Wednesday, 21 March 2018

Bias towards zero in maternity perceptions

We're all running estimations all the time. If we've a modicum of truth-seeking and are trying to figure out how the world works, we're all estimating implicit bivariate and multivariate relationships between things all the time. A lot of them are going to be wrong if we're not doing it properly in Stata, but we get a basic working model of how the world works.

Whether you're running that regression implicitly or explicitly, things go wrong if you're omitting variables. Fremling and Lott called this Bias Towards Zero in Aggregate Perceptions.

Suppose that youth unemployment rates are a function of the business cycle, the youth minimum wage, and regulations on ease of dismissing workers that don't pan out - and interaction effects among those.

And suppose further that everyone who considers all of those variables will get the relationships right, on average. There's variance around a true mean. So among the set of people running the implicit regressions, you've got a reasonable picture of the world.

But if half the population doesn't know that they need to think about the youth minimum wage, then that's omitted variable bias. It'll screw up the coefficients on the other variables on which the variation from the omitted variables will load, and it will bias towards zero the aggregate perception of the true relationship between youth unemployment and the youth minimum wage.

I don't want to argue about youth minimum wages and youth unemployment - the relationship could be as weak as you like. But if there's a relationship and you leave it out, you get the problem.

There's been a ton of press coverage around poor outcomes in maternity care in New Zealand. The explanations that have shown up thus far are around the number of beds in maternity care centres and overall funding.

There's another potential variable that should be considered. I don't know how big the effect of that variable is, but I know that aggregate perceptions of it are going to be biased towards zero because nobody's talking about it.

In the 1990s, maternity care shifted from being GP-led to being midwife-led. Lots of GPs left the market because the payment wasn't high enough to keep them interested, but was high enough to interest midwives with less training. One paper at the NZAEs in 2008 presented some evidence that this affected neonatal deaths. The paper's a decade old now and likely needs strengthening, but suggests there's something there in need of investigating.

When we had our kids, we knew that some midwives had only weak training. So we made sure to book in real early to get a midwife with proper nursing training as well - and we went for a shared-care arrangement with an obstetrician at the same time. The people who most need the best trained midwives will not know that they need to move early to get one. And then we get some rather poor outcomes.

Maybe the coefficient on the variable "We changed from GP-led to midwife-led maternity care" is small, maybe it's big. But aggregate perceptions of the coefficient are going to be biased towards zero. It would be nice if it were part of the mix when folks are thinking about maternity outcomes.

Thursday, 15 March 2018

The deregulated interregnum

E-cigarettes in New Zealand are in a really interesting spot right now.

Last year, the then National-government signaled that a new legal regime for vaping was on its way. At the time, importation of nicotine-containing vaping liquids for personal consumption was allowed, but retail sale within New Zealand was not. That wasn't to say it didn't happen, but you kinda had to be a Wellington hipster to know where to get the stuff.

I'm not sure that the de jure ban on retail sales was ever really enforced, but it was all on the down-low because retailers feared enforcement.

That's all changed. Here's what was on the counter at a local dairy just a couple weeks ago.

They're still operating in the shadow of regulation. They're restricting sales to those over the age of eighteen. 

But whatever regime winds up coming up will have a hard time being more liberal than the current de facto liberalisation. I don't know whether suppliers are currently holding back from advertising for fear of getting in trouble for it; allowing advertising would then be a step forward. Otherwise, the current de facto regime has sales at dairies and at specialty shops; some of the public health people would want to restrict access to pharmacies and only after a doctor's prescription. 

Peter Huber's work on regulation emphasised the difference in incentives facing agencies that seek to regulate against new risks, and those that seek to regulate existing risks. The FDA is a good example of the former. Because it would take stick for being too quick to approve an unsafe drug, and nobody would much protest if they killed a hundred thousand people by being too slow to approve a drug that was actually safe, the FDA will be too restrictive. There isn't an established constituency arguing for the release of the new drug because most of the would-be consumers of it don't know that they could be helped by it. He contrasted that with the EPA, which often goes after chemicals that are found to be more dangerous than previously thought. In those cases, there's a more careful balancing because the thing that's being regulated has a substantial group of users whose interests have to be considered - or they will raise heck about it. So there's then less likely to be excessively cautious regulation.

The upshot of all of that is that the longer this current de-facto legalisation runs, the more liberal will be the resulting regulation. Why? The more people who switch from smoked tobacco to vaping while the rules are easy, the more users who will be absolutely furious if the Otago public health people make them get a doctor's note for their next batch of nicotine e-liquid. Instead of a Select Committee process with the Otago restrictionists on the one side up against a few hipster vapers on the other, you'll have one with the Otago people up against thousands of rather more sympathetic community of users on the other.   

The vaping advocates wanting to hurry-up MoH on this stuff might think it through a bit more. 

We still don't know what will happen. If Parliament is determined to be stupid despite a huge existing set of users, like it did when it banned access to pseudoephedrine cold medications, it'll do it. But the longer the wait, the more likely we'll be to get something reasonable because the underlying balance shifts. 

Friday, 9 March 2018

Does a housing crisis have heritage value?

Ryman Healthcare wants to put up a retirement village in Karori, along with a pile of related services. They've bought the old Karori teachers' college.

It's a great site - easy land to build on, right on bus routes, an easy walk to Karori village and right next door to the Karori Swimming Pool.

But there's a problem. The old Karori teacher's college is apparently a great example of 1970 brutalist architecture, and New Zealand has a rather low threshold for applying heritage designations to buildings. Heritage New Zealand wants to impose a Category 1 listing on it.

Meanwhile, in Auckland, View West wants to tear down an old unreinforced soft concrete church in Epsom that is at risk of falling down onto passers-by and replace it with terraced housing - the kind of dense inner-suburb housing that Auckland is missing. Heritage New Zealand opposes that, noting "unacceptable adverse historic heritage effects."

I have a small proposal. Until the housing crisis ends, no new heritage designations. If a building or site is of sufficient value, Heritage New Zealand should solicit donations so that it might purchase the site. The good people at Architecture Centre might chip in to buy the Karori teacher's college as they seem rather fond of it. As owner of the building, Heritage New Zealand would then be able develop the site as it deems best.

For buildings with existing heritage designations, a Council refusal to grant consent to develop for housing should be treated as a Council offer to purchase the site at its rateable value. Submissions opposing the development should be treated as offers to contribute towards Council's purchase of the building. Submissions opposing development that do not include a pledged amount to contribute towards the site's purchase should have no standing. Council then can decide whether the weight of opposition, along with its own willingness to contribute towards buying the site, is sufficient to warrant purchase of the site.

Rational addiction and Naloxone

Suppose for a moment that addicts are rational and that we're really in a Becker-Murphy world.

What happens in that world if a new technology makes it safer to consume an addictive product?

If you're overdosing on opiods, Naloxone can save your life. The easier is access to Naloxone, the less likely addicts are to die if they have an overdose. Preventing death is good. 

If we're in a rational addiction framework, net effects on mortality will be lower than you might have thought though. The costs of taking up an addictive good go down, so more people will choose the addictive consumption path. The optimal quantity consumed on that path should be higher as well where the risks are somewhat abated. Addicts would have less need to build in safety margins against higher than expected potency, for example. 

Basically the A curve in the model shifts up. You then get an unstable low-consumption equilibrium at a lower amount of consumption than was previously the case, and a stable high-consumption equilibrium at a higher level than used to be the case. 

None of that says that improving access to Naloxone is a bad thing. It just says that, depending on all the relevant elasticities, you could easily see increases in consumption and increases in the number of consumers. Effects on overall mortality - hard to say. The number of overdoses will go up, but a higher fraction of overdoses will be countered by Naloxone.

What does the evidence say? Naloxone access laws vary state-by-state in the US, so we can figure it out. And Jennifer Doleac and Anita Mukherjee did. What did they find?
Policymakers have multiple levers available to fight opioid addiction, and broadening
Naloxone access aims to directly address the most dire risk of opioid overdose: death. Naloxone can save lives and provide a second chance for addicted individuals to seek treatment, but access to this lifesaving drug may unintentionally increase opioid abuse by providing a safety net that encourages riskier use. This paper shows that expanding Naloxone access increases opioid abuse and opioid-related crime, and does not reduce opioid-related mortality. In fact, in some areas, particularly the Midwest, expanding Naloxone access has increased opioid-related mortality. Opioid-related mortality also appears to have increased in the South and most of the Northeast as a result of expanding Naloxone access.

Our findings do not necessarily imply that we should stop making Naloxone available to
individuals suffering from opioid addiction, or those who are at risk of overdose. They do imply that the public health community should acknowledge and prepare for the behavioral effects we find here. Our results show that broad Naloxone access may be limited in its ability to reduce the epidemic’s death toll because not only does it not address the root causes of addiction, but it may exacerbate them. Looking forward, our results suggest that Naloxone’s effects may depend on the availability of local drug treatment: when treatment is available to people who need help overcoming their addiction, broad Naloxone access results in more beneficial effects. Increasing access to drug treatment, then, might be a necessary complement to Naloxone access in curbing the opioid overdose epidemic.
They frame their work in terms of moral hazard, and don't mention the Becker-Murphy model, but this is what we should expect out of a Becker-Murphy rational addiction model. The cost of consumption of an addictive good goes down, so consumption of it goes up. 
It may seem surprising that drug users respond to incentives in a sophisticated way. One may think that drug users are poor decision-makers or that addiction makes rational choices impossible. Addiction surely clouds judgement and makes policy in this area difficult, but there is substantial evidence that even drug users respond to incentives. A large body of empirical evidence documents that the consumption of addictive substances is sensitive to prices. For example, increasing taxes on alcohol reduces alcohol consumption (Cook and Durrance, 2013). Alcohol abuse also responds favorably to increasing the likelihood of punishment, as seen in evaluations of the 24/7 Sobriety program (Kilmer et al., 2013). Hansen, Miller and Weber (2017) show that marijuana consumption is price inelastic in the short run, but quickly becomes price elastic, with consumers reducing their consumption in the face of higher marijuana taxes. And finally, Moore and Schnepel (2017) show that a massive reduction in the heroin supply in Australia resulted in a long-term reduction in heroin consumption among those using heroin at the time, due to a spike in the price of the drug. These findings suggest that, at least on the margin, drug abuse may be sensitive to non-monetary costs such as the risk of death.
Rationality-based models. They may not be woke, but they work. Read the whole thing. Or, for Doleac's tweetstorm on the paper, check her feed here

Thursday, 8 March 2018

Dirty pool

Thanks to Newshub, we now know that the government-owned insurer Southern Response had private investigators snooping around after earthquake insurance claimants.

There are potential non-horrible explanations for some of this kind of thing. I was surprised that there weren't violent incidents involving severely aggrieved insurance claimants who had been treated very badly by EQC. If an insurer had received threats, it would have not been remiss in passing those along to police - and hiring some additional help wouldn't seem amiss.

And if an insurer thought that a claimant were doing something dodgy with respect to a claim, it's good to try to knock out insurance fraud.

But this smells more like attempts to silence critics by a state-owned entity.

RNZ adds more detail.

I'll look forward to seeing what the State Services Commissioner finds out.

Wednesday, 7 March 2018

A sunset solution?

There's no way that Labour would implement this now, but it's something that the next third-term government should seriously consider.

Any incoming government has a big platform it wants to implement. Some of it will be well thought-through; some of it will be off the cuff dumb promises made in the heat of an election campaign.

And policy passed in haste as part of an incoming government's 100-day plan will never get the kind of rigorous treatment in an RIS process that it should. That would be true even if Treasury hadn't diverted all computer power over to trying to measure happiness, much like the Starship Heart of Gold's computer being fully occupied in trying to figure out how to synthesize tea while missiles from Magrathea approached at speed.

A potential solution? Sunset clauses. Something like this:
All legislation and regulation must be accompanied by an adequate Regulatory Impact Assessment. Where haste prevents the an adequate RIS from being produced, that legislation or regulation must include a sunset clause voiding the policy two years after its implementation and mandating a Post-Implementation Review to be completed within eighteen months of the the policy's initiation. That would be followed by an amendment bill to the original legislation. A successful PIR would have the amendment bill simply remove the original sunset clause. But a PIR finding deficiencies, or failure to produce a PIR, could lead either to more substantial amendment or termination - with the default in case of no legislative action being the policy's termination under the sunset clause. 
A mid-term government could be pretty confident in doing this because it will impose far greater constraint on any incoming government than it imposes on itself. Mid-term governments have plenty of time to consider their legislative proposals; incoming governments try to make a pile of big changes in a hurry.

And an incoming government abolishing a requirement for sunset clauses could reasonably be asked by voters why it thinks its policies wouldn't stand up to post-implementation review.

Tuesday, 6 March 2018

Submission time

It sounds like religious groups have been sending through a pile of submissions opposing David Seymour's end-of-life bill on assisted suicide.

Today's the last day for submissions. Hit the link here to put in yours. If you don't have your own, you could just copy mine in and say you agree. Or improve on it. Remix as seems fitting.

  1. I am a New Zealand Permanent Resident who hopes to live in New Zealand until the end of his natural life.
  2. Death is abhorrent, but some ways of dying are more abhorrent than others. This bill allows those facing truly miserable ways of dying to find a better way of ending their lives.
  3. Failing to adopt this legislation will not abolish suicide. Whatever your moral views on suicide, people facing intolerable end-of-life conditions currently have suicide as a de facto option.
  4. Current legislation forces those facing intolerable conditions to pursue suicide in worse ways. Because it is forbidden to provide assistance to someone considering suicide, a person knowing that an intolerable end awaits must either suffer that end, or commit suicide while still in sufficient bodily control to be able to effect that choice. 
  5. Allowing assisted suicide allows those facing an intolerable end to defer suicide to a later date, when life has become - in their view - no longer worth living. This point may come well after the point at which suicide would be impossible without assistance. 
  6. The legislation imposes no obligation on any physician to provide assistance. Any physician with moral objections to providing assistance can simply not provide that service, leaving it to others who are willing. 
  7. The legislation places substantial restrictions on those wishing to pursue assisted suicide; the potential for abuse of the provision seems limited.
  8. Those with moral or religious objections to suicide, and to assisting with suicide, will continue to have recourse under this legislation to not pursue suicide or to assist in suicide. Nothing in this legislation compels those opposed to assisted suicide to assist in the process. Those with moral or religious objections to assisted suicide should not have the right to impose their views on others. 
  9. A continuation of the current ban on assisted suicide is morally abhorrent. My life is my own, and the choice to end it is mine.
  10. In the absence of a provision for assisted suicide, should I face terminal illness with severe diminution in capabilities and/or intolerable pain, I would have to end my life on my own while I still were able to do so. I would have to do so secretly, to avoid imposing legal risk on my family who might otherwise be thought to have been complicit in my decision. It will be lonely, and terrible. To those Members of Parliament who think it immoral to allow me a better end-of-life choice, should I ever wind up in that situation, know that you will not be preventing a suicide but only ensuring a worse one. Is that really what your God would want?

On the recommendations part, I only said that Parliament should pass the legislation and should not impose further restrictions on access to this choice.