Observing externalities

The existence of an externality isn't sufficient basis for declaring market failure; indeed, optimal Coasean internalization pretty often will involve what looks, to the outside observer, like one party imposing harms on another - the implicit side-payment is harder to see.

This isn't new news, or at least it shouldn't be, but Eli Dourado presents the logic pretty nicely in a Mercatus working paper. I'm likely to pick this one up for the week on externality in my Economics and Current Policy Issues class; the first section gives a lucid and tractable exposition of the basic theory I try to get through to students in lecture. We're both fans of Buchanan and Stubblebine's approach. Here's Eli:

Once again, when externalities are internalized in a complex fashion through firms and other transaction-cost-reducing arrangements, the visible external harm or benefit will often persist. We still observe external benefits from lighthouses and, at least in jurisdictions that still allow smoking in bars, we still observe nuisance externalities generated by smokers. But these are not market failures; these are problems that the market has solved despite the high transaction costs that plague the primary actors.

Market failure results from gains from trade that fail to obtain due to some impediment; alert entrepreneurs view such impediments as profit opportunities for those who are able to innovate around the blockage.

And Eli's extension to cybersecurity makes points that get missed by cybersecurity experts.

Pigovean paternalism

Frances Woolley reports on problems in her students' understanding of Pigovean taxes. After a set-piece question asking students to calculate equilibrium Pigovean tax in an externality case, students were told to answer this last bit:

Opponents of the tax on potato chips take a careful look at Dr. Economides’ study. It turns out that the only people harmed by potato chip consumption are potato chip eaters themselves, as potato chip consumption is associated with bad skin, weight gain and depression. Does this strengthen or weaken the argument for taxing potato chips?

Frances rightly notes that weakens the case for Pigovean taxation; I'd go farther and say that it darn near obliterates it. If the argument for taxation is consumer irrationality, then, as Seamus has noted before, we've stepped rather outside of the rational choice framework that's necessary for assessing costs and benefits in the first place. What does a demand curve even represent in the case where consumers aren't competent to evaluate net personal benefits? Maybe we can derive it from observing consumer behaviour, but revealed preference goes away and the welfare analysis then has to start from a rather different place. Some of the behaviouralists have started building frameworks for that kind of analysis, with multi-self Pareto criteria, but it's hardly canon.

...it seems that some students really don't believe that people are rational decision-makers, fully taking into account the long-term effects of their consumption choices. Even when people are only harming themselves, they support Pigouvian taxes on paternalistic grounds, to stop people from harming themselves.

It seems to take irrationality of a very particular form for Pigovean taxes to be a solution to internality problems. We need it to be the case that consumers irrationally discount health costs of monetized value x but to respond optimally to taxes of equivalent value. If consumers are also irrationally price-insensitive, you've doubly hurt them by imposing the tax. And we still have the problems that arise once revealed preference can't form a starting point for welfare analysis.

And, despite the explicit framing of the question - that consumers only harm themselves - some students read into it that they were hurting the taxpayer through the public health system:

Some students disputed the basic premise of the question, the idea that potato chip eaters are only harming themselves. Bad skin, weight gain and depression, they argued, are harms to others, because we have a public health system. ...

What interested me about this response was how "health" becomes a lens through which public policy issues are viewed, and a justification of policy choices. Perhaps, though, the students were just mislead by the wording of the question. Guessing that the specific details about bad skin, weight gain and depression must matter in some way, they figured that the question must be asking about health care. 

These are upper level undergrads in economics, and they're messing up the distinction between pecuniary and technological externalities. Costs through the public health system are only a transfer unless these consumers are eating more potato chips than they would in a private system; in that case, only the deadweight costs of the increased portion of consumption get to count as policy-relevant on an efficiency standard.

I hate to keep banging on about this [hit the fiscal externalities tab below]. There's no way that the folks at Carleton wouldn't have hammered home the distinction between technological and pecuniary externalities. But fiscal externalities - externalities that work through the budget constraint but via the tax system - keep seeming technological to students when they're really mostly pecuniary. Partially this is because folks don't start with Buchanan and Stubblebine, but I expect that it has more to do with that fiscal externalities have seemed perhaps an interesting sidebar not worth extensive class time.

We are graduating too many students who know that negative externalities are bad and that government should tax negative externalities, but who have little sense of which ones actually have efficiency consequence.

Pecuniary and Technological Externalities, and EQC

Recall that in standard micro, a pecuniary externality is one that affects you through your budget constraint, if you follow the standard Buchanan and Stubblebine definition, or one that is mediated through the market process, if you follow other texts. A technological externality is one that affects you through your utility function (Buchanan & Stubblebine) or operates outside the market process (others).

And standard theory says that pecuniary externalities are of no efficiency consequence. They don't result in sub-optimal decisions being made; further, we usually expect that they're offset elsewhere. If my bidding at auction forces you to pay more for your house, that's a loss to you but a gain to the seller and the house still goes to its most valued use. Technological externalities generate inefficiency. So we say that policy should generally ignore pecuniary externalities and focus on technological ones.

And in Friday's paper we read that New Zealand's Earthquake Commission withheld from neighbours information about dangers posed by neighbouring buildings. Why? Because of the potential pecuniary effects:

The Earthquake Commission (EQC) withheld information on a dangerous and ultimately deadly Christchurch building to protect privacy and property prices, a royal commission has been told.

Two women were killed in Wicks fish and chip shop in Worcester St after a brick wall from the neighbouring two-storey building collapsed into the shop during the February earthquake.

Natasha Hadfield, who owned the shop with husband Geoffrey, was serving Betty Dickson when they were both crushed by falling bricks.

The Canterbury earthquakes royal commission was told yesterday that on February 1 a commission inspection of the two-storey building found two walls were unstable and "in danger of collapse".

The EQC inspector urgently requested an engineering inspection of the building, but the request appears to have gone nowhere.

EQC chief executive Ian Simpson said yesterday the file had "gone into a hole", blaming a paper filing system that had since been replaced.

"I can't answer why an engineer was not contacted," he said.

After the September 2010 quake, the building sustained substantial damage. It had its eastern wall propped, and a hole in the roof was covered by a tarpaulin. It was cleared by at least one engineer as safe.

The western wall, which fell on to Wicks on February 22, was not braced.

EQC inspectors identified the western wall as dangerous, but the information was never passed on to neighbouring building owners or the Christchurch City Council.

Simpson said the policy had been to not release any information on inspections to third parties to preserve property prices and privacy.

"It was about bricks and mortar and property prices," he said.

The policy had reflected many residents' concerns that if information on quake damage was attached to their property, it could affect values, he said.

The policy was changed in October this year, largely because of the deaths at Wicks, he said.

The pecuniary effect is troubling enough; let's take that one first.

Let's start with the best case for EQC's information withholding being efficient. Suppose that a property's earthquake damage gets listed on the Land Information Memorandum (LIM) but the damage is fixed by the owner. If the subsequent repairs don't get noted on the LIM, or if buyers are irrationally risk averse, the owner may suffer a real loss despite having fixed all the damage. If buyers are rational, this only happens where property damage is a good indicator of likely damage in subsequent events despite the property's being repaired; in that case, we impose a cost on buyers if we fail to disclose. But if they're irrational, then we could perhaps view LIM listing as being akin to noting on the LIM that somebody was murdered on the property a few decades ago if everybody in town believes in ghosts. If ghosts don't exist, then both buyer and seller are better off by that the buyer never finds out that the property might have ghost-risk.

But earthquakes are real and ghosts aren't. Even if there aren't any technological effects, policy then effects a transfer that may not be desirable: buyers can't tell which properties are subject to greater earthquake risk and so we get a transfer to the owners of risky buildings from buyers and from the sellers of relatively safer buildings who then are pooled with the owners of risky buildings.

Bringing back the potential for technological externalities, things look worse. In order to avoid a pure transfer, EQC induced a technological externality.

It's not likely that different policy would have affected outcomes in the case noted in the paper; unless the inspector finding problems immediately went next door to warn people, the information would have been lost in the system until after the 22 February earthquake anyway. So it's wrong to blame this EQC policy for these two deaths. But that sure doesn't make it good policy to trade a pecuniary externality for a technological one.

Stories need scarcity, and unrelated bits

Bill Kaye-Blake has migrated over to a new blog, Groping Towards Bethlehem. I'll comment on a few of his recent posts here; I'm sure his blog will be worth following in future for those watching Kiwi econ policy.

• Bill's bored by a SciFi story which effectively finds the end of scarcity - the protagonists do away with constraints on space, energy, time and mortality.
  
  I've not read The Number of the Beast, but Bill didn't mention whether knowledge remained limited - if the multiverse already has a thorough and completely accurate Encyclopaedia Gallacticus. Presumably the protagonists could find challenges in expanding the domain of knowledge. Or, finding ever more creative forms of hedonism. 



• Does welfare reform posit $20 bills lying around?
  
  My quick read of National's proposal was that it didn't seem to do a whole lot; the biggest change was that additional children had while on the DPB won't extend the duration of the benefit by as much as it would have ex ante [see also Lindsay Mitchell]. I'll agree with Bill about the multiple barriers to work issues, and he's very likely right on National's overselling the benefits.
  
  But most outrage about the proposals thus far seems centred on the immorality of forcing women back to work before their kids are 14. This falls entirely on deaf ears for me - both our kids have been in daycare since they were 3 months old, in part to keep our family net income sufficient for raising the two kids adequately; were we not paying to keep a few single mothers home with their 12 year olds, maybe we could afford to choose otherwise. I suspect that more than a few double income families with kids in daycare get just a bit angry when called on to solve the injustice of that folks might not be able to choose to stay at home with kids at others' expense. If I had a button to push, it would be for 1) more daycare funding; 2) making childcare benefits of limited duration; 3) requiring Norplant (or other equally-effective equivalents) as condition for receiving the childcare benefit. I can accept that there are cases where it's more efficient that the mother stay home with the kids rather than pay for daycare given the woman's earning potential; I'm happier eating the losses from subsidizing daycare than the losses from screwing up the incentives. 



• Externalities and risky buildings. Bill's thinking about things in the right way: buildings with unreinforced masonry impose a risk on passers-by in case of earthquake - a negative externality - while providing uncompensated amenity value - a positive externality.
  
  Optimal response, I think, will have to involve Councils' putting some budget [hopefully supplemented by private donation] into paying building owners for heritage amenities. Rather than listing thousands of buildings on heritage registers and making it hard to fix them up, list the couple hundred most important and pay their owners for the heritage amenity provided. Then either add into a building's property tax assessment a fee for risks imposed by unreinforced masonry or require holding adequate liability insurance for damage potentially done to passers by in an earthquake. Some older masonry buildings will be torn down, but if their heritage value isn't sufficient to make them worth the risk, that's optimal. But we won't know it until we start putting some real prices on heritage value and paying owners for it, and charging owners for risk imposed.

I'm pleased to see Bill striking out on his own! We now have a decent selection of economics blogs by NZ-based academic economists: Offsetting, Anti-Dismal, and Sam Richardson; we've also another economics blog by non-academics that capably tackles academic issues: the Visible Hand. If we're not careful, we'll soon have enough folks around for reasonable conversations.

More things that aren't externalities

Roommates impose costs on each other; the decision to flat with someone and the agreements governing that arrangement internalize those costs. Nuisance costs roommates impose on each other are not an externality of any policy relevance.

I hope they don't come up with a policy solution to this one:

Using census data and analysis of an informal pricing survey of 114 users of [excised].com (a “share bills” app for roommates which I co-founded), I estimate that solving the loud sex problem alone would be worth $1.1-1.9 billion per year to the US market. Mitigating all unpleasant noises would represent a market of around $12 billion per year for the population we considered.

Ok, maybe they're just thinking of how much they could earn selling noise-reduction devices and taking a pretty high upper bound. But:

What is to be done about this cacophony of copulation? Noise machines and comfortable headphones would probably be the cheapest way to solve to problem, but these solutions are unlikely to be sufficient. Improved building codes in areas with high population density would put the burden on developers, but perhaps construction could be stimulated with tax incentives. Or perhaps creating a standardized rating system would be a good way to bring sound isolation issues out of the closet and let the market decide how much peace and quiet is worth.

Or, we could just leave people to choose their roommates based on a broad bundle of characteristics and to sort out amongst themselves how to deal with disputes.

Again, as helpful reminder: building owners install any feature, including noise-proofing, up to the point where the discounted value of the increase in expected rental flow matches the cost of the building feature. Tenants sort based on cost and on disutility of noise. If asymmetric information between landlord and potential tenant on noise issues were sufficiently large and if enough potential tenants cared, buildings would already be third party certified for noise characteristics.

It's a bit embarrassing that Forbes gave space to what's effectively an advertorial for the roommate bill-splitting website on the hook of a big dodgy cost number extrapolated from 114 users of a website designed for folks who can't otherwise figure out how to split their own bills without rancor.

Fence in or fence out?

Coase taught us that externalities are two-sided: there can't be an externality if nobody's around to experience it. And so we seek rules that make the lowest cost avoider of the externality be the one to bear the costs.

Today's application: open-plan offices. Canterbury's Department of Economics and Finance is now split between two open-plan barracks among a couple dozen such buildings in a muddy part of campus that once was a running track. Odds are that we'll be in the barracks for the next year and a half, but the right tail on that estimate is thick. They're gutting the old Commerce building looking for any problems in the cement floors (remember, we're still in earthquake-land); we have to entirely vacate that building, with most of our stuff going into long term storage as there's no way that the current facilities have room for everything.

But enough whinging; on with the problem at hand.

We all make noise in the course of our work: taking phone calls, chatting with colleagues, typing exceptionally loudly on a Das Keyboard Ultimate. Initial recommendations were that staff move to a small private room inside the open-plan space for extended phone conversations or chats with colleagues. The alternative is that everyone buy a set of noise-cancelling headphones. I've been pushing for the latter. Wearing headphones is pretty low cost, especially relative to having to run and transfer calls all the time. And everyone has a different idea about how annoying other folks' conversations are; that's more easily solved via the volume knob on a headset than by shushing colleagues. Added benefit: the heavy construction equipment outside provides only a visual distraction. A hundred bucks spent on a good set of headphones puts me into my own little world, open-plan office or not: money very well spent.

Two norms seem to be developing across the two barracks. In my barracks, the micro folks pretty much live and let live. Folks who don't like noise wear headphones (ok, it's just me so far, but folks are only starting to move in), others wear no headphones, and if I'm taking a long phone call, I head for the private room less for my colleagues' benefit than for privacy. In micro, we're basically Coaseans (as is right and proper for any Top-100 Economics Department)

In the pod where the macro, finance, and experimentalists live, a norm of being shushed all the time seems to be emerging. We'll see whether the two-norm equilibrium holds or whether those wacky macro folks see the light. Either could be optimal depending on folks' costs of wearing headsets and folks' noise abatement costs.

Final note: none of this counts as an externality for policy purposes precisely because we expect firms to optimally structure things within their own contractual nexus. The optimising employer will seek that the lowest cost avoider bear the costs and will allow sufficiently disaggregated decisionmaking to allow inter-departmental heterogeneity in different types of cost be reflected in different barracks' solutions.