Another improvement

Covid exposed, not for the first time, how New Zealand is short on timely official measures of the economic cycle. 

While that's true, let's acknowledge some qualifications. For one thing, it's not just us: many countries have struggled to figure out anything close to a real-time view of the state of the economy. And for another, we've had some progress in filling in the gaps, notably Stats' publication of monthly electronic transactions and monthly filled jobs. And where the official statistics haven't fronted up, there's a wide variety of timely private sector information, notably the ANZ's various monthly surveys and the BNZ/BusinessNZ ones.

But the general proposition stands: we haven't had a good enough range of near-real-time official data to give policymakers at the Reserve Bank, Treasury, or elsewhere a good enough feel for where we are and, from that, what needs to be done about it.

So let's put our hands together and applaud Stats, the Treasury and the Reserve Bank for cooperating to design the New Zealand Activity indicator (announcement, first results, technical note, Q&A). It's bound to become one of the go-to statistics for reading the shape of the business cycle. And another round of applause, please, for the decision to keep the thing going beyond the covid-inspired need. One of the things that's bothering me is that while people in various organisations have scrambled hard in difficult circumstances to provide quicker information on the economy, it's by no means certain that the new, useful data that Stats and others have been providing will be maintained post-covid. Daft, of course, that anyone would think of going back to where we were, but there you have it.

I really like this new Activity indicator, or NZAC as I suppose we'll have to get used to calling it. I'm a great fan of the statistical method used to put it together (principal components): it basically tries to identify what common factor is driving the variability across a whole range of different measure such as traffic count and electricity generation. The answer of course, is the overall level of economic activity, and with principal components analysis you can get a good handle on the behaviour of that underlying driver.

How good? Startlingly good, as it happens, as this graph shows (from p5 of the technical note) .

Let's face it. NZAC, and GDP, when shown a year on year percentage changes, are essentially the same thing. And no, I don't want any lectures from the people who put NZAC together, that NZAC and GDP aren't the same thing. I know that. You know that. And strictly speaking I suppose that Stats / the RB / Treasury are right when they say (p4 of the technical note) that "NZAC should not be interpreted as a 'flash estimate' or high-frequency measure of GDP" (though many angels could dance on what exactly 'estimate' means, and why something that gets remarkably close to the eventual outcome isn't for some reason an 'estimate'). So while the purists behind NZAC are technically right, and even as you read this there's probably some analyst at Stats going "See, I told you so! They're going to make a balls-up of it", I'm going to go ahead and say that NZAC is a pretty good advance estimate of GDP.

Which, by the way, you can figure out from the technical note itself. On p4 it says that the NZAC registered year on year changes of 1.5% (January), 1.8% (February), and -4.3% (March). Average those out and you get an average NZAC decline for the quarter of -0.3%. That compares with the official GDP figure of -0.2%. Pretty good, eh? The Q&A document (p3) shows that the NZAC in late 2019 registered 1.3% (October), 1.8% (November), and 1.7% (December), which averages 1.6%, again pretty close to the official 1.8% outcome. There are commercial forecasters who'd dine out on that level of accuracy.

Parking all that, let's use the NZAC to try and figure out what actually happened to GDP in the June quarter. It's a critical thing to know: it's not the only cost of covid, but it was a big one, and it informs how hard fiscal (and monetary) policy needed to push back. 

Here's the calculation. At the moment we've only got the April and May NZACs, so I'm going to make three guesses at June - continued improvement (a reading of -3), a stronger recovery taking us back to last year's activity level (a reading of 0), and a pent-up demand rebound that actually takes up well up last year's level (+5). That'll give three averages for the June quarter NZAC, which I'll then take to be the percentage change in GDP from Q2 2019, giving us Q2 2020 GDP. We already know Q1 2020, so we can then see the change from Q1 2020 to Q2 2020.

One wrinkle. The NZAC Q&A documents says (in Question 14) there's a good chance that the April NZAC may not have captured the full fall in economic activity in the month: "Many areas of activity that were severely hit under level 4 – such as tourism, hospitality, education, and construction (to name just a few) – are not necessarily well captured in NZAC ... We consider it likely that if indicators of such activity could be obtained, the magnitude of the drop in April shown by NZAC would be magnified. In fact, we fully expect the size of this drop to be revised as the index is refined". So as well as three possibilities for the June NZAC, I've run with three possibilities for the April NZAC: as reported (-19), a bit worse (-25) and a lot worse (-40).

And here's what comes out as the estimated decline in June quarter 2020 GDP.

The good news is that all of these estimates are less than Treasury's earlier take in the Budget forecasts "indicating a decline in real GDP of around one-quarter" (p4 of the BEFU). Nothing wrong with Treasury's view at the time, but it looks as if we came through the lockdown levels faster than earlier seemed likely, and possibly with lower hits to business and consumer confidence than originally feared.

These estimates are also a bit less downbeat than the median forecast of the bank forecasters: I collected their picks for the June quarter GDP decline a week or two back, and the median pick was -16.8%; my own first stab at the decline had been 18-19%. So there's a real chance that June will have turned out rather less bad than we'd all thought. The latest government financial statements, for May, say that GST revenue was "not as adversely affected by lower economic activity as assumed in the GST forecast" in the BEFU, again hinting at a better June GDP outcome than anticipated.

None of this is to minimise the extent of the counter-cyclical fiscal and monetary task ahead. Even after a better than expected June, we face umpteen challenges including some permanent scarring, the sectors shattered by closed borders (tourism, education), and our exporters' prospects in a soggy covid-riddled world economy. But it's something that things could have been a lot worse.