Monday, 20 May 2013

Job Guarantee buffer stock nonsense – Part 1.

Job Guarantee (JG) is a term often used to refer to an idea that has been around for centuries: the idea that instead of paying the unemployed to do nothing, they could be paid to do something useful. The idea has of course been put into effect in a hundred different guises over the centuries: the Work Project Administration in the US in the 1930s, various post WWII workfare or “job creation” schemes, and the “Work Programme” currently in operation in the UK.
Indeed, it doesn’t take a genius to work out that JG could, at least in theory, abolish unemployment. That is, and taking a very crude illustration, the unemployed could simply be told that their benefits are conditional on their walking up and down their street keeping it free of litter. Anyone refusing that work would be deemed to have turned down a job, and hence would not be counted as unemployed. Hey presto: unemployment vanishes.
A recent and largely useless addition to the thinking behind JG is the idea that JG employees can be regarded as a buffer stock. The buffer stock idea is set out in around 4,000 words here and here by Bill Mitchell. While I favour JG in some shape or form, the whole buffer stock notion strikes me as irrelevant. It’s hot air, and the paragraphs below demolish the idea in vastly less than 4,000 words.

What is a buffer stock?
A buffer stock, just to be clear, is a stock of commodity X, some of which can be sold into the market, given excess demand for X, and so as to stop the price of X rising, or rising too fast. And conversely, if the market price falls too far, then the operator of the buffer stock (normally government) can buy up X and add it to the stock. That of course stops the price of X falling below some level or other.
So in the case of JG, the big idea is that if the JG wage is $X/hr, that will allegedly stop wages in the regular economy falling below $X/hr. And if aggregate demand rises, then the stock of JG employees will supply the regular labour market with labour, which will tend to stop regular wage rates rising, and ameliorate inflation.
Now the first problem with the “JG buffer stock” idea is that JG labour is no more a buffer stock than are the unemployed in an economy where there is no JG. That is, if unemployment benefit is $Y a week, that puts a floor to the weekly wage: about $Y a week. So introducing JG to a typical developed economy adds very little “buffer stocking” because a buffer stocking effect is already provided by unemployment benefits.
And not only that, but most developed economies have minimum wage laws: another factor that ensures a minimum hourly wage for employees.  That makes the JG buffer stock idea even more irrelevant.

JG controls inflation?
As distinct from stopping the price of labour falling below some minimum, does the buffer stocking characteristic of JG stop the price of labour rising too fast, given a rise in aggregate demand?
Well hardly, and for the following reason. Assume there is no JG system, and assume increased demand by regular employers for a particular type of labour. What happens? Well, labour flows from the ranks of the unemployed into the vacancies that have recently appeared.
Now assume all the unemployed are doing JG work. And assume the same scenario: a rise in demand by regular employers for a particular type of labour. What happens? Well labour flows from JG employment into the vacancies that have recently appeared. Spotted the similarity?
Conclusion: when it comes to inflation control, there is no difference between the buffer stocking characteristics of the unemployed and the buffer stocking characteristics of JG. So given that the whole point of discussing JG is to see what benefits it might have as against unemployment, why mention buffer stocks? The buffer stock notion is plain irrelevant. (Incidentally, I’m fairly sure Warren Mosler, who advocates JG, has conceded the latter point about JG being no more of a constraint on inflation than unemployment, but I can’t find the relevant link.)  
Having said that there is “no difference” as between the buffer stock characteristics of JG and of the unemployed, there are a number of differences that MIGHT ARISE. For example, if the JG wage is much more generous than benefits, that would dissuade people from moving from JG to regular work. And clearly that would thwart the buffer stocking characteristics of JG. But all else equal, in particular if unemployment and JG are equally attractive, then there is no difference as between the buffer stocking characteristics of JG and unemployment.
Also, it’s possible that the QUALITY of labour while doing JG work might improve: e.g. JG might incorporate training. But then training can also be made available to the unemployed. So the training point is irrelevant.
Another possibility is that JG maintains work habits, and makes JG people more employable than those who have been unemployed for extended periods. But IMPROVING THE QUALITY of the commodity stocked is not a normal characteristic of buffer stocks. If anything, stocks of physical commodities actually DETERIORATE when stored for too long.
Put another way, if JG helps maintain work habits, why not just say “JG helps maintain work habits”? The buffer stock notion is simply irrelevant.

Let’s assume no benefits or minimum wages.
As distinct from the above assumption namely that JG is introduced to an economy which already has unemployment benefits and/or minimum wage rules, let’s assume neither of the latter two are in operation.
In that scenario, there’d be a tendency for the unemployed with no savings to take very low paid jobs pending the appearance of something better. And the introduction of JG to such an economy would certainly act as a buffer stock in that it would ensure a minimum price for labour (though it still wouldn’t bring about the “price rise ameliorating” effect that is inherent to the definition of the phrase “buffer stock”).
However, the fact that JG works as a buffer stock in the above scenario, i.e. maintains a minimum price for something, does not explain why JG works: i.e. why it raises aggregate employment.

So why does JG work?
Having hopefully established that buffer stocking has next to nothing to do with why JG raises employment, the next and obvious question is: why DOES JG raise aggregate employment?
Well the answer is simple enough – at least the answer SEEMS TO BE simple. It’s that no demand is required to create JG jobs. That is, and to illustrate, if someone on benefits of $Z/wk is told to pick up litter in the local park for $Z/wk, then aggregate employment rises, meanwhile aggregate demand remains unchanged, so there is no possible inflationary effect deriving from the “park tidying” job.
But there is a problem with that idea, and as follows.
It’s not just relatively low paid park tidying jobs which don’t require an increase in aggregate demand: the same point applies to ANY PUBLIC SECTOR JOB!!!! That is, for example, the output of the British National Health Service is GIVEN AWAY, rather than sold. So no extra demand is needed to expand jobs in the public sector as a whole.
But strange to relate, the typical Western country has brought about an ASTRONOMIC EXPANSION in the total number of “no extra aggregate demand” or public sector jobs over the last 150 years. Yet there has been no corresponding fall in long term average unemployment levels.
So it would seem that expanding the number of “no extra aggregate demand” jobs is not the explanation as to why JG raises aggregate employment.
The ACTUAL REASON why JG raises aggregate employment will appear here in a day or two.


  1. Ralph, can I have a stab at this? I feel that JG would raise aggregate employment for a couple of reasons. First, workers would retain working skills and not have to endure long periods of useless searching for non-existent jobs. The improvement in morale and confidence would be tangible. Employers would be able to choose from workers whose skills are not eroded when openings arise. Job training should be an essential part of the JG.

    1. JG could easily maintain skills (depending on how good JG is at allocating employees to jobs that actually use their skills). But if JG does have that effect, that doesn’t contradict my basic point above which was that the buffer stock notion is irrelevant to this whole argument.

      Incidentally, the above skill point is one of the arguments for extending JG to the private sector: doing so gives you a much bigger range of skilled or semi-skilled jobs to which to allocate JG employees.


Post a comment.