Monday, 14 February 2011

“Make work”, the Job Guarantee, WPA, etc: should they be limited to the public sector?


Make work schemes, like the WPA in the U.S. in the 1930s do not make sense. However they involve temporary subsidised work for the unemployed. And temporary subsidised work for the unemployed with EXISTING employers, as distinct from “specially set up employers” like make work schemes, DOES make sense. Plus there is no reason to confine this sort of work to the public sector.

Temporary subsidised work with existing employers actually amounts to something very similar to a totally free labour market, that is a labour market with no artificial interferences like minimum wage laws. While the price that employers pay for such labour can be “free market price”, i.e. very low, obviously the take home pay of the relevant employees must be up to socially acceptable levels.

If you find this article hard going, don’t worry. I estimate the number of people on planet Earth with enough brain and enough genuine interest in the subject to understand the article to be about three!


The idea that there are an almost infinite number of useful jobs the unemployed could do is as old as the stars. Pericles in Ancient Greece 2,500 years ago had the unemployed work on public sector construction projects. And more recently in the U.S. in the 1930s millions were employed on schemes of this sort: the “WPA” for example (which stood for Work Progress Administration).

And then there is the particular form of “make work” currently advocated by some Modern Monetary Theory enthusiasts, namely the so called “Job Guarantee”.

In principle, unemployment can be reduced to zero at the flick of a switch by schemes of this sort. For a very crude illustration, we could tell the unemployed their benefits are henceforth conditional on walking round their neighbourhood keeping it free of litter. Those accepting the work would be deemed to be employed. And those refusing would be deemed to have turned down work, and are therefore not unemployed. Hey Presto: unemployment vanishes.

Incidentally, I’ll refer to WPA, Job Guarantee, and similar schemes below as “make work”, for want of a better phrase.

Limit make work to the public sector?

The advocates of make work normally see it as being limited or largely limited to the public sector.

The reason (normally not set out too clearly by make work advocates) is that no additional demand is required. But there is a catch here, as follows.

Labour employed on make work can be employed alongside other factors of production (OFP) like permanent skilled labour, equipment, materials, etc. Or, secondly, such labour can be employed with little or no OFP. But if OFP IS employed, it has to be ordered up from the rest of the economy. I.E. ADDITIONAL DEMAND IS REQUIRED! Now there is a problem here.

If the economy can take extra demand, there is no place for make work in that employment can be created simply by raising demand. That is, raising demand is a better way of raising employment than implementing or raising the numbers doing make work.

Alternatively if OFP is NOT used, then the only input on make work schemes will be temporary and not very skilled labour: output per head will be hopeless.

Incidentally, make work employees were described as “not very skilled” in the above para not because they don’t possess skills. The point is that people tend to become unemployed when, even if they DO possess skills, there is no demand for those skills locally. Thus any job a person gets, whether make work or not, is LIKELY to involve them working in a RELATIVELY UNSKILLED capacity.

And there is a further nonsense here in relation to OFP. If a make work scheme DOES involve levels of OFP that are up to near those that obtain with normal employers, then what’s the difference between such make work schemes and a normal employer? The answer is “none”.

This nonsense actually obtained, more or less, with some 1930s WPA construction projects, where the amount of construction equipment, matrials, skilled labour, etc was almost up the level normally found with private sector contractors.

To summarise, make work is stuck between a rock and a hard place: use no OFP and be condemned to inefficiency. Alternatively, employ significant amounts of OFP, and accept two bits of nonsense: first make work has an inflationary effect, because OFP has to be ordered up from the rest of the economy. Second, the relevant make work scheme comes to much the same thing as a normal employer.

Make work and inflation.

The above point about make work being inflationary can be summarised as follows. If unemployment is above the level at which labour shortages exacerbate inflation (which I’ll call NAIRU for want of a better phrase), then make work will not be inflationary. But in this scenario, there is no place for make work in that employment is best raised by a straight rise in demand (plus, presumably, an equivalent expansion in the public sector).

Moreover, on a like for like comparison, there is little difference as between the inflationary effect of public and private sector make work. To illustrate, take two make work schemes, both of which the same OFP to unskilled labour ratio. The private sector scheme will be inflationary because any extra demand at NAIRU is inflationary. While in the case of the PUBLIC sector schemes, government will have to increase net spending so as to order up the required OFP. There is not much difference between the two!

Existing employers provide a decent OFP to temporary labour ratio.

As explained above, the problem that really stymies traditional make work schemes is the OFP problem.

There is actually a very simple way, at least in theory, of employing make work labour efficiently with ABSOLUTELY NO NEED for additional OFP. This is to make the unemployed available at a subsidised rate to EXISTING EMPLOYERS (as opposed to “employers” in the sense of “specially set up make work projects”).

If the unemployed (i.e. low priced and relatively unskilled temporary labour) is supplied to EXISTING employers, those employers will be induced to raise the amount of labour employed RELATIVE to the amount of OFP employed. This is for exactly the same reason as if the cost of say computing (or any other input) declines, employers will employ more computing power relative to other inputs.

Note that, as is always the case in economics, it is not the AVERAGE price of an input (or anything else) that is important: it is the MARGINAL price. That is, make a few EXTRA OR MARGINAL employees available to employer at a lower price, and (to repeat) more labour will be employed relative to OFP.

Now we have a piece of magic here: little or no extra OFP is required, yet the “make work employee to OFP ratio” is almost up the standard that normally obtains with existing employers (because it is existing employers who do the employing!). To put that in plain English, it is better to have an unemployed teacher work in some sort of peripheral capacity in a school, than have them do typical “make work” type work, e.g. weeding the flower beds in the nearest public park.

Incidentally, I’ll continue to describe the employees involved as “make work”, despite the fact that from this stage in the argument the assumption is that they are allocated to EXISTING employers rather than to WPA make work type schemes.

Would supposedly marginal employees really be marginal?

How do we ensure that make employees really are “marginal”. That is, how do we ensure that they are the employees that each employer regards as the least productive or most peripheral?

It’s easy. Just limit the TIME that each make work employee stays with a given employer. Employers don’t mind losing employees who are genuinely of marginal use, but they very much DO mind losing their more productive employees. If an employer knows an employee will leave in the near future, the employer will not classify the employee as make work, or claim the make work subsidy in respect of the employee. (Obviously it would be necessary here to have some rules to prevent attempts to “fiddle” the system: for example there is an obvious temptation for employers to fire and immediately re-hire make work employees who have reached their time limit.)

Make work with existing employers = free market!

Astute readers will have noticed by now that the system advocated above amounts to much the same thing a totally free and perfectly functioning labour market (a bit of an unrealistic and theoretical construct, but never mind). That is, a labour market where there are no artificial interferences like minimum wage laws or state funded unemployment benefit systems. In this free (and brutal) market, many of those becoming unemployed just have to take any old job however low paid if they want to avoid starvation. They are likely to move on quite quickly to something better paid as soon as they find it.

The system advocated above comes to much the same thing: certainly the cost of the relevant labour to the employer would be ultra low (or even free). However, take home pay would be up to minimum socially acceptable levels.

No limit to the number of potential public sector make work employees?

An objection that advocates of public sector make work schemes will probably raise is that there is in principle no limit to the number of public sector make work jobs that can be created. This is an obvious and crude truism (pointed out in the first para above – picking up litter, etc).

Incidentally this point applies both to traditional WPA make work schemes AND to “make work with existing employers”.

However, the above truism does not get public sector make work very far, and for the following reason. The marginal product of labour (or any other input) declines with increased numbers employed. I.e. it is true that limitless numbers can be employed on public sector make work schemes, but the bigger the number, the more likely it is that marginal product of the relevant labour is around zero, or even negative, e.g. the collection of non-existent litter – to echo the example with which we started above.

Indeed, the PRIVATE sector is BETTER at employing relatively unskilled labour than the public sector. On that basis, and assuming we want to expand the number of make work people in both sectors to the point where the output of the marginal employee is any specified amount (including perhaps zero), then then MORE make work people would be employed in the private than in the public sector.

The declining marginal product of labour is caused by macro as well as micro factors.

It might seem that the point just above about the marginal product of labour is flawed in that this marginal product point is invariably set out in the text books as being a MICRO ECONOMIC phenomenon. (I.e. where a firm employs increasing numbers relative to a fixed input of capital equipment and materials, the output or “product” of each succeeding person hired will decline.)

In contrast, in the above para, the point is being applied to a MACROECONOMIC scenario.

The answer to this apparent flaw is that there are also MACROECNOMIC forces at work which result in the marginal product of labour declining as unemployment in the aggregate falls. These “forces” are very simple and common sense: first, as unemployment falls, it becomes increasingly difficult for the nation’s employers to find labour that is suitable for vacancies. Second, employers in the aggregate always employ the best labour first. That is, as unemployment falls, lower quality labour becomes employed.

And therein lies the reason why temporary low cost labour raises aggregate employment both in a free market and under the system advocated here.

Simply pointing to the fact that an employment system has similarities to the free market is not a bad argument, but it needs to be strengthened to explaining the reasons, in theory, why the result is an aggregate increase in employment. The reason is thus.

As pointed out above, as unemployment falls, employers have to take progressively poorer and poorer quality labour. When this decline becomes sufficiently serious, employers tend to resort to bidding up the price of an increasing number of skills, until excessive inflation ensues. However, if relatively unsuitable labour is available for free or at a much reduced price, employers will employ more of such labour, rather than bid up the price of more suitable labour. NAIRU is reduced.


Hopefully the above paragraphs have demonstrated two points. The first it is that limiting make work to the public sector does not make sense. Second, there are strong arguments for allocating make work employees to EXISTING employers, rather than specially set up “employers” like the WPA. That is, it is better to allocate an unemployed teacher to a temporary and peripheral teaching post in a school (public OR private sector) than have the teacher pick up litter.

If the above arguments are fool proof, then there is NO case for “specially set up schemes” like the WPA. That is, ALL make work employees should be allocated to EXISTING employers, not specially set up schemes.

Possibly there are weaknesses in the above argument, which might result in it making sense to have WPA type schemes for a PROPORTION of make work employees, while allocating the rest to EXISTING employers.



  1. Why not a negative income tax? The old Friedman idea.

    I think I read somewhere that the concept goes back to Abba Lerner as a perfect tool to implement functional finance. To stimulate demand just raise the tax exemption. If the economy overheats just cut it back.

  2. Anon, You are right to say that negative income tax could be used to effect functional finance, but there are plenty of equally good alternatives, e.g. a payroll tax increase or reduction. So that’s not a strong point in favour of negative income tax.

    Negative income tax is similar to the above proposal of mine in that both involve cutting the cost (for employers) of employing the least productive members of the workforce. The advantage of my proposal is that it does not involve subsidising those who do not need subsidising, and the result of subsidising anything or anyone that does not need subsidising is to encourage inefficient use of the subsidised ite. I.e. the negative income tax idea is a bit of a shotgun approach.

    The negative income tax idea was also proposed by Edmund Phelps.

  3. I have a reason for this shotgun approach. Basically I asked myself the question who should “command” freshly created money?

    I’m a libertarian, I don’t like government. To keep government in check, we have the principle “no taxation without representation”. Now, MMT tells me, government doesn’t need taxes to spend, it needs taxes to fight inflation. Let’s assume MMT is right. I still prefer a system, where government has to tax before it can spend. For the same –strictly political- reason, I’m against government debt. That’s why I like your approach: functional finance without debt. But I think even better would be a system, where all the money -which has to be created to fight deflation - goes right into the pockets of each and every citizen. A negative income tax would be a way to implement this. After negative taxing (spending) enough to get demand up, any further spending would create inflation. This forces the government to tax if it wants to spend, right? So after enough spending by negative taxing we would again have the old “no spending without (positive) taxing”.

    My big problem with MMTlers like Bill Mitchell is that they don’t see the danger of socialism. The first guy to come up with chartalism was a German. (Ironically his name was Georg Friedrich Knapp, “knapp” means “scarce” in German.) Germany always had a strong tradition of idolizing the state and it didn’t end well. (I’m from Germany, I know what I’m talking about.) So even if you “socialize” central banks, I would keep the institution for “money creating” separated from the treasury. On the other hand, today –with QE1- we have the central bank in bed with private interest. This is socialism for the rich. Neither the state nor the banks, money should flow immediately into the pockets of each and every citizen.

    I like your Edison citation: “…but who is behind the Government? The people. Therefore it is the people who constitute the basis of Government credit. Why then cannot the people have the benefit of their own gilt-edged credit by receiving non-interest bearing currency… instead of the bankers receiving the benefit of the people’s credit in interest-bearing bonds?”

    Back to the subsidies for the unemployed. Since unemployment is high in recession, but in recession negative taxes are also high, I thought a negative income tax could solve a political and economic problem with one strike. (You still need help for people who are not employable because they are ill, too old, too young or disabled. I’m not a crazy libertarian, health care, schools and pension systems are necessary.)

    Sorry for the long post.

  4. Anon,

    I quite agree that new money should be the property of the people. Apart from advocating that on Warren Mosler’s site, I advocated it here.

    I don’t think negative income, an element of which varies with governments desire to stimulate the economy, is a good way of doing this because it would result in windfalls for the less well off during recessions – followed by nothing for the next few years when the economy got back to normal. That is a bit chaotic.

    There is a phrase of yours that rings alarm bells in my head: “I thought a negative income tax could solve a political and economic problem with one strike.” That phrase does not comply with the Tinbergen principle (named after Jan Tinbergen, an economics Nobel laureate). Coincidentally I did a post on this a few days ago (12th Feb). This principle states, roughly speaking, that “for each policy objective, at least one policy instrument needed”. In other words trying to hit two birds with one stone is normally a bad idea. The above “chaos” is an example. For more on this, see here.

    Thus there is nothing wrong with failing to hit two birds with one stone when it comes to government policies and “instruments”.

    Thus for the “policy objective” “subsidise the least productive employees”, I’d go for the “instrument” explained in the above article. And for the “policy objective” “feed extra money to households” I’d go for a payroll tax reduction, supplemented, as you say with something for the unemployed (plus pensioners, come to that).


Post a comment.