Interest rate hikes don't make houses more affordable

"Never reason from a price change." This is a very useful dictum, coined by Scott Sumner, that tells people to ask why a price change happened before they conclude it will have a given set of effects. Typically, a fall in house prices makes houses more affordable; therefore if interest rates go up and house prices fall, houses are more affordable, right? Wrong.

Think why house prices are falling. House prices are falling because the demand for houses is falling. The demand for houses is falling because higher (market) interest rates mean that for any given mortgage loan, the monthly mortgage servicing cost is higher, and the total you pay over the course of a loan is higher. Any given house price is less affordable for a prospective homebuyer. Can housing be more affordable because it's less affordable?

Yes, prices fall when interest rates rise, or, more specifically, when expected market real interest rates for the next forty years or so rise. But they fall precisely because these higher rates make mortgage loans more expensive and thus make housing at any given price less affordable. It falls to the point where it is about as affordable as before in terms of expected total repayment compared to expected earnings and wealth.

The case is slightly different with crime, graffiti, pollution or other disamenities. They do make specific houses cheaper, through reducing demand for them. But it doesn't improve the affordability of any given quality of flat or house. It just makes them all worse, which makes them cheaper, like if you decreed that four of the slices of a pack of bread had to come mouldy.

Land value tax advocates claim that a land value tax makes housing more affordable. But house prices rapidly adjust to take account of amenities, transport links, and expected taxes that the property owner will pay. These "capitalise in" so to speak. If you add £50,000 of expected taxes to a property (or, precisely, a net present value of that) its price will go down about that much. But it doesn't make the house more affordable! You pay £50,000 less for the house and £50,000 more in tax. No change. Once again, reasoning from the price change leads us astray.

What could make housing more affordable? Rent controls do work in the narrow sense of making rents cheaper. But without incredibly clever rules and micromanagement they're likely to have lots of side costs that end up frustrating the original goal: reduced quality, long waiting lists, rationing by other mechanisms than money, even lower construction, fewer new rental units, and switching units away from rental. Subsidies like housing benefit can work, but where housing supply is highly restricted, they bid up rents, going into landlord pockets. Help to Buy is an even worse version of this. They help those who get subsidies, but cost those who don't, as well as those paying.

Social housing can make housing cheaper. But where is the price reduction coming from? Either it's a subsidy from the government, and has all the problems of housing benefit, but is even stickier—once you've got a sweet deal you won't move. Or it's coming through steamrollering regulations and restrictions that stop private builders—well why not just let them do it? It's not a magic bullet, but a special way of applying other bullets. If council tenants in central London could, many of them would sublet their apartment, live further out, and pocket the difference. We can cut out the middleman by giving benefits in cash, not in kind.

All these clever solutions are dancing around the only thing that will dent affordability without costly side-effects, picking favoured groups, handing out to landlord, or devastating quality: increasing supply. If we regulated cars like housing there would be a cottage industry of commentators telling us that car banking was the cause of high car prices, or that an oligopoly of only 15 car companies caused them, or that we had a brick shortage. But these are actually outcomes of our system, not causes.

Flatten out the housing supply curve and when housing vacancies get low, prices get high, and rents become unaffordable you see builders fill the gap until these fall to the point they can’t profit. It’s happened before and could happen again.

Detaching the case for a basic income from post-scarcity nonsense

A popular argument for the introduction of a Universal Basic Income is that it will become necessary in the face of increasing automation. As machines can perform ever more tasks, the demand for human labour will fall, inevitably resulting in increasing unemployment. Taken to a purported logical conclusion, at some point the economy can be entirely automated removing the need for work entirely even as cheap mass production continues, resulting in post-scarcity.  Without some other major reform, such as Universal Basic Services, proponents suggest a UBI is necessary to ensure everyone can have their needs met in a world where working for an income is impossible.

This is somewhat fantastical. Post-scarcity may or may not be theoretically possible. Even if it were, though, it is not something policy should be concerned with in the here and now. Current technology and realistic projections suggest that although a significant number of jobs are at risk (around 30% in the UK), we are very far from automating all jobs.

It would be extremely difficult to successfully, let alone efficiently, automate certain jobs, particularly those requiring creativity, empathy, or a human face. This applies across a range of industries and professions. In many cases, AI will serve as a tool for workers, rather than their replacement. Just because doctors can use diagnostic machines, doesn’t mean the doctor is redundant.  In any case, the process will be incremental and gradual, rather than there being a shock moment where humans are suddenly economically redundant.

It is also probably false that the number of jobs will fall significantly. The myth that innovation destroys jobs has popped up intermittently at least since the Luddites raged against the Spinning Jenny. What actually happens is that whilst certain jobs are destroyed new ones replace them. The total number of jobs is unaffected. Ultimately, arguments to the contrary are reducible to the lump of labour fallacy.

There’s no reason to think that even mass automation would be any different. Although demand for assembly line or fast food workers may fall, demand for coders will increase. Further, new, and maybe unpredictable, demands will develop as existing needs and wants are satisfied at lower cost, leading to the expansion of (perhaps surprising) industries and the creation of new, yet to be conceived, ones. Consider that we no longer need switchboard operators whilst no one in 1950 had much interest in having a smartphone.

This does not undermine the case for a UBI. Even if automation were not happening, it would be part of a desirable reform of welfare. It would be more efficient to replace the current system of benefits with a UBI, especially if it were in the form of a Negative Income Tax. Going further, it would be desirable to replace as much state welfare provision as possible with simply giving people money. Asides from removing layers of bureaucracy it would allow for greater autonomy and responsibility than direct state provision.

Given that (some) automation is happening and that the labour market has been disrupted by developments like the gig economy, there is an additional impetus for a UBI. There are distinctions between contemporary automation and historic disruptive innovations. The speed with of job creation and destruction has increased (consider the recent emergence of cab hailing apps and the prospect of driverless cars in the not-too-distant future).  

The difference between jobs may have increased. Where labour is increasingly specialised, the difference between skills needed to perform different jobs has increased. Switching from one unskilled/semi-skilled position to another is (comparatively) easier and cheaper than switching from a semi-skilled or skilled position to a, perhaps radically, different skilled role. Whilst a former hand weaver required training to be able to work machinery, this could be done relatively straightforwardly and on-the-job. Transitioning from driving a taxi to coding requires more formal and theoretical training.

If there is a real threat from automation, it is the prospect of higher structural unemployment. One “solution” is stifling innovation, whether through attacking the gig economy or “taxing robots”. This would be unfortunate and unnecessary. Whilst other policies, such as updating education and expanding it for adults, may have a role, a UBI would be useful. Especially if it is structured as a NIT, ensuring that extra work always pays, the risks to both the individual worker and the economy of long-term unemployment can be mitigated, without undermining innovation and labour market flexibility.

Bookshop owners make very little - Adam Smith explains why

Running a physical bookshop these days is not, as we all know, a path to great riches. In fact, it's more akin to working for something less than minimum wage. At which point we've two points we might want to make.

The first being that we do and should run the economy in the interests of consumers, not producers. If there is now some less costly, more efficient, manner of providing reading material to the people then we positively desire that that flourish. For that means we are being economic with our scarce resources and we can have both the reading and more of something else. 

The second being that we can work out why people do still run such bookshops:

When I last toured indie bookshops two years ago, there was ebullience at the peaking of Kindle sales and popular revulsion at Amazon’s tax arrangements. But no conventional economist could grasp how 900 indies are still in business. They are, because so much bookselling is done out of love. That’s wonderful, but the rest of us – and publishers producing special editions – must love them back.

Absolutely every conventional economist on the planet could and can explain that. Humans do not maximise income, they don't maximise cash, profits are not the be all and end all of life. We maximise utility. As Adam Smith pointed out those 241 years ago: 

THE five following are the principal circumstances which, so far as I have been able to observe, make up for a small pecuniary gain in some employments, and counterbalance a great one in others: first, the agreeableness or disagreeableness of the employments themselves; secondly, the easiness and cheapness, or the difficulty and expense of learning them; thirdly, the constancy or inconstancy of employment in them; fourthly, the small or great trust which must be reposed in those who exercise them; and, fifthly, the probability or improbability of success in them.

A substantial number of people find their utility maximised, even if not their incomes, by running bookshops. And why shouldn't people do what they find makes them happy? 

As we really must insist running a bookshop earns little money not despite people loving doing so but because.

Theories - yes, even sociological ones - do have to be based upon some connection with reality, however tenuous

The world is wallowing in something called affluenza apparently. What you and I might call that hugely welcome climb up out of the absolute poverty which has historically afflicted the human race is instead the rapine of the planet in pursuit of trivial desires, transient pleasures.

Hmm, well, OK, possibly so. However, we do rather insist that any such analysis needs to make some connection with the observable reality outside the windows of those ivory towers. We might even be something appalling, right wing even, to so insist but there we are, we do.

Which gives us this headscratcher:

It is physically impossible for the production of stuff to grow exponentially for another thousand years. It’s probably impossible for it to grow exponentially for another hundred. And if the world is to avoid dangerous climate change, the trajectory of human consumption will need to change radically in the coming decade. It’s not complicated. Everyone knows that we need to change direction; the debate is about the timing.

Consider the following. Billions of tonnes of food are thrown away each year because fruit has spots on it, because leafy vegetables show signs of snails, or because producers put misleading “best before” dates on their packaging. Billions of tonnes of oil are transformed into plastic bottles, which, while lasting for thousands of years, are intended to be used once and then thrown away.

We do agree, entirely, with the idea that infinite physical expansion in a finite physical space is not possible. However, we also prepared an entire little book on the subject of how pressing those limits were.

Which leads us to a little consternation at the idea that billions of tonnes of oil are transformed into plastic bottles. For we cannot work out how anyone could possibly believe that.

Not that these figures are accurate but they'll do. Global oil consumption is of the order of 100 million barrels a day, 35 billion barrels a year. Some, -ish, 5 billion tonnes a year. Billions in the claim we take to be two or more. And we're really very certain indeed that 20% or more of global oil consumption does not go into making plastic bottles for water.  

One closer, but still hugely exaggerated (it is a measure of the production and transport of bottled water, not the making of the bottles), claim is one third of one percent of US energy usage. The equivalent, even including the energy costs of the entire system, of something more like 35 million barrels for that country. The feedstock for the bottles tends not to come from oil anyway, 85% of it from the bits of natural gas that can't be shoved down the pipelines, the other 15% from the detritus of oil refining that we can't put into cars and planes.

That initial claim is therefore nonsense. The very best we could say about it is that they've missed three zeros on their numbers and being out by three orders of magnitude really doesn't inspire confidence. A slightly less charitable reading would be that they've simply no idea of the subject under discussion. Or possibly just have no general idea of the size of the world and the quantities in it.

Very much more importantly though if they're going to get simple things like this wrong then what of the more complex parts of their analysis? Like the existence of this affluenza thing in the first place?

Sadly, Andrew Rawnsley won't understand why we agree with him

Andrew Rawnsley gets part of this right. It is indeed true that the 1930s were the last time that private housebuilding was putting up those 300,000 houses a year we think we need, or at least we'd like to have. However, he entirely misses the why of how this was so. Even though he links to a piece in his own paper which points it out very clearly. Concerning Neville Chamberlain:

The interesting question for today is how such a failure became prime minister in the first place. The answer is that he was once a success. He rose to the top on the back of a great reputation as a Tory social reformer. One thing he was particularly good at was housing. Planning for housing. Improving housing. Promoting social housing. Stimulating housebuilding by the private sector. He made his national name in the 1920s as health minister, a position he used to revolutionise planning, expand provision for the poor and get more homes built. His preoccupation with bricks and mortar began as mayor of Birmingham and continued when he was chancellor. The number of houses built during his time at the Treasury rose dramatically. Many of them are the 1930s semi-detached homes that still put a roof over the head of hundreds of thousands of people, particularly around London and southern England.

I commend that successful Chamberlain to Theresa May and Philip Hammond. As they wrangle with each other and cabinet colleagues about what to do with next month’s budget, the prime minister and chancellor should emulate what the Tory with the brolly did about housing.

We agree entirely, of course we do. But Rawnsley is insistent that this means government must be doing something about it which isn't the lesson to be taken at all. For as he links to, this is the reason it all worked back then:

Houses were cheap because the supply of land for housing was very elastic, which in turn meant that there was no incentive for developers to sit on large land banks. Underpinning the availability of land for house-building was an almost complete absence of land-use planning restrictions which applied to only about 75,000 acres in 1932; the draconian provisions of the 1947 Town and Country Planning Act were still to come.

In the 1930s those private housebuilders were free to build houses people wanted to live in, where they wanted to live. Absent those draconian provisions of the Town and Country Planning Act. As we've been saying for many a year now the answer is thus obvious - get rid of those draconian restrictions by blowing up the Town and Country Planning Act and successors.

Rawnsley's mistake, one shared with all too many others, is to think that the solution to a problem is that government must do something. When all too often the solution is that government do less, or stop doing something. As here, when government stops restricting housebuilding then more houses will be built.

And we do want more houses built don't we? Houses people want to live in, where people want to live? Why wouldn't we use the method which managed to achieve this last time?

This story about pensions isn't quite what it seems

Britain's pensions system is appalling apparently:

British workers can expect among the worst pensions in the developed world, according to a report from investment bank UBS, which compared the retirement outlook for a 50-year-old woman in major cities across the globe.

UBS calculated how much a country’s basic state pension, plus “mandated” pensions such as the UK’s automatic enrolment scheme and Australia’s “superannuation” scheme, would pay out as a proportion of the income of a 50-year-old “average Jane” living in the capital city.

This doesn't actually say anything at all about what pension can be expected. It only says something about the mandatory, state and other, pension that will be accumulated. Thus a system which deliberately and specifically encourages private pensions saving will do badly by this measure.

When we include such private pensions savings the UK comes second equal apparently.

Just to reiterate the point:

Private pension arrangements were not included in the study, which looked purely at mandatory arrangements in each country – in other words, the compulsory amounts that employers must set aside, plus the payouts from state pensions.

Why would a pension system be bad just because people voluntarily save for their own old age? 

There is an amusement here as well. For in all the wealth statistics (and yes, this is standard, entirely so) private pensions savings are counted as wealth. State pensions, including those owed to state employees but which are pay as you go rather than funded, are not counted as wealth. The argument being that governments can and do change their minds so we cannot regard such promised sums as being secure.

Thus pensions systems which use the private savings method, those which build real owned wealth, count as bad pensions systems. Ones which rely upon the caprice of future governments are good ones. Well, is that amusing, surprising or despicable?

Your call there.

We thoroughly approve of all of these community energy projects

Admittedly, we don't enthuse over them for the same reasons many others o but there's still something wondrous going on.

A wave of new publicly owned companies is taking on the big six energy suppliers, as local authorities search out new revenue and seek to restore faith in public services and tackle fuel poverty.

Publicly owned doesn't enthuse us, that faith in public services similarly.

No shareholders to worry about....(...)...The driving forces for these councils stepping into the complex, heavily regulated energy space are largely twofold. One is the need to create a new revenue stream in a time of austerity,

That it is a public revenue stream rather than payment to shareholders doesn't mean that the users of the energy aren't being charged to produce whichever of those two now, does it?

 Bristol Energy is biggest publicly owned energy supplier, with about 110,000 customers; Robin Hood has just over 100,000. Both have created more than 100 jobs locally, but neither has yet recovered their start-up costs.

As we've noted before covering those start up costs is rather the point of capitalists. Who lose everything if it doesn't work out - as the vast majority of business adventures don't.

However, as we say, we do thoroughly approve of the basic idea here:

Part of the reason councils are getting into energy is the barriers to market are not as great as they once were...(...)...“This is high-risk stuff. The sector is complicated. You [the council] probably don’t have the resources. You’ve probably underestimated the costs. And that’s with costs of entry falling dramatically,” he said.

Falling costs of entry can only have good effects upon the market as a whole. Major technological, even social, changes occur precisely because of the entry and exit of firms from a market. Thus a thing that we'd very much like, near a necessary precondition in fact, is ease of entry into the market. This has now been made much easier. Excellent.

That is, we've made market entry easier, people are entering the market - in whatever guise, more market players is still a good idea - and so at least one of the things we need to do about Britain's energy market has already been done. Isn't that great?  

The Great French Butter Crisis of 2017

Another entry into the anecdata bank convincing us that the last Frenchman who understood economics was Frederic Bastiat. France is suffering from a butter shortage, the croissant makers just cannot get enough of the stuff, the supermarket shelves are empty.

Quite why doesn't particularly matter - whether demand has risen or supply fallen doesn't make a difference here. What does matter is the reaction so far:

French consumers have not yet seen butter prices rise at the checkout because supermarket groups fix their prices once a year.

Sigh.

Which is, of course, to entirely miss the point of a market economy with a price system. To be able to vary prices so that supply and demand match.

All that is necessary to end the Butter Crisis is for the price of butter to change. On the commercial and wholesale side this is happening, industrial supplies have had price movements. But for something like butter retail demand is an important part of the whole. So, what happens if that retail price doesn't change? Well, facing those higher wholesale costs the supermarkets aren't exactly going to rush out to purchase stocks. Nor, given the higher prices available elsewhere will producers produce retail sized portions, will they? 

Further, facing no price rise themselves consumers won't substitute away from butter. Off to lard, or goose fat, margarine, or even allow that toast to go commando under the confiture.

When either supply or demand change it's entirely madness not to allow prices to change, for the point and purpose of the price system is to balance supply and demand as they change.

What the heck are they using as economics textbooks over there? 

We should be careful glorifying Denmark

Last week, Denmark was mentioned in a American TV debate between Bernie Sanders and Ted Cruz. Instinctively, as a Dane, I always feel a sense of pride whenever Denmark is mentioned on the big stage. However, in this instance I would like to take a step back and look at the facts because Denmark is far from the utopia politicians and pundits often claim it is.

Bernie Sanders is one of the advocates for implementing a welfare model similar to that of Denmark and for a good reason. In many ways, Denmark looks to be a very attractive country, which it is (except for the weather), but not always for the reasons stated in the public. Especially in this TV debate, Sanders mentions he is ready to raise taxes for every American in order to fund a system similar to that of Denmark. However, as it is mentioned by the moderators, Denmark pays the highest taxes of all developed countries in the world. Now, I don’t believe every country is able to raise their taxes the way Denmark has, unless they share similar values. A combination of strong work ethics and  high trust in one another is what have made the welfare state possible. This is not something that has happened concurrently with the creation of the welfare state. Actually, this predates the welfare state and is probably what has been the main factor for upholding it over time as well. Nima Sanadaji finds in his book, Debunking Utopia, that the explanation lies in the Danes and the Scandinavian work ethic, among other things.

Sanadaji also makes the claim that Scandinavians are actually doing better in America than their peers back home. For example, he presents a statistic showing the GDP per capita for the Scandinavian countries and their peers in America. The average GDP per capita for Danish Americans are $70,925 compared to $45,697 for Denmark. That is quite a difference that potentially comes from the fact that the features mentioned above didn’t just disappear when Danes travelled to America between the 1850s and 1910s. Something might therefore suggest that Danes aren’t doing well because of the welfare system. Rather, Danes appear to thrive better in the American economy because of the social institutions instilled in the Danish way of life. Milton Friedman was once confronted on this matter by a Scandinavian economist stating that there is no poverty in the Scandinavian countries. To that Friedman responded with “that's interesting, because in America among Scandinavians, we have no poverty either”.

This leads me onto my next point. Quite often it is argued that Denmark became wealthy because of the welfare system, not in spite of. But this is simply not true. Denmark was wealthy long before the welfare state came about. Combined with its social institutions, Denmark's early adoption of property rights led to a high degree of personal independence among Danes – a solid foundation for creating a wealth society.. Actually, Denmark was for a long period a low tax country and at one point in history its levels of taxation was on a par with the United States - until the middle of the 1960s where the development of the Danish welfare state began. The welfare state merely came about because of the wealth created before.

The above graph is in Danish, BNP translates as GDP and shows combined taxes.

In addition, Denmark has been able to uphold the welfare state, not only because of the social institutions, but because it has a capitalistic economy. Year after year, Denmark is among the most economically free countries in the world. Each year, the Fraser Institute measures the economic freedom of several countries - this year 159 countries are measured. They measure the economic freedom of the countries on a scale from 1 to 10 in five different categories where 10 is completely free: Size of government, legal system and property, sound money, freedom to trade internationally and regulation. This year Denmark ranked 15th overall - just four spots lower than the United States which ranked 11th. However, the one thing that always drags Denmark down the rankings is ‘the size of government’. Therefore, if we are to exclude the other categories and only look at ‘the size of government’, then Denmark suddenly drops down the ranks to a whopping 154th out of 159 countries. If we disregard ‘size of government’ completely and calculate their rankings based on the four remaining categories we come up with a almost completely different ranking:

As the picture shows, Denmark makes a giant leap towards the top landing on an overall 6th place. United States for instance drops two places. Denmark’s large levels of redistribution aside, Denmark has a very free economy, which perhaps is one reason it can afford so much redistribution in the first place.

The Fraser Institute describes what ‘the size of government’ measures in the following way:

Taken together, the four components of Area 1 measure the degree to which a country relies on personal choice and markets rather than government budgets and political decision-making.

Denmark obviously is one of the lowest ranking countries in this regard and a lot of the Danish population relies on the government in order to make a living. Such a large proportion of the population reliant on government for income means high and distortionary taxes, this means a smaller reward for extra economic activity and thus a smaller incentive to take on more work – resulting in a reduction in economic freedom. In addition, it is stated in the freedom index that an economy with a high politicisation of the redistribution of societal goods there typically tends to be a higher level of hostility towards immigration as well.

Just to give you a gist of how things are in Denmark: Overall, 2.1 million of the Danish population in one way or another get their income provided by the government. That means that for every 101 full time worker, there is 100 people who get government benefits. How will they react when a massive influx of immigrants happen? Because of the misperception that immigrants pose a threat by going on welfare, they will consider them as free riders who don’t contribute to the public purse which ultimately leads to welfare chauvinism.

This leads to several consequences of which I will mention a couple. First of all, immigrants have been shown to be beneficial for low-skilled workers and it could therefore lead those people to be worse off than they potentially could have been with immigrants coming to the country. Additionally, welfare chauvinism, among other things, has in Denmark led the anti-immigrant party, the Danish People’s Party, to gain 21.1 percent of the vote and thereby become the second biggest party in Denmark. Because of the way the Danish system works, their massive influence have made it difficult for the current liberal-conservative government to actually get through with some of their liberal reforms trying to cut spending in the public sector as well as lowering the marginal tax and taxes in general.

Denmark is indeed in many ways great, but we should be careful glorifying it, especially if we do it based on false information.

Cracking down on Craigslist put sex workers at risk

When I lived in the U.S. last year, I appeared on Al-Jazeera to criticize the U.S. government’s repeated attacks on Backpage.com—a classified ads site that also featured postings for erotic services. In January 2017, it shut down its adult section following years of intense government pressure, including a Sheriff threatening credit card companies providing services to Backpage and Congressional hearings. While opponents of the site erroneously argued that it facilitated child sex trafficking, internet freedom groups, several libertarian organizations and some anti-trafficking advocates united with sex workers to condemn the government’s bullying of Backpage.

Defenders of the site argued that the shutdown made sex workers less safe, created a chilling effect on free expression, and undermined anti-trafficking efforts by removing Backpage as a valuable resource for law enforcement. A working paper published last month by researchers at Baylor and West Virginia universities provides new evidence that classified ad sites like Backpage do indeed play a significant role in reducing violence against women—especially female sex workers.

The paper—authored by Scott Cunningham, Gregory DeAngelo, and John Tripp—focuses on Craigslist: a similar platform to Backpage that included an ‘Adult’ ads section in the U.S. from 2002 until 2010, when it was also shut down due to mounting pressure from the government. By analyzing female homicide rates over time in U.S. cities after Craigslist introduced an ‘Erotic Services’ (ERS) section and comparing them to cities without such a section at the time, the researchers were able to estimate the effect of an ERS section on women’s safety. They found that cities that got ERS saw their female homicide trend slide afterwards, averaging out to a rate 17.4% lower than control cities over the same period.

They argue that the introduction of an ERS section was effectively random from the perspective of Craigslist: in some places ERS was added long after Craigslist launched there and in others it was included at launch, it was unpublicised, and female violence was on similar trends in ERS and non-ERS cities before launch. Thus they judge that these drops are likely down to Craigslist and not other unobserved factors.

Why might this be the case? One explanation is that adult sections of sites like Craigslist and Backpage make it easier for sex workers to find clients online while working indoors, prompting a shift from more dangerous street-based sex work to indoor alternatives. Working indoors allows sex workers to screen their clients more thoroughly, and soliciting online “may have led to greater deterrence of client violence through the creation of a digital fingerprint that made detection of criminal offenses more likely”.

Another possibility is that Craigslist’s ERS section gave sex workers a viable alternative to working with potentially violent managers. The study provides evidence to support the claim that sex workers transitioned from agency work to independent work due to the introduction of a Craigslist ERS section:

Our second dataset utilizes reviews from The Erotic Review, which is a reputation website (similar to Yelp.com), and one of the largest prostitution websites in the United States (Cunningham and Kendall, 2016). Clients use The Erotic Review to share detailed reviews of prostitutes. We use these data to measure whether a prostitute worked for an agency or independently…In the first 10 months [after ERS section introduction], the probability a prostitute was independent rose 6.5 percent, which is 12 percent of the mean. This effect persisted in the long run, as evidenced by the positive and statistically significant 10+ month coefficient.

A potential challenge to the study’s headline figure (a 17.4% reduction in female homicides) is that it seems “to exceed the number of prostitute deaths by several orders of magnitude”. However, this could be due to a significant underestimate of the share of female homicides attributable to prostitution. The data in this area is extremely sparse. The paper references a 2006 study on a small number of cities that estimates “between the years 1981 and 2002, 2.7 percent of all female homicide victims in the United States were attributed to prostitution”. This is likely to be an extremely conservative estimate; the illegality of prostitution means that at the time of data collection, U.S. law enforcement are often unaware of whether female homicide victims were sex workers.

With a misguided online porn crackdown looming in the UK, there are legitimate worries that future governments will also adopt a more American approach to websites hosting adult ads. If that happens, at least this study has provided economic evidence to support what sex workers have been arguing for years—that online platforms like Craigslist and Backpage help make women safer.