Some Macro Effects of a Job Guarantee

A recent post considered one way of including a job guarantee in the income-expenditure model. Doing so makes it possible to represent various macro effects of a job guarantee within the model. An obvious effect is that the program would deliver a degree of demand stabilization. An effect that is perhaps not quite so obvious is the way in which a job guarantee would ensure supply-side changes in the economy automatically impact on demand, actual output and employment. Before illustrating a few of these effects, the modified income-expenditure model will be briefly outlined and tailored to present purposes. A fuller discussion of the model is provided in the earlier post.

The Model

In the basic income-expenditure model, total output (Y) converges on the level Y = A/α, where A is net autonomous expenditure and α is the marginal propensity to leak from the circular flow of income to taxes, saving and imports. Introducing a job guarantee into the model modifies this to:

where Gg is the government’s spending on the job guarantee. The main preoccupation of the earlier post was to find an expression for this component of government spending, which adjusts automatically to variations in the number of people who accept the standing offer of a job. It was shown that within the model:

In this expression, LF is the level of employment associated with full employment. This level of employment is determined exogenously by workers themselves in the sense that a job at the living wage can always be obtained in the job-guarantee sector. (A more sophisticated approach might also allow for endogenous changes in the level of LF by explicitly recognizing that some people unwilling to accept job-guarantee jobs will exit employment in a downturn, due to layoff, and re-enter in an upturn.) The cost of the job-guarantee program per unit of labor is wL/ϕ, where wL is the living wage paid per unit of job-guarantee employment and ϕ is the share of wages in total job-guarantee costs. Productivity in the broader economy is denoted ρb. As discussed in the earlier post, productivity in the job-guarantee sector can be interpreted as being equal to the living wage (ρg = wL). The difference in productivity between the job-guarantee program and the broader economy implies that average productivity for the economy as a whole (ρ) varies with the composition of employment between the two sectors (assuming ρb > ρg).

Since the present focus is on total output and employment, it is convenient to substitute the expression for job-guarantee spending given by (2) into (1). Rearrangement (eventually) gives:

Of the economy’s total output, Yb is the output produced within the broader economy and Yg = ϕGg is the amount produced in the job-guarantee sector. An expression for output of the broader economy can be obtained by subtracting ϕGg (= Yg) from both sides of (1), substituting for Gg and solving for Yb. An expression for Yg can be obtained simply by multiplying the expression for Gg in (2) by ϕ.

Expressions (3) to (5) show that, once a job-guarantee is introduced, Y, Yb and Yg can be affected by exogenous changes on both the demand and supply sides of the economy. On the demand side, there is the familiar effect of exogenous changes in autonomous expenditure (A) or changes in the marginal propensity to leak from the circular flow of income (α). On the supply side, an exogenous change in the level of total employment (LF) will directly bring about changes in Yg and hence Y and subsequently Yb. Exogenous changes in other supply-side factors – the living wage (wL), the wage share in job-guarantee costs (ϕ) and productivity in the two sectors (ρb and ρg = wL) – make themselves felt through their impact on the size of the multipliers applying to A and LF.

Demand Stabilization

Like other automatic stabilizers, a job guarantee would somewhat moderate fluctuations in total demand and output. In the following diagram, fluctuations in output are generated by arbitrarily letting private investment cycle between a high of 35 and a low of 25, beginning and ending at 30. The marginal propensity to leak is assumed to be 0.5, productivity in the broader economy 2, and the wage share in job-guarantee costs 0.67. Full employment is set at 100. Net autonomous spending other than private investment is set to 67 in the absence of a job guarantee, 64 when the job-guarantee wage is 0.5 and 58 when the job guarantee wage is 1. These values for net autonomous spending other than private investment are chosen to ensure that the initial and final levels of output are the same for the three scenarios, since the focus here is on illustrating the dampening of fluctuations.

Within the model, the stabilization effects are stronger when the living wage paid to job-guarantee workers is increased relative to productivity in the broader economy. This does not account, though, for the possibility that a higher living wage might encourage some employed workers to switch from the broader economy to the job-guarantee sector.

Fluctuations in the economy as a whole (represented by Y) are dampened more than fluctuations in the broader economy (represented by Yb). The greater volatility of Yb is illustrated in the next diagram. The values for the parameters and private investment are the same as in the previous diagram. Net autonomous spending other than private investment is set to 62 (rather than 58) when the living wage is 1, but given the same settings as before in the other scenarios.

To a significant degree workers would be sheltered from the volatility in the broader economy through variations in job-guarantee employment:

A Supply-Side Effect

From a Keynesian or Kaleckian perspective, a mere increase in potential output (such as through an increase in labor-force participation or an improvement in productivity) does not necessarily bring about an automatic increase in output and employment. The greater productive potential is only realized if demand happens to be sufficient to sustain the higher level of output.

The job guarantee introduces a mechanism through which an increase in labor-force participation directly initiates an expansion of demand, actual output and employment. The causation, in these instances, runs from individual decisions to enter the labor force to accept the job-guarantee offer, to an increase in employment, to an activated expenditure by government on the program and increased income and output, to multiplier effects in the broader economy.

Similar effects can be partially present under some other policy measures, though the effects are likely to be weaker and less direct. For instance, when government pays an unemployment benefit that is not contingent on prior employment (this occurs in Australia, for instance), a decision by a jobless person to enter the labor force activates, subject to administrative checks and with a delay, a benefit payment. This does not represent a direct addition to demand because the benefit payment is treated as a ‘transfer’ (or negative tax) in the National Accounts. Even so, to the extent the benefit payment is used for consumption expenditure (and mostly it will be), there will be an impact on demand with a subsequent multiplier impact on the economy. Such a policy therefore does enable a supply-side change to have an impact on demand, actual output and employment, though the channel is less direct. If the benefit is means-tested (this is the case in Australia), the channel will be narrower than otherwise.

Under a job guarantee, exogenous growth of the labor force would cause an increase in output and employment. The increase in output would be enabled by an automatic expansion of job-guarantee spending and an inducement of additional private consumption. Suppose, initially, that there is no change in the assumed behavior of other exogenous variables or parameters. In that case, although employment in the broader economy would expand, it would not keep pace with the expansion of the job-guarantee sector. This is illustrated in the next two diagrams. It is assumed that the level of employment associated with full employment (LF) grows by 1 percent each period. The other exogenous variables and parameters take the same values as in earlier examples except that the living wage paid to job-guarantee workers is always 0.5 in the following scenarios.

For employment in the broader economy to grow in line with total employment, it would be necessary for autonomous spending to keep pace with the growth in the labor force. In the next two diagrams, it is assumed that autonomous spending other than private investment grows by the same rate as the labor force. Growth in private investment is also assumed to occur, but with an arbitrary delay of four periods.

Concluding Remark

The automatic translation of extra supply potential into actual demand, output and employment seems quite an interesting feature of the job guarantee.

Academic Treatments of the Job Guarantee’s Macro Effects

A more sophisticated and comprehensive analysis of a job guarantee’s macro effects, which employs the Fair model (an econometric model developed in the tradition of the Cowels Commission by Ray Fair), is provided by Scott Fullwiler in a 2005 paper.

A recent paper by Warren Mosler and Damiano Silipo focuses on price-stabilizing features of the job guarantee.

Introductions to the Income-Expenditure Model

For newer readers, an introduction to the income-expenditure model is provided in:

Bill Mitchell – Spending Multipliers

Earlier heteconomist posts on the topic, with the easiest posts listed first, include:

Short & Simple 16 – The Expenditure Multiplier and Income Determination

Short & Simple 18 – Income Determination in a Closed Economy

Planned Investment/Saving and Keynesian Causation


7 thoughts on “Some Macro Effects of a Job Guarantee

  1. — where does your model account for the non-labor expenses of the JG, which are going to be substantial? [looking back at your earlier post it says “Of the government spending on the job guarantee (Gg), only wages will contribute to the measure of the sector’s output. If ϕ is the proportion of job-guarantee costs associated with wages, then job-guarantee output is ϕGg.”] So in other words, your model does not account for non-labor JG expenses? JG workers will work with their bare hands, outdoors? No materials will be needed, no tools will be needed, no vehicles or machines will be needed, no buildings will be leased, and no concrete or fuel will be consumed? Exactly what does one do with bare hands outdoors?
    — yes, any means-tested program will function as an automatic stabilizer.
    A JG is means-tested both in rule of law (both the New Deal and the Jefe job programs had means testing) and as a practical matter (people who already have a living wage of some sort are not likely to participate in a JG).
    — if functional finance is capable of maintaining full employment and price stability, as Abba Lerner theorized and as Australia demonstrated for 30 years (WWII until the 70’s oil spike), then why is a JG needed to maintain full employment?
    — this essay mentions “living wage JG” 8 times yet there is no reason to believe that any make-work job program will pay a living wage. The Jefe program was not a living wage, India’s job program is not a living wage, and some of the New Deal programs like the CCC were not a living wage (though the New Deal had the good sense to pay different wages for different jobs). In a capitalist society, there is zero chance the capitalists will allow any make-work program to compete with the private sector for labor.
    — In general, why is MMT fixated on the JG? Isn’t functional finance the most powerful tool in MMT’s kit? Yet talk to an MMTer and all you hear is “JG-JG-JG.” Here is this controversial program, that no one has ever satisfactorily explained how it would work, that offers little for unemployed skilled workers, that the capitalists who run capitalist societies will be opposed to, and that is simply unnecessary according to functional finance. I would like to see MMT move on to more productive subjects.
    — The real problem with any jobs program, or with functional finance, is political, as Kalecki pointed out 70 or so years ago. Capitalism does not have the slightest interest in full employment or a living wage. A capitalist society does not nurture the values that are required for a society that values “public purpose” above all else. MMT is yet another attempt to save capitalism from itself, but it is not clear that capitalism can be saved, or that it is worth saving.
    All that said, nice job on your modeling, Peter.

  2. “where does your model account for the non-labor expenses of the JG, which are going to be substantial? ”

    Ok, what private sector use of those goods and services is superior to ensuring that about 10% of the man-hours of the entire economy are not wasted?

    That’s really for you to show isn’t it.

    “yes, any means-tested program will function as an automatic stabilizer.”

    Nope. Money is valued partially based upon what you have to give up to get it. The idle rich don’t value money very much and that can cause inflation because they give too much away for too little in return. The same with income guarantees.

    The Job Guarantee puts the economy on a ‘labour hours standard’. It is a ‘proof of work’ system like the hashing algorithm in Bitcoin. The value of 10 tokens, say, is one hour given up. That means in exchange you will demand the equivalent in real terms or you won’t release the tokens. Since wide and enforceable taxation requires that others obtain the tokens they will meet your demand.

    “then why is a JG needed to maintain full employment?”

    Because if you don’t fix the Kaleckian bug in Keynesianism you get a wage price spiral and a collapse. You cannot rely on the private sector to ‘do the right thing’. The failure of the Beveridge approach was to rely on the private sector and constantly prop up failing public sector institutions because ‘arghh jobs’. Under the Job Guarantee you don’t do either of those. The Job Guarantee is the market under which your income is protected. That means private sector and public sector institutions can come and go as required.

    The Job Guarantee is instant, spatial and automatic. It requires no politicians to make decisions and it requires no unelected wonks in ivory towers to make decisions. It’s a pull process, not a push that just happens.

    And because people have to work for a living in a job *anybody can take if they want* you don’t get the resentment of ‘dole bludgers’ that causes the collapse of unemployment benefit.

    “et there is no reason to believe that any make-work job program will pay a living wage.”

    Yes there is. That’s what tax is for. If you have to get 10 tokens or go to jail, yet I make people work a whole week for one token, then the Job Guarantee job will be incredible valuable in real terms. But it would suck too much out of the system with those settings.

    So in reality you set the parameters on either side to the level necessary to drive prices to where they need to be. That guarantees that whatever the Job Guarantee pays will be what is required to buy enough for a decent simple life.

    “Isn’t functional finance the most powerful tool in MMT’s kit? ”

    Nope. You have to apply it. The Job Guarantee is the automatic stabiliser in both monetary and real terms because it puts the currency on a real world standard – labour hours. Which as Warren’s paper above shows is a superior anchor.

    “Capitalism does not have the slightest interest in full employment or a living wage.”

    Correct, and the JG fixes that bug. Once you put the JG containment vessel around the nuclear reaction of capitalism, then just like a nuclear power station you get the power without the boom.

    All of a sudden capitalists have to learn to compete, because we have no need of them. They become like horses after the invention of the car. You keep them for the entertainment value. But those that are really good get very well looked after.

    Fundamentally you don’t have a better system that is any more politically likely to get adopted. The JG gets the politics out of poverty.

  3. “that offers little for unemployed skilled workers”

    If a skilled worker is unemployed, they are not skilled. They may have obsolete skills, but they are not currently skilled or there would be a call for them.

    You don’t get paid in the real world just because you have a certificate. There’s rather too much of that from the middle-class liberal leftie.

    The only reason you are paid more than the living wage is because there is market demand for what you have to offer greater than the supply, or where socially we have determined to give up an amount of our income in tax to cover the higher wage of a valued professional (like a doctor, etc.).

    If you want more than the living wage, you have to justify it to your peers. A shiny certificate or the ability to shoe horses isn’t going to be enough.

  4. Dan & Neil: Very good discussion from both of you, in this and many other threads. Your contributions are greatly appreciated. Thank you.

    This comment is in response to Dan’s points:

    Non-wage costs. In the model, ϕ is the share of wages in JG costs; 1 – ϕ is the share of non-wage JG costs. In the examples, I set ϕ = 2/3, so it was assumed that 1/3 of JG costs went to non-wage costs. Another value for ϕ could have been chosen for the examples, but I chose 2/3 as an intermediate figure based on a passage from a paper by Bill Mitchell and Martin Watts (see p. 11 at the link below) which classifies JG roles as low capital intensive (in my notation, defined as ϕ = 0.75), medium capital intensive (ϕ = 0.6) and high capital intensive (ϕ = 0.5). The authors indicate that their research suggests low-skill roles will primarily be low capital intensive but that it varies by industry.

    Bill Mitchell & Martin Watts (2013) – Capacity Constraints and the Job Guarantee

    Means teasting. Yes, if the JG were means-tested (though I don’t think it should be), this would limit the demand stabilization as with any other auto stabilizer. I wasn’t really trying to score a point for the JG over other demand stabilizers so much as just illustrate the stabilizing effect of a JG, including the point that its effects will be different in the two sectors.

    On full employment with price stability requiring a JG. Actually, I’d say that, in principle, either a JG or an incomes policy could be combined with functional finance to achieve full employment with price stability. The JG is a way to achieve the two goals without an extensive system of price controls, and as Neil mentions, the effect is automatic once the program is in place. (I am not necessarily against price controls though. I think quantity signals – in the form of variations in inventories and capacity utilization rates – can be an effective indicator of where resources should be allocated when prices are administered.)

    Australia’s maintenance of full employment and price stability in the postwar period. My understanding is that, in Australia, government implicitly did play a role of employer of last resort in the postwar era. It’s true, though, that there was no explicit implementation of a JG. On the other hand, Australia had a tradition of centralized wage determination and an arbitration system for resolving disputes.

    Repetition of the term ‘living wage’. Stavros mentioned in the comments to the earlier thread that it doesn’t really affect the operation of the model whether the JG wage is called a living wage or something else, or whether the JG wage is a living wage or not. I call it a living wage because that is the intention of the JG advocates, and I share the view that the JG wage should be a living wage. I am trying to take on board some of the framing challenges pointed out by Bill Mitchell and others. 🙂

    Fixation on JG. The reason MMTers give so much attention to the JG is that it is the key mechanism, within the theory, to achieve full employment alongside price stability. I know you are talking about the leading MMTers and not this blog, but if anything the JG has received insufficient attention here in recent times. Prior to the recent two posts on the JG, there had not been a JG-focused post on the blog since September 2014. Mainly this is because I have pretty much stated my policy preference (a combo JG/BI or ‘JIG’), and this position and the reasons for it have not really changed.

    Political aspects and saving capitalism from itself. No doubt there is considerable political opposition to both a JG and a BI (and a combo), including from proponents of one or other of the policies. Kalecki thought that for full employment to be maintainable, a “fundamental reform” would have to be incorporated into capitalism. (For newer readers, I have discussed Kalecki’s famous argument in an earlier post.) Arguably, a job guarantee could be seen as such a reform. But I actually think a JG (or a JIG) could instead be a vehicle for moving the system toward socialism. That would require social struggle to broaden, over time, the activities that qualify as JG roles. In my view, a combo JG/BI would give maximal space for the broadening of activity, because then each individual’s activity could be shaped within a context of voluntary associations: the associations would be voluntary whether individuals chose JG roles or the BI, because they would always have the option to choose differently. But a BI by itself is not as empowering, because it does not ensure access to productive resources, materials, equipment or social networks. A standalone BI would leave some people who wanted a job without one, and thereby limit their opportunities to contribute to society in a way that best matched their preferences. (Of course, if we can move straight to socialism or communism through social revolution, I for one will be all for that!)

  5. How the actual magic money tree works

    The fundamental problem is that “capitalists” don’t want to compete because competition drives down profit share. They want to extract economic rent and rent is extracted by applying market power in asymmetrical markets.

    A JG is one way of reducing the rent in an asymmetrical labor market. Collective bargaining is another.

    However, the fundamental problem is “capitalism” as a system that favors one factor over others — ownership and therefore control of property over workers (most of the people in a society) and land (the environment or “nest”).

    Economic rent is an externality based on capitalizing the gains and socializing the losses in order to favor capital based on the assumption that the highest priority is growth regardless of distribution, with growth assumed assumed to result from capital formation.

  6. Oops, the quote lede to my comment above is wrong.

    It should be:

    All of a sudden capitalists have to learn to compete, because we have no need of them.

  7. Dumb question from a long-time supporter of ELR / JG and JG-oriented JIG:

    What out-of-the-box industries should be considered under the program?

    The historic example that job guarantee advocates like to put forward is FDR’s Works Progress Administration. The Bastard Keynesians love infrastructure spending (i.e., construction industry).

    The Post-Keynesians have put forward certain services as a more modern example.

    What about heavy manufacturing? Given all the populist talk about jobs being outsourced to China years ago:

    1) In China itself, today’s heavy manufacturing jobs substitute for Western capitalist social security; and
    2) Populist planks like to talk about national self-sufficiency in manufacturing, yet they approach this merely in terms of government incentives for business.

Comments are closed.