Friday, 27 January 2012

Checking for lumps

The 'lump of labour fallacy' is the erroneous belief that there is a fixed amount of work in the economy. This leads to a variety of claims: that a reduction in hours worked will free up more jobs; that if immigrants take jobs this must mean an equivalent number of natives lose them; and that automation means a reduction in labour (aka the Luddite fallacy).

This is believed to be a fallacy because the amount of work is not fixed. The economy is not a closed system (or "box economy"), so the specific change in hours, workers or automation may trigger other, countervailing changes, not to mention the simultaneous impact of other factors.

The specific claims made are dismissed on the grounds that: reduced hours often leads to increased productivity (same output in less time), which negates the need to increase labour capacity; immigration means more customers as well as workers, i.e. the size of the economy expands leading to more jobs overall; and technological substitution leads to workers moving to higher skill roles.

However, at a certain point, the lump of labour fallacy will turn out to be true. For example, marginal reductions in hours worked may not impact on job numbers, but larger reductions surely do, e.g. a job-share produces 2 jobs from 1.

So the lump of labour is a fallacy. Or is it? Consider ...

1) The gender distribution of jobs

Over the last 50 years, more and more women have entered the jobs market. In parallel with this, there has been a slow but steady increase in the percentage of men who are economically inactive (i.e. neither employed nor seeking work).

The numbers will be muddied by the impact of changes in work patterns (part-time and temporary) and the fluctuation in the rate of structural unemployment (between 1 and 3 million since the 1980s), but there appears to be a clear thread through the last 90 years: the number of people economically active, as a proportion of the population, has stayed broadly the same (just over 60%, according to England & Wales Census data), but it has become more evenly distributed by gender.

On the face of it, this would appear to support the lump of labour fallacy in relative terms, if not in absolute terms - i.e. the total economically active population has increased with general population growth, but as a proportion it has stayed pretty constant. 

In theory, an increase in women in jobs has meant an expansion in the economy and therefore greater demand overall, which should have led to a higher percentage of the economically active, however this has been offset by a tendency for male inactivity to increase. The loss of jobs by men is partly explained by the disproportionate impact of structural unemployment in the 80s (i.e. the collapse of heavy industry) and job polarisation in the 90s (see #3 below), though it's worth noting that the census data shows a steady decline since 1931 (they skipped the anomalous year of 1941 for obvious reasons).

The inference is that the relative decline in male employment and the increase in female employment is merely coincidental. The implication of the long-term data is that there may be a natural limit to the rate of economic activity.

There is also evidence that stagnation in median male earnings may have compensated for the increase in female participation, hence limiting growth. In other words, more working women may have served merely to maintain existing levels of consumption.

One oddity of the recent discussion about median wage stagnation is the finding that this started in the US in the 70s but didn't appear in the UK until the last decade. In fact, it looks like the better performance of female wages in the UK (relative to per capita GDP) from the 70s masked a comparable slowdown in UK male wage growth over the period, which then began to flatline after 2000. (See the graphs on page 19 of the Resolution Foundation's 2011 study Growth Without Gain?)

2) The impact on jobs of retirement

The increase in the UK state pension age has largely been discussed in terms of the imposition on the individual, i.e. working longer for the same. What has occasioned less comment is the impact that this, and the expectation of more people working beyond the state pension age, will have on employment.

The consensus view is that "there is no evidence that reducing the employment of older persons provides more job opportunities for younger persons" (see a 2011 UK government paper and an earlier 2009 US study.) However, most studies have focused on correlations in employment rates between young and old in the context of early retirement, i.e. would encouraging older employees to "free up slots" reduce youth unemployment.

There are fewer studies that look at the reverse, i.e. whether delaying retirement reduces hiring opportunities, largely because increasing the retirement age has only come onto the agenda recently. A study of the impact of a raised retirement age for women in Portugal in 1993, looking at specific firms, came to the conclusion that: "the new law had the effect of decreasing hirings by about one worker for each older worker retained in the firm."

There's an element of "no shit, Sherlock" to this. If an older worker defers retirement by 1 year, that means the resulting vacancy (let's assume the role is still needed) will be deferred by a year as well. During the period of transition to the new retirement age there must be a large number of such deferrals. As many of these would otherwise have created vacancies, there must therefore be a reduction in the total number of potential vacancies, adjusted for other factors, created during the transition.

The lump of labour fallacy, i.e. the explanation as to why it is wrong, would suggest that pushing the retirement age out creates a cohort who add further consumption to the economy and thus increase labour, but this assumes that people in their early 60s spend at a relatively high rate and then severely reduce their expenditure immediately upon retirement. Anecdotal evidence (I can't find anything better) suggests this isn't so. People reduce their expenditure well in advance of retirement, driven as much by lower need (kids gone) as by the looming carriage clock. It would be interesting to find some real data on this.

What this implies is that the change in the state pension age will result in a relative drop in vacancies during the transition years. These are absolute losses, i.e. there is no compensating above-trend increase in vacancies after the transition, so all other things being equal, this represents a net increase in unemployment.

Will this be significant? About 800,000 people turn 65 in 2012, which reflects the post-war baby boom in the late 1940s. Let's assume about 50% are in employment. People who retire early will probably still retire early, and it is expected that the abolition of the default retirement age will result in many extending their working life, so let's discount it by a further 50%. This still leaves us with a potential addition of 200,000 to unemployment, ceteris paribus.

The average retirement age is now increasing, partly due to poorer pensions and partly due to people electing to work beyond the state pension age. This will increase the working population, however it will probably be offset by fewer school leavers due to the falling birthrate since the 90s. In fact, while the working population may increase in absolute terms, it may decline as a percentage of the total population due to increasing longevity.

3) The change in the type of work

Technological changes since the 1970s have led to "polarisation", reducing the number of jobs in the middle income range (i.e. median pay). High skill jobs have grown, as have low skill jobs. The former implies a greater emphasis on creative and technical roles, the latter on personal services that can't easily be automated or offshored (e.g. care).

The increase in high skill jobs has been held up as a good thing, on the assumption that we are gradually upskilling as a nation and doing more technical/professional jobs in the "knowledge economy." This has been offered as an upside in tandem with the downside of the export of productive roles (particularly in heavy industry and manufacturing) as a consequence of globalisation.

I suspect that much of this increase isn't actually in high-skilled creative and technical roles, but is actually more "management". The ratio of managers to managed in business is roughly 1:10. I think this is understated as it doesn't include part-time managers (e.g. team leaders) plus "management support" staff who are effectively extensions of the formal management tier.

The ratio of officers to other ranks in the military is about 1:6. I suspect the true management ratio in business is not far off that, perhaps 1:8. While they're not run in quite the same way, (most) businesses employ the same basic command and control paradigm as the military.

It can be argued that today we have many more people who manage "things" rather than people, hence we need more managers, but that is often just a matter of job title status. A database manager is really just a skilled machine operative.

My personal experience of IT is that many so-called IT professionals know little or nothing about the subject. They have either mastered some arcane knowledge (i.e. they are an advanced user of a specific software package) or they are a business generalist (project managers, business analysts etc.) It was interesting to note how many refugees from banking in 2008 admitted they had never fully understood what they were doing (I'm currently reading David Kynaston's City of London: The History, which reveals that this attitude has been common among younger employees for almost two centuries).

Disciplines with professional qualifications, like accountancy, are no better. Most accountants don't do accountancy (it's not really that complicated and software does most of it for you). The qualification is usually a passport to a more general business role, i.e. a species of "management".

Professional qualifications, and the payment necessary to secure them, have always been an entry fee for access to high-pay jobs, as much as they have been about engineered scarcity or maintaining standards.

As wage inequality has grown, we have seen an increase in both the entry fee and entrant desperation for high-end careers. The minor furore over internships is an example of this: high-value roles being auctioned at charity events for £000s, and the well-off subsidising their kids to work a year or more for no pay.

What I think the increase in management, professional and technical staff is hiding is not grade inflation but "outdoor relief" for the middle classes (the phrase originates with John Bright, condemning the use of the Foreign Office to create jobs for aristocrats in the 19th century).


We are surrounded by robots, they just don't look like Robby from Forbidden Planet. That said, there are still plenty of jobs that could be automated. In other words, we are in the middle of a transition.

We seem to be managing this transition by creating more overhead roles - i.e. the productivity gains of automation are subsidising management and professional jobs that we had hitherto not felt the need for.

You don't have to believe we will reach a post-scarcity state to foresee a situation in which the majority of genuine production in the economy is automated. Most of us will then either be doing pseudo-productive jobs (i.e. disguised consumption), or we will have become artists who rarely sell a work.

It might be argued that modern capitalism would not stand for such a degree of inefficiency and waste, however another way of viewing it is that there is plenty of loot for all and we do need to circulate wealth to consumers in order to keep the show on the road. This is as good a way as any.

It was a popular trope last year to compare corporate looters with the rioters in London and elsewhere. This focus on the extremes conveniently ignored the degree of institutional looting undertaken by the silent majority in management.

It looks to me as if there is a natural (or preferred) rate of economic activity and that we will adapt organisations and roles to maintain it. In other words, we act as if there is a fixed lump of labour.

Structural unemployment (wrong skills, wrong place, wrong time) is less a failure to adapt productive labour and more a failure to adopt pseudo-productive work. Not every ex-miner can become a CSR administrator.

No comments:

Post a Comment