Basic income, again

This week's column for the Stuff papers covered the excellent new US work testing the effects of a UBI. 

From November 2020, 3000 low-income people were randomly assigned into two groups for three years. One thousand people each received $1000 per month in unconditional funds for three years. Two thousand people each received $50 per month.

Both groups filled in detailed surveys on how they spent their time, on their purchases, their health experience and more. Participants had blood tests to check health outcomes. Government administrative records were combined with the survey data to provide more detail.

The UBI amounted to about a 40% increase in recipients’ income – large enough to matter.

The researchers pre-registered their study design to guard against, well, fiddling. If you have dozens of potential outcomes that might be affected by the cash transfers, there are many ways for studies to accidentally find effects that are not there, and even more ways to put a thumb on the scales. Trial pre-registration says in advance how statistical testing will be run.

...

What did the experiment find?

Households receiving the UBI were able to enjoy more leisure than other households. Leisure is good and should not be underrated. But that was the largest actual effect. Their overall earnings, including the transfer, were higher – but labour force participation and hours worked fell. Households receiving the UBI earned about $5000 less per year than the control group, not counting the transfer, or about $6000 more than the control group when counting the transfer.

People receiving the transfer spent more time unemployed if they became unemployed, but more time for job search did not help. There was no effect on job quality – and a good study design meant they could rule out even small effects.

Remember that the funding for the study came from donors. Any real-world UBI would need to be funded by taxation that would have its own pernicious effects on work incentives. And a permanent UBI would have larger effects on work choices than a three-year programme.

What about health? In the short term, people receiving the transfer enjoyed reduced stress and greater food security. But those effects quickly faded. There was increased uptake of health services and some healthy behaviours. We can be more confident that a UBI benefit would not be blown on drugs and alcohol. But the study found not even minimal effects on physical health. And initial improvements in mental health disappeared after the first year.

While poverty is certainly associated with worse health outcomes, very large and sustained cash transfers did not improve health. Worth remembering when reading the next public health study asserting that more income redistribution would improve health outcomes.

They put a heck of a lot of work into this trial and its evaluation. To the extent that they got a law change so that the payments wouldn't be considered taxable or affect eligibility for other benefits - so they knew it was the same $1000/month increase for everybody. 

The disemployment effects were a bit larger than I'd expected, so I've updated my expectations on that. My priors on health effects were reinforced; in richer countries, the income-health gradient is going to be an artefact of an underlying correlate of both.

I was a bit more surprised by zero effects on job quality. I'd put some weight on that less time-pressure to match with a new job when unemployed could yield better job matches. But the 1.1 extra months' duration of unemployment spells relative to the control group cashed out into precisely estimated nil effects on a whole big range of job quality items. 

You can catch a twitter thread by Eva Vivalt, of the researchers on it, here.

But UBI-stans seems to be some of the stranger beasts out there. I mean, look at this. 

Eric Crampton must have been reading another study from the report he describes in his hostile critique of basic income (‘Putting a UBI to the test’, July 29). The US cash transfer project he eulogised was not a test of basic income.

There have been over 100 experiments which show positive results. The one he cites is not one of them. By definition, a basic income is a modest amount paid regularly to all usual residents, individually, without means-tests or behavioural conditions, regardless of income, gender, marital status or work status.

The study Crampton cites does not pretend to respect that definition. Several limitations make the results irrelevant for assessing basic income.

It was a means-tested benefit paid to 1000 individuals spread across 19 counties in Texas and Illinois, about 50 per county. That is hardly universal. The individuals self-selected. They only received the cash if they could prove they were poor. It was only paid to individuals aged 20-40. It was only paid to one individual per household, if nobody in the household was receiving disability benefits and they were not in publicly-subsidised housing. It deliberately over-sampled those from minority groups. Those features invalidate any claim to randomness, unconditionality or universality.

Ok. Where to start. If you want to run a UBI trial, you can try saturation where you enroll an entire town. In that case you get the second-round effects from the spending of UBI income and the like and interactions among people who all get the UBI, but you really need to set the thing so it's funded by the community receiving the benefits so you get the effects of imposing the taxes necessary to pay for the thing. That would get you more of a total effect. 

Or, you can do what this group did and test just the effects of the transfer on outcomes for those receiving it, leaving out both potential community-wide benefits from everyone getting it (whether social stuff or spending effects) but also the incentive effects of the taxes necessary to pay for it. 

They targeted the group whose responses are of most interest for this policy agenda: lower-income people of prime working age. Randomisation was achieved by randomly putting people from that group either into the treatment or control groups. Payment when assigned to the treatment group was unconditional. You didn't have to do anything to get the money. If you filled in the time use surveys you got extra payments for that - as did people in the control group. And payment was universal within the treatment group. 

I suppose you could take the position that the only real trial of a UBI is to actually implement it across an entire community, but we quickly get into No True Scotsman issues. Did the trial encourage inward or outbound migration? Well, gonna have to apply it to the whole country aren't we for it to be a real trial. 

I think we can pretty confidently say that a UBI providing an after-tax transfer on the order of 40% of non-UBI income for low income people will have effects like the ones found in this experiment, but very likely with larger labour market effects both because of the permanent income hypothesis (a permanent transfer will have larger effects than a known-to-be-time-limited trial), and because of the effects of tax rates required to fund the transfer. And that there won't be the hoped-for improvements in health or job quality. But that there won't be increases in problem drinking because of it, and that people choosing more leisure probably consider themselves better off. Anyway, here's one of the authors on this:

Since this study captures the pure income effect of giving cash, without any of the substitution effects that might occur with a means-tested program, it tells us how much people prefer more leisure on the margin, even without disincentives to work. 37/

— Eva Vivalt (@evavivalt) July 22, 2024

Leisure is a normal good, people. 

The experiment began at the height of Covid, so it is surely a marvel that a cash transfer only resulted in a drop of just over an hour a week in paid labour.

Covid affected both treatment and control groups. The control group wasn't in some no-Covid place.  Maybe the argument is that the transfers gave people the ability to better hide from Covid if they wanted? Remember too though that there were all kinds of work-from-home and other changes that will have hit both groups, as well as Covid payments. And that the trial ran for three years, finishing in the second half of 2023. 

Moving towards a basic income for every resident citizen is a matter of common justice, freedom and basic security. Economists have shown it is affordable without raising income taxes.

Guy Standing, Co-president, Basic Income Earth Network; Professorial Research Fellow, SOAS University of London

The first sentence here is a values assertion. 

The second one, well, Treasury in NZ showed that we'd need a flat income tax of more than 50% to fund it and it still wouldn't be enough to replace all other benefits. A small one could be done without raising taxes - say if you abolished NZ super and split that money equally regardless of age. The transfer would be pretty small though. Or maybe he's expecting it to be funded out of some other tax he's not mentioning.  

The basic tradeoffs in a UBI are the same as I'd pointed out in the Spinoff ages back. We now have some better numbers on the likely effects. 

Read the NBER papers for yourself though. They're here. 

Also fun to compare the study's results, with what people had hoped would be the results. I guess those whose expectations were most dashed and who were most committed to specific beliefs about outcomes are most upset about it? 

I asked GPT4-o to complete the UBI experiment abstract based on the first part: pic.twitter.com/shLDQDYMeG

— John Horton (@johnjhorton) July 22, 2024

But you can also check Table 16 of the paper comparing expert predictions of what the effects would be and the eventual effects - which is just so neat as method.

I was on with RNZ's The Panel last night on this one. Would have been far better for them to have had on one of the study's authors, but I'd written a column on it and I'm local. 

So I went through what I'd gotten from it. 

And as I got up to leave, the host (subbing in for Wallace Chapman) read out a text that came through. 

"We need to be reminded that the New Zealand Initiative is part of the Business Roundtable. It supports its own conservative right-wing ideology." 

Classy as always RNZ. 

RNZ remains on-form

There are a few basic bits of reality that I'd hope we could agree on.

Minister Costello has set a lower excise rate for heated tobacco products as a bit of a trial to see whether it proves successful in encouraging remaining smokers to flip to something less harmful.

Even without any change in tobacco excise, a smoker shifting to an HTP will result in a drop in tobacco excise. Or at least I'm pretty sure. I'm pretty sure that a heated tobacco stick contains less tobacco than a cigarette does. And I'm pretty sure that heated tobacco draws the lower excise rate that applies to cigarillos and the like. The combination of the two means less excise in a heated tobacco stick than in a cigarette.

If the trial proves successful and a lot of smokers shift from cigarettes to heated tobacco, tobacco excise revenue will drop. The more successful the trial is in encouraging shifts from smoked tobacco to heated tobacco, the bigger the drop in excise revenue. If a lot of current vapers or non-smokers take up heated tobacco, excise revenue will increase on that margin.

While Philip Morris is a dominant supplier of heated tobacco products, heated tobacco competes with smoked tobacco and with vaping. So PMI does not face a vertical demand curve for its product. And even if you model them as a monopolist in the market for heated tobacco, a monopolist will pass through at least some of an excise reduction. 

This is basic Econ 1 stuff. The monopolist sets marginal revenue equal to marginal cost looking across the quantity axis, then traces up to the demand curve to find price. That's the price that maximizes profit. If costs drop, the profit-maximizing price drops too. It's just math. 

I hope that all of this is completely uncontroversial. 

Some mornings I am more annoyed than usual about being forced to pay for Radio New Zealand. This morning was one of those mornings.

Guyon Espiner had his latest update on Minister Costello's reduction in heated tobacco excise.

Let's have a look here.

Govt set aside $216m to pay for heated tobacco product tax cuts

Ok, so RNZ's main focus is going to be on the potential excise losses. 

Now we know, from basic facts of the world, that there can only be substantial reductions in tobacco excise if the policy is incredibly successful in reducing smoking rates.

Let's continue.

The government has agreed to set aside $216 million it may need to pay for tax cuts for heated tobacco products (HTPs).

RNZ reported earlier this month that Associate Health Minister Casey Costello - who is also Customs Minister - had implemented a 50 percent cut to the excise tax on HTPs, where the tobacco is heated to a vapour rather than burned.

Costello's office had not publicly disclosed how much that would cost the government but a Cabinet paper, released without fanfare on the Health Ministry's website, shows Cabinet agreed in May to set aside $216 million as a contingency fund to cover the estimated lost revenue.

The excise tax cut is something tobacco giant Philip Morris has lobbied for in the past. Its IQOS product is a dominant player in the New Zealand HTP market.

The Cabinet paper, signed off by New Zealand First MP Costello, showed it was not even clear whether the tax break would be passed on to consumers.

"Because this product currently has a monopoly market in New Zealand, the extent to which a reduction in excise duty on HTPs would be passed on to consumers via lower retail prices is unclear," the paper noted.

I've highlighted a relevant bit here. Let's look at this section of the Cabinet Paper.

The Cabinet Paper isn't saying that there might be no pass-through. It's saying that they don't know whether there will be one-third pass-through or 100% pass-through. 

Officials could perhaps have been clearer that even a complete monopolist facing a downward-sloping demand curve is going to pass through at least some of a tax reduction, but they could be forgiven for thinking that it was awfully clear in context. They note one study finding a 31% pass-through rate, and that officials have advised to assume 100% pass-through. So the extent of pass-through is likely going to be somewhere between those. 

But RNZ put this as a question of whether there would be pass-through. Not the extent of it. 

If there is zero pass-through, and again this is just math, and it won't happen because even a monopolist will respond to a cost decrease by reducing price, there cannot be more than negligible effect on excise revenue. Why? The price of heated tobacco will not have dropped relative to smoked tobacco. It will not have driven any flipping from smoked tobacco to heated tobacco. Philip Morris would get a transfer equivalent to half of what they currently pay in excise. Excise from heated tobacco was just under $6 million in 2023, so the policy would transfer $3m to Philip Morris. $216 million is right out. 

But RNZ sets the framing to have people expect a $216 million tax cut benefitting PMI that might have no pass-through to consumers. 

Let's continue, again.

Costello declined an interview with RNZ and her office did not address questions about whether that monopoly position referred to Philip Morris.

In a statement the minister did say that she expected the industry to reduce the cost of its products.

"That means I'm expecting the excise reduction to pass to consumers, this is what we were advised would happen by officials and it is something we will also be monitoring," she said.

She also said she did not expect the cost to the government to be "anywhere close to what was modelled", as the tax collected on HTPs was only $3.62 million in 2022 and $5.97 million in 2023.

"Officials noted there is a lot of uncertainty around the modelling and fiscal impact because it was based on the very rapid increase in HTP use that happened in Japan, where vapes were unavailable."

Philip Morris did not respond to RNZ's questions.

Good context from the Minister here, and good that RNZ reported it. 

Unfortunate that they didn't think through what those numbers mean. If current excise from heated tobacco is only on the order of $6m, what would have to happen in the real world for a drop in excise on heated tobacco to result in a $216m excise loss? It would have to mean that there was substantial pass-through of the excise reduction and consequently substantial switching to heated tobacco, which means substantial reductions in exposure to the bad things that come out of combusted cigarettes. 

Another fun bit from the Cabinet paper. Guess which line of this paragraph RNZ chooses to emphasize? Hint: it's not the one saying that even the FDA recognizes that heated tobacco means less exposure to harmful chemicals. 

As always, RNZ's reporting on tobacco needs health warnings. 

Another pitch by the pharmacy guild

Recall that Chemist Warehouse found a structure that let them operate in New Zealand despite pharmacy guild regulations that had seemed aimed to block such entry. 

The pharmacists are having another tilt at it.

In Australia, a pharmacy that wanted to fill prescriptions could not set up within 200m, 1.5km or 10km of an existing pharmacy depending on whether it was in a shopping centre, suburb, or town.

Community pharmacists were calling for similar regulations here, warning if nothing changes, the community bond that came from knowing every customer by name could be a thing of the past.

Right.

"Hello, I am the owner of the local pub, where everybody knows your name. I worry about the potential entry of a new pub in a 10km radius, not because of its effect on my profitability!, but because it would erode the community bond that comes of having a single pub where everyone goes and where everyone knows your name. Please ban any such entry. Thank you."

And repeat for the local dentist. Or the local greengrocer. Heck, even the local bank. Could roll this one out for all kinds of things.  

Interchange fees

A few years ago, MBIE ran an inquiry into credit card interchange fees.

Most of the analysis seemed predicated on an assumption that retailers could neither impose surcharges for card transactions nor avoid accepting cards.

So I started taking pictures of EFTPOS terminals with tape over the credit card button or with obvious signs noting credit card surcharges to accompany my submission on it. The MBIE paper seemed to take "Well, anything's possible in two-sided markets so we should regulate" approach. 

Now it's ComCom that's proposing to regulate credit card interchange fees. 

But they did commission a couple papers. 

And this seems a key bit:

Moreover, it is theoretically argued that if the no-surcharge rule is lifted, interchange fee regulation is harmful for total welfare. Regulatory attention should in this case shift to merchants, rather than focusing on card networks. If surcharging is to be allowed, the optimal cap is equal to the merchant fee minus the merchant’s convenience benefit from card payments. In other words, the merchant should not surcharge more than his own incurred “transaction” cost of a card payment. This result is perfectly in line with the proposed “merchant indifference test” or “tourist test” to optimally cap merchants fees keeping the merchant indifferent between a cash payment vis-à-vis a card payment. Yet, recent cost-based surcharge regulations seem too lenient, as they allow surcharges up to the merchant fee – or even higher (Gomes and Tirole, 2018).

And remember that NZ retailers can set surcharges if they want. All of us see them all the time. Maybe John Small doesn't. 

It then goes on to note evidence from other countries about smallish proportions of transactions attracting surcharges.

But think about it for a minute. 

High volume retailers have some power in those relationships. You might expect that surcharges imposed on transactions at the grocery stores would be lower than the surcharges in other places. And it wouldn't just be about the supermarkets being big enough to have some heft. It would also be about the cost of providing the ancillary services that credit cards provide. You're going to be a lot less likely to see chargebacks for undelivered or unsatisfactory goods on a weekly grocery shop than you might for a mail order shipment that's gone wrong or, say, payment for your kid's trip out to space camp. Credit cards provide insurance; EFTPOS doesn't. That insurance is valuable, but more valuable in some cases than others. Those differences matter and I'd expect affect the charging structure that the card companies set. Nobody's doing a credit card chargeback if there's a broken egg in the darned carton when you get home. 

Another key bit, if you remember that NZ retailers very regularly set surcharges for credit cards.

Moreover, many payment networks have frequently imposed restrictive – and potentially “regressive” rules – on the merchant side, such as no-surcharge rules or honor-all-cards rules.31 Effectively, this implies that payment cards that are more expensive for merchants to accept, such as credit cards, will be cross-subsidized by cheaper means of payments such as debit and cash. As high-income consumers are the ones most likely to hold and use cards with higher reward schemes that are more expensive for merchants to accept, the cross-subsidies between the payment methods are regressive transfers from low-income consumers to high-income consumers (Felt et al., 2021; Wang, 2023).

I remember MBIE relying on this kind of argument in making its case, seemingly unaware that NZ retailers can and do set those surcharges. Hence my photography while out grabbing lunch. 

If ComCom tightly restricts credit card fees, expect a whole pile of services currently bundled with card transactions to disappear.  

It's annoying when there's a world of real problems that need to be dealt with and agencies like ComCom go off on these kinds of tilts. 

Fun antitrust application

David Harvey reports that AI scraping could wind up being part of the revised NZ Fair Digital News Bargaining Bill. 

Having defined what an AI system and an AI service is the Bill goes on goes on to link an AI system to news content for the purpose of training the AI system.

The focus is upon the way in which news content may be used to train a digital platform or AI system.

The first element is that an AI system must be trained using news content. This links to the definition of news content in the Bill. The training must generate outputs which happens if the AI system enables or facilitates the generation of outputs.

He continues through with technical elements on whether the definitions work and whatnot.

The better underlying question seemed to be why anyone thinks there's a problem here to be solved.

It's simple for a website to restrict against scraping. It would similarly be simple for a news site to licence its content for AI training, if anyone wanted to pay them enough to allow it. There is no obvious reason government needs to be involved in any of this. 

But there's a fun potential antitrust angle, and then conflicting priorities. It looks like Google has licensed Reddit content for AI training, and that part of the deal might mean that Google is now the only search engine that works on Reddit.

Reddit wouldn't have set an exclusivity deal unless the exclusivity deal gave it more money than licensing its content to multiple agents. 

Government-types have claimed to be deeply worried about news sites not being able to adequately monetise content, with consequences then for the public good aspects of journalism. And they've tried to punt the bill over to tech platforms rather than just fund a public good out of public funds.

Here we have a news content site that has made a voluntary deal with a platform for content access for AI scraping - the very thing that NZ's Parliament seems to want to legislate to force - and folks are worried that the exclusivity arrangements that mean more money for the news sites also give the platform too much power. 

Today in Antitrust: when we force (or even allow!) AI companies to sign deals w/ content producers for old content or API access, there is obvious way to "raise rivals' costs" and lock in market power, dynamically making things worse for content producers. https://t.co/5wvRIoE0mU

— Kevin A. Bryan (@Afinetheorem) July 24, 2024

I'm not saying there aren't potential competition issues here. But there are trade-offs. Ban exclusivity arrangements and you'll reduce the amount of money going to news content providers, and then you'll have other parts of government looking for convoluted ways to force additional payments. 

Afternoon roundup

The closing of the browser tabs:

• The Commerce Commission blocks NZ DJ software company Serato from selling out to Japanese DJ product company AlphaTheta. One lesson: don't do tech startups in NZ if your exist strategy is selling out to a larger company.


• New Zealand's bureaucrats have apparently been considering a harmonised set of plain packaging rules on infant formula with Australia, which would force an important part of dairy processing overseas. Hopefully the government stops this one while issuing a pretty hard "please explain" to the officials involved. 


• Aaron Smale's reporting on the Inquiry into Abuse in Care has been superb throughout, and is great again this morning at Newsroom. He asks the right questions about state accountability. I expect that those who helped cover up the state's abuses are more likely to earn knighthoods than to face charges. 


• Gas shortages have Huntly importing more coal. 


• The Department of Conservation has more opportunity than any other part of govt I can think of to realistically improve its own finances through user charges and concession auction. Newsroom reports this morning that they're facing "huge financial pressure". There is an obvious solution here.

Right to Repair

The Green Party has a Member's Bill up arguing for a consumer right of repair; Auckland University's Alex Sims has written a few columns in support of such a thing. 

I'd had an email asking about that legislation; figured I'd share my response here - tidied up a bit.

If it’s more expensive to produce a product that can be easily disassembled for repair, there will be trade-offs. Consumers could choose an offering with lower up-front costs, but hard to repair, or one with higher up-front costs, but easier to repair. There’s no reason for legislation to privilege one choice over another. 

If one car company makes vehicles that can only really be repaired by dealers, and another uses a more open standard, the latter could easily advertise that ease of repairability. I remember back on our farm we had a very strong preference for tractors made by Versatile, because field-repair was dead simple and you didn’t have to wait for a couple days for some tech to come out with a diagnostic kit in the middle of harvest. Folks who could afford fleets of John Deere tractors to cover twenty square miles of fields could have a couple in reserve; we couldn’t on 1000 acres. Trade-offs and consumer choice. The John Deere machines were great for folks in situations different from ours.

But even leaving that aside, New Zealand has to be a regulation-taker in this space. We import all this stuff. A bespoke rule could require separate production lines for products destined for the NZ market. That has to increase costs while sharply limiting the range of products here available. And if Europe or some other crazy kind of place sets rules requiring more repairable versions, nothing stops anyone from bringing the Euro-standard products into New Zealand. 

Also important to remember that avoiding putting things in landfill can itself be wasteful. Landfills charge people for dumping things. Important to make sure that user charges there are set to fully cover the cost of disposing of stuff in landfill. If the landfill charges are set properly, and it's cheaper to buy something new and dispose of the old one than it is to repair the old one, repairing the thing would be wasteful. It would take more real resources to effect the repair. 

And if there are competition issues around vertical integration in repair, that’s for ComCom right?

I hope the legislation does not progress. It could easily see a sharp reduction in the range of products offered onto the NZ market.

The current Consumer Guarantees Act amounts to an information requirement on this stuff. 

If a manufacturer does not undertake to provide parts or repair services, they inform the consumer. That's the exception provided at Section 42 and signaled in Section 12. It arguably increases consumer information and enables better-informed choices. I'd still argue that manufacturers of easily-repaired goods already have plenty of incentive to advertise that fact to customers, but it's harder to see that the Act does harm where that exception is provided.

Deleting that exception while extending the requirements placed on manufacturers willing to sell into the NZ market really wouldn't be good.