How High Will the Unemployment Rate Go?

Another dismal jobs report for December, with 34,000 more jobs gone (71,000 full-time losses) and the unemployment rate jumping a third of a point in a single month, has got everyone now wondering:  How high will the unemployment rate go in this recession?

The “consensus” view of the mainstream economics world is something like this: unemployment will climb to perhaps 8% over the coming year, then the “recovery” will start (everyone thinks in the second half of 2009).

But this story is incomplete, and misses an important aspect of the dynamics of the unemployment rate in a recessionary cycle.  The recovery in GDP (that is, the point at which is stops falling and starts growing again — which, by the way, is not how most individual Canadians would experience “recovery”) is not at all coincident with the eventual, subsequent decline in unemployment.  With population growth of something close to 1.5% per year, and ongoing trend productivity growth (abysmally low in Canada, but still positive), real GDP needs to grow at 2.5% per year just to maintain a stable unemployment rate.  I do not necessarily disagree with the prediction that real GDP will stop falling (and presumably start growing again) sometime in the second half of this year (although even that may be optimistic).  But when will the recovery gain enough speed to surpass that 2.5% threshold, and start the long hard job of bringing the unemployment rate back down again?

Given the nature of this crisis (and the role of balance-sheet problems in both the financial and the household sector in causing the decline in credit, spending, and hence production), I am with those who see an L-shape to this downturn (rather than the classic V-shape or U-shape).  Real GDP will fall, and then likely stagnate for some time until the conditions are recreated (through a painful process of write-offs, bankruptcies, and value-destruction) for eventual recovery and expansion.  Hopefully this trek to recovery will be aided by pro-active government measures to supplement demand: with public-sector invetsment, transfer payments to low-income Canadians, and other measures.  But even with an appropriate counter-cyclical response by government, this will not be a quick recovery.

In that case, the quick rise in unemployment over the next year will be just the beginning of a longer, more painful story.  In a scenario in which real GDP stagnates or grows very slowly (1-2%) from 2010 onward, then the unemployment rate will continue to grow by a full percentage point or more for another two or three years.  That could easily put the unemployment rate back into double-digit terrain for an extended period of time early in the next decade.

This trend will be buffered somewhat by two normal cyclical responses in the labour market: a likely decline in labour force participation (as discouraged job-seekers give up looking), and a likely decline in productivity growth below its trend.  In the former case, while the decline in participation is “helpful” in bringing down the official unemployment rate, it has no benefit (and is actually harmful) for the economic well-being and social engagement of Canadians.  In the latter case, it’s hard to imagine productivity growth getting any slower in Canada, but it likely will (for the familiar reasons that companies reduce payrolls more slowly than their sales fall, and encounter indivisibilities in capacity which produce a decline in capacity utilization and hence efficiency).  But again, while this helps suppress the rise in official unemployment, this is hardly helpful to our actual economic goals.

I have assembled an ultra-simple forecasting spreadsheet which starts out from the December 2008 labour force aggregates: working age population (27.1 million), labour force of 18.3 million (implying a participation rate of 67.6 percent), total employment of 17.1 million, and unemployment of 1.2 million (now 6.6 percent of the labour force).

Now apply whatever assumptions you wish regarding the decline in GDP (my illustrative case has a 1% decline for annual GDP in 2009, 0 in 2010, 1% in 2011, and 2% in 2012), productivity growth (I cut it in half for two years, to 0.5% per year, then recovering to 1% per year), population growth (1.5%), and participation (let’s assume it declines by 0.2 percentage points a year for the next four years).  I don’t think this is unduly pessimistic (in fact, on most counts it’s likely optimistic — although participation could certainly fall faster than I have assumed if Canadians abandon hope of work).  Yet it generates an unemployment rate above 9% by the end of 2009.  More worrisome, unemployment then continues to grow throughout the next three years — reaching almost 12% by the end of 2012. This isn’t dissimilar from what actually happened to Canada’s unemployment rate during the early 1990s.

(Just to clarify: this scenario is not my personal prediction: it’s just an illustrative example of how the various factors that determine the unemployment rate fit together, and how double-digit unemployment rates are certainly possible in the years ahead.)

Changes in any of these assumptions, obviously, change the forecast.  Play around with it yourself: it’s always instructive to see the interactions of economic growth, population growth, productivity, and participation that are the complex drivers of unemployment dynamics.  (You can assemble your own spreadsheet like this in 5 minutes — but if anyone wants a copy of mine, just e-mail me at

The key conclusions from this analysis are worrisome:

  • The rise in unemployment during the recession “proper” in coming months is just the beginning of the labour market damage resulting from this crisis.
  • Unemployment will likely continue to rise for at least 2 or 3 years after the onset of official “recovery” — and the increase in the unemployment rate during this “recovery” will likely be larger than the increasae in unemployment during the technical recession.
  • This will have an incredibly negative impact on household incomes and well-being for many, many Canadians.
  • The fiscal stresses on Canada’s EI system are going to be much, much larger than is anticipated in the Conservative government’s counter-productive new EI funding system (which allows only a $2 billion cushion, and then requires premiums to be altered annually to balance each year’s forecast expenses).  It is obvious that this system is going to result in successive EI premium increases beginning in 2010 — the last thing the labour market needs as it is trying to claw itself back from recession.  Instead, the government, with a stroke of a pen, should restore the former accumulated EI surplus (in excess of $50 billion) and allow the EI system to experience (appropriate) deficits during the coming tough years.
  • Better yet would be to improve access to EI (relaxing admission requirements, and extending benefits to cover longer-term layoffs) in order to cushion the blow of the weak labour market on Canadian households — not to mention sustain spending in these households and reduce the knock-on contractionary effect of unemployment on aggregate demand.

In short, I don’t think that the economics mainstream (expecting only a moderate and temporary bump in unemployment) has yet to grasp the likely scale of the rockin’ and rollin’ that we’re heading into.


  • I would suggest one minor adjustment to your model. Habitually labour force participation declines in response to the initial drop in economic activity. Then as households get really squeezed the participation rate trends up (during stagnant to low growth phase of the recovery) as households become more desperate to find paid work.

  • Agreed. And your analysis is along the lines of the European think-tank, LEAP, which predicted the current crash. I think we need to start working on non-GDP transactions, barter, sharing, helping each other.

  • This is an important piece of analysis Jimbo. Good on ya. The “official” definition of a recession was created by Martin Feldstein and the NBER. That would be the same Feldstein who just argued in the WSJ the U.S. should stimulate using the military budget, and has called for the UAW to give up wages, in other worlds disband in everything but name (BTW he can be reached at: Tel: 617-868-3900Fax: 617-868-2742 should anyone like to let him know what kind of an economics he practices).
    It would be more appropriate, following Brother Stanford’s lead, to define a Canadian recession as a period of x months, or x quarters, of successive rises in unemployment measured as full-time jobs lost, or decline in hours worked or whatever (I will leave the discussion of how best to measure it to the many more learned on these matters than I). GDP is such an imperfect measure as it is, and so artificial as compared to someone being told they are no longer needed, that it is misleading in a green world to base our idea of how things are going on how GDP estimates look.
    We could have a CCPA recession measure: define it, and announce it and explain why our measure is more appropriate than the one the NBER has hoodwinked people into adopting. Could we devise something more accurate, and meaningful? It might help policy debate to turn attention to the economy as people working, which is what it is, as opposed to (proxies for) market activity, which is what we are supposed to think it is.

  • Layoffs have been in the cards for a long itme. But I’m not convinced that the huge numbers can be justified. Maybe it’s time to require corporations, big and small to state reasons for terminating people. It’s done in other countries. “Everybody’s doing it” & “Times are Bad” are not reasons. The free for all has to end.

  • Jim, we were both quoted in the Globe’s piece on forecasting the unemployment rate for the recession, which was printed as part of the job-loss story in today’s newspaper. I thought that 8% was as good a guess as any on how high unemployment might rise during the recession, but was glad you got in the point that it could rise even higher during the recovery.

    Another factor that reinforces your point is changes in work hours. During a recession, employers cut costs not only by reducing the number of employees, but also by giving employees fewer hours. In December, for example, there was both a large loss of jobs and an equally large substitution of part-time jobs for full-time jobs.

    During a recovery, many employers will initially increase output by giving more hours to existing employees before hiring new employees. While this dynamic may limit the increase in unemployment during the recession, it will also tend to prolong high unemployment after the recession.

    Please do send me your spreadsheet.

    Duncan, officially defining recessions exclusively in terms of real GDP is a Canadian practice. In the US, the NBER has a committee (including Feldstein) that declares recessions based on a broader range of factors, including employment trends.

    I’m not necessarily defending Feldstein, but his group explicitly rejects the definition that Canada accepts as being official: “The NBER does not define a recession in terms of two consecutive quarters of decline in real GDP. Rather, a recession is a significant decline in economic activity spread across the economy, lasting more than a few months, normally visible in real GDP, real income, employment, industrial production, and wholesale-retail sales.”

    Indeed, the fact that the US defines recessions with reference to more than just real GDP gives credence to your suggestion that we develop alternative criteria for recessions in Canada.

  • First up I would like to say I am impressed that the Globe and Mail who quoted two separate labour sources in its search for a “how bad can this get” qualitative opinion poll. Very glad they feel obligated once in awhile. Who actually reads the National Post?

    Second, I know we as economists must somehow base our estimates on something, and having somekind of standards for measurement estimates is a good idea. GDP seems to be full of it, but hey its what they all use. GPI is much better as I stated previously, talk to Hans Messinger (sp) and he’ll tell yah.

    Third, in all these estimates, not once did I hear anything logic or rationale for what exactly was going to turn the whole thing around, hence I guess Jim’s Wiley Coyote- L – off the cliff projection. That long lonely fall off that cliff and length of the bottom of that elongated L.

    I think the article was focused too much on getting out of this magically, and forgetting about the fact that we are still falling, and very fast like. We need the focus to be on what can be done to prevent these scenarios.

    Even Harper’s headlines yesterday were seemingly just accepting that employment will hit some level and then we will magically come out of it and yes there will be job losses.

    Something is really wrong here with the discourse.

    I guess the right is now going into the mode that yes we are in a recession, and we must somehow just accept it and as the right wingers think, let the markets create lots of jobloss as that is what need to be done.

    The Tories seem to be in this shrug your shoulders and yes its going to get bad, but oh well. I don;t like the smell of the headlines, and even these projections seemed to be in a similar vein.

    Some kind of complacency that will never get us out of the ditch. I would have been projecting a lot higher numbers to get peoples attention. Thus I agree with Duncan, we need a diffferent kind of measurement. For example the employment numbers get all womky when we get into change like this. Discouraged workers start shooting up, and they are not included, the precariousness of work become more pronounced and there is no measure for that. And that is just a start.

    Why do we always have to fight on there terrain, it is time we started getting some of our own home field advantages.

    To me, putting numbers on the future that is so unknown at this time just affixes this whole shroud of don’t worry be happy, it’ll all work out kind of romanticized, veneer over the ugly mess that is developing. These are times that social unrest can get out of hand, these are time historically, when change for the good or the better, move out of their dormancy and discover a birth and rebirth. It is all correlated to how far off the cliff we fall. That is the kind fo message I would have sent. Scare the hell out of the policy makers as they all seem to complacent and faithful to the experts. This a not a time for such easy stalwart faces to paint some quick safe scenaros. This is a time to realize, that there is nothing out there right now that actually will be stopping the fall off the cliff. In fact if anything the measures in place, such as wages cutting by corporations is only accelerating the fall, continues unabated and seemingly rational behaviour.

    Okay so am I getting you scared. well you should be cause, if you look back in our history, the only thing that stopped the last fall that we went too far, was a war. Potentially we will not fall that far, but honestly, I do not see any reason to believe we will not.

    We would need some drastic idea changes to occur for any hope of a change in direction, and Obama, and the rest I do not see them having the political or economic capacity to implement these dractic changes. At least not yet. There is some hope, but we will need to have a lot more thinking outside the market box for this to happen.

    some long random thoughts

  • i should clrify, most of my comments above were in regard to the globe and mail article and to a lesser extent Jim’s comments here

  • Thanks for the NBER references Erin. My point was simply that the official definition was given by a private group, and not the government. Whether they do it monthly or quarterly is of interest but not my concern.
    What I do not like is the kind of loose talk we are getting about peaks and troughs and business cycles. I am pleased to see Jim focusing on the neglect of the real economy of people working for a living in the recession discussion. I would like to see it go one step further: how do we define a contraction in work? Why should we accept the Stats Can or NBER way of calculating, anymore than we accept the gold standard?

  • suggestion that we develop alternative criteria for recessions in Canada

    There aren’t any objective criteria for declaring recessions in Canada – or in the US, for that matter. The NBER reference cycle is the collective, subjective judgment of a committee of experts. No-one treats it as gospel (cf the Markov-switching literature), but it’s generally considered to be a decent point of reference – which is why they call it the ‘NBER reference cycle’.

    There isn’t a Canadian counterpart to the NBER business cycle dating committee, and no-one I know takes the 2-quarters-of-negative-GDP-growth criterion as the gold standard.

  • Except that it is the gold standard as far as the business press is concerned.

  • But when will the recovery gain enough speed to surpass that 2.5% threshold, and start the long hard job of bringing the unemployment rate back down again?

    Clearly, this is the $2^6 question: a scenario in which GDP growth rates are higher than the one you outline (-1,0,1,2) will generate lower estimates for unemployment.

    I would question qualifying your assumptions for GDP growth after 2009 as optimistic: the CBO’s projection is (-2.2,1.5,4.2,4.4), and it’s hard to see why the US will grow that much more strongly than Canada after 2009.

    Going back to previous Canadian recessions, y/y GDP growth was 6% in the four quarters following the trough in 1981Q4, and even the recovery from the 1991 recession (for which the word ‘sluggish’ seems to have been invented) managed to generate 3% growth by the end of the third year. And at least this time, we’re not coping with adjustments to NAFTA and a contractionary monetary policy.

  • Three years forward and it is anyones guess what monetary policy is going to be and a strong case can be made that some form of restraint will be in the works whether in monetary or fiscal form. I don’t think the last recessions are going to be a very good guide to, or even educated guess about what, we can expect over the next three to five years. My take is here.

    In nutshell I think Jim’s conjecture may be too optimistic.

  • The Globe article Erin linked has the following paragraph:
    “Canadian Auto Workers union economist Jim Stanford begs to differ – perhaps understandably, given that the auto and auto parts sector has been hard hit, losing 10,000 jobs in the past year alone and 35,000 since May, 2001, carving the work force back to about 120,000.”

    Are there a couple zeroes missing there? Shouldn’t the numbers be 100 000 and 350 000?

  • On a seasonally-adjusted basis, Canada has lost 380,000 manufacturing jobs since manufacturing employment peaked in November 2002. Most of these losses occurred outside of the auto industry, so zeros should not be added to the figures in that article. Although auto production drives (sorry for the pun) much of southern Ontario’s manufacturing activity, the auto industry itself accounts for a small fraction of Canada’s total manufacturing employment.

  • Ahh, that makes more sense. I knew I had seen a figure in the hundreds of thousands somewhere. Thanks for clearing that up.

  • thanks for this post here Jim, and to others for the dialogue.

    It seems that a large problem in doing forecasting, or even addressing the noble goals in the Gutierrez article, is that with the Harper government we don’t even have reliable accounting. It’s rather non-transparent.

    There are also lots of related numbers i’d like to know. Which companies and public sectors are investing how much in which real renewables (like sun, wind, geothermal) and how many good jobs (full-time with benefits) are they creating in the process?

    Maybe folks on this blog have access to, or have already posted, answers to these questions which are Canada-specific. If so, please post/re-post.
    I’ve seen some items on potentials, but not on actual existing comparative numbers. For the employment part, and probably for every other aspect of the necessary shift, it would help us get
    into more useful territory.

    but again, until we can regain access to valid numbers, on the investment and finance aspects, this won’t be possible.

    So in terms of priority, I still think the coalition is #1.
    That way, labour-friendly people in cabinet can at least have a better chance of accessing a more accurate picture.

    I realize this is all very global, and financiers can hide their numbers in a myriad of ways. But again, having labour-friendly people in government can help clarify the picture at the broader level as well.

  • sorry, to finish the connecting thought;
    if we don’t get accurate #s and start tallying increasing creation of good eco jobs, i most definitely agree with Jim’s projection of an ‘L-shaped downturn’, or perhaps a backwards ‘J’ shape, with the economy taking a flying leap off into the abyss after the downhill run.

  • after my long winded comment on this l should say the following short winded one.

    measurement vehicles and the numbers during sharp change are problematic. as stated, definitions, segments of the various populations and the sampling plans enter into a quite complicated morass of the practical and the theoretical.

    However, as stated, the numbers changing within the employment dynamic, definitely start singing different tunes when crisis hits.

    Here is a good example of such from my friends Sherry and John and the Group at the Youngstown Center for Working Class Studies (what a cool site and such nice people).

    The De Facto Unemployment Rate: 25.12%.

  • Useful post to get one thinking about the dynamics, Jim. Not sure what the relevant assumption should be on trend productivity growth. As you have noted, has been negative of late .. indeed over last year of slumping growth October 2007 to October 2008 employment rose a bit faster than output.
    One might also note that albout half of immigrants coming in last while have been temporary migrant workers – whose permits will expire in 2009 and 2010 – and entry of echo bay boomers into work force is now slowing down rapidly , so labour force growth may slow down below 1.5%. On other hand, older worker part rates may well continue to rise as older boomers stay in workforce longer than expected to rebuild retirement savings..
    Key point is indeed that recovery in the job market will long lag an output recovery. I suspect we will hit double digit unemployment in 2010 if only because I see no significant recovery of output on the horizon.
    See December piece by Godley et al from the Levy Institute for a very pessimistic view of impacts of even a major US stimulus package – US savings rate now soaring

Leave a Reply

Your email address will not be published. Required fields are marked *