In a vague sense, productivity means the ratio of outputs to inputs. Improvements in productivity arise from improvements in technology or new and better ways of organising production. Productivity can also decline if productive effort is diverted to wasteful ends, which may happen for a variety o reasons.
In m yexperience, most of the time, reported short-term changes in productivity are spurious, in a sense I’ll try to make more precise below. In thinking about this, I came up with the following proposed test/definition for an increase in productivity (it would also apply, with modification, to a reduction)
A change in productivity is sustainable if additional inputs of labour, capital etc could be added and would generate additional output at the new, higher level of productivity
The idea of this definition is to rule out as many as possible of the sources of spurious productivity gains, including
- factor composition biases, such as the gain in average productivity from closing the weakest plants in an industry/economy
- relative factor intensity biases, such as gains in labour productivity from an increase in the capital-labour ratio
- factor use intensity biases, such as labor hoarding during recessions and work intensification arising from microeconomic reform
The most famous case of factor composition bias was the Thatcher productivity miracle, achieved primarily through plant colsures. Factor intensity bias is a chronic problem, which has been addressed through the construction of measures of multifactor productivity. Increased work intensity was the main source of the Australian productivity miracle in the 1990s.
(Since miracles are invariably spurious, I’ve avoided putting scare quotes around the term.)
John Quiggin 
I thought the Thatcher ‘batting average’ effect was shown to be not true, that is, there really was a genuine increase in productivity in Britain in the 80s. (I don’t have a reference.)
I found this paper (PDF) which says that plant selection effects are important. It’s focus is 1994-97, though, and it doesn’t really say whether they were more important under Thatcher than at other times.
I’d read the evidence as saying that any period of rapid contraction in total employment in a sector is likely to produce spurious gains in productivity in that sector. At least in the first half of the 1980s, this was clearly the case in the UK. It’s also relevant in assessing recent performance of the US manufacturing sector.
I’ve found a reference that says the opposite
Bean & Symons, ‘Ten years of Mrs T.’, NBER macroeconomics annual 1989
Let me selectively quote
” … the batting average effect looks increasingly less plausible as time passes. This produces a once-and-for-all change in the level of productivity, and this should have come to a halt after 1982 as the economy entered the recovery phase. … Furthermore, while the level of productivity is … higher in large plants, the 1980-81 recession was associated with a shift in employment away from large plants. Consequently, there was, if anything, a reverse batting average effect in which the high productivity producers were eliminated.”
If I have my timing right, 1989 was just before the Lawson boom fell in a heap. If you look at Australian publications around the same time (that is, pre-recession) they were very positive, but this was seen in retrospect as a period of very poor productivity performance.
As noted in the Economist report I linked not long ago, British productivity is still well below French, which runs against the idea of a sustained productivity improvement under Thatcher (though it’s scarcely proof.
Pr Q makes a marginal increment in the productivity of the economics profession:
I like it – it seems clearly testable and consistent with economic concepts.
Productivity increases when, for a given factor input, you are maintaining a higher marginal output.
This complements the notion of production increases which, by economic axiom, are subject to diminishing marginal returns.
But I would be more interested in theoretical explanatiopns, rather than conceptual analyses, of productivity.
In particular, what causes the productivity to increase rate to remain at a steady ~2% increment over time.
Does this somehow measure the rate at which improvements in economicly relevant knowledge are embodied in economic activity’s:
– technical tools
– social rules
Is the growth of productive economic knowledge the ineluctable result of submitting industrial hypotheses to the test of the market, pace Hayek?
And how can we make productive economic knowledge accumulate ten times faster without working too hard?
ie If productive economic knowledge is the ulitmate free lunch, where is the tuck shop?
Handy post, John. So would you agree the latest US productivity numbers constitute ‘bad productivity’ (an unsustainable sackings-plus increased-work-intensity scenario)?
As a non-economist I do like the definition of the sustainable change in productivity.
One aspect of factor composition bias, i.e. the increase in productivity that comes from throwing the least productive people into the dustbin, is also an aspect of French relative success : –
> As noted in the Economist report I linked not > long ago, British productivity is still well
> below French
which partly has to do with the sort of enterprise you channel your labour into, which is probably discussed at length in economic texts, which I don’t read, but seldom makes the business pages, which I do.
Specifically the UK has very little of the French luxury production. I imagine our farm labourers are as productive as French ordinary farm labourers on their prairies, but probably not as productive as the labourers at Chateau Petrus, as an extreme example..
It appeals to me too John … and I liked Milton’s second quote, if I understand it correctly (isn’t 1981-82 a little early for measuring Thatcher?).
About 15 years ago I had a mini-epiphany at a seminar on productivity. It’s economics without prices! This cynicism was reinforced by watching Swannie do a study of productivity at (then gov’t owned) Telecom. Naturally, he found productivity was dismal. I asked “what if you had found it was great?”. Peter was honest enough to say “well, then we would know they were getting all sorts of underpriced inputs”. A few minutes of giggling and silence before he admitted that, no, it really was not a test after all….
John,
As you say, productivity is an example of a vague concept that can be defined in different ways. And this vagueness can be exploited for political purposes to mislead the gullible.
But as a general principle can you really solve this problem by insisting on a particular definition? It seems to me that you then rule out other definitions that may in fact be useful in certain contexts.
In the case of productivity, it is in fact sometimes useful to consider ouput per unit labour. Suppose for example I observe that country A has higher long-run productivity growth than country B by this measure, and proceed to account for this in terms of A’s faster-growing capital-labour ratio. Do you really want to disqulaify me from using the term productivity here? Do you want to protest that this is an invalid use of the term precisely because capital intensity differs between the two coutries?
Surely it’s better just to insist that people define their terms carefully and not make unwarranted claims about the results of policies. Bob Rowthorn did this very energetically and effectively in the case of the Thatcher productivity miracle, and did so without trying to mandate a particular definition.
In any case your definition is a bit too inelegant to catch on, I would think, since it’s an implicit one. That is, instead of saying ‘productivity is …’, you are forced to say ‘productivity increases if etc etc.’