The Great Debate©: The Job Guarantee Brouhaha

by Hugo Heden and John Lounsbury

John Carney, a senior editor at CNBC.com, has been writing a series of articles about the job guarantee (JG) aspect of what is known as Modern Monetary Theory(MMT). The JG is also known as the employer of last resort (ELR), a concept often associated in modern times with economist Hyman_Minsky.

On December 29 Carney posted an article which has raised a vigorous response from many MMT protagonists. The specific three issues Carney raised were really presented as statements of fact as follows:

  • It’s massively inflationary.
  • It’s a bureaucratic nightmare.
  • It’s economically stagnating.

There has been a great deal of discourse in informal discussion groups on the internet, including one that we have participated in hosted by Roger Erickson.  A flavor of some of the problems that the proponents of MMT have with Carney’s three statements can be sampled in the comment stream following the article.

In this article, which developed from our discussion in the Erickson group, we hope to bring together some of the salient points related to Carney’s three statements and attempt to establish some positions that address arguments against their validity.

1. Inflation

Carney suggests that a JG program would be “massively inflationary”.

MMT proponents acknowledge that at the point in time the JG program is introduced there could be price distortions, but these would be less than would occur with the same level of aggregate demand without JG.  See Wray and Mitchell, page 7.
If there is considerable slack in the economy (businesses have problems with low sales and operate below full capacity etc.) an increase in aggregate demand should in fact not be inflationary according to MMT proponents and other economists.

Businesses running below capacity are likely to meet increasing demand by increasing output — not prices.  See “Painstaking, dot-point summary – bond issuance doesn’t lower inflation risk” (Mitchell). (Apparently, this is a conclusion that follows partially from results presented by neo-classical economist Alan Blinder in his book “Asking About Prices”.) The rationale for opting to increase output is to not lose market share – or even increase it; raising prices (instead of output) while competitors keep their prices constant is risky for any business.

Anyhow, for the sake of argument say instead that JG is introduced when the economy is running at “full capacity” (with say 5% unemployment) at a point at which an increase in aggregate demand would indeed be inflationary. Bill Mitchell has called this the “inflation barrier” (Mitchell, ibid).  This is somewhat akin to an unemployment level the mainstream calls NAIRU.  See Wray and Mitchell, page 8.

Say that the introduction of JG results in an increase in (nominal) aggregate demand stressing capacity constraints. In that case, taxes could be raised somewhat to compensate — and inflation would not result.

There is also the option of letting aggregate demand increase, accepting a bit of inflation. After a while, prices should stabilize. The inflationary episode should be a once-off, continuing only for a limited period of time. Why? Because during an inflationary episode, prices and wages are rising — but the JG “pool” of workers that businesses can hire from remains at a constant floor wage level. As the discrepancy between nominal wages in the non-JG sector versus JG-sector increases, it will be cheaper and cheaper for businesses to find good labor from within the JG pool to hire.

So what happens then? The size of the JG “pool” decreases while the capacity of the private sector expands. That is — the government deficit spending decreases which dampens aggregate demand, except to the extent that production increases. Ultimately, prices should fluctuate around a new level supported by increased private employment and with reduced government JG support.

This is, say the MMT proponents, the way in which the JG program acts like a “nominal price anchor” and “automatic stabilizer”, much like unemployment benefits do.  During an economic slump, the JG program causes government deficit spending to automatically increase.  In a boom, deficits are automatically decreased – The JG program automatically (partly) offsets fluctuations in private sector spending (commonly known as “business cycles”). This is similar to how unemployment-benefits work as an automatic stabilizer for the economy.

Carney’s argumentation rests on the statement that the JG program adds to aggregate demand, while the “work product is largely waste” – and an increase in demand without a corresponding increase in “supply of desired goods” is inflationary.

We will note that the JG output should be useful, and absolutely not “largely waste”. Examples would be environmental restoration, urban development, social work or infrastructure repair, maintenance, development.

But ignoring that, MMT does not accept the theory of inflation invoked by Carney.  As indicated above, MMT proponents hold that an increase in aggregate demand is likely to be inflationary (like Carney suggests) when the economy already operates at full capacity. But when there is slack in the economy (aggregate demand is considerably lower than the “inflation barrier”), an increase in aggregate demand would not be likely to cause prices to increase.  Instead, to meet increased demand it is output that would increase – in a move toward greater capacity utilization. This is an entirely good thing.

2. Bureaucracy

Carney indicates that a Job Guarantee scheme would be a “bureaucratic nightmare”, as that creating 13.5 million (the number referred to as the currently unemployed) jobs would be “impossible”.

First, an MMT aware regime would use other means to stimulate the economy — all the way to the inflation barrier. The end result would perhaps be a JG pool fluctuating around a few million.

Yes, there will be bureaucracy. The current system has its flaws too though. Mass unemployment results in huge permanent losses every day in foregone output and income. Add to that the depreciation of human capital, increasing family breakdowns, child abuse, crime, medical costs, skill loss, psychological harm, ill health, reduced life expectancy, loss of motivation, racial and gender inequality and loss of social values and responsibility. The personal, family and community losses are enormous and persist across generations.  See MMT Wiki.

In the light of this, a bureaucratic burden seems a reasonable price to pay.  And the bureaucratic structure is already in place – unemployment insurance benefits administration which would largely be replaced by JG administration.  Unemployment benefits would not be paid to the unemployed, they would be paid to those temporarily entering the JG programs.

It should be noted that the idea is not for the federal government to “central plan” the millions of jobs that may be needed. This is best done locally by charities, environmental organizations, other non-profit groups and possibly by local businesses. The role of the federal government is solely to finance the program and not to define what job training assignments are needed.

3.  Economic Stagnation

Carney claims the jobs program would inhibit movement of people into productive new lines of work.  He wrote:

Economic booms are often characterized by malinvestment–people dedicating capital to projects that turn out not to be economically sustainable. This is true not just of financial capital—it is true of human capital as well. People learn trades and develop career networks that turn out to be worth far less than they expected.

Unemployment encourages those who went into trades that turn out to lack adequate demand to give up those trades and seek another. This is economically productive because it brings stagnate resources—people who can do things no one will pay for—out of stagnation.

The Jobs Guarantee would eliminate this process. The government would buy the labor of people who hold skills not demanded by the market, preventing those people from seeking out new skills. Stagnant human capital would just continue to stagnate.

Carney’s statements are assumptions rather than facts.  He assumes that the jobs program would provide no avenues to new skills for those with skills no longer needed. He could just as well assume that the JG program would provide employment that would be directed toward areas were skills are lacking and that part of a JG position would actually involve retraining.

Indeed, the JG could be integrated into a coherent training framework to allow workers (by their own volition) to choose a variety of training paths while still working in the JG. The retraining could be in skill areas that private industry specifies have shortages of qualified workers. If there are no such shortages in private industry, then Carney’s stagnation would be occurring in spades, with or without a JG.

If private industry has no shortages in any skill category, we would argue that feeding these unemployed a stream of unemployment insurance benefits and then allowing them to drop off the end after 26 weeks, 52 weeks or 99 weeks offers no opportunity for re-employment at all.  Contrarily, the capacity and skill level of unemployed workers has been shown to deteriorate over time.

There is also a hysteresis effect in business cycle downturns, meaning that yet more workers finally become unemployable or give up seeking work.  In that sense, they are dropping out of the “pool of unemployed” at “the bottom”. This reservoir at the bottom becomes the real stagnation risk of a downturn, where the long-term unemployed have no way out.

Since a certain number of job-seeking unemployed (or JG-workers if a JG program is implemented) is required to discipline wage demands, this leads to ever larger numbers of fresh unemployed or underemployed being required — in order to maintain the inflation fighting effect of the “unemployment pool” when demand starts to increase.  See MMTWiki – Full Employment along with Price Stability.

A JG program, on the other hand, would offer skill maintenance or retraining, considerably increasing likelihood of re-employment, and of finding skilled job-ready workers as needed for business.

We thus argue that a mass unemployment system should be far more economically stagnating than a JG based system.

4.  Conclusion

John Carney is not a lone voice in raising the issues he has articulated.  From Wray and Mitchell, ibid there is a list of those who have raised the same issues.  Here is a list from that paper:

  • Aspromourgos, A. (2000) ‘Is an Employer-of-Last-Resort Policy Sustainable? A Review Article’, Review of Political Economy, 12(2), 141-155.
  • Kadmos, G., and O’Hara, P. (2000) ‘The Taxes-Drive-Money and Employer of Last Resort Approach to Government Policy’, Journal of Economic and Social Policy, 5(1), 1-
  • King, J.E. (2000) ‘The Last Resort? Some Critical Reflections on ELR’, Journal of Economic and Social Policy, 5(1), 72-76.
  • Kriesler, P. and J. Halevi (2001) ‘Political Aspects of ‘Buffer Stock Employment’, in E.
  • Carlson and W.F. Mitchell (eds.), Achieving Full Employment, The Economic and Labour Relations Review, Supplement to Volume 12, 72-82.
  • Mehrling, P. (2000) ‘Modern Money: Fiat or Credit’, Journal of Post Keynesian Economics, 22(3), 397-406.
  • Ramsay, A. (2002-3) ‘The jobs guarantee: a Post Keynesian analysis’, Journal of Post Keynesian Economics, 25(2), 273-292.
  • Sawyer, M. (2003) ‘Employer of Last Resort: Could It Deliver Full Employment and Price Stability?’, Journal of Economic Issues, 37(4), December, 881-907.

The Wray and Mitchell paper presents a thorough analysis of these issues in response to the papers listed above.  The arguments we have presented here are a non-academic discussion of some of the factors as we perceive they have evolved since the 2005 paper by Wray and Mitchell with a few links to discussions presented since 2005.  We have not tried to make a thorough search of the literature.  If a reader recognizes research since 2005 that would bear adversely on the Wray and Mitchell results we hope they will point such work out to us.

Our conclusion is that Carney appears to be a captive of his assumptions rather than a critical analyst when it comes to the JG (Jobs Guarantee) concept.  We do not feel that we can prove his three statements are wrong under all scenarios but we think it is clear that he is not correct in many realistic situations.

Related Articles

Analysis Blog and Opinion Blog articles about MMT


About the Authors

Hugo Heden is Swedish and lives in Stockholm, Sweden. He has worked for nine years at the Swedish Defense Research Agency, mostly with computer science issues, air warfare and software warfare simulations. His free time is spent way too much on trying to understand and explain MMT. There is also time for the occasional (or rather frequent) visit to the local coffee shops.

John Lounsbury is Managing Editor of Global Economic Intersection.


15 replies on “The Great Debate©: The Job Guarantee Brouhaha”

  1. This assumes that governments can produce a net positive result by ‘managing their economies’. What evidence is there for that assumption?

    There is not a single ‘large government’ nation that is standing tall. All are staggering under enormous debt loads, their politicians experiencing crushing dissent. Almost all have declining ratios of workers / retirees, almost all due to declining populations of the native population.

    It is the same wherever ideologically-based government, mercantilism, Progressivism, … has controlled the system.

    We 99% are finally catching on, and the catchy combination of Keynesian policy for jobs won’t reverse that. If it ever could be implemented, this plan would merely give governments more power that would be abused against us. The 0.1% oligarchy would go on getting ever-richer.

  2. “All are staggering under enormous debt loads … ”

    All this reveals, Lew, is that you aren’t bothering to sample even 10% of the data accumulating. The very definition of “fiat” – as in fiat currency – means that there is no debt except “fiat debt”, which is an accounting definition describing the increase in fiat currency required when population and/or distributed economic activity occur (i.e., as in almost every year for the past 200K years?).

    The message that should be catching on is that if you want better public governance, then get the hell involved. Opting for less self-governance in the face of population growth isn’t going to cut it. Once way or another, we have to constantly discover better/faster/leaner way to self organize. Otherwise, it’s back to the mindless war that you implicitly call for.

  3. Hugo + John, Good article. I’ve got a few criticisms.

    1. Under the “Inflation” heading you say that if the economy is at NAIRU when JG is introduced, the inflationary effect can be nullied by raising taxes. True. But the net result is the creation of JG jobs AT THE EXPENSE OF regular jobs, which is not the objective of JG: plus that casts doubt on whether the introduction of JG raises GDP.

    Next (para starting “There is also the option..”) you suggest letting inflation run for a bit, while the pay of JG workers (and presumably also those on unemployment benefit) remains constant in nominal terms. And the net result would be JG employment rising with no decrease in regular employment and no long term inflation.

    Agreed. But the outcome is obtained by moving the goal posts. I.e. you’ve broken the all else remaining constant rule, I suggest. That is, if one cuts the pay of JG workers and the unemployed, it’s pretty obvious there’ll be an increased incentive for both to find regular work, and GDP would probably rise.

    Under the “Economic stagnation” heading, you argue that JG could involve training. Unfortunately there is a clash between the characteristics of a training course and JG, which is thus. Nearly every worthwhile training course lasts for a SPECIFIC PERIOD. That is incompatible with the fact that JG employees need to be available for regular jobs as soon as a suitable vacancy appears (otherwise aggregate labour supply is constrained, which is inflationary).

    No doubt you could have JG people leave their courses half way thru the courses, but that would make the money spent on such courses bad value for money.

    There is however a way round that clash which would normally work: and that’s evening classes at some college. (Most jobs involve working roughly 9-5, rather than in the evening.)

    Finally, re inflation and JG, I suggest inflation kicks off when unemployment is so low that employers cannot find suitable labour amongst the unemployed. They therefor spend ever more on advertising vacancies, which bids up the price of labour generally. However, if the unemployed (i.e. relatively unsuitable labour) is made available to employers at a subsidised rate, they’ll take on such labour and abstain from bidding up the price of labour generally. I.e. how’s about making subsidised labour (i.e. JG labour) available to EXISTING employers?

  4. Thanks Ralph!

    Good stuff. I’ll respond with one comment per question – using a lousy handheld here.

    > 1. Under the “Inflation” heading you say that if the economy is at NAIRU when JG is introduced, the inflationary effect can be nullied by raising taxes. True. But the net result is the creation of JG jobs AT THE EXPENSE OF regular jobs,

    Yep. If it turns out that aggregate demand does indeed increase (when JG spending replaces UE spending) and also that the economy is sufficiently close to the inflation barrier (NAIRU) – then yes. Then we have this unfortunate situation that you’ll either accept a tax-hike or a once-off inflationary episode.

    Both amount to the same thing though – a “cost” is imposed on the economy – either as inflation or a tax hike.

    And yes, a JG program is likely to consume more resources – in a direct sense – than a UE system. Longer term however, this should be offset by the declining social costs in a JG system, otherwise present in a UE system.

    However, if the economy operates at a demand level considerably lower than the inflation barrier, there shouldn’t be much of a problem. This is arguably how most of the economies in the world operate today.

    > … which is not the objective of JG: plus that casts doubt on whether the introduction of JG raises GDP.

    Initially, yes.

    But don’t forget the vast social costs inherent in a UE system that JG should get rid of.

    Also, MMTers argue that the size of a JG “pool” (NAIBER) should eventually stabilize around a smaller value than NAIRU – the numbers in “regular” employment should increase. Also, there is actually output coming from those outside of regular employment.

  5. > But the outcome is obtained by moving the goal posts. I.e. you’ve broken the all else remaining constant rule, I suggest. That is, if one cuts the pay of JG workers and the unemployed, it’s pretty obvious there’ll be an increased incentive for both to find regular work, and GDP would probably rise.

    Agreed. The JG wage would decline (in real terms).

    And yes, that’s how it’s supposed to work – the goal post is standing firm (and up the Villa!)

    The JG won’t be allowed to follow along with every fluctuation in the general wage level. If it was, JG wouldn’t function as intended as a price anchor.

    But still, you need to slowly and evenly adjust the JG wage upwards to follow the general wage level over a whole business cycle (or more).

    So how do you actually perform this balance act? That’s a story for next time.

  6. Good thoughts on training!

    We were partially influenced by Bill Mitchell’s blog posts there, but should probably go deeper into the litterature than that.

  7. > However, if the unemployed (i.e. relatively unsuitable labour) is made available to employers at a subsidised rate, they’ll take on such labour and abstain from bidding up the price of labour generally. I.e. how’s about making subsidised labour (i.e. JG labour) available to EXISTING employers?

    Yeah, that’s Ralph Musgrave version of JG!

    Haven’t been able to fully think this through yet, but refer interested readers to Ralphanomics! – currently dealing mostly with financial stability.

  8. I would like to add to Hugo’s last comment.

    I see no reason why JG could not have a subsidized labor feature within its framework. Call it a public-private partnership? After all who is better able to define and execute a satisfactory training program for skills that are not available in the market than the employer? So one subset of JG activities could be a 3- 6- or 12-month subsidy to the employer to partially pay for the compensation and training costs.

    Ralph’s suggestion that retraining subsidies to employers for existing employees could be part of JG is a little more problematic. There would need to be two means of control to prevent undesired behavior and consequences:

    1. The use of JD funds by companies to simply reduce their labor costs and not actually do any retraining should not be tolerated.

    2. Dismissal of existing employees (who could have been retrained for the new skill requirement) just to get the lower cost JG supported new employee-trainee needs to be avoided.

    I see a major problem to be the possible abuse of the JG system. After all, there are those who will look for any way they can to skim money for themselves and/or their company without actually producing any net gain to production, capacity or work force skills improvement. An example I can cite for comparison is the Medicare fraud that constitutes several percent of the program cost.

    Ralph, thanks for taking the time to comment.

  9. Hugo + John, One of the things I liked about your article was the assumption that unemployment is at NAIRU. Great minds think alike: I made the same assumption in a paper on JG / WPA here:

    http://mpra.ub.uni-muenchen.de/19094/

    The logic of this assumption of course is that when unemployment is well above NAIRU, a straight rise in demand is a far better cure for unemployment than JG. Thus the real niche for JG is in dealing with NAIRU level unemployment. (Not that I’m ruling out JG where unemployment is above NAIRU).

    So when you say “However, if the economy operates at a demand level considerably lower than the inflation barrier, there shouldn’t be much of a problem…” you’re breaking the above assumption: very naughty, etc etc.

    Re your final para and NAIBER stabililsing at a level lower than NIARU, I’d want to see the detailed arguments for that idea before agreeing. One of the arguments put by JG advocates here is JG work enhances the employability of JG people. But I’m afraid there is empirical evidence that contradicts this. A study done in Switzerland found that while subsidised temporary jobs in the private sector IMPROVED subsequent employment outcomes for those concerned, temporary subsidised jobs of the JG type (i.e. in the public sector) actually IMPAIRED subsequent employment outcomes. That’s one reason I favour JG taking the form of subsidised work with EXISTING EMPLOYERS, public and private. See:

    http://ideas.repec.org/p/iza/izadps/dp606.html

    If existing public and private employers can hire JG labour on the same conditions, the bulk would actually to the private sector because the latter is better at employing relatively unskilled labour than the public sector.

  10. Ralph – – –

    I have read comments of others that refer to a JG reservoir which becomes very small at full employment (NAIRU or any evolution of that concept) and expands when unemployment increases above the “optimimum level” in order to act as a labor market “shock absorber.” I like this conceptually, much the same as I like the “welfare to work” idea, because it requires the performance of work to receive “benefits” rather than just keeping registered with the employment office. It also has the potential of creating the opportunity to retain or gain new skills that does not exist with unemployment compensation benefits as currently operating. Additionally it meshes very well with the support by JG of temp private sector employment that can also interact with skills improvement programs. The second link you provide seems to support the value of the JG participating in temp placements. The ideas you develop in your article (your first link) follow right along with public-private interaction theme.

    I have been quickly through both articles and intend to return to them later.

  11. You apparently understood how to use the “Reply” link. Hm.

    > So when you say “However, if the economy operates at a demand level considerably lower than the inflation barrier, there shouldn’t be much of a problem…” you’re breaking the above assumption: very naughty, etc etc.

    Aha, I see your point now! Ok, point taken.

    > Re your final para and NAIBER stabililsing at a level lower than NIARU, I’d want to see the detailed arguments for that idea before agreeing.

    Right. What I personlly could to is to explain what the MMT’ers believe should happen — you know the very technical stuff that motivates the statement that NAIBER should be less than NAIRU. (I hope I could explain that — I’ll try to add that to the MMT wiki and let you know.)

    But that’s probably not what you’re looking for? The Swiss study is clearly something for the MMT academia to deal with.

Comments are closed.