Sea level rise, coastal flooding, and house prices

In my ECONS102 class last week, one of the things we discussed was hedonic pricing - the idea that the price of some goods (such as houses or land) reflects the sum of the values of all of the characteristics of the good. In the case of property, if the property includes a dwelling, the price will reflect the quality and size of the dwelling, number of bedrooms, bathrooms, whether it has off-street parking, and so on. But the price also reflects the access of the property to local amenities, such as good schools, public transport, and so on (for example, see this post from 2017), as well as the property's risks of damage due to environmental disasters such as earthquakes or floods.

In the case of risk, properties that have a higher risk profile should have lower prices - a higher risk profile is a negative characteristic for a property. Two new research articles provide some relevant evidence.

First, this article by Allan Beltran, David Maddison, and Robert Elliott (all University of Birmingham) published in the Journal of Environmental Economics and Management (sorry I don't see an ungated version), looked at the impact of floods on property prices in the UK. They used data on over 12 million property transactions and nearly 5 million properties over the period from 1995 to 2014. Interestingly, their method looked at 'repeat sales'. That means that they essentially looked at property's prices before, and after, a flood event. Some properties were directly affected by flooding, while others weren't. They found that:

...in the immediate aftermath of inland flooding the average price of property in a postcode entirely inundated is 24.9% lower. For incidents of coastal flooding the corresponding figure is 21.1%. These results moreover emerge from a comparison of inundated and non-inundated properties all within the floodplain. Such discounts are however short-lived; property affected by inland flooding typically recovers after 5 years and in just 4 years for coastal properties. The time for price recovery differs markedly for properties in different price-quartiles. For properties affected by coastal flooding in the highest price-quartile, the property price discount disappears after only 1 year whereas for properties in the lowest price-quartile the discount remains statistically significant for up to 6-7 years.

So, floods reduced house prices, but the prices rebounded so that there was no net negative effect within several years. Interestingly, the effect was slightly lower for coastal flooding, and disappeared quicker. That is, people were quick to return to demanding coastal property soon after coastal flooding. That should be a bit of a worry to us, given that sea level rise is likely to be one of the enduring effects of future climate change.

Which brings me to the second article, by Asaf Bernstein (University of Colorado at Boulder), Matthew Gustafson (Pennsylvania State University), and Ryan Lewis (University of Colorado at Boulder), published in the Journal of Financial Economics (ungated earlier version here). This article provides more direct evidence on the effect of sea level rise on house prices, using data from over 460,000 property transactions of properties in the US that would "be inundated following a 1-6 foot increase in average global ocean level". Their analysis is not based on repeat sales, and neither is it based on actual sea level rise (it is projected future sea level rise). The latter point means that, if there are negative impacts on property prices, then buyers are factoring in future sea level rise in their decisions about buying. They find that:

...SLR exposed properties trade at a 6.6% discount relative to comparable unexposed properties. We further break this into exposure buckets, with properties that will be inundated after one foot of global average SLR trading at a 14.7% discount, properties inundated with 2-3 feet of SLR trading at a 13.8% discount, and properties inundated with 4-5 and six feet of SLR trading at 7.8% and 4.4% discounts, respectively.

Interestingly, it is non-owner-occupiers would are more likely to apply a discount to the property:

We find that the SLR exposure discount is concentrated in the non-owner occupied segment of the market. On average, exposed non-owner occupied properties trade at a 10% discount, relative to comparable non-exposed proper- ties, while exposed and unexposed owner occupied properties trade at similar prices.

In other words, owner-occupiers likely underestimate the negative impacts of sea level rise on their homes. They also found that owner-occupiers with stronger beliefs regarding climate change did apply a discount in buying coastal property.

These two papers, taken together at face value, should probably worry anyone who is concerned about the future impact of sea level rise and coastal flooding on people living near the coast. Coastal property is at risk in many (perhaps most) areas. Holding all other factors constant (such as the quality of housing, access to amenities and services, etc.), the value of these properties should be decreasing relative to less vulnerable property (or at least, not rising as quickly). It appears that is not the case, and in fact following flood events (which should make it abundantly clear to potential purchasers that these properties are vulnerable to coastal flooding and sea level rise), property prices are rebounding quickly to their previous levels. On top of that, it appears that it is owner-occupiers (and in particular climate-change-naive owner-occupiers) who will face the brunt of these future impacts.

I don't know that this leads to a strong case for regulation of coastal property in some way, but at least it suggests that coastal property owners (and potential buyers of coastal property) must become better informed about the risks. The specific vulnerability of coastal property to inundation and flood events probably needs to be communicated to potential buyers for every coastal property transaction.

Book review: The Everything Store

At the end of 2017, I read and really enjoyed Brad Stone's book The Upstarts (which I reviewed here). So I looked for other books by the same author. It turns out, I already had one on my shelf that I hadn't read - The Everything Store. The book tells the story of Jeff Bezos, and in particular of the rise of Amazon, initially as an online bookstore, and later as an online purveyor of almost everything (hence the title).

As he did for Uber and AirBnB in The Upstarts, in this book Stone does an excellent job of chronicling the history of Amazon. The key players (and there are many) are all included and their contributions to Amazon are noted in some detail. Stone includes lots of anecdotes that help you to really feel the relentless pace of development of the company over the years including, surprisingly, that they briefly toyed with the idea of calling the company Relentless (indeed, the web address relentless.com still takes you to the Amazon homepage today!).

Many of the details in the book, I had already read elsewhere over the years. However, there were still many stories that were new. I hadn't quite appreciated Bezos's total dedication to low pricing, but this quote in relation to Amazon Web Services (AWS) captures it well:

Bill Miller, the chief investment officer at Legg Mason Capital Management and a major Amazon shareholder, asked Bezos at the time about the profitability prospects for AWS. Bezos predicted they would be good over the long term but said he didn't want to repeat "Steve Jobs's mistake" of pricing the iPhone in a way that was so fantastically profitable that the smartphone market became a magnet for competition.

The comment reflected his distinctive business philosophy. Bezos believed that high margins justified rivals' investments in research and development and attracted more competition, while low margins attracted customers and were more defensible.

In its determination to capture customers through low prices, Amazon has driven hard bargains with its suppliers, and the book contains lots of stories to that effect. It has also driven hard bargains with its employees, and if the book is missing one thing, it is the lack of stories from the front line. I guess that many readers would not have appreciated it in a business book, but I wanted more than a few isolated anecdotes about "top-grading of employees" (where managers grade employees along a curve, and "dismiss the least effective performers"). Even those few anecdotes only appeared towards the very end of the book. However, despite their absence, there is no doubt from reading the book of the challenges that working for Jeff Bezos presents. He appears to be a hard taskmaster.

The book doesn't lack humour though, such as the story of the missing pallet of Jigglypuff Pokemon toys in one fulfilment centre:

The group was looking for a single box inside an eight-hundred-thousand-square-foot facility. "It was very much like that scene at the end of Raiders of the Lost Ark," Rachmeler says. She dashed out to a nearby Walmart to buy a few pairs of binoculars and then passed them out among her group so they could scan the upper levels of the metal shelving....

After three days of exhaustive searching, at two o'clock in the morning, Rachmeler was sitting, spent and dejected, in a private office. Suddenly, the door flew open. A colleague danced in, and Rachmeler briefly wondered if she was dreaming. Then she noticed that the woman was leading a conga line of other workers and that they were jubilantly holding above their heads the missing box of Jigglypuffs.

I guess it wasn't all bad news for the employees of Amazon. It has made a large number of them (not least Bezos himself) fabulously wealthy. For a company that was originally "conceived in 1994 on the fortieth floor of a midtown New York City skyscraper", Amazon has certainly come far. And this book does a great job of showing you just how far it has come.

You can pay a subsidy with a tax, but it won't eliminate the deadweight loss

The government's new plan for incentivising a switch to electric vehicles (EVs) is interesting, as reported in the New Zealand Herald last week:

The Government is signalling its intention to slash the price of imported electric and hybrid vehicles by up to $8000 in a bid to make greener cars cheaper for Kiwis.

But it is also planning to slap a new fee of up to $3000 on the import of vehicles with the highest greenhouse gas emissions.

The Government has today opened a six-week consultation period before it introduces new legislation in Parliament later this year...

The Government is proposing discounts of up to $8000 for zero-emission new imported vehicles, such as electric vehicles (EVs).

That number would be $6800 for plug-in hybrid electric vehicle (PHEVs) and $4800 for hybrids.

The level of the discount depends on the total net emissions of the vehicle...

A used Mazda Axela, which is one of New Zealand's most popular imported vehicles, would cost $7200 after an $800 discount.

But a new Land Rover Sports V8 would be slapped with a $3000 high-emissions fee.

A $22,000 Toyota Hiace would cost an extra $1400 after the fee was applied.

Genter said the policy would be cost neutral – meaning the money gained through the fees from higher emitting vehicles would offset the subsidies provided to the lower emission cars.

A specific excise tax on the sale of a good, such as high-emission vehicles, will raise revenue for the government, but it also creates a deadweight loss - there is some economic welfare from the market for those vehicles that is lost, because fewer of them are being traded. This is illustrated in the diagram below. If the market were left alone, it would operate with a price of P0, and Q0 high-emission vehicles would be traded. When the excise tax is imposed, we represent that with the new curve S+tax. The price the consumer pays for a high-emissions vehicle increases to PC, but the effective price for the seller decreases to PP (which is the consumer's price PC, minus the amount of the tax paid to the government). The quantity of high-emissions vehicles decreases to QT.

However, now think about economic welfare. Consumer surplus is the difference between the amount that consumers are willing to pay (shown by the demand curve), and the amount they actually pay (the price). In the diagram, at the equilibrium price and quantity, consumer surplus is the triangle AEP0. Producer surplus is the difference between the amount the sellers receive (the price), and their costs (shown by the supply curve). In the diagram, at the equilibrium price and quantity, consumer surplus is the triangle P0ED.  Total welfare is the sum of the two areas (consumer surplus and producer surplus), and is equal to the triangle AED.

Once the tax is imposed, the consumer surplus decreases to ABPC, while the producer surplus decreases to the area PPCD. The government gains the area of tax revenue, which is the rectangle PCBCPP (this rectangle is the per-unit amount of the tax, multiplied by the quantity of taxed vehicles). Total welfare is the sum of all three areas (consumer surplus, producer surplus, and government revenue), or ABCD. Notice that total welfare with the tax is lower than it is without the tax, by the area BEC. That is the deadweight loss of the tax - lost economic welfare as a result of the tax reducing the quantity of high-emissions vehicles traded.

So, we lose economic welfare in the market that is taxed. Does that mean that we gain welfare in the market that is subsidised? Actually, it doesn't. The diagram below shows the effect of a subsidy on the market for EVs. If the market were left alone, it would operate with a price of PA, and QA EVs would be traded. When the subsidy is introduced (and assuming it is paid to the importers of EVs), we represent that with the new curve S-subsidy. The price the consumer pays for a high-emissions vehicle decreases to PE, but the effective price for the seller increases to PF (which is the consumer's price PG, plus the amount of the subsidy paid to the seller by the government). The quantity of EVs increases to QS.

Now consider the areas of economic welfare. Without the subsidy, consumer surplus is the area FGPA, and producer surplus is the area PAGH. So, total welfare without the subsidy is the area FGH. With the subsidy, the consumer surplus increases to the area FJPG, while the producer surplus increases to the area PFKH. The government subsidy is the rectangle PFKJPG (this rectangle is the per-unit amount of the subsidy, multiplied by the quantity of subsidised vehicles). The subsidy is negative welfare - it reduces total welfare, because the government could instead use that subsidy money to pay for schools, roads, etc. So, it has an opportunity cost (it is not free money). Total welfare with the subsidy is the sum of consumer and producer surplus, minus the area of the subsidy. This is tricky because all the areas overlap, but if you work it out you'll find that total welfare is now FGH-GKJ. So, total welfare with the subsidy is lower than without the subsidy, by the area GKJ - the subsidy also creates a deadweight loss.

Now, combining the two markets, it is clear that the government could use the revenue that it raises from the tax on the high-emissions vehicle market, to pay for the subsidy on the EV market. However, that only pays the subsidy - it does nothing about the deadweight loss in either market. [*] So, while the policy may be cost neutral from a government fiscal standpoint, it clearly isn't cost neutral for society as a whole.

*****

[*] Now, you could argue (rightly) that the high-emissions vehicle market has a negative externality, and so too many high-emissions vehicles are traded relative to the welfare-maximising quantity. So, a tax on that market would actually increase total welfare (once you factor in the externality). However, that still leaves the deadweight loss in the EV market.

You could also argue that the EV market has a positive externality, since EV use reduces the number of high-emissions vehicles, and so too few EVs are traded relative to the welfare-maximising quantity. So, a subsidy on that market would actually increase total welfare (once you factor in the externality).

However, you couldn't argue that both of those things are true, since you would be double-counting the externality. Either there is a negative externality of high-emissions vehicles, or a positive externality of EVs, but there can't simultaneously be welfare increases for both of those things.

Mark Kleiman, 1951-2019

Normally on this blog I only mark the passing of famous economists, but I'll make an exception for Mark Kleiman, who was a professor of public policy at NYU, and was most famous for his work on drug policy. In fact, I referenced a blog post by Kleiman in a post last year on the economics of fentanyl.

Gabriel Rossman at National Review has an excellent article that captures some of Kleiman's key contributions:

Market failure and high transaction costs are policy successes when the commodity is poison, and so good policy means encouraging bad market design. For instance, Kleiman favored a noncommercial approach to marijuana decriminalization precisely because he expected nonprofit or state-operated dispensaries to be less efficient than for-profit firms, and in particular less likely to grow the user base through advertising and make intense use more convenient. The billboards advertising dispensaries, and even marijuana delivery, that saturate Los Angeles are exactly what Kleiman thought sensible decriminalization should avoid.

But just as good market design has to be careful, so does deliberately bad market design. A major argument in Against Excess is that if you make selling drugs risky by locking up drug dealers (or encouraging them to shoot each other over territory), you build in a risk premium to the price, which draws in suppliers who don’t mind risk. The better approach is to create a deadweight loss so you don’t encourage more supply. For illegal drugs, make it a time-consuming hassle to score. For legal drugs like tobacco and alcohol, impose stiff excise taxes. In both cases the consumer faces costs that do not benefit, and therefore encourage, sellers. These costs might not discourage addicts in the short run, but long-run demand is relatively “elastic”: Increased costs from hassle or taxes can discourage potential users from starting and encourage existing addicts to quit.

Those are insights that I have used in my ECONS101 and ECONS102 classes. Rossman also highlighted some additional contributions that I probably should make more use of:

Jointly tackling mass incarceration and crime was the aim of his most famous book, When Brute Force Fails. At a theoretical level, the book is an argument that Gary Becker’s economic theory of crime must be radically reconceptualized in light of behavioral economics. Becker argued that deterrence was the expected value of punishment, defined as the probability of punishment times its severity, which has the practical upshot that we can achieve deterrence by punishing infrequently but severely. However, behavioral economics suggests that people aren’t good at reckoning unlikely-but-severe outcomes — and if ever there were a group of people who live for the moment and ignore the future, it would be those who are either intoxicated or addicts looking to score. (Contrary to popular myth, relatively few prisoners are incarcerated for non-violent drug offenses, but many violent and property offenses are committed while intoxicated or to acquire money for drugs.) This implies that a ten-year prison sentence won’t have much more of a deterrent effect than a five-year sentence would. In practice, extremely long sentences serve not to deter crime, but to induce plea bargains and incapacitate criminals throughout their prime-offending young years — and beyond.

When Brute Force Fails is an important book not just for contributing to a theoretical dispute, but also for its empirical evidence and practical solution: mild but extremely consistent punishments, the opposite of Becker’s approach.

Given that I make a lot of use of Becker, and of behavioural economics, the juxtaposition of the two seems like something that would work well in future.

You can read more about Kleiman in this article by German Lopez on Vox, and here is the New York Times obituary.

[HT: Marginal Revolution]

Creating and capturing value, Spark style

One of the key things that I teach in my ECONS101 class (which is a business economics class, rather than an economics principles class) is that pricing strategy is all about creating and capturing value: first you create value for your customers, and then you find (often creative) ways to capture that value back from them in the form of higher revenues (and profits). Spark provided a great example this week, as reported in the New Zealand Herald:

If you're not scratching your head over how to stream the Rugby World Cup on Spark Sport, the telco will now come to your home and hand-hold you through the process - for a price.

A new $149 Spark service includes a Spark rep walking customers through their broadband connection, testing their connection speed, setting up any streaming devices and demonstrating Spark Sport.

The service is for Spark customers only.

For those wanting to get connected but are not Spark customers, Noel Leeming, Harvey Norman and Geeks on Wheels have all set up dedicated in-home Spark Sport solutions, the telco says.

Each have teams that will arrange a time to visit the customer's home to discuss and set up Spark Sport. This includes internet speed tests, making sure the households tech is set up correctly, and then teaching the customer how to easily watch Spark Sport on their preferred device to ensure they will be comfortable to do it on their own.

So, Spark creates a valuable service (Spark Sport) that consumers are willing to pay for (which is accentuated by the fact that Spark Sport has the rights to the Rugby World Cup later this year). Consumers sign up for the service, but they want to make sure that it is going to work well for them (or maybe they aren't very technically savvy), so Spark offers them assistance. That assistance comes at a price, so Spark can extract additional profits from those consumers.

Add-on services, or up-selling, are a feature of many markets (e.g. "Is that a large combo?", "Would you like fries with that?"). Once a firm has attracted consumers to buy from them, the goal is to extract surplus from them, and this is a very effective means of doing so (if it wasn't effective in increasing profits, then firms would quickly stop doing it).

In this case, the additional service itself may be valuable, but $149? That seems expensive, especially when you consider that the price for the subscription to Spark Sport for the whole Rugby World Cup is just $90. It could be that Spark is also trying to price discriminate here - if consumers who are less price sensitive are also those who are less technically savvy, then this would be an optimal strategy, since you would want to set a higher price for the less price sensitive group of consumers. However, I would take some convincing that this is the case.

A more cynical reading of this situation might be that there is now an incentive for Spark to ensure there are reasonably frequent technical 'glitches' that disrupt coverage between now and the start of the Rugby World Cup (an even more cynical reading might suggest that this is already happening). Then Spark Sport customers start to become worried and start shelling out for the in-home assistance.

Far be it for me to be that cynical. I'm sure Spark wants to develop a long-term relationship with its subscribers, and gouging them for additional charges in the first few months is clearly not the way to achieve that.

[Update]: We had a great discussion on this over on the EDG Facebook Group, which highlighted one way that this could be price discrimination. If a consumer is not tech savvy, and they don't have friends or family who are tech savvy, then there are few substitutes for Spark Sport for that consumer. That is, they are also the sort of person who won't be able to find an illegal stream or torrent of the matches. So, there are fewer substitutes for having Spark Sport, which means that non-tech-savvy consumers will have more inelastic demand for Spark Sport (while tech-savvy consumers will have more elastic demand).

Therefore, if Spark is price discriminating here, the optimal pricing strategy is to set a higher price for the non-tech-savvy consumers (and a lower price for tech-savvy consumers). Since Spark can't directly ask their consumers "Are you good with computers?" and price on that basis, having an additional service that only the non-teach-savvy consumers would use is a covert way of raising the price for those consumers.

Superstar effects, the Black Caps and the Silver Ferns

New Zealand has seen a bit of team sporting success over the last two weeks, with the Black Caps runners-up in the Cricket World Cup, then the Silver Ferns winning the Netball World Cup. However, the two successes were not equally remunerated, and this has created a bit of a media storm today, after it was revealed that the Black Caps will share $3 million in prize money for their second place, while the Silver Ferns receive nothing. On the surface, that seems grossly unfair, and the head of the International Netball Federation was interviewed on Radio New Zealand on that point. However, it isn't clear what the INF could possible do to address the situation.

In my ECONS102 class today, we discussed one reason that explains the discrepancy in prize money between men's cricket and women's netball - 'superstar effects' in labour markets. In the early 1980s, Sherwin Rosen identified that some workers earn much more than their peers do, or more than similar workers did in the past. [*] The reason is that top performers are paid (in part) based on the amount of value that they generate for their employer (e.g. for the national or international sporting body). If a top performer generates a lot of value, they will be paid more. This explains much of the rise in salaries over time for top sportspeople - as television (and more recently internet) viewership has grown, the value generated by a top sportsperson (in terms of the number of viewers they attract) has grown, and their salaries have grown as a result.

Cricket is a sport with a massive global following, albeit concentrated in a small number of former British colonies. The ICC has revenues in excess of US$1.5 billion per year, which is shared among the member countries. In contrast, the International Netball Federation barely breaks even, with revenues of around US$1 million per year (see page 15 of this report). The cricket players clearly generate much more value for the ICC, than the netball players generate for the INF. And they receive compensation that is commensurate with the value they generate.

The comparison across two very different sports, with different levels of spectator and sponsorship support, is clearly problematic and should be avoided. If your goal is to highlight gender differences in remuneration, it would be much better to instead concentrate on the disparities within given sports, such as the difference in prize money between men's and women's FIFA World Cups. Superstar effects can still explain this difference, but at least then you could argue for a redistribution or equalisation of incomes within a single sporting body, rather than arguing that a relatively poor sporting body should be able to match the prize money paid by a much wealthier organisation.

*****

[*] This was not a new insight, as Alfred Marshall had made a similar point as early as 1875.

Read more:

• Footballer earnings, superstar and tournament effects

Street lighting and crime in New York City

This year in Waikato's Economics Discussion Group (EDG), we've adopted a new format. At each session, we discuss some recent research paper. At the most recent session, it was this recent NBER Working Paper by Aaron Chalfin (University of Pennsylvania) and co-authors, on street lighting and crime.

This paper was interesting, because it reported on a field experiment in New York City:

The field experiment described in this paper was conducted in the Spring and Summer of 2016 in NYC. Through a unique partnership between NYPD and MOCJ, we randomized the provision of street lights to the city's public housing developments, allowing us to avoid the potential challenges that could result due to spurious time trends as well as selection bias...

In order to select developments for the study, NYPD provided a list of 80 high-priority developments based upon their elevated crime rates and perceived need for additional lighting from among the 340 NYCHA developments in NYC. From this list, we randomized 40 developments into a treatment condition that would receive new lights and 40 developments into a control condition via paired random sampling, stratifying on each development's outdoor nighttime index crime rate and size in the two years prior to the intervention; treatment developments were then randomly assigned a lighting dosage.

Most prior studies have simply compared areas with more lighting with areas with less lighting. However, there is selection bias that is not accounted for, because areas with more lighting may differ from areas with less lighting, in ways that also affect crime. This study gets around that problem because the amount of lighting increase is randomised.

Chalfin et al. looked at the impact on night-time crime, and found that:

Accounting conservatively for potential spillovers, lighting reduces outdoor nighttime index crimes by approximately 36 percent and reduces overall index crimes by approximately 4 percent in affected communities, an outcome which is likely to be cost-beneficial, should the impact of lighting persist over time.

To be clear, the effects noted above are for a doubling of street lighting intensity. So, if you double streetlight intensity, you reduce outdoor night-time crime by 36 percent.

Our discussion in the EDG session highlighted a potential source of bias in the analysis, which the authors have only partially addressed. If you increase lighting in some areas, but leave other areas darker, some criminals will relocate their criminal activities from the well-lit areas to the less-well-lit areas. This will decrease measured crime in the well-lit areas, but increase measured crime in the less-well-lit areas. If the less-well-lit areas are the control areas for your analysis (and the well-lit areas are the treatment areas), then this displacement of crime will bias your comparison of treatment and control areas upwards. In other words, the impact of street lighting on crime will be overstated in this analysis.

I hadn't picked up the full implications of that point in my reading of the paper, so well done to the EDG students. Chalfin et al. did attempt to account for 'spillover effects' in their analysis:

Estimates are reported for on-campus crimes and, in order to test for spatial spillovers, for crimes that occur within a radius of 550 feet (two standard NYC blocks) from campus. While we do not detect evidence of spillovers...

So, they also test for whether crime increase outside of the treatment areas (but within two city blocks), but the effects they find are tiny. However, this attempt to detect spillovers will only pick up spillovers into neighbouring city blocks. So, if criminals relocate further than a block away from the well-lit area, the analysis isn't going to pick it up.

Finally, they find that day-time crimes also reduced (by 25% net of spillovers) in the treatment areas, although the impact is pretty imprecisely measured and is not statistically distinguishable from zero. This last point should make us very wary of over-interpreting the importance of the results from this paper. So, while the field experiment is a good approach, and certainly an advance on the previous literature, this doesn't provide strong evidence for the impact of street lighting on crime.

[HT: Marginal Revolution]

Religious competition and witch trials

Witch trials were a common feature of late medieval Europe, up until the Renaissance. These trials were less common in the early Middle Ages, and became much less common through until modern times. Why were witch trials so common during this period?

A recent article by Peter Leeson and Jacob Russ (both George Mason University), published in The Economic Journal (ungated here), provides a compelling answer. Leeson is known for using economics to investigate somewhat unusual research questions (such as those reported in the book WTF?! An Economic Tour of the Weird, which I reviewed here, or The Invisible Hook: The Hidden Economics of Pirates).

In this case, Leeson and Russ investigated the factors associated with European witch trials over the period between 1300 and 1850 C.E. They argue that the witch trials were a response to intensified religious competition (predominantly between the Protestant and Catholic churches) following the Reformation:

Europe’s witch trials reflected non-price competition between the Catholic and Protestant churches for religious market share in confessionally contested parts of Christendom... By leveraging popular belief in witchcraft, witch-prosecutors advertised their confessional brands’ commitment and power to protect citizens from worldly manifestations of Satan’s evil...

The idea here is that the churches were competing for followers. One way that church leaders they could encourage followers to join their church was to convince them that their church was more active and successful in banishing evil. To do this, the churches engaged in witch trials. Leeson and Russ argue that witch trials should therefore be more common in areas that were more contested (that is, where there was more religious competition), and should become less common after the Treaty of Westphalia in 1648, because the Treaty permanently fixed the 'confessional geography' (that is, which areas could be claimed by Protestants, and which by Catholics), so that:

After 1648, it was no longer possible for Catholic or Protestant religious suppliers to change the denomination of any of the Empire’s territories, greatly reducing their motivation to compete.

Leeson and Russ assemble an impressive dataset of the dates and locations of European witch trials of over 43,000 people across 21 European countries, as well as the dates and locations of 424 religious conflicts (which they use as their measure of religious competition). They find that:

...each additional confessional battle is associated with an approximately 8% increase in the number of people tried for witchcraft; each additional confessional battle per million, with an approximately 11% increase in the number of people tried for witchcraft per million.

They then go on to test other theories from the literature. These included that weather shocks such as the 'Little Ice Age', or shocks to incomes, created incentives for scapegoating. Witches make handy scapegoats, so witch trials could increase when scapegoats are needed. Leeson and Russ are able to show that these alternative explanations are not as strong in predicting witch trials as religious competition is.

This is a really nice paper, underpinned by simple economic theory. For instance:

Of course, prosecuting witches was not free; it could be very expensive... The intensity of a religious supplier’s witch-trial activity thus depended on its benefit, which depended on the intensity of the religious market contestation he faced. The more intense the contestation, the higher the benefit of conducting witch trials, hence the more he would conduct.

However, one aspect was missing from the theoretical underpinning of the paper. Religion is, by its very nature, what economists call a credence good. With credence goods, some of the characteristics of the good (the 'credence characteristics') are not known to the consumer before they purchase, and they are still not known to the consumer even after they have consumed the good. [*] Health care is another example (because consumers can't know what the outcome would have been if they hadn't been treated, so they can't evaluate how much better the treatment is than the alternative).

In this case, it can be very difficult for a potential church follower to evaluate which church they should follow. Presumably, they want to follow whichever church will make them safer, 'in this life and the next'. The church leaders try to signal that their church is better through witch trials, but their signals cannot be credible, even to highly superstitious potential followers. To be credible, a signal must be costly (witch trials are costly to the church), but must be costly in a way that makes them unattractive for the lower quality church to attempt. Witch trials are not credible as a signal, because both churches presumably face the same costs of engaging in them. Instead, we see an escalation of intensity of witch trials, since whichever church engages in the most intense trial behaviour will be the church that appears to be the 'highest quality' church. Some further exploration of this point would have been interesting in the paper.

Leeson and Russ also note other periods of religious competition have led to increases in witch trials, such as the Salem witch trials in seventeenth-Century Massachusetts, where there was competition between Puritan ministers. Given that we are currently in a new period of religious competition, it makes you wonder whether we might see witch trials reappear. And indeed, that does appear to be the case, as The Economist notes (in an article about the same Leeson and Russ paper):

The persecution of vulnerable folk on trumped-up allegations of witchcraft may sound like a horror story from a history book, but the practice is on the rise in modern-day Africa. The prime victims are now children, with orphans, the disabled and albinos particularly at risk. In 2010 Unicef, a charity, estimated that 20,000 children accused of witchcraft lived on the streets of Congo’s capital, Kinshasa. Areas of intense religious competition between Christians and Muslims are hot spots. In Nigeria, for instance, Pentecostal Christian preachers fight for converts by offering protection from child witches.

The witch trials in Europe only ended after a treaty between the competing religious factions. What hope for such a treaty today?

[HT for The Economist article: Marginal Revolution]

*****

[*] Credence characteristics are therefore different from search characteristics, which the consumer can find out before purchase, or experience characteristics, which the consumer finds out after they purchase the good.

Criminalising prostitutes, or their clients?

One of the examples that I have used for many years in illustrating supply and demand is the different impacts of enforcing penalties on the sellers, vs. the buyers, of illegal drugs. Last month, The Economist had an article (maybe paywalled for you) on a similar point, related to the market for sex services:

In 1999 Sweden banned the purchase—but not the sale—of sex. A curious coalition of feminists and Christians backed the law. They argued that it would wipe out prostitution by eliminating demand, and that this would be a good thing because all sex work is exploitative...

Over the past two decades the Swedish model has been taken up by nearby Norway and Iceland, and beyond, by Canada, France, Ireland, Israel and Northern Ireland. In 2014 the European Parliament urged EU members to adopt it. Spanish lawmakers are in the process of doing so. In America politicians in Maine and Massachusetts are calling for a similar approach.

In areas where prostitution is illegal, the supply of sex services is lower than in areas where it is legal. That is because the sellers in this market face higher costs, such as the costs of fines or other penalties for engaging in an illegal activity. This is shown in the diagram below, where the supply curve is lower in the market where prostitution is illegal (S1, compared with S0). The price of sex services will be higher in the market where prostitution is illegal (P1, compared with P0), and the quantity of sex services traded will be lower (Q1, compared with Q0).

If instead the purchase, but not the sale, of sex services is illegal, then the penalties are imposed on the buyers of sex services. This reduces the demand, because the net benefit of sex services to the buyer is lower (because of the risk of fines or other penalties). This is shown in the diagram below, where the demand curve is lower (DB, compared with DA). The price of sex services will be lower in the market where buying sex services is illegal (PB, compared with PA), and the quantity of sex services traded will be lower (QB, compared with QA).

Notice that the key difference in these two markets is what happens to the price. The price is higher when selling sex is illegal, while the price is lower when buying sex is illegal (if both of these activities are illegal, then the overall effect on price will be ambiguous).

If your goal is to "wipe out prostitution" (as per The Economist article), which of these approaches is better depends on what you think happens next. If the price is higher, then the potential income from sex services is higher, so that might encourage more sellers to enter the market. That would simply shift the supply curve back towards where it started, meaning that the policy of penalising sellers has little overall impact. That suggests that penalising buyers may be a better approach to eliminating sex services. However, as The Economist article notes, that doesn't mean it is all good news:

Supporters of the Swedish model claim it protects prostitutes by giving them some power over clients, who will be worried about being shopped to the police. Prostitutes say it has the opposite effect. Face-to-face negotiations are more hurried. Kate McGrew of Sex Workers Alliance Ireland says that fewer sex workers are heeding what used to be red flags. For example, a trans woman was beaten up after taking on a client who asked if she was alone. Clients are more likely to insist on assignations in remote places. And because men refuse to reveal identifying information, prostitutes have little recourse if they are attacked.

In a study of more than 500 sex workers in France, nearly 40% said their power to negotiate prices and insist on condoms had diminished since buying sex was banned in 2016. Nearly 80% said their earnings had fallen, and almost 90% did not support the law. In Ireland violence against prostitutes shot up by almost 80% in the year after buying sex was banned, according to Ugly Mugs, a group that encourages sex workers to report attacks.

Yet the number of sex workers in Ireland who tell the police about such crimes has fallen. France has seen similar shifts.

If buyers are being penalised, then they become more cautious. They want to meet in more secluded locations, where the sellers of sex services are more isolated and unable to easily summon help if something goes wrong. This increases the risk of harm to the sellers. Penalising the buyers also reduces the price, which makes the sellers of sex services more willing to take risks (to their safety, but also to their health) to maintain their income.

Perhaps the best approach is actually to decriminalise (or legalise) both sides of the market:

Advocates of a more liberal approach point to New Zealand, which treats selling sex like any other job. An official report says that “the vast majority” of sex workers are safer and healthier since prostitution was decriminalised in 2003. Those working on the streets report that their relationship with the police has improved. Likewise, in the Australian state of New South Wales, where selling sex is legal, prostitutes’ use of condoms is higher than in other Australian states where it is banned.

If the market isn't 'hidden', then sellers of sex services can more easily be targeted for public health, they can have better relationships with the police. Buyers and sellers are safer as a result. And police resources can be diverted from policing the sex services market to other tasks. This seems like an easy win-win for all parties, and indeed has proven to be the case in New Zealand. The main argument for penalties on the buyers (or sellers) in this market is ideological.

Who are the happy economists in the UK?

In a new article published in the journal Economics Letters (gated; I don't see an ungated version online), Karen Mumford (University of York) and Cristina Sechel (University of Sheffield) used data from 443 academic economists in the UK to investigate the factors associated with job satisfaction. The method was essentially similar to happiness studies, with the question people were asked being: "Overall how satisfied are you with your job these days?", and responses measured on a scale of 1 (low) to 10 (high). If you substitute the word "job" for "life", then you have one of the standard questions for measuring life satisfaction.

Anyway, they found that:

Our most consistent results occur with the workplace characteristics. Job satisfaction is significantly related to working with proportionately more women (negative); working in London (positive); or working in a co-operative environment (positive). The latter relationship is particularly substantial. Never having had a mentor is negatively related to job satisfaction for these academics. Having a network available for professional advice is positively associated with satisfaction for women, but not for men.

The last points, about mentoring and professional networks, are important and accords with other research that has suggested the importance of mentoring for emerging female economists (for example, see this post from earlier this year). However, professional networks (one of the few differences between men and women in this study) becomes statistically insignificant if you control for self-reported salary (though this reduces the sample size down to 306, and results are not shown separately by gender for this reduced sample).

Overall, it seems that the characteristics of the workplace dominate in terms of their association with job satisfaction. Academic economists in the UK are most satisfied in a collegial and cooperative environment, with access to a supportive mentor. Who would have guessed that?

If you think you can score a point off Serena Williams, you're not purely rational

In the first week of my ECONS102 class, we discuss behavioural economics. In particular, we discuss a range of behavioural biases and heuristics that create deviations from 'purely rational' decision-making, and lead to what 2017 Nobel Prize winner Richard Thaler has termed 'quasi-rational' decision-making. One of those biases is positivity bias, or the Dunning-Kruger effect (both related to what some psychologists call self-enhancement), where people overestimate their ability.

There are lots of real-world examples of positivity bias. If you've ever watched professional darts or poker and thought, 'that doesn't look too hard; I could totally do that', then you've been subject to it. And there was an excellent example reported in Newsweek last week:

A recent YouGov poll suggests 12 percent of men, or about one in eight, think they could score a point off Serena Williams.

The poll, conducted of 1,732 adults in the UK on Friday, revealed that only 3 percent of women thought they could get one past the 23-time Grand Slam champion.

Maybe women are more realistic about their chances, but I don't think even 3 percent of anyone is going to score a point in a game against Serena Williams. According to this article, her average first serve speed at the 2014 US Open was 108 mph (174 km/h) - that's the average serve speed. And her return game is pretty good too. Unless you're a top professional tennis player, you're not scoring a point. Although, maybe if you played the game on a dark night, with no lights, and matt black tennis balls...

Positivity bias makes us more likely to believe that we can achieve things, whether or not those things are realistically achievable. You might think that would be pretty benign in its effect. So what - we have good feelings about ourselves? However, this bias can lead us to invest in activities that are underproductive (because we aren't as talented, or productive, as we think we are), wasting resources in the process. In other words, it may cause us to underestimate the opportunity costs of some activities, making us potentially worse off than if we had never attempted them. Like these for example:

Purely rational decision-makers have realistic understandings of their strengths and weaknesses, and can accurately judge their ability to successfully undertake activities. Unfortunately, we're not purely rational decision-makers, but that doesn't mean that Serena Williams will be any easier to take a point off.

The adjustment of the egg market to increasing supplier costs

This week in my ECONS102 class, we are covering supply and demand. Understanding how the market adjusts from one equilibrium to another (which we call comparative statics) is an important component of that topic. There are loads of real-world examples. For instance, in a very timely article, the New Zealand Herald reported last week:

Gilmour's, the country's largest supplier of wholesale food and beverages, is warning that the price of eggs is set to increase and the breakfast favourite may be harder to come by as egg farmers move to meet changes to the law.

In an email sent to customers today, the retailer owned by supermarket giant Foodstuffs, said "huge investment" was required by the industry to meet the Animal Welfare Code of Practice for Layer Hens which in turn would drive up the price of eggs.

"There is currently uncertainty around supply as farms struggle to gain resource consent for new production whilst other suppliers exit the supermarket sector and/or industry altogether. This is resulting in a shortage of eggs which is expected to continue over the short to medium term as the industry readjusts," the notice outlined.

Gilmours said due to higher production-related costs colony eggs would be sold at a premium as cage eggs are phased out.

Egg producers are facing increasing production costs. When costs of production increase, that results in a decrease in supply. As shown in the diagram below, the supply curve shifts up and to the left, from S0 to S1. If egg prices were to remain at the original equilibrium price (P0), then the quantity of eggs demanded (Q0) would exceed the quantity of eggs supplied (QS) at that price, because egg producers are only willing to produce QS eggs at the price of P0, after the supply curve shifts. There would be a shortage of eggs.

When there is a shortage, we expect the equilibrium price to increase. This is because some buyers, who are willing to pay the going price (P0), are missing out. Some of them will find a willing seller, and offer the seller a little bit more, in order to avoid missing out. In other words, buyers bid up the price. The result is that the price increases, until the price is restored to equilibrium, at the new (higher) equilibrium price of P1. At the new equilibrium price of P1, the quantity of eggs demanded is equal to the quantity of eggs supplied (both are equal to Q1). We can say that the market clears.

Saving the elephants, only for the hippos to be hunted

When two goods are substitutes, if the price of one of them increases or it becomes less available, consumers switch to the other. So, this report from The Telegraph (gated, ungated version from the New Zealand Herald here) last week should come as no surprise:

The elephant ivory ban is killing hippos, conservationists have said, as poachers and hunters take advantage of a loophole in the new law.

The Ivory Act, which will come into force later this year, was championed by Michael Gove, the British Environment Secretary, but conservationists argue that it puts hippos at grave risk as the import of their tusks will still be legal.

Hippo ivory, which resembles that of an elephant, is being increasingly traded globally with 12,847 hippo teeth and tusks, weighing 3,326kg, bought and sold in 2018. Trade increased from 273 items in 2007 to 6,113 in 2011...

Campaigners have called on the Government to close the loophole to ensure the ban applies to all ivory-bearing animals. They have also warned that it is nearly impossible to tell whether a tusk is from a hippopotamus that was slaughtered recently or many years ago, and whether it was poached or legally killed.

Will Travers, president of the Born Free foundation, said authorities were "shifting pressure" on to hippos by only banning ivory from elephants.

Banning elephant ivory doesn't completely shut off the supply of elephant ivory, but it does decrease it. This is because the costs of supplying ivory have increased (due to the penalties for supplying an illegal product). This decrease in supply shown in the diagram below, by the shift from S0 to S1. The equilibrium price of elephant ivory increases from P0 to P1.

Elephant ivory and hippo ivory are close substitutes (in the article, Will Travers says that "I sometimes can't tell the difference between different types of ivory and I've been in this for 35 years"). Elephant ivory has now become relatively more expensive, so more price-sensitive consumers switch to hippo ivory. This increases the demand for hippo ivory, as shown in the diagram below, from DA to DB. This increases the price of hippo ivory (from PA to PB) but importantly, it also increases the quantity of hippo ivory traded from QA to QB.

I have to admit, I hadn't realised the number of animals that produce ivory, including walrus, wart hogs, and narwhals (as well as elephants and hippos). But if you're going to ban ivory to save the elephants, it probably pays to ban all of it.

NZ ranks 18th in new ridiculous healthcare ranking

Last month, the New Zealand Herald reported:

New Zealand's healthcare system is ranked 18th place out of 24 countries - lagging behind a number of countries including Japan, Germany and even Australia.

UK healthcare recruiter Medical ID has ranked 24 OECD countries on their healthcare systems.

The ranking was based on the amount of GDP spent on healthcare, the number of doctors and nurses, how many hospital beds they have and the average life expectancy.

New Zealand was ranked 18th with a score of 60/100 and spending 9 per cent of its GDP on healthcare. It has 12,821 hospital beds and 62,843 doctors and nurses and the average life expectancy is 81.45.

It shared its 18th placed ranking with the UK which, despite having the 13th highest spending on healthcare, was brought down by being placed 22nd for the number of doctors and beds per capita.

This shouldn't even be news. Why? Because the ranking method is ridiculous, for a couple of reasons.

First, it is based on a mashup of both inputs (healthcare spending, doctors/nurses, and hospital beds) and outputs (life expectancy). This is basically double-counting, since inputs get turned into outputs. But also it double counts some inputs, since spending on doctors/nurses and hospital beds depends on the number of doctors/nurses and hospital beds.

When you want to know if a health system is good or not, it matters most to you what the output of the system is - does it keep people healthier, for longer? It doesn't matter so much how the health system achieves those outcomes. That is, it doesn't matter how much the health system spends, or what inputs it uses, if all you care about is whether it keeps people healthier. In that case, why not simply look at life expectancy, or even better healthy life expectancy, to work out which health system is better?

Alternatively, you might be interested in how much health outcome you get per unit of inputs (which might be per doctor/nurse hour, per dollar of spending, per hospital bed, etc.), or the amount of health inputs per unit of health outcome (for example, the cost per year of additional life expectancy). In the latter, you would be measuring the cost-effectiveness of the health system. In both cases, more output (using the same amount of inputs) is good, but more inputs (to get the same amount of output) is bad.

In contrast, in this ranking system, if two countries have the same number of doctors/nurses and hospital beds, and the same life expectancy, but one spends more than the other, the country that spends more is ranked higher. WTF? If Country A is spending more but only achieving the same outcome (the same life expectancy) as Country B, then Country A has got a worse health system. It is wasting healthcare resources, relative to Country B.

Second, and related to the previous point, higher health spending is not necessarily better. If that were true, a country could improve its ranking by simply contacting Big Pharma, and offering to pay them double for medicines.

So, it should be easy to see that this is a ranking system that is complete rubbish. I guess that's what happens when it is produced by "UK healthcare recruiter Medical ID", which has a vested interest in having a ranking system where the number of doctors and nurses, and healthcare spending, are indicators of a better healthcare system. In fact, they could instead be indicators that the healthcare system is simply wasteful.

Rational avocado thefts are on the rise

Last month the New Zealand Herald reported:

Tauranga growers Liz Pratt and Neville Cooper have caught the latest thieves on film after two separate night break-ins on June 8 and June 12.

Cooper said the sole thief caught on film on June 8 stole avocados worth about $1250...

The couple believes the thieves are selling the fruit on the black market to sushi shops, where the fruit is used immediately and can't be traced...

Police said provisional figures recorded 130 avocado thefts in the six months to last December, up from 110 on the same period of 2017.

There were 210 reported thefts in the full year to December, mainly in the Bay of Plenty, Northland and Eastern districts...

Police Senior Sergeant Alasdair Macmillan said police "have seen a rise in reporting these types of thefts in recent months".

"Avocados are a target for thieves due to availability and price," he said.

This relates to previous posts of mine on honey thefts and onion thefts. At the risk of repeating myself, Gary Becker (the 1992 Nobel Prize winner) identified that rational criminals would weigh up the benefits and costs of their actions, in his economic theory of crime (see the first chapter in this pdf).

A similar way of thinking about it is represented in the diagram below, where Q is the quantity of avocado thefts. Marginal benefit (MB) is the additional benefit of engaging in one more avocado theft. In the diagram, the marginal benefit of avocado thefts is downward sloping - the more avocado thefts a criminal engages in, the less they can sell their stolen avocado for (because it is harder to 'fence' greater quantities of stolen avocados - there are only so many sushi shops that will accept stolen avocados). Marginal cost (MC) is the additional cost of engaging in one more avocado theft. The marginal cost of avocado theft is upward sloping - the more avocado thefts a criminal engages in, the higher the opportunity costs (they have to give up more valuable alternative activities they could be engaging in, and as well, they are more likely to get caught and it becomes harder to 'fence' their stolen avocado). The 'optimal quantity' of avocado thefts (from the perspective of the thief!) occurs where MB meets MC, at Q* avocado thefts. If the criminal engages in more than Q* thefts (e.g. at Q2), then the extra benefit (MB) is less than the extra cost (MC), making them worse off. If the criminal engages in fewer than Q* thefts (e.g. at Q1), then the extra benefit (MB) is more than the extra cost (MC), so conducting one more theft would make them better off.

Now consider what happens in this model when the value of avocados increases. The benefits of avocado crime increase. As shown in the diagram below, this shifts the MB curve to the right (from MB0 to MB1), and increase the optimal quantity of avocado thefts by criminals from Q0 to Q1. Avocado thefts increase.

It is incentives that lead to an increase in avocado theft, and avocado theft can also be reduced by changing the incentives. The New Zealand Herald article gives some suggestions:

Pratt and Cooper had four thefts last year, prompting them to spend about $2500 on security cameras and $1700 on electric fences along all their road frontages.

"Up to then we hadn't had any trouble at all," Cooper said.

"Other growers have had similar experiences. I was just talking to one the other day, he's just bought 10 of them [cameras] to put around his orchard. He's had a lot of break-ins."

He said he had heard that at least one sushi shop owner had been prosecuted for receiving stolen avocados. But that didn't seem to have stopped the practice...

[Police recommended that] "Orchardists can help prevent thefts by taking action to secure their properties and crops. Measures include installing boundary fences, and CCTV and hidden cameras to catch offenders.

"Such measures can be highly effective, and the information captured through CCTV can be extremely helpful as the more information residents can pass on to police, the more likely it is that we can make an arrest."

Installing CCTV and security fences increase the marginal cost of engaging in avocado theft, shifting the MC curve up and to the left. The optimal quantity of avocado theft will reduce. Prosecuting sushi shop owners who receive stolen avocados will make shop owners less willing to receive stolen avocados, reducing the marginal benefit of avocado theft. This shifts the marginal benefit curve down and to the left, and decreases the optimal quantity of avocado theft.

Read more:

• Why honey thefts are on the rise
• When avocado demand outstrips supply
• Rational onion theft

For economists, talking with sociologists

Have you ever (as an economist or economics student) found yourself talking to sociologists, and wondering what on earth they are talking about? Sometimes they seem to be speaking a foreign language. Maybe they are. Or maybe, they can't understand you and you wish you could say things in a way that the sociologists could understand?

Fortunately, there is a solution. Back in 1990, Jeffrey Smith and Kermit Daniel (both PhD students at the University of Chicago) compiled the Economics/Sociology Phrase Book:

...to help economists adjust their way of speaking in a manner that will make it comprehensible to Sociologists.

Why sociologists? Smith and Daniel explain that:

We chose Sociologists rather than Political Scientists because the latter tend to be unpleasant, emaciated people with glazed eyes, while Sociologists are often entertaining and cute. Unlike Anthropologists, they can be invited to parties without much worry for the safety of the silverware, and their rhetoric, when treated like background music, has a pleasant, lyrical rhythm.

The phrase book contains very helpful translations, such as the sociologists' use of "is correlated with", "determines", and "is caused by", all of which translate to economists as "is correlated with". Harsh, but fair. Enjoy!