Suppose a tear in the spacetime continuum had delivered you a copy of December quarter 2020’s employment statistics one year early. You had received the key graphs on the employment rate, the unemployment rate, labour force participation rates and underutilisation rates for the year to come – but none of the accompanying discussion.

What would you have concluded must have happened over the course of 2020?

A mild downturn in the employment rate began around March quarter 2020, with a bit of a rebound in the December quarter. The drop was sharper for women, but only returned women’s employment rate to where it was in 2016. For men, it’s hard to see anything other than noise around a post-2017 mild softening.

Something bad clearly happened to the unemployment rate in the September quarter – the single-quarter jump from 4% to 5.3% topped the worst quarter of the Great Financial Crisis, but also only brought unemployment rates back to where they were in 2016. And things started righting again in the December quarter.

And the labour force participation rate was only 0.2 percentage points lower than December quarter 2019, after a mild dip earlier in the year.

In February 2020, I was starting to freak out about Covid and had started stockpiling essentials. Had I then seen the year-ahead’s employment statistics, I’d have concluded that Covid must not have amounted to much: either it somehow fizzled out on its own, or a cure was quickly found.

The September quarter spike? Maybe, before fizzling out, Covid pushed a couple of Italian banks over and triggered a minor European financial meltdown.

It would be hard to look at the past year’s employment figures and see that the government had shut down most of the economy for most of the second quarter and shut the border for the year.

Back in April, every economist, myself included, expected Very Bad Things To Come in the employment statistics.

International tourism had been about 5.5% of GDP. The multiplier effects of that spending, best ignored in normal times, could matter during a giant pandemic recession.

Closing the border should have provided an obvious hit to the employment statistics.

Disruptions abroad made a mess of supply chains. And, of course, we had a couple of lockdowns.

It’s hard to see much of it in the employment data.

Reality is always consistent with itself. It is our models of how it all hooks together – the toy models in our heads and the formal models on economists’ whiteboards – that must be the source of any confusion.

What did we fail to foresee back in April?

Clearly, policy has helped. But the main bits of policy were already in place by late April when we all still expected worse outcomes.

Wage subsidies let viable firms reboot quickly after lockdown with existing teams in place, rather than needing to struggle back to life after receivership. Maintaining those links also prevented what economists call ‘scarring’ – the effect of being unemployed on future employment prospects.

But while the scheme was important stopgap, it could hardly shield against the loss of nearly 4 million annual overseas visitors who fill the country’s restaurants, bars, and tourist venues.

Monetary policy matters but should have been priced into our expectations in April; the Bank was very clearly going to be printing an awful lot of money. And if the Bank had been more active than we all expected in April, why has the exchange rate appreciated? Perhaps monetary policy abroad has been even more active, helping sustain demand for exports from New Zealand?

So what have we missed?

Are labour markets just far more flexible than we had expected, perhaps assisted by government’s flexi-wage programme? Flexible labour markets make it easier to redefine roles as circumstances change. They also make it easier for firms to take a punt and find out whether someone who has spent over a decade as an airline steward might easily take on a very different role at a Covid-testing laboratory. But training to shift from hospitality to construction takes time.

Perhaps employment losses in tourism and hospitality disproportionately affected those on working-holiday visas, with less hit to the employment statistics? But even in that case, the loss of flow-on spending should still have been important.

Did the closed border mean more jobs for Kiwis? There are substantial problems with that account too. Migrant workers also buy things here. One usual story suggests migrants affect demand before supply. Those workers not spending on goods and services here should have meant less employment in the places where they would have spent money. And many workers from overseas are complements to local workers, enabling firms to hire more locals as well.

Or perhaps the real pain is still yet to come, as firms update expectations about how long it will be until the border can re-open safely.

Having a crystal-ball view into New Zealand’s employment statistics, a year ago, would not have shed any light on the Covid pandemic ahead. But a thorough dissection of those statistics could yet prove illuminating. It is wonderful that our predictions have thus far proved wrong, but it would be exceedingly helpful to really know why.