployment benefits being a response to high unemployment during the recession, high unemployment during the recession is due to expectations by employers of the impact of future benefits on the wage needed to retain employees.

As if unemployment fell rapidly in a jurisdiction that paid benefits, the interpretation would be that it was initially high due to the expectations.

31 Jon January 27, 2015 at 10:38 am

Kevin: On your comment—unemployment benefits are a fraction of what people earned prior to being unemployed. The penalty for not working is even higher for higher wage employees.

The ones who would be unemployed and receiving benefits are more likely those who cannot obtain a job paying close to their old position and are trying desperately to avoid accepting what could be a permanent pay cut.

32 derek January 26, 2015 at 1:55 pm

Unemployment benefits allow a large number of people to be unproductive, not working, and with little extra cash. Demand for employment is a factor of the economic activity as a whole, the more activity, the more demand for employment. So one is pushing towards less economic activity and you find it odd that when that push is removed, there is more economic activity that is evident by an increase in labor demand.

33 JonFraz January 26, 2015 at 3:14 pm

How is there more activity BEFORE the people are hired? Seems to me there is even less, since fewer benefit checks are being issue and cashed and less money is being spent.

34 ChrisA January 26, 2015 at 8:04 pm

Jason – yes there was a general improvement in the economy during this period which lowered unemployment. They use the data on a state level to determine how much of the lowered unemployment was due to reduction in unemployment benefits versus the general improvement in the economy. They didn’t just say that cutting UB reduces unemployment. Perhaps you should read the paper first before commenting.

35 rayward January 26, 2015 at 11:48 am

In November 2011, there were 3.1 million job openings (BLS). In November 2012, there were 3.7 million job openings. In November 2013, there were 4.1 million job openings. In November 2014, there were 4.9 million job openings. The authors of this study include some rather foreboding but impressive looking formulas for confirming that cutting unemployment at the end of 2013 resulted in an explosion in the number of employed. What you won’t find are BLS statistics on job openings. That’s not to suggest there aren’t a few malingerers out there, but only to point out the obvious.

36 rayward January 26, 2015 at 12:01 pm

Maybe I don’t appreciate the supply siders perspective: that more people willing to work because their unemployment was cut off (i.e., the increase in supply) created its own demand for them.

37 Bill January 26, 2015 at 12:08 pm

Says law–supply creates demand.

Maybe we need to cut tenure to create more demand for professors as this is otherwise a sticky market that would otherwise be slow to adjust.

38 derek January 26, 2015 at 2:24 pm

That would probably work.

39 Mike Sankowski January 26, 2015 at 12:17 pm

it is probably the money they did not spend which created the demand for more products, which caused entrepreneurs to add capacity to fill the demand for additional products.

That’s it.

40 JonFraz January 26, 2015 at 3:16 pm

However this could just be a “post hoc, propter hoc” error. Every year since the crash there’s been a few months where it looked like things were picking up, then things ran out of gas. In 2014 the economy finally achieved liftoff velocity (we hope). Attributing that to cutting back UI may be the classic error of attributing dawn to the rooster’s crowing.

41 WC Varones January 26, 2015 at 12:09 pm

Time to do something about the disability slackers.

The Social Security Disability “trust fund,” by the way, is set to run out of money next year, requiring significant, immediate benefit cuts to all recipients. Get ready for an epic exercise in can-kicking by the Boehner Congress.

42 JonFraz January 26, 2015 at 3:18 pm

Next year is an election year, and the GOP will be extremely vulnerable to charges of cruelty to the sick and disabled. The Tea Party (who never met a pauper they didn’t want to kick to the curb) will howl, but the adults in the room will fix the problem as they must.

43 TMC January 26, 2015 at 8:21 pm

Almost always it’s the Tea Partier that IS the adult in the room.

44 JonFraz January 27, 2015 at 2:38 pm

Defining “adult” down I see. The Tea Party is nothing but a primal scream uttered by the selfish, the racist and the clueless.

45 Careless January 26, 2015 at 10:17 pm

the adults in the room will fix the problem as they must.

By which he meant “they’ll borrow more money they will never be able to pay back”

In which case, I wish I were as youthful as my high credit score would predict

46 JonFraz January 27, 2015 at 2:39 pm

Huh? Young people have high credit scores? That’s a claim I have never heard.

47 datroof jackson January 27, 2015 at 9:27 pm

Jon, you’re like Donny in Big Lebowski.

48 rayward January 26, 2015 at 12:33 pm

Speaking of questionable associations, here’s one that Mankiw has on his blog (approvingly quoting Ed Lazear): “The share of the private workforce employed in the BLS-defined industries “financial activities” and “hospitals” decreased by about 5% between 2010 and 2014. Jobs in these industries pay 29% and 24%, respectively, above the economy mean. Because a smaller share of labor is working those high-wage industries, the typical job in the economy is now lower-paying than in 2010…. So what accounts for the relative decline in jobs in high-wage hospitals and finance? One obvious possibility is increased regulation. The Affordable Care Act for hospitals and Dodd-Frank for finance both passed in 2010, the year real wages began to decline.” It was in all the papers, but perhaps Lazear and Mankiw, tenured college professors, missed an important event that occurred in 2009-14.

49 prior_approval January 26, 2015 at 12:48 pm

And I do wonder how a comparison with Kurzarbeit and its results would look – but then, why use unionized and strictly environmental regulation oriented Germany as a counter point when it comes to OECD unemployment rates. (After all, it was that noted socialist Bismarck that implemented things like worker health care, workmen’s compensation, and something that resembles social security, though 50 years earlier than the U.S.)

But hey, it isn’t as if links to such things as this are likely to appear here – http://www.nber.org/papers/w20694

It tends to be another aspect of American exceptionalism to pick and choose what the U.S. compares itself to.

50 The Other Jim January 26, 2015 at 1:01 pm

>”this is another area where the Keynesian analysts simply have not proven a good guide…”

It just kills you to say “Krugman is wrong yet again,” doesn’t it?

Do you get outright banned from the New York Times for saying it, or is it just a matter of being denied cocktail party invitations?

51 Floccina January 26, 2015 at 3:10 pm

Who goes to cocktail parties?

52 Jamie_NYC January 26, 2015 at 5:42 pm

Nobody. They are too crowded.

53 ThomasH January 26, 2015 at 6:23 pm

It would be interesting to see how Tyler derives a supposedly “Keynesian” policy recommendation about unemployment insurance from a Keynesian economic model. If Krugman is wrong, he’s wrong because of bad microeconomics, not bad macroeconomics.

54 derek January 26, 2015 at 1:33 pm

Unemployment benefits don’t strike me as having a lot of support from a Keynesian viewpoint, as it is easy to see how the distortionary cost could outweigh whatever multiplier effect we are trying to create. Wasn’t the political case for unemployment benefits at the time one of social justice?

55 Donald Pretari January 26, 2015 at 1:53 pm

In the state of Washington, the beginning of 2014 coincided with a large number of uninsured getting health care insurance. Perhaps, now that they had health care taken care of, some unemployed workers were willing to accept jobs without health care benefits. I don’t know that, but I think a bit more research needs to be done about what else might have coincided with the end of UI benefits for some unemployed people. I do think that that the end of benefits, depending upon other things going on in the economy, can cause an increase in employment. But, as I say, if you say that nothing else out there explains it, you’d better be sure nothing else was out there.

56 Kevin Erdmann January 26, 2015 at 2:05 pm

I’ve been on the anti-EUI train, but that’s a good point.

57 JonFraz January 26, 2015 at 3:21 pm

This could also help pull some people off disability: under the old system anyone with a disability was going to have trouble getting health insurance. If they were out of work too long (> two months) even HIPAA regs would not help them. The ACA removes a major distortionary incentive for people with health problems to go on or stay on disability.

58 JonFraz January 26, 2015 at 3:05 pm

I am having trouble seeing a real-world mechanism by which cutting benefits would lead to job creation. Most job seekers: yes. But how would this have any feedback on actual employers? If any, the loss in the purchasing power of the unemployed might adversely impact demand, though I;m willing to consider that this effect might be small.

59 Whatever January 26, 2015 at 5:40 pm

It isn’t free to recruit people. From that perspective it makes financial sense to do your hiring when there are more people looking and the cost of recruitment goes down.

60 JonFraz January 27, 2015 at 2:36 pm

Do you really think there was a shortage of people looking for work before 2014? Most HR departments would laugh at that assertion. They were inundated with resumes.

61 ChrisA January 26, 2015 at 8:10 pm

JonFraz – the mechanism is that people on UB have an alternative to working so the wage they need to persuade them to work is too high for a employer to afford. Once their UB is cut then they will take a job at a much lower wage that an employer can afford. Note that actually the wage needed to persuade people to work might be significantly higher than UB pays. Working has costs, like owning a reliable car, organizing and paying for child care or for other dependents, maintaining a professional wardrobe and so on. There are also psychological costs, you have to do what you are told, you can’t get drunk and sleep it off the next day etc.

62 JonFraz January 27, 2015 at 2:35 pm

Re: the mechanism is that people on UB have an alternative to working so the wage they need to persuade them to work is too high for a employer to afford. –

I am extremely skeptical about this claim.
Most unemployed people are willing to work for a somewhat lower wage than they were pre-job loss. Really, all that is required is that the new salary be substantially higher than the unemployment benefit is, since that is the other option they face. UI pays extraordinarily poorly: the average benefit is c. 35% of what the unemployed person was making before– so if a new salary offer is even just 50% of the previous salary it’s still a step up from UI in most cases.
I am a case in point since I went back to work for 8K less than I had been making (with some fudging there around the question of benefit costs). And why can’t employers afford to hire them at something close to their original wage? Yes, yes, there may be individual situations where that isn’t possible, but a mass effect across the entire economy? That would require that the whole economy is getting poorer. Which was true for a short while in 2008-09, but hasn’t been for a while: the economy has been growing not contradicting.

63 Lord January 26, 2015 at 3:40 pm

One could as well look at this as monetary policy failure. Lower wages are no cure for low demand.

64 foosion January 26, 2015 at 5:15 pm

Dean Baker re-runs the analysis using a more reliable data set and comes to another conclusion. http://www.cepr.net/index.php/blogs/cepr-blog/did-cutting-the-duration-of-unemployment-benefits-lead-to-faster-job-growth-in-2014

65 Rich Berger January 26, 2015 at 8:27 pm

So a one page blog post by Dean Baker reaches a conclusion you like and the authors of a paper that took great pains to lay out the data and methodology should have to respond to what is essentially a broad brush back of the envelope estimate?

66 ckb January 27, 2015 at 11:22 am

fooison did not say what the NBER authors were obliged to do, and it’s unclear why you find Baker’s response so unacceptable. He referenced four prior studies that reached different conclusions, went through a number of methodological concerns, and took a short look at an second dataset which seemed to cut against the NBER paper. That’s not bad at all for a quick response…especially when people who like the conclusion that Hagedorn, Manovskii, and Mitman reached are basically circulating their work as the final word on the matter. Tyler’s post seems to regard the HMM paper as nothing short of definitive. It’s plainly not.

67 TallDave January 29, 2015 at 9:43 am

One of Baker’s links doesn’t work. One says UI increased unemployment, just not as much as earlier surveys, and one is a 2011 study which only looks at post-2008. On the whole, it’s not very persuasive.

68 TallDave January 29, 2015 at 9:49 am

Also, notice that the original authors were testing bordering counties, not states. It seems safe to assume that (for instance) the counties bordering Iowa are more similar to Iowa than, say, North Dakota or New York. Baker is importing a lot of confounding factors into his analysis.

69 TallDave January 27, 2015 at 12:02 am

Dean seems to be unaware of other studies data that support the same conclusion, e.g. http://www.nber.org/papers/w19499

Also, some of his complaints don’t hold water — e.g., he complains that LAUS is sometimes modelled. Well, guess what, so is CES state-level data.

“The majority of State and area estimates are produced using direct sample-based estimation. However, published area and industry combinations (domains) that do not have a large enough sample to support estimation using only sample responses have been estimated using modeling techniques”

Given that employees receive the benefits, it seems logical one would survey them about what they did when the benefits ended. Employer surveys have their shortcomings, too. “The data exclude proprietors, the unincorporated self-employed, unpaid volunteer or family employees, farm employees, and domestic employees. Salaried officers of corporations are included. Government employment covers only civilian employees; military personnel are excluded. Employees of the Central Intelligence Agency, the National Security Agency, the National Imagery and Mapping Agency, and the Defense Intelligence Agency also are excluded.”

70 ckb January 27, 2015 at 11:38 am

Dean did not attempt to analyze the establishment data at the county level, so it’s shortcomings (whatever they may be) at the county level are not relevant to the state-level comparison that he performed. Agreed that every survey has weaknesses, but on the whole, if the establishment survey is saying one thing and the household survey another, my inclination is to listen to the payrolls. I think Dean has made a sound criticism here, not to be brushed aside.

71 TallDave January 29, 2015 at 9:32 am

“State and area” means state and smaller (e.g., county).

Again, it seems logical to assume households are a better source of data about the effect on unemployment on members of households. I haven’t seen any persuasive arguments as to why payrolls would be better.

72 TallDave January 29, 2015 at 9:36 am

But hey, if Dean has a serious criticism to make, then let him do a similar study, where he actually shows his data and methodology. At a glance, his complaints don’t hold water.

73 ThomasH January 26, 2015 at 6:03 pm

The Keynesian position on unemployment insurance, like other expenditures, is that during periods of low real interest rates it makes sense to increase expenditures that have present costs and future benefits. It’s microeconomics — cost benefit analysis, not “Keynesian economics” — that tells us whether any specific expenditure is worthwhile or not. The only insight a Keynesian, qua Keynesian, would have on whether to extend unemployment benefits or not would be to point out that “to reduce the deficit” is NOT relevant to the decision.

Keynesian economics is basically the insight that governments that can borrow in their own currency should not during recessions not behave like credit-constrained households.

74 mulp January 27, 2015 at 4:21 am

Of course, nothing you describe has anything to do with Keynes who simply argued the only way to put money in people’s pockets is with a job and if the private sector won’t give someone a meaningful job, then government must give them a job to put money in their pocket.

Lots of work should be done but never is because the work only benefits the public while conservatives argue only the private should pay.

If a street is filled with trash, and burned out buildings, only the desperate will stay and they probably won’t sell for a low price to a rich developer, and that goes double for the owner of the vacant lot with thousands of back taxes owed who will fight to keep the city from taking it for taxes, waiting to sell it to a rich developer.

Why not put the idle worker to work getting paid to clean up the neighborhood, paid for with higher taxes? Every penny in higher taxes goes right back into the local economy as the worker spends what he earns.

75 Nate January 26, 2015 at 6:11 pm

1.8 million jobs.

Simple maths, because I’m no economist:
America created ~200,000 jobs a month. 1.8e6 / 12 months… ~150,000 jobs.

Wow. A full 3/4ths of job growth was due to cutting unemployment benefits. Who knew the economy was so simple?

Which brings me to my next question: if the economy is really that simply, what have all these economists been doing all these years?

76 ChrisA January 26, 2015 at 8:40 pm

In most of the rest of the world the vast majority of people are willing to work for much less than the UB pays in the US. Why therefore is it surprising that people are willing to stay on UB rather than work? And that if they don’t have UB anymore they will move back to work even if the compensation is much lower than what they wanted? This is not difficult economics.

77 Kevin Erdmann January 26, 2015 at 10:27 pm

Nate, I think they are using different data than you are referencing, which has a different scale of change over 2014. Also, at the end of 2013, many sources, including the Fed, were projecting an unemployment rate much higher than we ended up with at the end of 2014.

Just because something arguably has a large effect doesn’t mean that agreeing about the effect is simple.

78 Grad student January 26, 2015 at 6:11 pm

They really need to respond to Dean Baker’s criticism immediately or else lose all credibility. My first reaction when reading the paper (before even seeing Baker) was “why on earth are they using LAUS rather than CES?”. That’s incredibly sketchy and must be explained.

79 TallDave January 26, 2015 at 11:43 pm

Incentives matter.

80 mulp January 27, 2015 at 4:07 am

Not having read the paper, I’d question its objectiveness.

1) The long term unemployment number did not fall after the end of extended benefits.
2) the labor force shrank significantly after the end of extended benefits.

What did happen at the same time as the end of extended benefits was hikes in the tax rates on capital as the Obamacare tax kicked in.

Thus I argue tax hikes boosted hiring of new workers. Just like Reagan hiking taxes Jan 6 1983 to pay for creating jobs saved him from defeat in 1984 by creating 7 million jobs in six years,

81 Jon January 27, 2015 at 10:25 am

The last paper with these authors was so severely flawed that it is reasonable for me to ignore this one.

It made rather fantastic claims that were close to asserting that almost all of the unemployment of the recession was due to the extension of benefits, even though in their sample it appeared that the counties states that extended benefits saw faster decline in unemployment.

They accomplished this by using a model in which the expectations of future benefits causes the initial high unemployment. I believe it was Arindrajit Dube who did a pretty thorough debunking of that.

These authors are out to prove a dogma, not test the limits of theories or policies.

82 Economist January 27, 2015 at 1:08 pm

Very interesting work. It’s consistent with what Scott Sumner was saying after the last jobs number: you shouldn’t make inferences about aggregate demand from a quantity change in the labor market when we’ve had a shift in the supply curve. I see one of the authors (Kurt Mitman) is answering questions. Is there any estimate for how quickly this boost to job growth will fade? If your estimates are correct, payroll growth would be running 150,000/month lower if extended benefits were still in place. In light of your research, do you guys have any sense of how 2015 nonfarm payrolls will come in vs consensus?

83 glasnost January 27, 2015 at 2:00 pm

Mike Konczal demolishes this propaganda:

http://www.nextnewdeal.net/rortybomb/did-ending-unemployment-insurance-extensions-really-create-18-million-jobs

Reihan Salam is ‘inclined to agree with him (Mike)’.

Reihan Salam @reihan · 2h 2 hours ago
I’m inclined to agree with @rortybomb on the new UI extension working paper: http://www.nextnewdeal.net/rortybomb/did-ending-unemployment-insurance-extensions-really-create-18-million-jobs

84 Baban January 28, 2015 at 3:35 am

I don’t really prefer trading on news as I prefer normal trading which helps me gain decent returns and rewards but we need to be very consistent and patience for it. At the moment I am just trading with a long term approach that’s helping me maintain my trades and not get into desperate mode and it’s only possible because of having the license by OctaFX broker which allow me and all to have swap free account for long term trading.

Comments on this entry are closed.

Previous post:

Next post: