Wednesday, 24 February 2021

Border Costs

Cecile Meier walks us through some of the costs of a border system that has neither been able to safely scale up to meet need, nor able to find any reasonable way of prioritising entry into those scarce MIQ spaces.

When Zane Gillbee hugged his family goodbye in South Africa before moving to Wellington, his daughter Lyla was still a baby and his son Callum a happy seven-year-old.

Lyla is now a potty-trained, walking, talking two-and-a-half-year-old and Gillbee has missed it all.

Callum, who is about to turn 9, has been diagnosed with separation anxiety and is on medication for it.

Zane Gillbee is one of the hundreds of skilled migrants who moved to New Zealand for a better life before Covid-19 hit, expecting his family to follow.

There are a lot of people in this situation, but not so many that it would be impossible to fix. 

A Facebook group for families split from New Zealand migrants counts 1600 members sharing increasingly desperate stories.

Immigration lawyer Katy Armstrong conducted a survey through the group, which 700 odd people completed, including 500 who had children.

She estimated the total number of split families of temporary visa holders in New Zealand to be about 2000.

The MIQ system has 4500 rooms. If there are 2000 split families, then that would be about half of the system's fortnightly capacity to get the job done.

The government provided hundreds of MIQ slots for the Auckland boat race. As of late January, Immigration had issued 753 visas for the Cup, including at least a couple hundred dependents of Cup workers. But MBIE can't say how many rooms they've taken up. I suspect it's because they don't want anyone to know. It would be feasible to know. The Immigration side of MBIE knows to whom visas were extended for the Cup. The MIQ side of MBIE knows who was in each room. All it would take it running the one list beside the other to see if the names from the one list match up with names on the other list. 

There simply is no good way for the government to be picking and choosing who should get these scarce spaces. Lots of Kiwis just hate foreigners and are happy for migrants with families abroad to get the shaft, so there's little political pressure to solve this. Lots of Kiwis like boats. So we get the outcomes we get. 

But there is a good way to expand capacity in the system so that more people can travel safely. Testing before travel will knock back the number of positive cases arriving here. If the government ever gets around to implementing the most obvious way of strengthening safety in the border system, that too can help enable greater capacity. 

Daily saliva-based PCR testing makes each person in MIQ far less risky because they can quickly be shuttled to JetPark if they're found to be infected. That means far less risk of infecting an MIQ worker or another MIQ visitor. Combine it with daily testing of every border worker, and there's far less risk of the virus getting out of the border system. You just won't get into spots where you're not sure where it came from, if every single person in the border system is tested daily. 

Because saliva collection doesn't require nurses, another bottleneck in the system is eased.

Because each infection is less likely to be transmitted, some facilities that might not have been safe as MIQ facilities could be safe when combined with daily testing, with pre-flight testing, and with triaging of visitors by Covid-rates in their country of origin. Australia is going to be low risk even when they're having a minor outbreak. America is going to be higher risk for a while. Safe the more secure facilities for the folks coming in from riskier places, open up a few more facilities for people coming in from less risky places, and run daily testing to keep the whole thing in order. 

Safely expanding MIQ capacity matters. We're likely to be stuck with this for the rest of the year, if vaccine roll-out will only be happening in the second half of the year. The costs of keeping things as they are for families like the ones Meyer has spoken with - they're just too high. 

Tuesday, 23 February 2021

Blessed are they that have not seen the model, and yet have believed

The Climate Change Commission's recommendations span the breadth of the economy. They are required to come up with sector-by-sector climate budgets consistent with getting New Zealand with net zero emissions under the Zero Carbon Act. 

The sector-by-sector budgets rest on underlying models. The models build predictions about what will happen as ETS prices rise, and what will happen when some additional constraints are put into the system. Some of the CCC's recommendations then mandate what they think are their best guesses about what a carbon price would do, subject to those constraints.

The scope is vast. The entire economy, really. 

And the Government has already signaled that it will just do whatever the Commission says to do. 

So getting things right seems to matter and is rather high stakes. 

In that kind of situation, you'd think that the underlying models would be available for checking and testing. Getting bits wrong could be really really expensive, whether you want to frame it as economic costs, or as carbon mitigation forgone. 

But the Commission is not in a sharing mood. Here's Kate MacNamara.

Critically, the commission has not provided either sensitivity analysis nor the marginal abatement costs, broken out by industry.

That data matters. Sensitivity work helps economists to understand just how precarious that "less than 1 per cent of GDP" figure is. Will it alter significantly with slight adjustments to inputs? And the industry data for abatement cost would allow interested parties to properly test the assumptions the commission has made.

A commission spokesperson said it was unable to answer questions before the Herald's deadline on Monday. But one reason Carr has given for withholding information is the use of some US$6000 worth of proprietary global trade data from the Department of Agriculture Economics at Purdue University. It isn't clear why this data can't simply be stripped out.

In an emailed response to the letter signatories, Carr also said the board of the commission will consider the request for more time. But he didn't sound hopeful.

"As you know the commission has a deadline to deliver its final advice on or before 31 May 2021. Evaluating submissions and determining the impact of submissions on our draft advice also needs adequate time," he wrote.

You might have hoped that plans that have potential to re-engineer the entire economy would have more provably robust underpinnings. 

A couple years ago, our shop started putting out some new measures on school performance, based on work in the SNZ datalab by our excellent Joel Hernandez. We put up a technical report to go with it to show our workings, and we made all of our code available to anyone else in the datalab to check over and build on if they wanted to. 

If Joel's work turns into policy, we'd be really happy about it, but there isn't a direct channel there.

The Government has signaled they'll just do whatever the Commission tells them to do. But the Commission has done less to make their workings available for testing and checking than we did for a report on school performance. Our remit was small. Theirs is, potentially, the re-engineering of the entire economy. 

Blessed are they that have not seen the model, and yet have believed the results as described by their analysts in a series of webinars. 

Friday, 19 February 2021

Covid loans?

The excellent Richard Meade makes the case for Covid loans instead of wage subsidies. You can read the journal article on it, or his column over at The Conversation

Richard and I independently came up with the idea way back in March/April. I'd included it in our first comprehensive pandemic response policy paper, 26 March, with a bit more detail in a second short policy piece on 27 March. and he emailed me his two-pager on it about a week later. He'd not seen my version of it beforehand; his was better worked-up than mine was. Richard cites our version in his journal article. 

I'd liked it as a complement to the wage subsidy programme. We'd pitched a slightly different version of the wage subsidy scheme, based on Germany's short-time work scheme (see our short piece of 27 March), but given the time pressure it's impossible to fault the government for putting out the system it did put up. They made it work incredibly quickly, and the officials who got it through deserve medals. It avoided a pile of scarring. 

At the time, we all expected far worse employment outcomes, even with the wage subsidy programme. You can't just turn off tourism for what looked likely to be at least a year without having a pile of problems right? 

I'd figured that opening the student loans scheme to non-students, for borrowing up to a capped amount, with penalty interest applying ex post to any amount of borrowing that was above realised income losses, would likely work. 

There would be huge admin cost in trying to expand the student loans scheme; I'd thought this way of backloading a lot of the administration would make it more feasible. Easy approvals up front, but knowledge that penalty interest (recouped through the tax system, just like student loans) applies if you borrow more than your income loss would discipline. And it would mean that a lot of the admin would come at the end of the tax year, rather than while the bureaus were up to their eyeballs in trying to sort through the wage subsidy scheme. What those folks got done late March - just amazing. 

The big advantage in relying on something like extensions of the student loan scheme is that it automatically targets assistance to where it is needed, if you apply penalty interest for borrowing above realised income losses, and because it lets the government deliver more assistance in a hurry for any amount of cost it is willing to front in providing that assistance. If the government gives out a dollar as a grant, that costs a dollar plus the deadweight cost of taxation. If the government gives out a dollar as a loan, it gets some of that back. And that means it can loan out rather more than a dollar for the same expected one-dollar cost. 

Richard has it up as a substitute for the wage subsidy scheme. And I expect that's where policy has to turn over the longer haul: credit support rather than wage subsidies if we get more lockdowns, but government-funded Covid leave (now in place!) so workers and firms don't bear the costs of workers staying home while waiting on test results. 

Here's Richard:

Furthermore, since any firms and households borrowing against their own future incomes will ultimately be repaying their debt, COVID loans represent an asset on government balance sheets.

This offsets the extra liabilities governments take on by borrowing to finance these loans — something wage subsidies do not do. This increases the affordability of a loans-based approach from a government perspective (even allowing for defaults and subsidies implicit in student loan schemes).

Using illustrative data for New Zealand, my paper shows COVID loans are 14% cheaper than wage subsidies (and small business loans) in terms of their impact on net government debt.

More importantly, they are almost 2.5 times as effective in terms of the level of support they offer. And since 67% of the cost of COVID loans ultimately falls to those who make use of them (allowing for defaults and implicit subsidies), they place less of a burden on future taxpayers than deficit-funded wage subsidies.

The point of the wage subsidy scheme was to make sure firms could have a rapid reboot after lockdown. Making it the go-to in any outbreaks risks providing too much support in places like tourism where resources should shift to other uses. 

Hopefully this will all soon be irrelevant. But it is fun looking back through the email correspondence on it from back then. 

Thursday, 18 February 2021

MIQ and the America's Cup

I was curious how many spaces in New Zealand's Managed Isolation and Quarantine system were taken up by folks coming in for the America's Cup.

It looks to be unknowable, at least for now.

Immigration NZ has a list of people who were invited to apply for entry visas for the Cup, and another list of people who subsequently applied, and another list of those who were approved. There were 753 people whose entry visas were approved as of late January, with another 16 applications then under consideration. 

But we don't know how many rooms that takes up. A lot of those 753 will have been dependents of arriving America's Cup workers. Sometimes dependents travel with the worker. Sometimes they arrive at a different time. If they come in together, that would be one room. If they come in separately, it would be more than one. 

And some of those 753 might have decided not to bother coming through after having gotten their visas.

And some will have self-isolated on boats.

On the other side of MBIE, the MIQ people have lists showing which people were in which rooms and when.

But they've never looked to compare names across the two lists. It would seem likely to be pretty easy; you could even brute force the thing by just sorting the two different lists alphabetically and then eyeballing it. 

Anyway, Immigration told me their numbers, and I wrote on them in last week's Insights newsletter in our 3 slot - the one we reserve for satire or dark humor

The Aristocracy of Pull

If you want to get into New Zealand during the pandemic, it’s not that hard.

The government just needs to consider you to be a priority for a scarce managed isolation space.

Simplest is to be attached to a subsidised government programme. If a government programme falls over, there are real consequences. Political consequences. If private projects fail for want of critical staff, that is obviously less important.

Back in August, being a racetrack specialist mattered. The Provincial Growth Fund considered horse tracks critically important, so those workers got in.

But government racing priorities change with new coalitions, and the America’s Cup is on in Auckland. Its importance is as obvious as the dollars government and Council have thrown at it – about a quarter of a billion of them, regardless of business cases that struggled to show benefits in excess of cost even before Covid.

The argument for subsidising and prioritising the Cup is so powerful, they don’t need to explain it in a business case that stacks up, and they don’t need to revisit it when circumstances change.

Immigration New Zealand tells me that, as of the 21st of January, 753 people had been granted visas for travel into New Zealand for the Cup and a further 16 applicants were waiting on decisions. That is roughly equivalent to one in seven arrivals over one MIQ cycle – though MBIE has not yet told me how many MIQ spaces have been needed.

Being vital for projects involving racing or films could help your prospects for getting a space. A film about a horse race on a boat could tick all the right boxes.

Alternatively, if you worry that urgent need to enter the country might arise, cultivating a sympathetic media profile could be helpful. When 1.2 million Kiwis were born abroad and another million Kiwis live abroad, the number of legitimate urgent travel cases will always exceed capacity. Selection will be impossibly hard. Being featured on the news could help you while making the government’s selection job a bit easier.

Finally, if all else fails, being elected as a list MP might help, but it’s hard to tell. One MP took a spin on that roulette wheel in desperate circumstances and won, despite long odds. Lucky spins do happen.

Fortunately, scarce spaces in MIQ are allocated according to democratically chosen priorities. A bit of political pull is all you need to ease things along.

The Herald's Tom Dillane asked about the numbers; here's what I told him. 

Priority at the border has been and should be returning Kiwis. The government consequently reserves only a small number of spaces for workers essential to a wide variety of industry and business needs. Taking up hundreds of MIQ spaces for what is fundamentally a large government project – a boat race that would not be occurring in the absence of some quarter of a billion dollars in public funding – means fewer spaces are available for workers critically important for other projects. That is the real cost of these spaces: the opportunity lost when the government prioritises workers and dependents for its own projects over anyone else’s. We regularly hear not only of essential skilled workers unable to get a spot in MIQ, but also about critical skilled migrants in New Zealand deciding to leave because they have not seen their families for a year and have no hope of bringing them in. Prioritising the entry of hundreds of dependents of boat race workers, over workers already here who have not seen their families for ages, suggests the government’s priorities are out of whack.

Dillane's story is here. The headline isn't right - it's over 700 entries, with at least a couple of hundred of those being dependents. But Dillane wouldn't have written the headline. He goes through the numbers; I sent him through the OIA material. 

I still have a request in with MBIE asking whether they can compare across those lists. 

Wednesday, 17 February 2021

Test test test

My column over at Newsroom this week points out the fairly obvious.

The government can add daily saliva testing for everyone at the border to the existing testing regimen. 

If daily testing winds up proving the swab tests to be redundant, ditch the swab tests when we find that out. 

If it turns out they're both useful, keep both.

And in the very unlikely chance that the saliva tests wind up being redundant, they won't have cost much.

From the column:

Saliva-based PCR testing is a game-changer. It is at least as accurate as the swab tests currently used, but has several advantages over the swab test.

Because it relies only on saliva collection, it does not need scarce medical professionals to gather the samples. Availability of nursing staff is a substantial constraint in the system – and especially since the government shifted border nurses back onto the lower pay rates available through the district health boards.

Swab-testing can make people sneeze in reaction to having a swab stuck up their nose. Saliva testing does not come with that risk. Sneezes are far riskier than some drool on a disposable spoon.

And, finally, because providing the sample is far less inconvenient and invasive for the person being tested, and because it comes at only a fraction of the cost of the swab tests, it is rather more feasible to require it as part of a daily testing regimen.

Suppose the tests cost on the order of $35/pop (the Milne article doesn't say, but notes that large-scale runs would be less than the $40-$50 range advertised for smaller runs. And suppose there are 5k people in MIQ and 10k people in the border system. 15,000 * $35 = about half a million a day in test costs. So about fifteen million a month. 

If we run it and find, after a month, it was useless, we might have wasted twenty million. Less than a tenth of what the government decided to throw at some stupid boat race in Auckland. 

If we run it and find it to be useful, we get to spend maybe twenty million a month for substantial reductions in the risk of the virus getting out into the community and future lockdowns. 

And, just as obviously, daily testing gives us way more information about what the heck is going on. Are the cases that are currently found at Day 12 ones that just took a long time to incubate, or are they cases that were transmitted within MIQ? If we knew what was going on at every day along the way, we'd see it right? The PCR tests would make it way less likely that people could transmit within MIQ. They'd reduce the inherent riskiness of potentially infectious people sitting in hotel rooms not designed as quarantine facilities and spreading infection through ventilation systems. Right now, somebody could be infectious and asymptomatic in MIQ for days and days before being caught. Daily testing would fix that. 

My working hypothesis has been that MoH is just a wall of "Computer Says No" because the whole system's held together with bailer wire and they know they can't trust themselves to try to adjust anything. But some moves reduce the riskiness of the whole shambles. Daily testing in MIQ makes the whole thing less risky. 

Tuesday, 16 February 2021

Forests and intertemporal equilibrium

I'm a bit of an ETS-absolutist. Or at least a carbon-pricing absolutist, in a place the size of NZ. I think the Weitzman reasons for preferring a carbon tax to an ETS are second-order relative to political economy considerations, and any weight at all put on switching costs makes it ludicrous to want NZ to flip from an ETS to a carbon tax. 

But if a carbon tax were already in place, I'd be an absolutist about supporting the carbon tax.

There are fun and interesting arguments around tech-forcing, but NZ's not at the scale where that even really comes into play except maybe around ag biotech. So, prices are the way to go. 

One of the big things I hear from non-economists pushing back on it is around future generations. 

Here's my version of what I get a lot on Twitter, and as pushback from others. I hope I'm not providing a strawman treatment of it here. 

The ETS is fine in principle, but it is inequitable. In particular, it does not consider the interests of future generations. If we allowed the ETS to do its work without hindrance, a lot of the countryside would be planted in pine trees. Let's leave aside all of the biodiversity issues with that for now and just focus on one additional problem: intergenerational equity. If you plant trees now, that's a one-off. The land you've planted in trees has to stay in those trees forever as a carbon sink. If you harvest the trees, you have to re-plant them if you don't want to have the carbon released. That locks the land into forestry, unless some future person buys the ETS credits necessary as part of surrender obligations if they take the land out of forestry. So you're locking a future generation out of other uses of that land. Worse, you're only offsetting some current emissions. If gross emissions do not come down, then future generations either have to do more to reduce their gross emissions, or come up with new carbon sinks that could be even more difficult than forestry. This is intergenerationally inequitable. 

I hope that's an accurate representation of that particular concern.

I used to cover something similar in my 200-level policy course at Canterbury on intertemporal equilibrium. I don't think that this is at all commonly appreciated outside of economics, although the core paper establishing it is almost a century old.  

Imagine that you own a piece of land that's currently a sheep paddock but could be turned into a forestry block. There are only so many of these in New Zealand; it's a big country, but it is finite. When should you run a forestry conversion? 

Let's think through the owner's incentives and decision, and then relate it back to the equity issue that non-economists like to beat me up about. It isn't that economists haven't thought about it, it's rather that it's already kinda taken care of by the system.

If I keep the land as a paddock, I can keep earning paddock-related income from it. I can maintain that as long as I like. Carbon prices in the ETS are increasing over time. If I keep the land in sheep, eventually I'll be facing a carbon price for emissions from those sheep. That cost will be increasing, but is reasonably predictable. The current carbon price is about $38/tonne and is capped at $50/tonne plus 2% per year after that. Some price will attach to methane, but that's yet in flux. 

If I flip the land to forestry, I face a one-off conversion cost that won't change a lot over time, unless somebody figures out far less costly ways of planting forests. I will get a one-off set of ETS credits for converting the land. I can hold onto those credits, or sell them into the system. I can sell them into the system now, or I can sell them into the system at some later date. 

From now until time t when I convert, I get paddock net earnings per year. From time t onward, I get some lumber earnings as forests mature and get replanted, but I have to keep the land in forest. As soon as I fail to replant, I have to surrender credits comparable to the bundle of credits that I got when I planted. I have a bundle of ETS credits when I flip; if I sell them, I get the interest rate as return on that investment. If I hold them, ETS prices could rise or fall; I take on that risk and potential return. 

How should I decide when to flip to trees? What sorts of things will matter? This is a rather standard application of the old Hotelling problem. 

Howard Hotelling wanted to figure out time paths for prices for non-renewable resources. He showed that those futures prices should be coordinated through the interest rate. How? Let's take a simple example.

You own a paddock that's currently in sheep. You can keep running it in sheep, but methane prices will eventually hit your returns. If you flip it to forestry, you'll get a wad of ETS credits for the forestry conversion, and thereafter you get forestry returns from the land. If you ever flip out of forestry, you have to redeem a pile of ETS credits to do so.

Suppose you think that carbon prices are going to be far higher in future than they are now. If you believe that, you do not want to flip to forestry today, unless the returns from running a forestry block are comparable to the returns in sheep, leaving aside the ETS credits. Why leave aside the ETS credits? You can get those anytime. The opportunity to flip from sheep to forest will be there as long as you own the land. Once you convert, you get that bundle of credits. The credits don't do you any good, really, until you sell them. And, once you sell them, you can earn the interest rate on the money you get from selling them. 

Suppose that the returns in sheep are the same as the returns in forestry. The only reason to flip to forestry now, instead of later, would be if you figured that you could earn more by banking interest on your ETS credits than by holding the option until later. And that will only be the case if you're expecting the value of ETS credits to rise more slowly than the interest rate. 

And that's an equilibrating process. 

Whenever it seems like carbon prices will rise more rapidly than the interest rate, people will want to hold off on forestry conversions until later. If you wait until later, you get a more valuable bundle of ETS credits for the conversion. Fewer forestry conversions now means fewer ETS credits now, which pushes up current ETS prices relative to future ETS prices, which pushes that time path to look more like the interest rate.  

Whenever it seems like carbon prices will rise more slowly than the interest rate, people will want to flip to forestry now so they can sell the credits, invest the money, and earn interest. That process means you get more ETS credits now, which pushes down the ETS price now relative to later, which again makes the time path look more like the interest rate. 

Interest rates provide intertemporal coordination. They're a law-of-one-price for the future, holding constant riskiness and other stuff. 

Expectations about technological change will also matter. If you think new equipment will make future forestry conversions more cost effective than they are now, that will have you leaning against current conversions. If you think regulatory changes will make future conversions more difficult, you'll want to bring those conversions forward before councils ban them or before the Climate Commission mandates something weird about what kinds of trees you're allowed to plant. 

So back to our opening worry about relying on tree conversions. The underlying premise there is not just that a lot of land will be flipped to forests as a way of generating ETS credits, but that doing so imposes an uncounted cost on future generations. But if people expect carbon prices, and consequently the value of ETS credits, to be a lot higher in the future, they'll want to wait to exercise that option. They only will want to bring forward that option if current ETS prices seem high relative to futures prices, given interest rates. So the equilibrating process encourages preserving the option to plant later in scenarios when future scarcity in carbon permits seems high, but exercising the option to plant now in scenarios when future scarcity in carbon permits doesn't seem high. And that's exactly the set of incentives that you want, right?

Now there are ways that that can all blow up. If you expect that somebody's going to ban forestry conversions, then you can distort that intertemporal decision. That will make you want to convert sooner rather than later, because later might not be allowed. If you expect that the government will blow up the ETS in future, then you'll want to plant now, sell the credits, and hope that whatever mess of a policy we have in future doesn't hold you to the surrender obligations if you flip out of forestry. Both of those point to the importance of credible reliance on the ETS, and of not continually making noises about future bans on forestry conversions. Talking up future bans brings forward forestry conversions. 

If your objection is that investors can err about whether to make the conversion, because they're underestimating future ETS prices, the solution to that isn't banning conversions, it's making sure that the binding cap tightening over time is credible and advertising likely price paths consequent to those conversions. 

The usual critique is that people don't think enough about the future because they're too selfish, so they don't think about the opportunities they take away from future generations. But every agent in the Hotelling setup is a pure profit maximiser. They only care about making money. And holding the option to convert later can be better than converting now, if you think that ETS prices in future will be a lot higher, and depending on the interest rate. 

And if you think that everybody else is wrong about the likely price path for NZU and that ETS credits are going to be way more expensive in future, you could go and trade in ETS credits to make a lot of money out of it. 

I worry a lot about policy initiatives where the underlying problem definition requires that markets somehow can't handle long term investment decisions. It winds up with an arbitrariness - it gets rolled out whenever a market outcome doesn't correspond to some bureaucrat or politician's vision of the good.

Anyway, bottom lines: 

  • intertemporal prices already provide incentive to consider the value of maintaining the option to convert to forestry later; 
  • these decisions wind up being guided by and coordinated by overall interest rates;
  • expectations about tech and cost paths can also matter, but all else equal it's going to be expectations around future ETS prices and interest rates;
  • you can break a lot of that setup by threatening to abolish the option to convert in future, as that will have folks rush to convert earlier. And if you're threatening to abolish that option, it's probably because you don't want them to be converting now, right? 

Tuesday, 9 February 2021

Mismanagement of the highest order

Josh Gans's newsletter on Covid, testing, and vaccination continues to be excellent. 

Here's the latest from Josh, on failures in the Australian system that led to the most recent outbreak there.

However, let’s look at the testing. First of all, quarantined travellers are tested just twice (usually a few days after they arrive and a few days before they leave). Why aren’t they screened daily using rapid antigen screens? It costs very little and can ensure that cases are picked up quickly and extra precautions are put in place so the workers aren’t exposed.

Second, what about the employees? The workers are tested but not often. This one had symptoms before he got a test. From here the plot thickens. From WA Today:
Daily testing of hotel quarantine workers, which could have identified WA’s first community transmission case of COVID-19 in nearly 10 months several days earlier, was only rolled out on Friday, Premier Mark McGowan has revealed.

Following a National Cabinet meeting on January 8 all states and territories agreed to roll out daily tests of hotel quarantine workers after a Brisbane hotel worker was infected with the highly transmissible UK COVID-19 variant.

Victoria had already been doing daily saliva testing and Queensland enacted their testing regime three days later but Mr McGowan said on Sunday that WA had only just finished testing its new regime at the Novotel hotel last week.

“We put in place the saliva testing as quickly as we could using the health department and appropriate protocols, unfortunately, it didn’t start until late this week,” he said.

“It’s not easy, it is a big exercise to roll out.”
Well, that really cheeses me off. Yes, it is hard to get screening in place (I know it more than anyone). But it is not hard to do it in a rough and ready way while you work out best practices. So in WA’s case, they missed it by that much. But if this is your weakest link why did the Australian government wait until this year to decide to roll out daily testing? It boggles the mind. It is mismanagement of the highest order.
If it's mismanagement of the highest order that Australia only just recently started daily testing of workers in the border system, how should we describe a country that hasn't even considered implementing it yet?

Rapid antigen tests would be a fine addition, but the University of Illinois has had a saliva-based PCR solution in place for months, and it's at least as accurate as the nasal swab tests. 

It would be entirely feasible to add daily testing here for everyone in the border system, including those in MIQ. 

Back to Josh on Australia:
Nonetheless, this time around, you might be tempted to say, “it is what it is.” But it is February. A vaccine has been available since December. The Australian government has not rushed to procure vaccines. Nor should they necessarily do so as Australia is at Covid-Zero and doesn’t need mass vaccination as urgently as other places. However, Covid-Zero is a tough policy to maintain. You cannot tell me that it wouldn’t be possible to pay top dollar and procure vaccines for quarantine workers who, despite doing all you can, can still get infected and pass it through to the community. To be sure, if you have regular screening of everyone, there is a good chance leakages can be caught. But, there is a vaccine. The Australian government has sensibly placed these workers at the very front of the queue to get vaccines when they eventually arrive. But these workers and the country need them to have the best vaccines right now! The Australian government needs to be held to account for lockdowns and restrictions that occur that could have been mitigated by obvious, economically sensible actions.

Rather similar here.

New Zealand should have used advance market commitment orders to help secure an earlier spot in the queue while providing the funding that helps vaccine manufacturers scale up for everyone. But from where we are now, whatever your arguments around holding off on mass population vaccination in a spot where the virus is not prevalent, surely there should have been provision for enough early doses for workers in the border system. 

Wednesday, 20 January 2021

Afternoon roundup

 The afternoon's closing of the browser tabs brings the following worthies:

  • Superb news! The police have taken an operational decision not to waste resource sending helicopters out looking for cannabis plants. Or at least National Headquarters isn't going to resource it any longer. Lots of things are illegal; police (rightly) have limited budgets and so have to make decisions about where to focus their efforts. Flying around in helicopters on gardening operations makes far less sense than putting resource into preventing crimes that actually have victims. 

  • Audrey Young on the slow pace of getting the vaccine out.

  • In any other circumstance, I'd be a bit nervous about Otago Public Health recommendations around smoking policy. But I'm in complete agreement with Baker/Wilson on this one. Shared spaces in MIQ seem crazy risky, and especially so with more contagious forms of the virus coming through. I'm also with them, sadly, on the desirability of reducing intake from risky places - at least until we can get to far more frequent testing of everyone in the border system. It totally can be done. Daily testing of all border staff, through daily saliva-based PCR tests, would mean any infection of MIQ staff would be almost certainly caught before it could turn into community spread. Oh, and Michael Baker also wants rapid testing at the airport pre-departure. It's been feasible for a long time. We could still do it. The Abbot rapid antigen tests are cheap and could be rolled out for use at the departure gate. The government could buy thousands of them, send them out with every outbound flight so that they can be used at the gate pre-departure at every gate departing to New Zealand. Or pick a different test if Baker prefers a different one. At least now that Labour has signaled support for pre-departure testing, folks can talk about it without being attacked by a pile of Labour twitter partisans. 

  • Luke Malpass thinks we can still be aiming for a broader travel bubble, and has this encouraging news about Auckland airport.
    Once the Australian bubble is opened, it is understood that the international terminal will basically become green lane only, while “red lane” flights will land at a separate building and passengers will be bussed to be processed for quarantine, or to a separate area if transferring to another flight. The airport will need two weeks’ notice to get this working.

    The other big issue, which Covid-19 Response Minister Chris Hipkins has repeatedly raised in public, is how to manage repatriation of both New Zealanders and Australians in the case of a big lockdown. Here, Australia’s federal system should make things easier.

    Should there be an outbreak in a certain part of Australia, say Queensland, the other states would likely close their borders to Queensland. Meanwhile, the bubble could continue in the other states. In addition to this, one option being batted about is getting travellers to Australia to sign a form on departure acknowledging that, in the event of an outbreak, they may have to hunker down where they are for 14 days, should exit flights not be able to be arranged. Vice versa for Australians on these shores. 
    All of this has been entirely obvious as the way forward, for months. Government moves slow.

  • This is fun. SocialBubble provides you with typical twitter feeds as seen by people of various ideological persuasions. You can look at Twitter as though you were a socialist, leftist, progressive, liberal, centrist, moderate, conservative, right-winger, or alt-right person. FWIW, my twitter feed looks closest to the one they identify as centrist. Or at least I recognise the folks in that feed, and a lot of them are the ones I also follow: Neoliberal, Noah Smith, Conrad Hackett, for example. I recognise a lot of the feeds in Socialist too, like Sanders and Jacobin and Existential Comics, but choose not to follow them. I had followed Existential Comics because the comics are often superb, but the twitter feed is just too tedious. 

  • Eden Park will be allowed to operate as a stadium. This is good, not least because it reduces the prospects of anybody throwing money at a new downtown waterfront stadium. 

  • Tyler Cowen asks that you start from your estimates of labour demand elasticities and be consistent about things.

Tuesday, 19 January 2021


Getting to Browser Tab Zero so I can reboot the computer is awfully hard when the one open tab is a Table of Contents for the Journal of Environmental Economics and Management, and every issue has more stuff I want to read.

A few highlights:

Vaccines are cheap

Israel chose to pay a bit over the odds for the Pfizer vaccine to get earlier access. 

Here's The Times of Israel from 16 November.

American government will be charged $39 for each two-shot dose, and the European bloc even less, but Jerusalem said to agree to pay $56.

Israel has now vaccinated more than 80% of their elderly population and is getting the second doses into arms. 

I do not know what New Zealand is paying. But suppose we're paying the same as the US. For only $85 million USD more, or just under $120m NZD more, we might have also had early access. 

$120m sounds like a lot in normal times. But if it meant that we could have everyone vaccinated from mid-year, instead of starting to roll-out vaccination from mid-year, we'd be able to open the border, right? Surely being able to open the border half a year earlier is worth at least $120 million. Like, if someone had offered the government a quarter of a billion to close the border for half a year, no sane government would have taken that offer.

I've heard a lot of rationalisations on this one. Some would argue that we'd have bid it away from places that needed it more. But done properly, through advance market commitment orders, it would also have helped to fund the capacity building that allows more production for everyone. 

Ex ante, it would be impossible to know which vaccine would be best. We'd have had to have paid above the odds for each of several vaccines. But even that is still incredibly cheap as compared to having to keep the border closed for longer than necessary. Even if it added half a billion all-up to the cost of vaccination, the costs of lockdowns are in the billions and so too are the costs of a closed border. If MIQ is about $6k per stay and handles 10k returning Kiwis a month, then that all by itself is $60m a month. Vaccines are cheap, even paying well over the odds. 

Meanwhile, the new online vaccination register is still under development.

Monday, 18 January 2021

Aristocracy of Pull continues

RNZ reports on continued arbitrariness on decisions at the border. 

British comedian Russell Howard is about to tour New Zealand and other acts allowed in through managed isolation this summer include drag queen RuPaul and musicians at Northern Bass in Mangawhai and the Bay Dreams festival.

The vice-president of the Promoters Association, Gray Bartlett, said despite being an approved promoter with Immigration New Zealand, he was offered no explanation on why acts such as the Music of Cream and American speaker Michael Franzese were turned down.

No-one had spelt out what the criteria were for approval nor who was making the decisions.

"What I really don't like is where governments begin to start with favouritism and choosing who they like, or getting people to choose who they like to come in. That's not right. And and it can be done quite easily in our business because they can formulate reasons why someone may be important to have here. But in reality, it doesn't stand the sniff test, I'm afraid. And we can prove that with some of our applications."

One project he was involved with would be employing about a 100 local workers if their application was approved, but he said he did not want to put that at risk by talking about it.

"I don't want to affect the lives of these Oscar winners and Grammy winners who want to come out here in a month or so's time to give great enhancement to New Zealand as a country, but it's all a game a favourites, it's who you know in government, and this particular government is particularly poor at this particular task."

It is an impossible task. Nobody in government can really know what is the most highly valued use for the small number of MIQ spaces held for non-residents' entry. Is is an entertainer, or an engineer? If it's an entertainer, which one? How can they decide?

An Immigration New Zealand spokesperson said all requests for a border exception for individuals in the arts and entertainment industry are assessed against the same 'other critical worker' criteria as any other request for a worker as set out in immigration instructions.

The criteria is based on whether the skills or experience the person has are readily obtainable in New Zealand or whether the worker is undertaking a time-critical role in specific areas, they said.

"INZ can confirm that since 18 June 2020, 66 requests have been received relating to the arts and entertainment industry, which includes all artists and festival performers but excludes film and television. Of those, 39 requests have been approved, 23 requests have been declined, and 4 requests are still being assessed.

"In regards to the request relating to Music of Cream, based on the information provided INZ was not satisfied that the performers met the criteria, particularly in regards to demonstrating that the role was time-critical to work which would bring significant benefit to the national or regional economy as there was limited information provided to demonstrate what the economic benefit would be. INZ has engaged with Mr Bartlett to advise him of his options following the decline in August, however INZ is yet to receive any new information or a request for reconsideration. Should Mr Bartlett request a reconsideration or make a new request, INZ would consider this."

It would be surprising to me if a Cream cover band were the highest valued use of MIQ spaces, but how would I know? How would INZ know? 

The government runs essential workers spaces generally on something like cost-recovery. But cost-recovery is nonsense in this kind of environment. The cost of the rooms isn't whatever the government is spending on them. The cost rather is the value of the room in its highest-valued use - the opportunity cost of the room. 

In normal scenarios, the numbers wind up being about the same thing. In competitive markets, the price is the cost is the opportunity cost and it all works out. MIQ rooms aren't like that. There are far fewer of them available, at the administrative charge imposed for use, than there is demand for them. 

The only way of really knowing the highest valued use for those rooms is to put them to auction - the same way that we find out who the highest valued user of a house is. 

In July, we'd suggested what I still think would be a far better way of allocating MIQ spaces. 

Allow MIQ facilities to charge inbound visitors directly. The government would charge the facilities for the services provided by government workers - nurses, security, and everyone else. Instead of allocating rooms to returning Kiwis free of charge, with those Kiwis then having incentive to make multiple bookings just in case, provide eligible returnees with a voucher that could be applied toward the cost of a stay in MIQ. The value of the voucher would be equivalent to the cost of a stay in a basic MIQ facility during an off-peak time. If returning Kiwis wanted to stay at a 5-star facility instead of a more basic one, they could pay the difference. And the government could impose a surcharge on stays by non-Kiwis to help fund the cost of the vouchers provided to returning Kiwis. 

What's changed since July? The virus has become far more contagious. The epidemiologists at Otago, like Nick Wilson, have been recommending pushing MIQ back to the source country. They're right. Setting MIQ facilities overseas, and having visitors stay there before flying here, reduces the likelihood of cases making it here. If we do wind up with facilities set overseas, running the kind of set-up I've suggested would make it easier to allocate spaces and would help in funding the thing. Non-Kiwis coming in would wind up funding the MIQ stays of returning Kiwis. And it's easier to scale that system abroad than it is to scale it here. 

If we can ever get rapid PCR saliva tests available for MIQ staff and for returnees, ramping up to daily testing of MIQ staff and returnees could substantially reduce the risk inherent in the new variants of the virus. A more contagious variant is more likely to be transmitted to a border worker. If a border worker gets it, it's more likely to be transmitted onwards to someone in the community. Daily testing would mean that it just couldn't get very far. The test would be far more likely to catch cases before others were infected. Without either far more frequent testing of border workers, or MIQ being pushed back to the source countries, we're taking an awful lot of risk. 

Friday, 15 January 2021

Mom's Time

Dan Hamermesh always takes on the fun projects. 

A decade ago, he did a pile of work looking at returns to beauty

Now, he's looking at time-diary data. Here's the abstract from his latest at NBER

Using time-diary data from the U.S. and six wealthy European countries, I demonstrate that non-partnered mothers spend slightly less time performing childcare, but much less time in other household activities than partnered mothers. Unpartnered mothers’ total work time—paid work and household production—is slightly less than partnered women’s. In the U.S. but not elsewhere they watch more television and engage in fewer other leisure activities. These differences are independent of any differences in age, race/ethnicity, ages and numbers of children, and household incomes. Non-partnered mothers feel slightly more pressured for time and much less satisfied with their lives. Analyses using the NLSY79 show that mothers whose partners left the home in the past two years became more depressed than those whose marriages remained intact. Coupled with evidence that husbands spend substantial time in childcare and with their children, the results suggest that children of non-partnered mothers receive much less parental care—perhaps 40 percent less—than other children; and most of what they receive is from mothers who are less satisfied with their lives.

After adjustments for education, age, children's age and so on, married mothers spend about 40 minutes more per day in household production than do similar other mothers. But most of that time difference is in non-childcare activities; married mothers spend about 6% more time providing childcare. But their husbands spend a lot of time in childcare as well. Consequently:

Together with the slight amount of additional time in childcare by married women, this suggests that children of married mothers receive over 3 hours per day of care from their parents, compared to about 1-1/2 hours per day that children receive from their single mothers.

The biggest gaps are for children aged 3 to 12; 

Widows differ substantially from other groups of mothers without a spouse in the home:

One sub-group of non-married mothers uses time differently from the others—widows. They account for only four percent of non-married mothers ages 25-54 in the sample, but they show statistically significant differences in the time they spend on various activities compared to other non-married mothers. They exhibit much more home production time than others; indeed, they differ only minutely (five minutes less per day) in this dimension from married mothers. They make up for this extra time by working and sleeping less than the other non-married mothers. Overall, except in their leisure time widows behave more like women with a husband present than do divorcees, separated mothers or those whose spouse is absent.

And marital status has substantial effects on self-reported life satisfaction:

The estimates of the impact of marital status in regressions describing this indicator of life satisfaction are shown in the bottom panel of Table 7. For all four countries the same vectors of covariates that have been used throughout are included. If there is no spouse/partner in the household, the mother is significantly less satisfied with her life—by 16, 43, 23 and 16 percentage points in the U.S., France, U.K. and Italy respectively. There is a very large difference in this measure by partnership status in all four countries. (The effects of being non-partnered are even more significant statistically in ordered probits describing the entire range of responses to the questions about life satisfaction.22) While feeling only slightly more rushed for time than partnered mothers, non-partnered mothers are much less likely to be satisfied with their lives. This difference is essentially unrelated to how they allocate their time across different activities—the results hardly change if the mother’s time allocation is included in the estimating equations.

Note that while the Table 7 estimates adjust for the effects of education (among other covariates), income doesn't seem to be included.  

Hamermesh concludes:

The results suggest that children of non-partnered mothers not only receive less parental time than others. The attention that they do obtain is from mothers who feel more stressed for time and who are less satisfied with their lives, a concatenation of time and possible interest that may on average disadvantage their children even more. Overall, our findings imply the need for even more attention and concern to the difficulties facing children in single-parent households. With non-married mothers in the U.S. being disproportionately less-educated and more likely to be from minority groups than married mothers, this conclusion takes on special importance.

Hamermesh doesn't draw policy conclusions; I expect folks will form them based on their prior preferences. 

Some conservatives may take it as an argument for strengthening families, discouraging out-of-wedlock births, and encouraging marriage counselling over rapid divorce. 

Others might take it as argument for very substantial investments either at school or before school to try to make up for the hour-and-a-half difference in parental time per day. There will be a substantial cumulative difference by the time a child enters school. 

It may not be particularly controversial to suggest that the government's proposed Equity Index, which would replace school decile funding formulas, include marital status as one of the variables. I don't think it's currently in there.

Tuesday, 12 January 2021

Cleansing the Twitters

I'm less than convinced by arguments that platforms like Twitter should be subject to common carrier regulation preventing them from being able to decide who to keep on as clients of their free services, and who they would not like to serve. It's much easier to create competition for the network in this case than it was for Telecom in the 1950s.

There has been some concern about the coordinated action by a lot of platforms against a set of conservative platform users in the US. It has been taken as suggesting some leftist conspiracy against right-wing views. 

There are, of course, multiple hypotheses consistent with the available data.

Here are some of them, with some very thumb-suck probabilities.

  1. The terms of service always barred what Trump et al have been up to. But the platforms have been cowed by fear of sanction by Trump’s executive branch, until the combination of Trump supporting and encouraging insurrection AND Trump’s having little time to retaliate against them meant they could finally enforce said terms of service. But this hypothesis doesn’t explain coordinated action across platforms where Trump and his people weren’t as active. (1 chance in 8)

  2. Intelligence agencies warned the combination of platforms that Trump’s people (Q crazies, Covid-crazies, etc) were using the platforms to coordinate insurrection, and that Trump’s tweets would easily help coordinate those. The platforms listened and blocked out of fear of criminal liability for supporting terrorism. (1 chance in 8)

  3. As above, but not out of fear of liability. Instead, out of existential worry. If America goes full-despot, the platforms can no longer exist. Barring crazy people from the platforms to prevent that reduces that risk. Both 2 & 3 explain coordinated action. (1 chance in 10)

  4. They’re pandering to an incoming Democratic administration with a tipped Senate, in hopes of more sympathetic regulatory treatment. The GOP losing in Georgia provided the trigger that coordinated action across platforms, not any intelligence tip-offs. (1 chance in 8)

  5. They were always a bunch of commies just looking for an opportunity to knock out conservatives; the timing follows from (1) but the ToS bit is just an excuse. (1 chance in 40)

  6. Other explanations (1 chance in 2) 
What's your preferred hypothesis? Have I missed any obvious ones?

New Zealand's Privacy Commissioner also, bizarrely, weighed in with his worries about the platforms being able to knock out those users without some regulatory democratic oversight. If hypothesis 1, 2, or 3 is what's going on, a Trump appointee overseeing the administrative agency regulating the platforms would not have helped. If it's 4, then that also doesn't help. Maybe it would have guarded against 5.