In my post on Monday I drew attention (again) to the fact that New Zealand has made no progress at all in reversing the decline in relative economywide productivity (relative to other advanced countries) since what was hoped to be a turning point, with the inauguration of widespread economic reforms after the 1984 election. If anything, the gaps have widened a bit further, and more countries (most former Communist ones) have entered the advanced country grouping, first matching and now overtaking us. Despite being so far behind the OECD leaders there are also clear signs that labour productivity growth has slowed further in the last decade or so.

All that discussion proceeded using simple measures of labour productivity (real GDP per hour worked).  The data are readily available for and are more or less comparable across a fairly wide range of countries, and there is meaningful levels data. Labour productivity is a common measure in such discussions, even though total (or multi) factor productivity (TFP or MFP) is the in-principle preferred measure. It is the bit of growth in output or output per capita that can’t be explained just by the addition of more inputs (labour or capital). Some decades ago the late Robert Solow, recently deceased, observed that in modern economies perhaps 80 per cent of the growth in output per capita had been attributable to TFP.

It is a line that should be taken with several pinches of salt since in practice (a) TFP is an unobservable residual, and b) much of the innovation and new knowledge often thought of as the basis for TFP growth is probably embedded in better human and physical capital and the disaggregation is a challenge (to say the least).  Thriving economies are likely to have better smarter people, better tangible and intangible capital, all used in better smarter ways etc.

But with all these caveats I thought it might still be worth having a fresh look at the OECD’s MFP data for the last few decades. They only have MFP (growth) data for a subset of (24) member countries (mostly the “old OECD”, and including none of the central European countries). For New Zealand, the first MFP growth data is for 1987, and with the annual data available only to 2022 that gives us 36 years of data.

There is a lot of year to year noise in the series, but for illustrative purposes I simply split the data in two, to compare the record for the 18 years to 2004 with the 18 years to 2022. As it happens, the global slowdown in productivity growth in leading economies (US and northern Europe) can be dated to about 2005.

New Zealand averaged annual MFP growth of 0.9 per cent in the first 18 year period, and only 0.2 per cent per annum in the second 18 year period to 2022. It is a pretty dire picture. (All data in this post use arithmetic averages, but using geometric would not make any material difference.)

Now, champions of the reform story might be tempted to look at that simple comparison and say something like “yes, you see. In the wake of the decade of far-reaching reform New Zealand made real and substantial economic progress, but then after the reform energy faded and drift took hold it all faded away to almost nothing.

Unfortunately for that story, here is how New Zealand MFP growth record compares (on the OECD’s particular methodology) for New Zealand and (the median) for the other countries (most of them) for which there is a complete set of data.

We all but matched the average growth performance of those other advanced OECD economies in the earlier period, in the wake of our reform process, but even then didn’t do well enough to begin to close the large levels gaps that had opened up in earlier decades. And then in the more recent period, we’ve done worse again: the comparator group (typically richer and more productive, nearer the productivity frontiers) slowed markedly, but we slowed a bit more still. When you start so far inside productivity frontiers there is no necessary reason why New Zealand could not have made some progress closing the gaps even if the frontier countries themselves ran into difficulties. But no. (Over that second 18 year period when New Zealand averaged 0.2 per cent per annum MFP growth, South Korea - also well inside productivity frontiers on an economywide basis – is estimated to have averaged 2 per cent per annum MFP growth).

It is only one model, and only one set of comparators but there is simply nothing positive in the New Zealand story. There is, and has been, no progress in closing those gaps, and our living standards suffer as a result.

And what of Solow’s 80 per cent? In New Zealand real GDP per capita increased by an average of 1.7 per cent per annum between 1987 and 2004. MFP growth averaged 0.9 per cent over that period. For the period 2005 to 2022 average annual growth in real GDP per capita increased by an average of 1.4 per cent per annum, but over that period MFP growth averaged just 0.2 per cent per annum. (The comparisons are no more flattering if one uses the OECD “contributions to labour productivity growth” table as the basis for comparison.)

Whichever measure of productivity one looks at the New Zealand performance is poor. Champions of reform 40 years ago would, I think, have been astonished if they’d been told how poorly New Zealand would end up doing. I hope they’d be even more alarmed at the indifference to that woeful record that now seems to pervade official and political New Zealand.

[And since I’ve already had one past champion of the reforms objecting to my characterisation in this post and Monday’s post, I’m equally sure that all serious observers now - ie excluding our political leaders and officials - have their own story about what else should (or in some cases shouldn’t) have been done over recent decades. That doesn’t change the fact - on my reading and my memory – that if asked in say 1990 most would have envisaged several decades of catch-up growth as the decline of the previous decades was slowly reversed. It is quite clear from the documentary record that that was the goal, and the intense political disputes of the era were not about that goal.]

Earlier in the week a journalist asked me for some thoughts on which political party in government had managed the economy better – in overall macroeconomic terms – over the years since we moved to MMP.

My initial response was that the answer would be pretty dull. Pressed to write something anyway, I outlined briefly why really there was not a great deal between them, at least without a great deal more in-depth study. And that shouldn’t be very surprising. After all, external shocks happen (overseas, or physical/climatic ones here), and cyclical macroeconomic management has been outsourced to the Reserve Bank over all that period, with very similar targets set by successive governments. Crude partisans might point out that (core) inflation went outside the target zone at the end – or so it appears likely – of both Labour governments over that period, but those failures are first and foremost on the Reserve Bank. Other crude partisans might point out that unemployment has been at its lowest right towards the end of both Labour governments, but……since that is basically the same phenomenon as the overheated economies that gave rise to the inflation problem, you are back with it being the Reserve Bank’s mistake again. One can argue, say, that the current government should have done more to sort out, punish, or even support (via fiscal policy) the Reserve Bank, but…..it is a sample of one event.

When I wrote something a bit fuller on this topic a few years ago, I noted that fiscal policy had also largely been a bipartisan success story. We might not have had a very successful economy, but when deficits have emerged governments of both parties have restated their commitment to surpluses, and had delivered. You could argue that National deserves a better rating there, having inherited large deficits in 2008 (as I’ve argued here before Labour had been badly advised by Treasury and did not think it would be leaving deficits) and returned to surplus. But actually if you look back at the 2009 and 2010 Budgets, contemporary Treasury estimates were that – starting from a deficit – they were expansionary. I’ve been quite critical of this government’s fiscal stewardship over the last couple of years, but…..nothing much in the campaign suggests any more urgency for or conviction about a return to surpluses from the other side.

But the backdrop to it all was that while, until quite recently, New Zealand – under governments led by either party – had done reasonably well on the stabilisation side of things (monetary and fiscal, and even with structural policies that kept the non-inflationary rate of unemployment fairly low), productivity (or the lack of it) was the elephant in the room. It has been for a long time, and still is (or should be).

We don’t have an official quarterly labour productivity series, but it is easy enough to construct one’s own. In this chart, I’m showing the average of production and expenditure GDP measures, divided by hours worked from the HLFS, all normalised and expressed in log terms. Expressing things in logs means that a slowdown in the growth rate is mirrored in a flattening of the curve. We don’t have long runs of official historical data in New Zealand, but this goes back to 1987Q2.

If you can easily see any great difference from governments of one party to governments of the other you are more eagle-eyed (or perhaps “motivated”) than I am.

But I did check anyway. One could go from the last quarter of the previous government to the last quarter of the next one, but….there is clearly noise and measurement error in the data, and nothing is that precise, so although I checked both, this little table uses annual data (eg average change from 1990 (last year of that Labour government) to 1999 (last year of that National government) and so on. Now, no one really believes that changes of government make a difference immediately, so this is illustrative more than anything.

Average annual growth in real GDP per hour worked (%)
1990-1999 National 1.1
1999-2008 Labour 1.4
2008-2017 National 0.9
2017-2023 Labour 0.7

Much the biggest story isn’t the difference between the parties, but the difference over time. Productivity growth in the last decade or more – under both governments – has been materially lower than it was earlier in the period – under governments of both parties. This is consistent with the factoid I’ve thrown around a few times in recent weeks: in OECD league tables for labour productivity we dropped six places – in a club of only 37 – in the last decade.

Here is the deterioriation illustrated graphically. Eyeballing the data it looked to me as though there was a break around mid 2010. So what I’ve shown is (a) the actual data per the previous chart, and (b) an extrapolation to now of the trend in the data from 1987 to 2010.

Roughly speaking the gap between the two lines as of now is equivalent to a 10 per cent loss of productivity (growth we would have seen if the previous trend had continued).

Note that all of this is simply New Zealand data. I have repeated often charts showing our deterioriation – or at best lack of catch-up – relative to other advanced countries. But this is us. And remember that we are so far behind the productive frontier economies – it would take perhaps a 60 per cent increase to catch them – that even to the extent world productivity growth slowed down (and it did, in the US from about 2005) there is no necessary reason why New Zealand productivity growth needed to slow. Our slowing was about New Zealand policy choices, passive or active.

It is depressing how little serious attention has been paid to these failures – and challenges – in the election campaign, and since politicians mostly display little interest our bureaucratic institutions don’t bother doing or supplying the hard analysis. Some are simply emasculated to that end – one could think most notably of what the current government has done at the Productivity Commission. Productivity really matters for our future material living standards, and even for the shiny baubles both main parties try to woo us with.

I’m not an ACT supporter – on quite different grounds – but here I would give that party some credit. Their policy document on productivity evinces a degree of seriousness about the issues that nothing from any of the other parliamentary parties has even hinted at. I don’t agree with all the specifics, and would probably disagree substantially on some, but….they write as though it matters. And that isn’t nothing (even if it can’t overcome my scruples about the party leaders’ values etc). In fact, a week or two back a reader not otherwise known to me got in touch and asked who I thought they should vote for if it was housing affordability and productivity that mattered most to them. Making clear that I was definitely not an ACT supporter myself, I nonetheless gave them an analyst’s answer: probably ACT, on both counts.

The recent quarterly national accounts data prompted me to update a couple of charts that I used to run here regularly, one of which I hadn’t bothered updating since the start of Covid.

The first is for labour productivity. In New Zealand, we don’t have an official series (level or index) of economywide labour productivity (real GDP per hour worked) but it is easy enough to construct one (or several) using the two quarterly measures of real GDP (production and expenditure) and one or both of the HLFS hours worked and QES hours paid data. I used to simply turn all the series into indexes and divide the average of the two GDP series by the average of the two hours series.

Covid messed up those estimates. For example, under the wage subsidy scheme a lot of people were being paid (as a matter of policy) for hours they were forbidden to, or were otherwise unable to, work. And on a more enduring basis, we now have a higher baseline level of sickness, which comes and goes with waves of Covid, and so for the time being I’m just using the HLFS measure of hours (since it is measuring the hours people tell SNZ they were actually working for). It also then allows for a more-direct comparison with the official ABS real GDP per hour worked series for Australia.

With series for both countries indexed to 100 in 2019Q4, the last pre-Covid quarter, this is how the productivity indicators have done over the last few years.

There is quite a lot of “noise” in the New Zealand numbers, and I still have very little confidence in the any of the numbers for lockdown quarters themselves (in early 2020 and late 2021), but taken over the entire 13 quarters (from before Covid was on any horizons in either country, to now where there are few/no regulatory disruptions) both countries end up with – on current estimates – next to no productivity growh at all over the full Covid period.

I don’t find that particularly surprising (although who knows what the numbers will eventually be revised to) – Covid was a really big disruption and costly dislocation in all sorts of ways – and was much more puzzled by the earlier indications that a reasonable level of productivity growth had been maintained (even allowing for compositional points – eg low productivity motel cleaners and cafe waiters were some of the disproportionately most likely to have lost their jobs in 2020). Perhaps too there is now some cyclical and compositional effects at work: both labour markets have recently been substantially overheated and almost anyone who wanted a job could get one, probably averaged labour productivity a bit downwards.

But they are hardly numbers to be complacent about. For what it is worth, UK official numbers on growth in economywide real GDP per hour worked are (a bit) less bad than these Australian and New Zealand numbers, and US non-farm labour productivity (thus not strictly comparable) while weak in recent quarters still looks to be not far from a pre-Covid trend line.

My second chart has, over time, raised more hackles. Almost 20 years ago a visiting IMF mission team were looking at how the (then newly) rising exchange rate was affecting the economy, and one of their people put together a chart crudely illustrating the relative performance of the tradables and non-tradables sectors of the economy. Tradables here was represented by the production GDP components for the primary and manufacturing sectors, to which was added (in a way that makes statisticians wince, but which isn’t without meaning) exports of services from the expenditure GDP accounts. They were, loosely, the bits of the economy either producing for exports or facing direct international competition. The non-tradables component was the rest. I think I later hit on the idea of expressing the series in per capita terms, to cope with longer runs of time.

This is roughly what the chart looked like around the time it was devised.

If anything, the tradables sector had been growing faster than the non-tradables sector in the 1990s and early 00s, which was sort of what the opening-up narratives had led people to expect. But after about 2002, activity in (this proxy for) the tradables sector seemed to be going sideways. All else equal that might not have been too surprising for a cyclical rise in the exchange rate (back then, the New Zealand real exchange rate experienced really big cyclical swings).

This is what the chart looked like the last time I bothered updating it, just prior to Covid

The non-tradables sector had continued to grow quite strongly (in per capita terms), while this proxy for per capita tradables sector output had had its ups and downs but was by the end of 2019 no higher than it had been in 2002. A common narrative on this blog through its first few years was that the economy had become increasingly inward-focused, even though sustainably successful economies tend to be ones with rapidly growing tradables sectors (not, to repeat myself for the umpteenth time, because exports are special, but because there is a big global market out there and successful competitive firms will tend to find global customers).

I stopped looking at the data for several years because when the government says people can’t travel then of course services exports (notably tourism, but also export education) were going to dip quite sharply, and all it was doing was reflecting the priority placed on Covid control, nothing about underlying competitiveness issues.

But now things are returning to normal. We don’t have restrictions on people coming, and nor do most countries, and even China is opening up (for outbound travel) again. So it seemed worth coming back to my indicator chart.

I was quite surprised by what I found.

Non-tradables real per capita output has still been running well above trend, consistent with an overheated economy (high inflation, large current account deficit), as we’d seen on a smaller scale in the mid 00s. But the proxy for the real per capita output of the tradables sector hasn’t yet recovered much at all, and is only about 10 per cent higher than it was 32 years ago.

Something didn’t seem right. After all, there were a lot more tourists around.

And sure enough the data (seasonally adjusted but not per capita here) show that services exports have recovered a lot as the Covid restrictions were lifted. But notice the blue line, the GDP components of the primary and manufacturing sectors. It has its ups and downs but the level of the latest observation was first reached in 2004Q1 (when the population has only just passed four million).

It is an increasingly inward-focused economy, where policy (such as it is) is only tending to reinforce such developments. It doesn’t have the feel of the foundations for a prosperous and highly productive economy for New Zealanders – this generations, or our children and grandchildren.

But while the big political parties fight over the keys to the Beehive offices and cars, it would be most surprising if one sees any serious or sustained engagement with data like these (or the policy choices that have given them rise) in the election campaign over the next few months.

In yesterday’s post I drew attention – yet again – to New Zealand’s continued drop down the international productivity league tables. There are all sorts of caveats to the details – PPP comparisons are inevitably imprecise, and the data are subject to revisions – but few seriously doubt that we do much worse now relative to other advanced countries than we did just a few decades ago.

But it is easy to lose sight of what the numbers actually mean for ordinary New Zealanders, so I thought today I might do just a short stylised illustration.

In yesterday’s post – as on various occasions in the past – I’ve contrasted our outcomes with those of a group of highly successful OECD countries (but excluding Norway (oil), Ireland (even their own authorities don’t use GDP as a measure of Ireland’s outcomes) and Luxembourg): Switzerland, Denmark, Belgium, the United States, Sweden, Austria, France, Netherlands and Germany. They can be thought of these days as some sort of “leading bunch”, at least as regards labour productivity.

In 1970, when the OECD data start, our real GDP per hour worked was about 82.5 per cent of the median for this group of countries. By 1990 that proportion was about 64.5 per cent.

In 1990 the confident hope – among officials and ministers, and more than a few outsiders – was that we were on the brink of turning things around. I’ve shown before this photo of then Finance Minister David Caygill’s aspirations/expectations.

Let’s suppose it had worked.

The fall in our performance relative to that OECD leading bunch had taken 20 years, so suppose that over the following 20 years we had steadily improved such that by 2010 our real GDP per hour worked was again about 82.5 per cent of that of the leading bunch, and then held at that ratio subsequently. That wouldn’t have made us a stellar performer, but at least on the OECD’s numbers we’d be doing about as well as Canada and a little better than Australia. Since we’d more or less tracked Australia for many decades earlier it wouldn’t have been an unrealistic aspiration at all.

How much difference would it have made?

There are lots of possible moving parts, but I did this little exercise by taking as my starting point actual real GDP per capita for New Zealand each year over the last 30 years and scaling it up by the ratio of the assumed productivity performance to our actual. Fortunately the official GDP per capita series starts in 1991.

This is what the chart looks like.

By the end of the period, the annual difference – per man, woman, and child – is about $20000.

But what does it add up to? After all, every year since 1991 our incomes could have been higher than they were. And $1000 extra from 1991 invested even just at a real government bond rate adds to quite a lot by now (especially given New Zealand’s interest rates over that period). Applying a (stylised) series of real interest rates – falling over time – and applying them to each year’s difference in real GDP per capita, the total difference came – in today’s dollar terms – to a bit over $500000 per man, woman, and child. For almost everyone in New Zealand that would be serious money.

You could produce a different number with different scenarios. Slower convergence would, of course, produce a lower number. On the other hand, using the sort of discount rate The Treasury requires government agencies to use – rather than just a long-term real interest rate – would value past gains more highly and produce a materially bigger number.

The point also is not to suggest that if somehow economic policy had been run better and produced these stronger productivity outcomes that everyone would have banked all the proceeds and be sitting today on an extra nest-egg of $0.5 million. That wouldn’t have happened at all. In a more productive economy, people would have been able to – and no doubt would have- consumed more. Government revenues would have been stronger, and better public services might have followed even at the same tax rates. Some might have chosen to work less (a real gain to them too). The half million is a way of putting a number to the options that much better performance would have created. And the gains would be mounting further every single year. Another way of putting that is that every single year, the failure of successive governments means a median family of five is missing out on another $100000.

To repeat, this exercise is entirely stylised. Depending on one’s view of which set of policies might have delivered these better outcomes. some other things might have been very different. Perhaps our terms of trade would have been different (since probably a somewhat different mix of products). Perhaps our real interest rates would have been closer to world levels. Perhaps…perhaps. The point is simply that decades of economic failure adds up to really large amounts of income (and potential consumption) just lost. In New Zealand’s case, half a million per capita will do to be going on with, mounting at $20000 per capita with each year we languish so far behind the bunch.

And just think of how much better off our country would be – avoiding all the systematic and deeply unjust redistributions – if over the same period successive governments also had not so badly messed up the land market, in a way that has delivered us such extraordinary house prices.

What might have been……

But what still could be if there were to be a government of courage and vision.

Next Monday morning in Paris, the OECD’s Economic and Development Review Committee will be gathering to discuss the draft of the latest Economic Survey of New Zealand. These mid-level public servants will pose some questions, offer some observations (some insightful, most probably not), and their French and Hungarian members will lead the questioning and – if the past is any guide – get to enjoy a very nice lunch at the New Zealand Embassy. Some weeks later, after the text has been haggled over line by line, we will get to see what the OECD has to say about our economic performance and policies.

But in many respects, the numbers speak for themselves, and the OECD does a pretty good job – one of the useful things it does – of collating and making accessible a wide range of data for its member countries.

The OECD’s series of labour productivity data (real GDP per hour worked (in PPP terms)) starts from 1970. New Zealand joined the OECD in 1973 when it was a club of 24 countries. For all but two (Austria and Greece) of those countries there are labour productivity estimates back to 1970. By 1970, New Zealand had already dropped a bit below the median OECD country, but the countries either side of us were France and the United Kingdom, and we were just slightly behind Germany.

In relative terms, New Zealand’s economic performance was particularly bad in the 1970s. Of those 22 OECD countries for which there was a consistent series of data, over the course of the 1970s New Zealand’s real GDP per hour worked fell from about 95 per cent of that of the median country to about 75 per cent. By 1980 only four of those countries had lower labour productivity than we did, and we were no longer between the UK and France but between Ireland and Finland. To younger readers, Ireland may not sound too bad, but Ireland in those days was still an underperforming economic backwater.

Things didn’t look too different in 1990. The number of countries for which there is data had increased a bit and of the 1973 membership we now stood between Japan and Greece.

The 1990s was an era of opening-up. A variety of countries re-emerged from the ashes of the Soviet Union and Yugoslavia, other countries formerly under the thumb of the Soviet Union began turning themselves into market economies. And the OECD itself started to broaden its membership, looking to Asia and Latin America for new members. That process has continued and there are now 38 OECD countries, and (more importantly for this post) labour productivity data for all of them since at least 2000.

In 2000, 17 of those 38 countries had average labour productivity lower than New Zealand’s. We were still between Greece and Japan.

By 2010, 16 of those 38 countries were below New Zealand. By then we stood between Greece and Slovenia, the first of the former eastern-bloc states to match us.

So far, so mediocre.

But what about the more recent period? 2020 marked the end of another decade.

It takes many many months to get a complete set of annual data (and, of course, many of those numbers – including our own – are still subject to revision). But the full set of 2020 estimates is now available.

I have been quite hesitant about using 2020 data for anything other than Covid purposes. Countries had quite diverse Covid experiences – some good (bad) luck, some good (bad) policies etc but most of any differences probably not telling us much about the longer-term economic performance story. And measurement was a real challenge – both GDP and hours worked. Bear those caveats in mind in what follows, but looking at the data (and checking against trends to 2019) the distortions seem less than I might have thought/feared.

Here is how things looked in this snapshot.

Only 9 countries now do less well than us, and we are bracketed between Korea and Poland (for all the hype around Korea’s economic performance – and it is impressive over several decades – only now has their average labour productivity matched that of underperforming New Zealand). Former laggards Turkey, Slovakia, Slovenia, Lithuania, the Czech Republic and Estonia have now moved past us – some well past us – and mostly just in this last decade.

And if you think this is just a story about other countries doing really well – which we shouldn’t begrudge for a moment, it is something to celebrate – bear in mind that on these estimates New Zealand’s productivity growth rate in the 2010s was less than it was in the 1970s. That’s the New Zealand of Key, English and Ardern – oh, and record terms of trade – managing to underperform the New Zealand of Kirk, Rowling and Muldoon.

If that chart is bleak enough about New Zealand’s standing in the OECD league tables, just think where it might be a few years from now. It is good of the OECD to have welcomed in the four Latin American members – the diversity hires – who will keep New Zealand off the very bottom of the league table for a long time. If Chile is managing pretty good productivity growth. Mexico in 2020 had real GDP per hour worked no higher than it had been in 2000.

But of the five countries below us last year, three are likely to go past us in the next few years.

Portugal could pass us too, but they’ve been just slightly behind us for 30 years now.

2025 was once – briefly – the year when we were supposed to have caught up with Australia by. Instead, most likely, on this metric the only OECD countries that will be doing worse than us will be Portugal, Greece, and the four Latin American members that don’t even make most lists of advanced economies.

What a woefully bad set of outcomes successive waves of politicians and officials have delivered for New Zealanders. (All while they’ve delivered us some of the most expensive housing anywhere.) And what is perhaps most depressing is that none of this seems to bother either side of politics, and neither political parties nor (so far as we can tell) officials seem to have any real interest in reversing decades of underperformance. These outcomes aren’t some exogenous given, but the outcome of repeated sets of policy choices. And if the policy choices of the previous government in this area were bad (and they were) those of the current lot seem materially worse (exemplified in recent days by the tens of billions of taxpayers’ money they want to spend on glorified trams, as if money were no object).

Incidentally, lest anyone be under the illusion that Australia itself is some sort of success story, Australia (7th in the OECD in 1970) is now close to the middle of the league table – much better than New Zealand, but not exactly a multi-decade success story.

There is a group of countries near the left hand side of the first chart – from Switzerland to Germany – that I’ve often contrasted us with. Even in 1970 we were behind all of them on this measure (basically level-pegging with France). But now, to catch up with the median of those countries it would require a 69 per cent lift in New Zealand average labour productivity.

It could be done – it would take productivity growth averaging 1.3 per cent per annum faster than in those countries for 40 years – but it isn’t going to happen by simply ignoring the issue, hoping for different outcomes, or by adopting sets of policies that are only likely to continue our decades of relative decline.