Cross-country comparisons are often enlightening, even if they just throw up more questions. In matters economic I’ve often highlighted New Zealand/Australia comparisons – both distant, resource-dependent, economies, with a similar legal and political background, market-oriented economies etc.
Ever since the coronavirus issue started to come to the fore it has also seem natural to try to make a little more sense of the New Zealand experience by looking at what is happening in Australia. After all, both countries took the China travel ban route (us dragged along a day later by the Australian choice), both are islands (to the extent that matters at a time of more generalised movement restrictions), both have just come off summer (to the extent that may, or may not, matter) and both typically prided themselves on being open and democratic societies. In both countries, (lots of) returning travellers seem to have accounted for most confirmed cases.
For quite a while, the raw numbers in New Zealand looked quite good relative to Australia. The first Australian case was more than a month before the first New Zealand one (by the time our first case was confirmed they had 25 confirmed cases). When I was running some charts on the comparisons in March they were still perhaps two weeks ahead of us (eg on 21 March we had 13 new confirmed cases, the number (adjusted for population), Australia had had on the 6 March). Quite why there was that difference wasn’t (still isn’t, that I know of), clear, although Australia had been doing more testing per capita. But the gap was beginning to close.
By mid-March some mix of government urgings and conscious voluntary unease and precautions was already inducing quite a lot of behavioural change, and physical distancing (I recall a quick trip to Auckland on 19 March, and coming back through a largely empty Auckland airport domestic terminal with a sense of the world closing down around us) but the extent of official restrictions was still a few days away from being beefed up dramatically. The following week, New Zealand was put into a partial lockdown that went significantly further, in important respects, than the restrictions put in place by state and federal authorities in Australia.
Physical distancing works. There seems little or no doubt about that. Our government seems to have insisted on quite a bit more of it than Australian governments have. And so over the last few days I have been intrigued, perhaps even a little puzzled, to watch the evolution of this chart.
It is an updated version of a chart I’ve shown before, showing new confirmed cases for the two countries, with the Australian total divided by five, to adjust for their much larger population.
You can easily see how their new confirmed cases numbers were well ahead of ours, right through until 23 March. Then they more or less level-pegged for a week or ten days, but for the last week or more their confirmed new cases (per capita) have been running well below the new confirmed case numbers in New Zealand.
I don’t know quite what to make of that chart. I’m genuinely intrigued by it. But a few points are worth making:
- first, confirmed cases reported today do not reflect actions/choices made today, but those made up to, say, 10 days previously (take time for people to get infected, develop symptoms, get tested, and get the results reported),
- second, I’m not sure we can yet put much weight on the idea that Australia’s position is now still worse than ours because they have more deaths and more in ICU. Clearly, their position was worse than ours (the early weeks of the chart), but it takes time for cases to get serious enough for people to get to the ICU stage and/or die (see, most, prominently, the Boris Johnson cases) and three weeks ago we had very few new cases,
- relative to many countries, our confirmed cases so far were for a long time more concentrated among younger people (travellers), where the effects of the virus are generally less serious (I haven’t seen the Australian data on that particular point),
What of testing? Australia has done more tests per capita than New Zealand, but in the last couple of weeks our per capita rate of new tests has been higher than in Australia. In both countries, the rate of confirmed cases to tests done is fairly low (a good sign about how much testing is being done).
Presumably in both countries not all actual cases are being captured in the numbers (this seems to be standard line worldwide), but whatever the “true” situation, all indications are that both New Zealand and Australia have, for now at least (and in some sense), got on top of the situation, and new cases numbers have not only not escalated, but fallen away (sharply so in Australia’s case). In both countries, the reduced flow of travellers from abroad must have helped and (residually at least) the quarantine of the few remaining travellers. Private choices to distance must have made a difference. And so must government restrictions and injunctions.
But it must still be a bit of an open question quite what things explain which bits of the respective slowdowns in the number of new confirmed cases. A detached observer two weeks ago might have looked at the respective regulatory regimes and controls in the two countries and reasonably have expected a much sharper slowing in case numbers in New Zealand than in Australia (that would probably have been what I would have expected). But it hasn’t happened. Perhaps it still will: time will tell.
Is it then possible that, given the extent of the virus in both countries, our partial lockdown – and partial it is – went further than was really necessary, well beyond the point where there were material marginal gross benefits at all? I don’t know, and I don’t suppose anyone does at this point. But in thinking about appropriate next steps it is a possibility that looks as though it at least needs to be explored, teased out and tested. After all, there are gross costs to take into account before any sort of net benefit assessment is possible.
At very least there seems to be very little basis for the sort of all-or-nothing thinking embodied in the Prime Minister’s spin the other day – waving around a chart of case projections from mid-March, contrasting it to the actual new case numbers, and suggesting – at very least by implication – that the difference is all about the extent of New Zealand’s stringent partial lockdown. In principle, only any superior performance relative to Australia should be able to be attributed to the more stringent regime in place here. It is hard to get any sort of clean read on the “true” relative performance, but the prima facie numbers suggest reason to doubt we’ve done better at all.
Finally, and to be clear, I am not championing any particular approach to policy over the next few weeks. There are details of the current regime that I regard as frankly barbaric and inhuman, unworthy of the sort of Prime Minister who endlessly exhorts people to “be kind” (as if we’d ever take our moral lead from Prime Ministers or MPs), but personally I remain somewhat uneasy about any very substantial easing of the economic restrictions in the near future. That is visceral more than analytical. It is also easy for me to say – my life hasn’t changed much through the lockdown (once a week to the supermarket rather than once or twice a day, and that’s about all), I like having the kids around (boring as it may be at times for them), our household income seems secure, and even family members with businesses seem able to run them more or less. I might well feel differently if my livelihood were at stake. My only point here is that we need to keep testing and challenging interpretations, including against what is happening in other otherwise similar countries. And relative to Australia, there are some real unanswered questions and intriguing puzzles.
croaking cassandra 
What’s really bizarre is that the virus apparently behaves differently in NZ. We have 1283 identified cases and two deaths. That’s a tiny case fatality rate of 0.15 while Oz’s is 0.85, or 5.6 times higher (and with similar testing criteria).
Our hospitalisation rate is also way lower than Australia’s.
Prof Michael Baker tried to explain it as a result of NZers’ returning travellers being young and fit… Are we meant to believe our travellers are younger and fitter than Australia’s?
Something is very odd here… either in the way we are coding deaths or how less willing people here are to go to hospital… unless Australia has a more virulent strain, which seems unlikely…
LikeLike
Two things:
– Aus had a lot of cases from the two cruise ships (one in Japan, and the Ruby Princess), where people will have been older
– it seems to take 15-18 days on average from infection to death. we still had few cases 18 days ago.
LikeLiked by 1 person
What’s really bizarre is that the virus apparently behaves differently in NZ
———————–
To which, we need to add amazingly rapid recoveries from what, elsewhere is projected as a deadly disease.
What is very disappointing though is the extent to which the old people homes are affected. With all our officially sanctioned “kindness” and “caring” we failed to protect those most at risk.
LikeLiked by 1 person
My mother seems to have many more visits from her nurse. Before lockdown the nurse was very busy and only able to see my mother once a month. Now it is 2 to 3 times a week. It does appear to me that many older people have lost the home visit health care perhaps the nurse making a personal health choice where households appear to be higher risk to be avoided and lesser risk more regularly visited. Great for my mother but not so great for many who seems to have lost their care worker.
LikeLike
Thus Emergency Level 4 Alert is getting rather embarrassing. Watched the news tonight and the Emergency Response Unit and the hospital is pleading for more patients. Patient numbers are down more than 50%. We have doctors and nurses twiddling their thumbs staring at empty beds.
The government really needs to understand the meaning of acceptable risk versus the cost of shutting down the entire country. This is a really really dumb and stupid government.
LikeLiked by 1 person
I can add an anecdote to Getgreatstuff’s comment. I have spent two spells in the onco ward at Waikato during the shutdown. My treatment wasn’t affected as I was seen as high priority treatment. However, near half the beds in the ward were empty as a lot of the other patients had their treatment deferred. This ward is for in-patient treatment of people with actual cancer, not just there for tests. They had near normal shifts of nurses, though the doctors were doing week about. Now some of these deferrals will have serious consequence for their survival options. So there is just one downside of the lockdown.
LikeLiked by 1 person
It has behaved differently in a lot of places. Northern Italy hit hard and very localised. Southern Italy, not much.
One conjecture is that air pollution is a very significant factor. Another is that it is very local to specific hospitals.
LikeLiked by 1 person
Thanks, Michael.
What is is interesting that epidemiologist Dr Tony Blakely yesterday in the Herald attributed the difference to NZ detecting more mild cases, which can’t be true since both nations have roughly the same (restrictive) testing criteria.
“What I take from that is New Zealand is doing a better job testing and finding those people out there who are asymptomatic or have very mild disease, and we must be missing them in Australia,” he said.
https://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objectid=12324069
LikeLike
Yes I had those comments in mind when I wrote.
There must be more than a few aspects of “luck” in various bits of many countries’ experiences. Both countries seem to have been v fortunate that no China-sourced cases clusters resulted from travel before the China ban.
LikeLike
The real test is when the cold depth of winter arrives. Perhaps what is forgotten is that over there in the worst hit areas, geographically it is winter and over here and Australia it is still summer. As their summer approaches their cases are starting to level off and fall. No doubt lockdowns do help to reduce the spread but I think weather does play a major part as it does with the ordinary flu.
My thoughts are that the covid19 virus in NZ and Australia see still in hibernation. They are not really super active. It does take a fast replication by the virus to be able to overwhelm a human bodies white cell defences. If the disease is mild it does mean a much slower rate of replication. The higher UV radiation in summer weather does slow down replication.
We are all hoping this lockdown works. But if covid19 behaves like the flu then the worst is yet to come.Hopefully this winter remains warm and mild.
LikeLike
We can’t be finding more asymptomatic cases, because the rules for testing specifically say that asymptomatic patients must not be tested (except if instructed by a regional Medical Officer of Health).
Click to access covid-19_case_definition-8apr20.pdf
In addition, the tests that we do conduct are finding absurdly low rates of the disease, even though the group being tested consists entirely of patients who have symptoms and are (largely) suspect because they are connected to other cases. Yesterday, 2,421 tests were conducted and 14 new cases were discovered, a rate of 0.6%. Yes, I know that yesterday’s tests are not the same as yesterday’s results because of lags of a few days in reporting the results, but the rate has been less than 1% for the last 5 consecutive days, and that is not a artefact of lagged reporting.
So we are testing only people where there is a very strong presumption that they are infected, and yet 99% of those tested do not have the disease. In European countries, the rate of positive results seems to be in the range of 10%-30%.
My immediate reaction was that the NZ tests are horribly flawed with a very high false negative rate, so that most true cases are being missed. A counter-argument is that if we were missing lots of cases, then the missed cases would be causing more outbreaks in the community, and we are not observing that; but since everyone is isolated regardless of their test results then perhaps this would not occur.
So I have a serious concern that the way the tests are being conducted in NZ labs is wrong, and that most genuine cases are being missed. This implies that the number of cases reported is badly understated – the European comparison suggests that the true number of cases may be 10-50 times what we are being told. Presumably sick or dying people who get a false negative covid test result are reported as being ill or dying from pneumonia of non-covid causes. The death rate is unreliable because the numbers are so tiny anyway.
Does anyone have an explanation for the very odd NZ data which does not imply a high false negative rate in our tests?
LikeLike
I wonder if the data is indicating that the most effective part of the virus management strategy is the banning of large social gatherings. As we have seen from the largest clusters here that one person can spread the virus to 80 odd people in 18 days if they attend a St Paddy’s party or wedding reception. Australia seems a lot looser in their restrictions on work etc, and as you point out, they are not doing obviously worst than us at this time.
Both countries may also have cultural factors working in our favour. Having been to Italy and Spain and seen just how social the locals are (as well as the cheek kissing greetings) it is not too hard to see how they would have a faster virus spread than us hand-shaking antipodeans.
LikeLike
The early banning of large social gatherings does seem to be a big part of explaining the mystery of Japan.
LikeLike
Hi Michael
NZDAUD currency cross has fallen back quite sharply over the last few days and I think a good part of the explanation is the market is starting to think Australia is going to come out of this faster than NZ due to their less restrictive approach and the sharp improvement in Chinese data (if you believe it!).
Take care Regards Conal
________________________________
LikeLike
Thanks Conal. Of course sharp improvement in China still looks like activity levels running perhaps 20% lower than usual.
LikeLike
Thanks for the post, just a point of detail, your totals for New Zealand include ‘probable cases’ whilst the figures for Australia don’t.
For a better cross-country comparison, you would use New Zealand’s confirmed daily cases as Australia is only providing confirmed (i.e. resulted in a positive laboratory test) daily case numbers. Or you could make an assumption that Australia has a similar rate of ‘probable’ cases to New Zealand and increase the Australian numbers accordingly.
If you were to remove New Zealand’s probable daily cases the curves actually look remarkably similar.
Another perhaps more interesting analysis would be to look at the state of Queensland versus New Zealand in isolation. Queensland’s population is about the same as New Zealand’s both have one major population centre/region as well as a bunch of other mid-sized towns and cities that are rather geographically dispersed along a thin population corridor.
WIth either the updated numbers or the comparison with Queensland I suspect the same point would be made. It appears that both countries/regions have done well in managing the spread albeit Australia appears to have done so without quite the same degree (albeit still very significant) disruption in daily activity and liberties.
As you say quite intriguing puzzles to consider.
LikeLiked by 1 person
Thanks for the point about the probables
LikeLike
Rascally Rabbit. I don’t think NZ Emergency Level 4 lockdown can be considered managing anything. A lockdown is somewhat of a last resort and in NZ case we have gone absolutely bonkers with our response. That’s not managing. That is like killing a mosquito using a sledgehammer and bringing the house down in the process without only some chance of actually hitting the mosquito.
If this covid19 flu behaved anything like our normal flu then with our current warm, bright high UV radiation drawn out summer and drought means the virus is hibernating, viral replication and transmission rates are much lower in summer weather anyway. The lockdown helps but when the virus is in hibernation transmission rates are already low. The real test would be in about another 5 weeks when we are out of lockdown in the bitter damp cold and covered by the usual winter long white cloud and the virus activates.
LikeLike
Getgreatstuff; agree. The NZ approach will see a significant test come winter. Keeping people inside is a great way to maintain the health of the virus.
LikeLike
Heading off down another trail, I found the results of this Harvard study interesting, suggesting as they do that there are more cases in places where air quality is compromised … this might be part of the explanation for what we are seeing so far in this part of the world?
https://projects.iq.harvard.edu/covid-pm
LikeLike
Reblogged this on Utopia, you are standing in it!.
LikeLike
Another thought may be seasonal factors… just chucking this into the mix…
It seems that coronaviruses as a group like it dry and cold, which is why, generally speaking these things pop up in winter more often than summer… So Australia as a warmer climate than NZ, which acts are something of a barrier, and parts of Australia are way more warner and humid than NZ….
Both countries are heading into winter so it will be colder here sooner and last longer than Australia..
That won’t explain all of the difference between the two countries, but it could be a factor. Worth a regression model or two??
Also getting plenty of sunshine and generating vitamin D (rather than tablets) seems to also be effective in helping to protect yourself generally from coronaviruses… More sunshine in the burnt country over there than here…
LikeLike
A friend told me Western Australia has very few cases (thanks to the desert).
LikeLike
When my brother lived in Perth he always said WA stood for Worlds Apart. It may be their saving grace now.
LikeLike
I was once told there’s a military dictum something like “In any situation you can make the right decision, the wrong decision, or no decision. The worst response is no decision.” My take on it is that in highly uncertain situations (like a battle, or policy to deal with a new and poorly understood virus) lots of mistakes are being made. People will die, or suffer because of those mistakes. The worst response to that knowledge is being indecisive and allowing each bit of analysis and advice to change what you do.
I suspect like most of your readers i have a guess what lies behind the graph, but that’s the wrong way to understand it as advice for decision makers. As you say, it shows social isolation works, here as elsewhere, but what lies behind the detailed trajectories between countries are unknown (and for decision purposes in the next couple weeks, probably unknowable).
We are in the midst of Knightian uncertainty. Ardern has taken a decision within the bounds of what we do know and acting on it (unlike a certain D Trump). The rest is for historians.
LikeLiked by 1 person
RNZ has had a go this morning at deciding why NZ has a low death rate.
I’m hoping Dr Bloomfield has been misquoted here because the rate of deaths can’t be due to how early we reacted (although the number might be):
“When asked about New Zealand’s low death rate earlier this month, Director-General of Health Ashley Bloomfield said he put it down to two things. One was that we reacted much earlier in the outbreak than other countries and had put in measures much sooner.”
https://www.rnz.co.nz/news/national/414057/covid-19-why-is-nz-s-mortality-rate-so-low
LikeLike
RNZ has had a go this morning at deciding why NZ has a low death rate
————————
What’s even more remarkable (and baffling) is a very low hospitalization rate. This, in my view, is more important indicator of the severity of this disease in NZ so far. It would be good to see an enlightened comment on this particular aspect.
LikeLiked by 1 person
Maybe not enlightened this is my explanation. Deaths are predominately elderly people. NZ has roughly the same number of elderly as other developed countries but we are different – young people in NZ do not meet the elderly as frequently as other places. This isn’t because we are especially unfeeling but because our immigrants have mainly left their grandparents back home in country of origin and our native born Kiwis have had small families most of whom have emigrated overseas or at least to Auckland. Our elderly are concentrated into retirement villages. The exception to my point would be Pacific Island and Maori who do prefer multi-generational household but just as NZ had an advantage in getting the epidemic a week or two after most countries (which helped close down retirement villages before the national lockdown) the Pacific Islands have largely managed to avoid infections.
LikeLiked by 1 person
Taieri, how fast the virus replicates determines how severe the disease is. In order to get past the human body natural defences it must reach critical mass numbers before being killed off. Therefore it is most severe in winter conditions rather than in summer conditions. The NZ and Australian long warm, drought ridden and cancer level high UV radiation summer has meant that the virus is hibernating.
The real test is the cold wet and cloudy winter when we will see full activation. But by then deaths from suicides from job and business losses and domestic violence should also be counted. Also don’t forget the 50% drop off in patients that had their operations cancelled for other diseases while our hospital beds sit empty, doctors and nurses twiddling their thumbs with nothing to do.
LikeLike