Some Policy Implications of the Interdependence of Cyclical and Structural Unemployment

I have decided to forego discussion of the potentially heavy burdens faced by households with incomes in excess of $250,000 should the tax cuts not be extended for income in excess of $250K (see the poignant story here), and focus on the challenges of the unemployed, and what challenges persistent unemployment in turn poses for macroeconomic policy. (Side note: our assessment of the plight of the +$250K income households should be tempered by the knowledge that even those households with income in excess of $250K will see a tax cut under the President’s proposal, since household income below the $250K threshold would be taxed at the current lower marginal rate [0])

From the introduction to a paper presented at the joint ILO-IMF conference on Growth, Employment and Social Cohesion, entitled The Human Cost of Recessions: Assessing It, Reducing It, by Mai Dao and Prakash Loungani.

Recessions leave scars on the labour market; the Great Recession of 2007-09 has left gaping
wounds. Over 200 million people across the globe are estimated to be unemployed at present.
Among countries with unemployment data in the IMF’s World Economic Outlook (WEO)
database, there has been an increase of over 20 million unemployed people since 2007. The
ILO estimates that globally the increase is over 30 million. As shown in the left panel of
Figure 1, three-fourths of this increase in the number of unemployed people has occurred in
the ‘advanced’ economies (the term used in the WEO to denote high per capita income
countries) and the remainder among emerging market economies. The unemployment rate
has increased by 3 percentage points in advanced countries since 2007 and by 0.25
percentage points in emerging markets…

The paper documents a wide variety of effects, but one striking implication is the impact on mortality, as shown in this graph:

Figure 5 from Dao and Loungani (2010). Notes: Marginal effect of displacement on odds of mortality, with 2 SE bands. Source: Sullivan and von Wachter (QJE, 2009).

In terms of implications for macroeconomic policy, I thought this graph was of relevant. It shows that the longer the duration of unemployment, the lower the probability of employment in the next period. Depending upon the interpretation of this correlation, there are important public policy implications. If the extended duration of unemployment implies depreciation of skills relevant to the labor market, then this implies short term (cyclical) and long term (structural) unemployment are related phenomenon.


Figure 5 from Dao and Loungani (2010).

The (unoriginal) idea that long term cyclically induced unemployment can result in highly persistent quasi-structural unemployment implies that there is no sharp distinction between cyclical and structural unemployment. Econometricians might like to think of this point in the context of the distinction between the Beveridge-Nelson and Unobserved Components approaches to separating trend and cycle (see here for e.g.). I’m not saying the correlation of shocks to the cyclical and permanent components is one, but the above implies it’s bigger than zero.


The authors write:

…With hysteresis, government policies
that stimulate demand confer larger benefits than otherwise if they can counteract the
increase in the natural rate and bring output back to potential. . A cross-country study by Ball
(2009) presents evidence that long-lasting demand expansions in the past have indeed
permanently driven down the natural rate of unemployment. Under some illustrative
calculations, the increased tax revenues from the return to a higher potential can lower the
debt-to-GDP ratio over the medium-run.
There is a risk of hysteresis in some countries, particularly in the United States and Spain,
given the sharp increase in the duration of unemployment and the persistent nature of the
shocks (e.g. to the housing sector) that lie behind the cyclical weakness in the economy and
hence the increase in unemployment (see Benes et al., 2010 and Vitek, 2010). Hence to the
extent that countries have fiscal space, exploiting it when there is a risk of hysteresis may
create jobs in the short run without hurting the medium-run fiscal outlook.

What this, and other research, suggests to me is that one shouldn’t just throw up ones hands, when confronted by the possibility of structural unemployment. First, one needs to assess the magnitude of structural unemployment (which in my view is at the moment not the majority of the 9.6% unemployment rate [1], but more like 1-1.75 percentage points). Second, one has to then consider whether some efforts to reduce cyclical unemployment might also reduce cyclical structural unemployment.


Stephanie Guichard, Elena Rusticelli at OECD have examined the links from unemployment to long term unemployment to structural unemployment. From this, they estimate the change in structural unemployment arising from the crisis and great recession:

Figure 7 from Guichard and Rusticelli (2010).

More on the structural/cyclical unemployment debate: [3], [4], [5], [Econbrowser/Hamilton], [DeLong], [Economist symposium].


I would note that the two statistical estimates of the increase in structural unemployment (IMF, OECD) based upon two different econometric approaches are substantially less than Minneapolis Fed President N. Kocherlakota’s estimate of more than 2.5 percentage points [6] (which is the sum of a calibration using Shimer’s matching model, and the SF Fed’s estimate of UI-induced unemployment).


27 thoughts on “Some Policy Implications of the Interdependence of Cyclical and Structural Unemployment

  1. Robert Bell

    Menzie: minor typo in this sentence about 2/3rds of the way through?
    “whether some efforts to reduce cyclical unemployment might also reduce cyclical unemployment”
    I think you meant “structural” for the latter “cyclical”?

  2. Buzzcut

    Using a marginal analysis, your comment regarding the tax cut for people making over $250k under the Obama plan is irrelevant (it doesn’t change the marginal tax rate for the next unit of work).
    Regarding Figure 5, does that not argue against unemployment insurance in and of itself? Again, at the margin, UI discourages one to seek work. The consequences of doing so, according to Figure 5, is that you risk NEVER working again. That’s a frightening government policy.
    But I guess the real question that inquiring minds want to know, since Menzie brought it up, is are U of W profs as well compensated as those at U of C?
    I don’t know about Cheezheadland, but my state posts government employee salaries online. I may have to do some investigating…

  3. Buzzcut

    Well, the information is a little out of date (’07 to ’08)…
    Not sure if Menzie is married, but assuming that he has the same marital situation as Henderson, I’d guess that he is in the household income situation being targeted by the Obamaniks.

  4. rayllove

    I live near Killeen/Ft. Hood in Central Texas. This area has no shortage of jobs. The classifieds throughout central Texas list a long and varied assortment of low-paying jobs.
    However, by example, I know journeyman carpenters who are unable to find jobs at half of their previous wages in capacities ranging from janitorial to maintenance etc. What little work that is available to construction workers in this area is being done by Hispanic crews, mostly by contract. These crews do not consist of undocumented workers exclusively, but fluent Spanish is a must and segregation is rare.
    I live in a rural area where it is made obvious that the competition with undocumented workers intensifies as the distance increases outwardly from metropolitan areas. The diversity regarding job types also expands into occupations that are essentially the entire spectrum of jobs that do not require some form of documentation. For instance, undocumented mechanics have small businesses doing auto/tractor repair etc. Put simply, undocumented workers are not just workers, they have a wide range of skills and they offer those skills at a price level that is nearly impossible for citizens to compete with due to the significant differences in overhead costs. And they have become entrepreneurial in all of the ways open to bidding for work, by commission etc., but mostly as subcontractors in competition with each-other more than with legal businesses.
    Immigration, both illegal and otherwise, is causing a significant amount of ‘structural’ unemployment that seems to be ignored in the macro-assessments. This subtle type of structural unemployment would also be much worse if it were not for military recruitment. In our rural areas those coming out of high-school but choosing not to attend college have very, very, few opportunities as things are, and if those being recruited were added to the existing surplus the shortfall would be extreme. The graduating class that included my son two years ago had about 15% to 20% of the males going into the military ‘before’ they graduated. How many others who either joined after graduating or after dropping out I don’t know, but… of all of my son’s immediate friends (12), 4 of them are currently serving. My son and 3 others out of that 12 are still in school and the remaining 5 are still at home, unemployed, but not being counted as unemployed. These young people ‘are’ however structurally unemployed at least to a degree, or as compared to when I was their age and living in commensurable circumstances (hysteresis), and… it may well be that the economic framework has left us with a choice between a militaristic culture or increased levels of structural unemployment.
    This has nothing to do with party politics. Our nation, and our culture, are simply in decline and I worry that macro economics as a tool is too blunt for the problems at hand. But I mean no disrespect.
    Ray L. Love

  5. Menzie Chinn

    Robert Bell: Thanks for catching that — fixed now.

    Buzzcut: But welfare should depend on overall consumption feasible within the intertemporal budget constraint, which depends on the average tax rate. Even retaining the marginal perspective, which pertains the to the labor/leisure tradeoff, if one is concerned with supply-side issues, it seems to me an open issue. At that income level, it’s not clear to me the relative price effect dominates the income effect. (This is micro/textbook analysis of incentive effects — at least it was textbook when I was an undergrad, 30 years ago).

    I don’t think Figure 5 makes the argument for eliminating UI. I believe the sample pertains to a period in which the extraordinary extensions did not exist, so would fall far below the 1 year span; and more importantly, we don’t know what the counterfactual would be — what would be the mortality rate in the absence of unemployment insurance; it could very well be even higher.

  6. Cedric Regula

    Jim Rogers advises us they are hiring in China, but Chinese is the language of choice there.
    He’s such a wise guy.

  7. Buzzcut

    Menzie, I leave the welfare effects aside, my only concern is supply side issues. I think that you make things muddled when you repeat that this only applies to people with more than $250k in income.
    The $250k figure applies to households, does it not? Individuals will be targeted at the $200k level.
    And there are more than marginal tax rate increases on order. There are accelerated phaseouts of every deduction and credit. No child tax credit, no mortgage interest deduction, no state and local tax deduction, nothing.
    So the two income couple with children, with a household income of more than $250k, will face marginal tax rates well in excess of 39.5%. My guess is that this will “encourage” (force?) many high income women out of the workforce.

  8. Menzie Chinn

    Buzzcut: OK, if one ignores welfare (not surprising), then I still I don’t understand your point — I didn’t mention individuals in my post or my comment. You did. But to clarify, households with married individuals, filing jointly, at $250K, households with single individuals, at $200K.

    Will this encourage “many” high income women out of the workforce? I don’t know. As I tried to stress, when relative prices change for labor, then one has offsetting income and substitution effects, which is not the case in other markets. Perhaps it will return female participation rates to those that obtained pre-2001, although that would require imposing cet. par. Somebody with a detailed model will have to answer that question (although given the current composition of employment, perhaps it would be less sexist of you to consider the possibility that the second wage earner is the male, and that this would induce changes in the participation rate of males — which as I say could go up or down).

  9. ReformerRay

    In the poignant story, the U. of Chicago law Professor says he is not rich enough to hire expensive lawyers, so he must pay at the prescribed rate.
    The tactics used by the super rich to escape paying minimum income taxes is more offensive to me than their objections to raising the legal rate 3 or 4 percentage points. People with equal earnings are not treated equally.
    Instead of raising the top rate, Obama should change the tax system to a flat rate for all income over $250,000 regardless of martial status. If all who filed reported the amount of Gross income over $250,000 (regardless of source, whether tax free municipal bonds or capital gains), the rate could be lowered from 34% to something like 31% amd still collect more revenue.

  10. jazzbumpa

    Buzzcut -
    Do you seriously believe that an increase in the top marginal rate of 3% will drive people (of any gender) making in excess of six figures out of the job market?
    I’d love to see that marginal analysis.
    And your contention that UI discourages someone from seeking work shows just how out of touch with reality you are.
    Now, how about posting your own income statement?

  11. bernard

    very interesting how the discussion instantly shifted to the story Menzie Chinn did not want to discuss…It looks like the plight of the super-rich are of more interest than unemployement hysteresis. Coming from hysteresis territory (eg europe)where we have been discussing this for 30 years, I would suggest that people ought to get more interested in the subject of this blog post. You see, this is the first time in my memory at least where we have preceded the US. Study Europe, now.

  12. endorendil

    On the salient issue of long-term unemployment translating to future employability.
    I don’t think that this is dominantly a cause/effect thing, i.e. while for some hiring managers a long spell of unemployment may disqualify candidates, the main issue is more likely that someone who has not been able to find a job with their current credentials, will have a hard time finding one at any point in the future. As our economies change to reflect new realities, the harsh truth is that people may have to change themselves to match the new reality – or be left behind. Without continuous education and a willingness to completely change what you do and where you live, individuals cannot possibly expect to stay employed.
    When you find yourselves unemployed with no immediate prospects, you have to consider changing careers. Not after two years, not after one, but within months of losing your job. You may have to consider moving – if you’re renting – and you may have to think about leaving the country – if you speak other languages -, at least temporarily. That is how companies think, and if you want to be employed by one, well, you have to think like one.
    All of this runs squarely against what people generally want. But we’ve known for over a century that what people want is generally at odds with what companies want.

  13. Ricardo

    Menzie wrote:
    I have decided to forego discussion of the potentially heavy burdens faced by households with incomes in excess of $250,000 should the tax cuts not be extended for income in excess of $250K (see the poignant story here), and focus on the challenges of the unemployed, and what challenges persistent unemployment in turn poses for macroeconomic policy.
    It would be more interesting fo ryou to discuss th epotentially heavy burdens facing households making less than $250K if taxes are raised on those making over $250K. Since those making over $250 are paying almost all of the federal income tax when they stop making income they stop paying taxes. The tax burden will shift directy to those making under $250K.

  14. tj

    If I am reading the BEA tables correctly then federal expenditures and state/local expenditures are roughly equal ~ $3T each, so total G spending (fed + state + local) is around $6T in 2008, 2009.
    Total national income, as measure by nominal GNP was around $14T in 2008, 2009.
    That means the average ‘total’ tax rate needs to be such that $6T = Rate*(14T), so the average tax rate would need to be ~43% of all national income, if all G revenue came from income, which it ultimately does.)
    Some say we should cap Federal spending at 20% of GDP, or GNP in this example. Apply the same rate to state and local G spending and we still have a need for a tax rate 40% of all national income.
    This is similar to what I pay when I add my income tax, payroll tax, local tax, etc. It seems I bring home about 60% of my middle class gross pay.
    The point here, is that even if my numbers are not exact, the national debate over taxing the ‘rich’ is simply a diversion. This will not solve the real problem.
    The problem is that we have too much government spending and we pay too much our income in taxes.
    Think of it this way, if we eliminated all taxes and replaced them with a single national sales tax, it would need to greater than 40% of each dollar of national spending.
    We need smaller government and lower total taxes.
    Wake up people! Do we really need to surrender 40% of our annual income to the government? Really?

  15. rayllove

    I’m a little confused about this:
    “which in my view is at the moment not the majority of the 9.6% unemployment rate [1], but more like 1-1.75 percentage points”
    Does that mean that you are comfortable with the notion that cyclical unemployment represents 8.6%- 7.85%. Does your contention give any consideration to what % of structural unemployment may have existed before the downturn?
    And, if I haven’t already used up too much of your time, I am also a little confused about which of these 2 categories would best suit ‘technological unemployment’?

  16. rayllove

    bernard, presumably, with some odd accent, says:
    “very interesting how the discussion instantly shifted to the story Menzie Chinn did not want to discuss…”
    If there is one thing that pisses me off more than any other thing, it is being offered advise on ‘labor’ issues, by someone who, presumably (again), does not so much as know how to spell the word “labour” CORRECTLY!
    What “Europeans” don’t get is this: we are obsessed with taxes because taxes are what determine how the booty is split-up. But that is because we know that any little inefficiencies that might limit the bounty from our economy, could, and will be more than made-up-for by the taking of any efficiency gains from YOUR economy.
    So, instead of interfering in our affairs, get busy finding new and clever ways to exploit some poor nations somewhere. If you know so much about ‘structural’ unemployment, apply that knowledge to all of the grimy workers in all of the little crappy nations, and get all you are able to out of your own smelly workers too while you are at it. Then… buy some T-bills and pay your share of our taxes as we focus on a way to inflate our debt your way. We will of course provide any needed military support.
    As for this bit of irony:
    “Study Europe, now.”
    It just so happens that we ‘have’ been studying Europe ever since we realized the limits on what we could take from the rest of world. When we ‘Americans’ (yes we will take nomenclature too) say: “Be careful what you wish for”, we are giving fair warning.
    I would explain further but that is about everything that you need to know, and ‘alas’… YOU MUST GET BACK TO WORK.

  17. Buzzcut

    Jazzbumpa, the increase in marginal tax rates is far in excess of the 4.5% bump in the top rate. There are a lot of phaseouts of deductions and credits as well.
    As for UI discouraging the unemployed from seeking work at the margin, I’m sorry, I thought that multiple regression type analysis between states and countries with different UI systems had determined that to be true. I must be mistaken.
    BTW, I don’t make the argument that taxes must not be raised. If I were dictator for a day, I’d phase out credits and deductions, particularly the mortgage interest deduction and the deduction for state and local taxes. Then I would make health insurance a taxable benefit. These moves would raise taxes on high income earners in a way that doesn’t discourage them from working.

  18. Menzie Chinn

    rayllove: The figures generated by the IMF and OECD are for changes in the level of structural unemployment. My view is that structural unemployment prior to 2007 was probably not large, and hence, of the increase in unemployment since then, no more than 1 to 1.75 ppts of the increase is structural in nature.

    I would think of technological change as primarily structural in nature — i.e., not due to demand deficiency, but rather to changing of match between labor skills and firm needs, as we computerize away, for instance, manual data entry.

  19. jazzbumpa

    Buzz and Ricardo -
    You are ignoring the hugely disproportionate benefit to the wealthy of preserving the middle class tax cut, which Menzie mentioned in the post.
    Further, the reason those at the top pay most of the tax, is that they have most of the income. How much they make collectively is a bit more important than their percentage of the population.
    And yes the top 5% pay a greater share of total taxes than their share of total income. It’s called progressive taxation, and it’s not particularly onerous.
    You have still made no case that raising taxes in any fashion discourages anyone from working. Taxes during the Reagan admin. were higher than now. Of course, we also had all those welfare queens . . .
    Re: UI causing a marginal disinclination to work – I’m perfectly willing to be wrong. Show me some facts.

  20. Buzzcut

    Menzie’s last comment is interesting.
    What, instead of technological change, the increase in unemployment was due to the burst of the housing bubble? And what if much of the pre-bubble employment was housing related? And what if that employment was not “real”, but directly related to the bubble?
    And what are we going to do with all those construction workers, realtors, mortgage brokers, Harley-Davidson salespeople ;) etc.?

  21. rayllove

    I may be putting to fine a point on this but, this from Wiki is the cause of my confusion:
    “Much technological unemployment (e.g. due to the replacement of workers by fewer workers who use machines) might be counted as structural unemployment. Alternatively, technological unemployment might refer to the way in which steady increases in labor productivity mean that fewer workers are needed to produce the same level of output every year. The fact that aggregate demand can be raised to deal with this problem suggests that this problem is instead one of cyclical unemployment. As indicated by Okun’s Law, the demand side must grow sufficiently quickly to absorb not only the growing labor force but also the workers made redundant by increased labor productivity. Otherwise, we see a jobless recovery such as those seen in the United States in the early 1990s, in the early 2000s and in the two year period after the 2008 economic meltdown.”
    However, the most important criteria for structural unemployment would seem to be the hysteresis aspect. This because the remedy for technological unemployment and that of structural unemployment would correspond. The remedy for cyclical unemployment on the other hand would be too inaccurate and therefore require excessive and wasteful stimulus, unless of course I am missing something here.
    Thanks, Ray.

  22. Menzie Chinn

    rayllove: I think I see your point. Easier for me to think about it in terms of output. Define y* as log potential GDP, y as GDP, (y-y*) as the output gap. Higher productivity growth due to technological change could both raise y* — since less labor is required per unit of output — or lower it — since the match between labor stock characteristics and labor requirements might have changed. Consequently, even holding y* constant, (y-y*) could move either way. Since (u-u*) is linked to (y-y*) by Okun’s Law, then you can see the amount of cyclical unemployment could move in either direction, even holding u constant.

  23. rayllove

    Most of what you said I am only able to understand by assuming that your explanation is some sort of algebraic or ancient Egyptian way of saying what the ‘Wiki’ quote says in English. I am unsure whether I should feel flattered because you have overestimated my skill-set here, or whether I should feel ashamed for not knowing more than what I do. But I think I agree, with what I think that you think that you said, if what you said is what I think it is!
    That said, I still believe that the best standard for judging the subtle differences between cyclical and structural unemployment is that which separates jobs into 2 simple categories: Those which are temporarily lost and those which are gone for good. Maybe this is awkward for those weird algebraic efforts to bend reality but it seems more important to bend stimulus to what allows it to do the most good.
    Anyway… my mind prefers things kept simple and clear and so perhaps my categorizing of unemployment and underemployment types is oversimplified. My ‘temporarily lost’, or ‘gone for good’ standards though would lead to an understanding of how to stimulate an economy in different circumstances with improving accuracy.
    However, as my comment above regarding undocumented workers suggests, underemployment especially, but some unemployment too, becomes a very complicated dynamic to quantify, no ‘neat fit’ possible perhaps?
    My only agenda is to eliminate the algebra.

  24. Buzzcut

    You know, jazzbumpa, I was going to tell you to do your own damn Google search (in the nicest way I could say it ;) ), but then I decided to look to make sure my own recall on the issue was accurate.
    Google: “unemployment insurance” marginal economic analysis regression “cross country”
    Among other things, I got this:
    The time spend searching for a job is inversly related to the generosity of the unemployment benefit, among other findings.

  25. jazzbumpa

    Buzz -
    Thank your graciously supplying a link.
    The sentence you quote above, from the abstract, is the most relevant statement listed there.
    So, saying, “at the margin, UI discourages one to seek work,” has some validity.
    But, to say “The consequences of doing so, according to Figure 5, is that you risk NEVER working again. That’s a frightening government policy,” is a wildly speculative extrapolation.
    Further, the paper goes on to say (emphasis added): “UI is expected to affect the duration of unemployment through its effect on the amount of effort devoted to searching for a job and the reservation wage of the unemployed, yet these variables have rarely been studied directly.
    So it’s not nearly so cut and dried as you suggest, especially in a severe economic downturn when job-seekers outnumber available positions by about 5X.

  26. Scott

    Ummm. How are any of the policy proposals considered tax cuts? More accurately we are talking about raising taxes or holding rates at their current levels.

Comments are closed.