New Zealand’s weak productivity performance has been an on-and-off theme of discussion for decades. We’ve been falling behind for 70 years now, something that was recognised by expert observers almost 60 years ago. In all that time, there has never been any sustained period when we’ve made any progress in closing the gap. A typical cycle seems to involve Opposition politicians – of whichever party – suggesting that they will do better, and that under their stewardship we’ll catch Australia, get back into the upper half of the OECD, or whatever. Once in office, the rhetorical concern often lasts for a year or two, and then typically nothing much else is heard or – worse – there are attempts to twist the data to try to render our underperformance less stark.
There was a bit of focus last year on New Zealand’s latest run of poor productivity outcomes. I and others had noted that we seemed to have had no productivity growth for almost five years. And, sure enough, Opposition parties picked up the issue to some extent, and the then-government attempted to play distraction and pretend everything was fine.
And then there was a change of government, and a couple of months later a new annual update on the GDP numbers. The new numbers saw estimates of GDP for the last few years revised up a bit and – since estimates of hours worked didn’t change – that translated through into a lift in estimates of real GDP per hour worked. In some quarters, a sigh of relief was breathed. And, to be sure, in this context more was undoubtedly better than less.
But when I dug into the numbers it still resulted in this chart
We’d gone from having no labour productivity growth at all (actually marginally negative) over the last five years to total productivity growth over that period of 1 per cent (ie about 0.2 per cent per annum). It is a little better than the previous iteration of data has suggested, but……it wasn’t anything to boast about. It shouldn’t have made anyone much less uncomfortable. And on the updated data this was the New Zealand vs Australia comparison.
Once a year, Statistics New Zealand release official estimates of annual labour productivity growth for what they label the “measured sector” of the economy, which covers around 80 per cent of the economy (total GDP). The latest release, including data for the year to March 2017, was out last week. The “measured sector” includes about 80 per cent of the economy, where Statistics New Zealand is reasonably comfortable about its real output measures (the main exclusions are education and health).
Unsurprisingly, there were some upward revisions in these numbers as well. The numbers don’t get a great deal of commentary, but the reaction seemed encapsulated in this chart from one of the banks.
Not only were the numbers for the last couple of years revised up, but if you eyeball the chart productivity growth in the last decade doesn’t look much different than that for the previous decade.
That certainly looks more encouraging.
This is how productivity growth in (a) the measured sector and (b) the whole economy compare, based on the latest SNZ releases. Here I’ve used just production GDP – since it is production sectors SNZ uses to do the measured sector numbers – and have shown the data in log form, in which a constant slope of the line means a constant growth rate.
Recall that the measured sector is about 80 per cent of the economy. And for a gap of that size – the measured sector productivity growth was 17 percentage points faster than that for the whole economy over the 21 years – to have opened up it suggests that productivity in the non-measured sector (the rest of GDP) must have done very badly indeed.
This little exercise is purely illustrative. I’ve deducted measured sector productivity from the total, assuming the measured sector is indeed 80 per cent of the economy, and then multiplied what was left by 5 [(100/(100-80)] to produce a proxy residual index of implied labour productivity in the non-measured sector. This is the result.
It is only a proxy, calculated residually, and the precise numbers are sensitive to the (changing) exact share of the non-measured sector industries. But the proxy suggests a pretty calamitous picture for productivity (especially, if summary numbers SNZ includes are to be taken seriously, in education) in the non-measured sector. Disgruntled parental consumer of the education system that I am, something doesn’t seem to entirely ring true – these aren’t, after all, quality-adjusted numbers. When, as a matter of policy, money is being thrown at education and (eg) teacher/pupil ratios are being raised, one might expect crude measure of education sector productivity to fall. But that doesn’t seem to have been the story of the last nine years – whether the educational lobbies, or the former government, are to be believed.
One problem with the measured sector data is that there is no ready way to compare New Zealand productivity growth numbers with those for most other countries. There are no standard compilations of such data and it would take a huge amount of painstaking effort for an individual to attempt to replicate the numbers for a reasonable range of other advanced countries. Given the importance of common global (or at least advanced country) trends in productivity, that severely undermines how much use can be made of the aggregate data.
However, the Australian Bureau of Statistics publishes some very similar data, for what it calls the “market sector”, also around 80 per cent of the economy. And SNZ themselves highlight the comparisons between the New Zealand and Australian productivity growth numbers in each of their annual releases. As they note in the latest release, over the period since 1996 (the period for which the two countries have comparable data), labour productivity in the measured/market sector averaged 1.5 per cent per annum in New Zealand and 2.2 per cent in Australia (Australian data is for June years). Over 21 years, those differences multiply up to big numbers – and, in levels terms, we had already fallen well behind Australia by the mid 1990s.
You might have hoped that all those differences were early in the period. Unfortunately, the SNZ/ABS data suggest not.
Starting from just prior to the 2008/09 recession (downturn in Australia) the relative performances of the two measuered sectors look depressingly similar to the pattern in the aggregate (real GDP per hour worked) chart I showed earlier. Over nine years, Australia’s market sector managed total productivity growth a full eight percentage points fast than New Zealand’s measured sector managed.
All these charts just use hours worked as the relevant input measure. Usually SNZ also publish (as do the ABS) the data using composition-adjusted labour input measures (eg if the amount of human capital per worker is increasing, as people get more skilled, that represents more inputs not higher productivity). I’ve become a bit more sceptical of such measures in recent years, since proxies for human capital are often educational achievement numbers, and much of what Bryan Caplan writes about – that much of formal education is about signalling rather than skill acquisition – I find increasingly persuasive. But this year I can’t even show you the numbers as SNZ ends their release noting plaintively
The composition-adjusted productivity in the measured sector data did not meet our quality standards for publication..The absence of this data does not affect any other data published in this release. We don’t have an expected release date for this data
In sum, there isn’t much in the recent waves of productivity data/estimates that should give anyone serious about economic performance much comfort at all. There are, no doubt, countries that have done worse than us on this score in the last decade, but – starting well behind – we’ve made no overall progress in closing the gaps to other more advanced countries, and have continued to slip (quite materially) further behind our closest comparator, Australia.