In the last few weeks, presumably simply by coincidence, I’ve had various comments and emails about productivity growth in the agricultural sector. The most recent one finally prompted me to dig out the official data and check that my impressions were still supported by the data. They were. Agricultural sector productivity growth was very strong, but has been much more subdued for some time now.
There are two main measures of agricultural sector productivity: labour productivity (in effect, output per hour of labour input) and multi-factor productivity (in effect, the residual after what can be attributed to growth in labour and capital inputs has been deducted). In principle, MFP is superior. In practice, estimates rely more heavily on the assumptions used in the calculation (although – diverting briefly – to the various readers who have sent me a recent piece by GMO on TFP/MFP, I reckon there is less to that critique than the authors claim).
Agriculture is one of the sectors for which we have consistent productivity data back to year ending March 1978. By New Zealand standards, that is a long run of data. Note that there is quite a bit of natural volatility from year to year because of fluctuations in the climate (droughts and all that).
Here are the charts showing growth in labour productivity and in multi-factor productivity (both expressed in log terms, so that the slope of the line indicates the growth rate).
Really rapid growth in the first 20 years or so (up 150 per cent in 20 years), but much slower growth since then. It isn’t disastrous growth by any means – averaging just under 1.5 per cent per annum from 2005 to now – but nothing startling either.
Once again, really strong productivity growth for the first 20 years, and much less since then (none at all in the last decade). If one had confidence in the data, that might be more concerning.
Perhaps – and it is an obvious question to ask – you are wondering if the slowdown in agricultural productivity growth is just part of the general slowdown in labour productivity growth in New Zealand (seen mostly starkly in the overall economy figures, where there has been very little growth at all in the last five years or so).
So here are the same two charts – labour productivity and MFP – but this time showing agriculture relative to the overall “former measured sector” (the series SNZ publishes that also goes back to 1978, covering the 70 per cent or so of the economy where productivity is relatively less hard to measure). All series are indexed to a common base in 1978 (so none of this is about productivity levels, it is all about cumulative growth rates).
Labour productivity first.
There is clearly something in the story. Growth in the “former measured sector” productivity itself has slowed – labour productivity growth averaged 2.4 per cent from 1978 to 1998 and 1.6 per cent for the 20 years since 1998. But whereas agricultural sector productivity growth was (far) outstripping that in the rest of the “former measured sector” in the earlier years – the upward-sloping bits of the last two charts – in the last 20 years (and even more so in the last 10) productivity growth in the agricultural sector has been a little slower than in the rest of the “former measured sector”.
I don’t have a policy point to make here. Sector-level productivity isn’t my area. Relative optimists might look at the data and suggest liberalisation played a big part in the earlier productivity surge, and that liberalisation wasn’t ever going to be repeated. Perhaps, but bear in mind that our agricultural sector was always the part of the economy most heavily exposed to international markets (by contrast, a sector like “Transport, postal, and warehousing” – with similar total labour productivity growth over 40 years – was almost certainly much more sheltered). Pessimists might focus on the role of unpriced externalities (particularly around water pollution, but also methane emissions) and wonder how agricultural sector productivity might respond as regulation bears more heavily there (although it is possible we could see higher average productivity and lower total output).
Others might reasonably wonder about these productivity measures for agriculture in particular: in most sectors labour and capital are the critical inputs, but land is a huge input in agriculture and it isn’t directly accounted for (something which matters – much – more in looking at levels estimates). [UPDATE: A SNZ person got in touch and pointed out that they do take account of land in their capital services estimates.]
As for me, my only motivation was to update my impressions. Sadly, the latest data confirmed them.