Bryan Caplan's Blog, page 64

Koyama Responds on State Capacity, by Bryan Caplan

My colleague Mark Koyama responds to my recent critique of state capacity research.  Reprinted with his permission.  Enjoy!

Hi Bryan,

 

My response is below. I agree
with Noel on measures of rule of law.  When one looks at how these indexes
are constructed it is hard to put too much faith in them.  At least
estimates of tax revenues collected are a "thing". 

Bryan is kind enough to link to
the article Noel Johnson and I wrote on state capacity.  Bryan's critique
is aimed less at the literature per se (at least or our survey of it) but at
the concept itself.  Bryan argues that a sleight of hand is
involved. 

As I understand him, Bryan is
saying we don't know that state capacity causes outcome x (say growth). This is
because state capacity is itself correlated with many characteristics that
might be associated with growth. 

This is no doubt correct.
 It is why our survey article is also a call for more historical research
on the origins of highly functional states. But this is also true of all
research in the empirical study of economic growth. Moreover, it neglects the
extent to which recent research including papers by Melissa
Dell, Sascha
Becker, and their coauthors do undercover specific causal channels.

In absence of an RCT, how much
weight should we put on these findings?  When confronted with any one
paper, a skeptic can always claim "I believe it is x rather than state
capacity" where x is any of one of many variables. In my view, this is a
potential issue for any single paper, but much less of an issue for a
literature, if many papers find comparable evidence that state capacity is
associated with growth and development and that this link is plausibly casual,
then this is something we should take seriously.

If someone finds counter
evidence, this is great too. If the evidence is persuasive they can probably
get a great publication.

Bryan is a stickler for
conceptual clarity.  Hence I understand why he is dissatisfied by a
concept like state capacity that aggregates a number of different things. 
Other scholars have suggested alternatives.  Michael Mann writes of infrastructure
power.  Mark Dinecco uses the term "Effective
States". Nevertheless, state capacity is the most widely used term and
one that we have inherited. Rather than rejecting it and using alternative
bespoke concepts, I prefer to refine and clarify it.  

Moreover, I think Bryan maybe
looking for something that may not exist: the cause of economic
development.

Approaching the question of state
capacity from the perspective of economic history gives me a different
perspective.  Spurts of economic growth occurred due to trade, commerce,
markets, and division of labor in many preindustrial settings as my colleague
Jack Goldstone has argued
(ancient Greece, Rome, Song China, the Italian city-states).  The
transition to sustained modern economic growth occurred after the Industrial
Revolution largely due to the combination of innovation & invention with
sophisticated market institutions (see Mokyr and McCloskey).  

So state capacity didn't cause
growth.  But state capacity can help to explain why British prosperity was
not destroyed by warfare as occurred in Renaissance Italy or Song China.

Moreover, when it comes to the
importation of new technology and institutions why were some non-Western
countries able to do this successfully (such as Japan) where others were not
(such as China)? 

Bryan suggests that our answer to
this question is tautological. But I've written a paper
with Chiaki Moriguchi and Tuan-Hwee Sng on precisely this question. Suffice to
say that the argument doesn't rest on the simple observation that Bryan quotes
but on abundant qualitative evidence drawn from the historical literature.

If the literature is a fad then
it should be possible to publish a decisive scholarly takedown and the returns
from doing so should be substantial.  

(0 COMMENTS)

The Underbelly of State Capacity, by Bryan Caplan

Here are some highlights from Johnson and Koyama's "States and Economic Growth: Capacity and Constraints" (Explorations in Economic History, 2017).  Remember: Since this piece is a literature review, you should continue to take my critical remarks as a criticism of the literature, not as a criticism of my esteemed colleagues.

1. How do researchers actually measure state capacity?  One of the most common measures is simply per-capita taxes.

Recent work by Besley and Persson (2011) has drawn attention to an important correlation between per capita GDP and measures of state capacity, usually defined as per capita tax revenues. This positive correlation is robust and holds over a wide range of definitions of 'state strength'.

High per-capita taxes plainly indicates capacity for collecting a lot of taxes; after all, you can't do stuff beyond your capacity.  But isn't the whole point of this literature that a low-tax state could still have very high capacity?  The standard claim, after all, is that you want a "strong but limited" state: One that's capable of great things if and when they're needed. 

So what?  Well, if you believe that per-capita taxes genuinely cause growth, you've effectively switched to the simpler theory that big government is good for growth.  State capacity then becomes a red herring.  If you object, "Perhaps it's politically easier to have high taxes in a rich country, so prosperity causes taxes rather than the other way around" you should also be worried about reverse causation from prosperity to state capacity.

2. North and Weingast famously argued that the Glorious Revolution established a strong but limited government, paving the way for the Industrial Revolution.  Johnson and Koyama point readers to a long list of challenges to North-Weingast:

North and Weingast (1989) saw 1688 as making a constitutional watermark leading to the better protection of property rights and the rule of law. Subsequent research has overturned many aspects of this thesis. In the century and a half following 1688, rules and laws did become more general. But the process was saw numerous setbacks. Although as Brewer (1988) documented, the excise was modernized and bureaucratized in the period following 1688, the rest of the British state remained patrimonial in organization and modernization was a slow and gradual process--an open examination for entrance into the civil service was only introduced in 1870. In the sphere of business organization, the ability to form joint-stock companies was severely limited by the Bubble Act of 1720 and incorporation laws only began to be liberalized in the 1820s with the full legalization of joint-stock companies waiting until 1844 (Harris, 2000).

3. What's up with China? 

[R]ecent scholarship has established that taxes in China were low (Ma, 2011, 2012, 2013; Rosenthal and Wong, 2011; Sng, 2014; Vries, 2015; Ma and Rubin, 2016). This had positive effects as the policies of the Qing state did not impede the effectiveness of the market in goods and services (Pomeranz, 2000; Shiue and Keller, 2007; Li et al., 2013). However, many aspects of the effectiveness of the Chinese state in the early modern period remain subject to debate. It is a matter of some contention whether or not the low taxes collected by the Chinese state reflected low fiscal capacity or a reliance on Confucian ideology. Kent Deng refers to 'under governance' in terms of the reigning political ideology observing that '[h]]eavy taxation remained politically taboo' (Deng, 2015, p. 328). Rosenthal and Wong (2011) write that the 'Chinese logic for successful state maintenance ... emphasized light taxation and generally tried to avoid interfering with commerce' (174). Sng (2014), on the other hand, compellingly argues that the low amounts of tax revenue collected by the central government in Qing China reflect the fragile political equilibrium that the Qing rulers faced in governing such a large empire using premodern technology. The Qing state relied on a land tax based on a fixed amount due each year based on the value of land. The collection of this tax gave a large amount of discretion to the local officials who extracted bribes or otherwise manipulated the process for their own benefit. Sng (2014) develops a formal model which predicts that where the principal-agent problem facing the ruler was more severe, the weaker was the ruler's ability to tax. Consistent with this model, he presents evidence that taxes were significantly lower in areas further from the capital and that they declined significantly from the mid-eighteenth century onwards. Vries (2015) similarly interprets the low taxes collected by the Qing state as reflective of a low-state capacity political equilibrium.

4. Here's a seemingly unobjectionable comment on China versus Japan.

Japan also ceased to face geopolitical competition after pacification at the end of the sixteenth century and as a result there was little incentive to build a fiscal-military state during the early modern period. However, unlike China, Japan was able to centralized its governing institutions and invest in fiscal capacity once it awoke to the threat posed by the West in the mid-nineteenth century.

What's the problem?  Sleight of hand, again.  The fact that Japan "centralized its governing institutions and invested in fiscal capacity" shows that Japan was able to do so.  But it does not show that China was unable to do so - merely that it did not.  Organizations fail to do all sorts of things of which they are capable.

P.S. Bardhan's "State and Development" (Journal of Economic Literature, 2015) is a good complement to Johnson and Koyama.  Most notably, they question even more of the North-Weingast narrative:

Historically, however, England has indeed been a successful case where political centralization and pluralism have fit together. But, contrary to North, Weingast, Acemoglu, and Robinson, economic historians like Epstein (2000), Clark (2007), and Allen (2009) have expressed doubts if the economic success of England can be mostly attributed to the constitutional changes that came with the Glorious Revolution. Even some of the more recent defenders of North and Weingast, like Cox (2012) and Pincus and Robinson (2011), agree that neither cost of capital nor enforcement of property rights improved significantly after that Revolution, even though it represents an important constitutional watershed (Cox) or an institutional change shifting the balance of power from the king to the new manufacturing classes (Pincus and Robinson).

(3 COMMENTS)

Name Our Graphic Novel!, by Bryan Caplan

The non-fiction graphic novel I'm writing with Zach Weinersmith is tentatively titled All Roads Lead to Open Borders: The Science and Ethics of Immigration .  But we're considering other titles, including just Open Borders: The Science and Ethics of Immigration and Tear Down These Walls: The Science and Ethics of Open Borders.

Question: Do you have any other title suggestions?  If we use your idea, we'll honor you in print!

P.S. If you're reading this before the morning of June 7, you can still participate in the initial Twitter poll on the title.

(28 COMMENTS)

"State Capacity" is Sleight of Hand, by Bryan Caplan

Suppose I proposed the following theory of economic development:

The root cause of economic development is social capacity.  Highly-developed societies have immense social capacity, as evidenced by their countless achievements.  Backwards societies lack social capacity, as evidenced by their countless problems. 

You'd probably be unimpressed.  At first glance, this is a circular theory that explains nothing: Why are things so good in X?  Because X has great social capacity.  How do we know X has great social capacity?  Because things are so good in X!

On closer consideration, though, the social capacity theory is not utterly empty.  It makes one grand empirical prediction: Good social outcomes correlate.  Income, wealth, happiness, health, culture, entertainment, leisure, and safety will all go together.  And in point of fact, this prediction is true: Good social outcomes do tend to go together.  Saying, "Society X has good health because it has high social capacity" is comparable to, "Person A is good at learning math because he has a high IQ."  In both cases, we legitimately use good general outcomes to predict good specific outcomes.

Still, using social capacity to explain the world seems deeply unsatisfying, for two main reasons:

First, informed observers are already well-aware that good outcomes correlate.  A conceptual framework that predicts the standard findings is better than nothing, but it's hardly a triumph of the human intellect.

Second, once you understand the concept of social capacity, you can immediately articulate an obvious critique: "The question we really care about is not whether social capacity varies widely between countries, but why!"

Why bring this up?  In recent years, many social scientists - including several of my colleagues - have fallen in love with the concept of "state capacity."  GMU economic historians Noel Johnson and Mark Koyama provide a nice overview in Explorations in Economic History:

State capacity describes the ability of a state to collect taxes, enforce law and order, and provide public goods...

State capacity can be thought of as comprising two components. First, a high capacity state must be able to enforce its rules across the entirety of the territory it claims to rule (legal capacity). Second, it has to be able to garner enough tax revenues from the economy to implement its policies (fiscal capacity). State capacity then should be distinguished from either the size or the scope of the state. A state with a bloated and inefficient public sector may be comparatively ineffective at implementing policies and raising tax revenues. Furthermore, historians agree that the eighteenth century British state had high state capacity even though it played a very limited role in the economy. Similarly, state capacity requires a degree of political and legal centralization, but it should not be identified with political centralization per se. The rulers of feudal society in which many legal and fiscal choices were devolved to local lords indeed had low state capacity. But the concentration of political authority in the center may cause inefficiencies and thereby undermine state capacity...

This seems like a fine description of what state capacity researchers think they've discovered.  But I reject Johnson and Koyama's approving tone.  Read them closely.  You can't measure "state capacity" by the size or scope of government; that's a no-no.  If the public sector is "bloated and inefficient," that's not high state capacity either.  Concentration of political authority is causing inefficiencies?  Well, you just undermined state capacity in the process of trying to increase it!  To my mind, this is scarcely better than saying, "Good government is good; bad government is bad."

Matters would be different, admittedly, if the state capacity literature showed that good government is the crucial ingredient required for success.  But researchers rarely even try to show this.  Instead, they look at various societies and say, "Look at how well-run the governments in successful countries are - and look at how poorly-run the governments in unsuccessful countries are."  The casual causal insinuation is palpable.*

More careful researchers, to be fair, make the extra effort to try to explain the origins of state capacity.  Johnson and Koyama name "state antiquity," "culture," and "civil society."  But at this point, why not just ditch your premature focus on "state capacity" in favor of an open-minded exploration of social capacity?  Good government might be the crucial ingredient for success.  But maybe good government is a byproduct of wealth, trust, intelligence, freedom, or some cocktail thereof.  Or maybe there's a complex feedback loop.  We can readily verify that all these good things tend to go together.  Untangling that tangled causal web of success, in contrast, is a labor of Hercules.

If I'm right, why has "state capacity" become such a fashionable explanatory concept despite its pervasive conceptual weaknesses?  Here's a short list of stories.

1. One simple theory of development gives overwhelming credit to government.  But a theory this simple is open to devastating counter-examples like Stalinist Russia and Maoist China.  "State capacity" has the same psychological appeal without this empirical vulnerability.  Why not?  Because while we can observe states, we can only infer their "capacity."  Thus, anytime government leads to disaster, you can say, "Too bad they didn't have higher state capacity."  Anytime government leads to good results, in contrast, you can say, "Hooray for state capacity!"

2. "State capacity" sounds a lot less tautological than "social capacity," even though the latter concept is much cleaner.  "State capacity causes social success" sounds profound.  "Social capacity causes social success" - not so much.

3. "State capacity" splits the ideological difference between libertarians and statists - and thereby appeals to a broad academic spectrum.  State-sympathizers can emphasize the wonder of government.  State-skeptics can emphasize the danger of badly-run government.

4. Don't forget sheer faddishness.  Academia's full of it.

I freely admit that the state capacity literature has much to teach us.  But that's largely despite its conceptual framework - not because of it.  Readers must constantly guard against the intellectual sleight of hand that pervades this work.  While good social outcomes all tend to go together, the state capacity literature fails to show that government is the crucial factor that makes all the others possible.  Indeed, as far as I can tell, existing empirics are quite consistent with Sutton's Law that people rob banks because "that's where the money is."  Perhaps rich societies have big governments because it takes a colossal host to sustain colossal parasitism.  Think of San Francisco or New York City before you scoff!

* That's not a typo.







(5 COMMENTS)

Szasz and the Statistics of Rare Events, by Bryan Caplan

Many years ago, Thomas Szasz largely convinced me that mental illness is radically different from ordinary physical illness.  In economic terms: People with physical illness have unfavorable constraints; people with mental illness have socially disapproved preferences.  Physical illness is about what you're able to do; mental illness is about what you want to do. 

Yes, it's generally bad manners to loudly call attention to this distinction.  Even though "I can't come to your party" usually means, "I would rather do something else with my time," it's impolite to say so.  The same goes for "I can't stop drinking" or "I just can't manage to show up for work on time."

But what about really weird cases?  Perhaps 95% of all alcoholics simply value their favorite beverage more than they value their families.  But every now and then, you read a compelling first-hand account of someone who persuasively insists, "I just can't help myself."  What about multiple personalities?  Severe delusions?  Can Szasz explain those?

Maybe not.  But Alex Tabarrok's post on the statistics of rare events got me thinking. 

The CDC asked 12,870 individuals about defensive gun use over the
three samples.That's a relatively large sample but note that this means
that just 117 people reported a defensive gun use, i.e. ~1%. In
comparison, 12,656 people (98.33%) reported no use, 11 people (0.09%)
said they didn't know and 86 people (0.67%) refused to answer. People
answering surveys can be mistaken and some lie and the reasons go both
ways...

The deep problem, however, is not miscodings per se but that
miscodings of rare events are likely to be asymmetric. Since defensive
gun use is relatively uncommon under any reasonable scenario there are
many more opportunities to miscode in a way that inflates defensive gun
use than there are ways to miscode in a way that deflates defensive gun
use.

Imagine, for example, that the true rate of defensive gun use is not
1% but .1%. At the same time, imagine that 1% of all people are liars.
Thus, in a survey of 10,000 people, there will be 100 liars... Adding it up, the survey will find a defensive gun
use rate of approximately (100+10)/10000=1.1%, i.e. more than ten times
higher than the actual rate of .1%!

Notice that Alex's point generalizes readily from defensive gun use to extreme mental illness.  Suppose, for example, that you study the prevalence of multiple personalities (now called Dissociative Identity Disorder).  In a group of 10,000 people, ten insist they have it.  Taken individually, each of the ten seems credible.  But if .1% of respondents would energetically lie, the discovery of ten believable stories is perfectly consistent with the complete non-existence of the disorder. 

Why oh why though would anyone tell such a lie?  Perhaps to be the center of attention - one of the most ubiquitous of all human motives.  As Szasz puts it:

[W]hen a grisly, unsolved crime is reported by the press and the police look for the person who did it, innocent people often come forward and confess to the crime. Such a confession is never accepted on its face value as true; on the contrary, it is treated with the utmost skepticism. On the other hand, when a person lodges a psychiatric complaint against himself, it is not investigated at all.

Do the statistics of rare events prove Szasz right?  No, but they do tip the evidentiary scales further in his favor.  If X almost never happens, basic numeracy urges us to question whether the few purported cases of X are genuine - especially if many of us feel a temptation to claim X regardless of the truth.

Isn't it desperate, though, to use Tabarrok's insights to treat admittedly rare events as absolutely non-existent?  Perhaps.  But then again, isn't that just what the great David Hume did in his "Of Miracles"?

When anyone tells me, that he saw a dead man restored to life, I
immediately consider with myself, whether it be more probable, that this
person should either deceive or be deceived, or that the fact, which he
relates, should really have happened. I weigh the one miracle against
the other; and according to the superiority, which I discover, I
pronounce my decision, and always reject the greater miracle.

To my mind, a person who genuinely "can't stop drinking" is almost as miraculous as a dead man restored to life.  If you dismiss the latter, you should at least be open to dismissing the former as well.

(20 COMMENTS)

Teach My Books, by Bryan Caplan

Now that June has started, many professors are ready to update their syllabi.  While "keep it unchanged" is the easiest route, there's always room for pedagogical improvement.  Contrary to what you may have heard, I love education... as it might and ought to be.  And as a firm believer in self-promotion, I think one of the best ways to make education more lovable is to adopt one (or more!) of my books as secondary course texts.  A quick tour of my menu:

The Case Against Education ,
my latest book, is perfect for any class on the economics of education,
educational psychology, or sociology of education.  It also works well
for any class on cost-benefit analysis, information economics, public
policy, and philosophy of education.  The Case Against Education isn't just a dogged, interdisciplinary defense of the signaling model of education; it clearly presents the standard social science of education prior to criticizing it.

The Myth of the Rational Voter is perfect for any class in public choice, political economy, or political psychology.  It also works well for any class on democracy, political theory, and even intro econ.  Myth of the Rational Voter earnestly defends the anti-democratic, pro-market vision of public choice that Nancy MacLean's Democracy in Chains falsely attributes to James Buchanan.

Selfish Reasons to Have More Kids is perfect for any class on economics of the family, population economics, sociology of the family, or behavioral genetics.  It also works well for any class on population ethics, risk analysis, or demography.  SRtHMK provides a parent's-eye view of the science of nature and nurture - and hammers home the connection between how you raise your kids and how many kids you'll want to have.  And along the way, you get a lively update of Julian Simon's analysis of the neglected upsides of population.

Both The Case Against Education and The Myth of the Rational Voter are published by Princeton University Press.  If you're interested in course adoption, Julie Haenisch of PUP is happy to chat.

(2 COMMENTS)

Franckly Absurd, by Bryan Caplan

Every human being, including me, inclines to self-serving bias.  But I was genuinely taken aback by Raphael Franck's attempt to resolve our French terrorism bet in his own favor.  Why?

1. Franck counts the 150 fatalities on Germanwings Flight 9525 as terrorism, even though virtually no one else does.  When I challenged this determination, his response was: "The crash of the German flight was cheered upon by radical
Islamists. So I guess they viewed it as a success for their cause."  But radical Islamists presumably cheer every tragic event in the West, so why not count death itself as "terrorism"?

2. Franck counts virtually all people killed by French police as "rioters."  Wikipedia lists 50 people killed by French police from 2008-2012.  Franck counts 49 of these as killings of rioters.  His rationale:

For
simplicity I used the lacunary data on people who were killed by the police (it
is only this year that the France started tracking the number of individuals
killed by the police - otherwise the job is done  by leftist activitists).
Usually these people are not killed at random by the French police, they are
killed in small-scale riots that occur in the  suburbs of the major
French towns.

I agree that French police kill few people "at random."  But that hardly makes them presumptive rioters.  Perhaps they're just common criminals resisting arrest, like in virtually every other country?  This is especially absurd because standard accounts mention no fatalities for any of the major French riots in 2008-2018.

3. Franck and I bet on "the total number of deaths in France from riots and terrorism."  But Franck gives himself a further buffer by claiming that he could legitimately count deaths in former French colonies! 

For the
record, to win this bet, I did not even have to include in the final count the
French soldiers who were killed by jihadists in the former French colonies in
sub-Saharan Africa (which are only former colonies by name, g��ven that the
French army still seeks to control those areas to protect French financial and
industrial interests). But I could. And then I would also count the jihadists
that they killed.

Gee, why not count everyone killed in the former French colony of Syria, too?

4. Franck refuses arbitration:

I do
not see much point in asking for a neutral arbitrer. At the end of the day, the
neutral arbitrer will have to decide whether terrorism motivated by radical
Islam has become a common occurrence in France. For some reason, I have yet to
meet someone who is neutral when it comes to the relationship between radical
Islam and terrorism.

Actually, a neutral arbiter would only have to determine who won according to the literal terms of the bet, using standard English.  Most people can do this despite their political views.  The real story, I have to think, is that Franck doubts that any credible arbiter on Earth would take his side.

5. I can sympathize with someone who thinks that mainstream media is irredeemably biased against the truth.  But I can't sympathize with someone who makes a bet without mentioning that they will reject evidence that virtually everyone else would consider decisive.

Some my friends have suggested that I count this bet as "contested"
rather than a clear win.  If I thought there was a reasonable doubt
here, I would comply.  But frankly, Franck's efforts to salvage his side
of this bet are absurd.  And that's a word I don't use lightly.

(22 COMMENTS)

Franck on the French Terrorism Bet, by Bryan Caplan

Raphael Franck claims that he actually won our ten-year bet.  With his permission, I reprint our email exchange on the topic.

Franck

Dear Bryan,

I attach to this email an excel file which lists the number
of people who were killed in terrorist acts and riots in the past 10 years. I
undertook a thorough research job (not limited to wikipedia of course) and won
our bet because my count, there were more than 500 people who died in terrorist
acts and riots. 

Clearly, this is a lower bound estimate: the French
politicians and media have tried to minize the occurrence of terrorist acts and
prefer arguing that some crimes are not undertaken by terrorists. This is
possible because many terrorists are "lone wolves" who are just armed with knives or
ram crowds with their car. As such, the media can attribute the growing number
of attacks to mental problems. For instance, politicians and journalists have
been denying that the driver who rammed his car through a crowd in December
2014 while shouting "Allahu Akhbar" and who killed at least one
person (https://www.tendanceouest.com/actualite-88705-nantes-une-victime-succombe-la-justice-enquete-pour-assassinats.html)
was a terrorist. He was, as the French say, "d��s��quilibr��
" (i.e, unhinged).  

It was difficult to find data on victims for the many
small-scale riots that often occur in France. For simplicity I used the
lacunary data on people who were killed by the police (it is only this year
that the France started tracking the number of individuals killed by the police
- otherwise the job is done  by leftist activitists). Usually these people
are not killed at random by the French police, they are killed in small-scale
riots that occur in the  suburbs of the  major French towns. I left a
few links that document the deaths of people killed in those riots (in French,
but there is nothing that google translate cannot do).

The other thing is that I counted the Islamic militants
killed by the police in the final bodycount (listed under the column
"Shaheed") alongside with their victims (listed under the column
"Kuffar"). Like the French citizens killed by the police in small
scale riots,  I am sure you will agree that the French police used
excessive force and should have negotiated more with these militants who were
for the most part French citizens who were often born and bred in France. It is
a bit irrelevant for our bet (although completely in line with what I had
predicted) that those individuals were domestic terrorists who were motivated
by an ideology that is alien to French culture and history. 

For the record, to win this bet, I did not even have to
include in the final count the French soldiers who were killed by jihadists in
the former French colonies in sub-Saharan Africa (which are only former
colonies by name, g��ven that the French army still seeks to control those areas
to protect French financial and industrial interests). But I could. And then I
would also count the jihadists that they killed.

I am not interested in the money. I suggest you donate it to
a charity that takes care of children with special needs.

Best regards,

Raphael Franck

Caplan

As you might suspect, I find your methodology
biased. Who besides you counts the
German plane crash as "terrorism"? 
And you want to count 100% of people killed by the French police as
deaths from "riots"?

And of course deaths outside of the borders of France are
explicitly ruled out by the wording of the bet, though I'm willing, given the
wording to count the deaths of the perpetrators.

I'm open to resolution by a neutral arbiter.

Franck

Dear Bryan,

My methodology may be biased. But I have one and it is very
reasonable. 

The crash of the German flight was cheered upon by radical
Islamists. So I guess they viewed it as a success for their cause.

As for the riots, the first link in the excel file documents
that about 10 to 15 people are killed every year in France by the police; the
deaths that I counted in the dataset were therefore a subset of all the deaths
caused by the police and they occurred in the suburbs of French towns where
small level riots are a common occurrence. Hence the numbers in the excel file
are not 100% of the deaths caused by the police. 

As I wrote before, quite a few crimes in France motivated by
radical Islam may be reported as a-ideological  violent crimes because the
motivations of the attacker are hard to tell.  If anything, I did not
include in my count the March 2017 murders committed by a 31 year old Frenchman
who slit the throat of his brother and then of his father (http://www.lepoint.fr/societe/paris-l-egorgeur-du-11e-etait-fiche-s-17-03-2017-2112697_23.php).
I chose to do so,  even though this murderer had a "fiche S"
(i.e., "S file" where S stands for security and designates people
whom the French police view as potential terrorists). In fact, I was wondering
whether you would try to deny that the car ramming in Nice in July 2016 was a
terror attack because the attacker had a history of mental illness (as if
everyone who has issues decides to kill people at random, like him or like the
German pilot).

I thus stand by my earlier statement that the estimates on
terror victims that I gave are a lower bound. It is totally fine with me if you
do not want to acknowledge that my estimates are reasonable.

I do not see much point in asking for a neutral
arbitrer. At the end of the day, the neutral arbitrer will have to decide
whether terrorism motivated by radical Islam has become a common occurrence in
France. For some reason, I have yet to meet someone who is neutral when it
comes to the relationship between radical Islam and terrorism.

I am fine with your posting this conversation online. And
for the record, I oppose terrorism by radical Islamists and take very little
satisfaction in my correct predictions. After all, if more people had listened
to me, more lives could have been saved.

Best regards,

Raphael

(12 COMMENTS)

I Win My French Terrorism Bet, by Bryan Caplan

Ten years ago, I made the following bet with French Israeli professor Raphael Franck:

If the total number of deaths in France from riots and terrorism is less
than 500 between May 28, 2008 and May 28, 2018, Franck owes me $50. If
the total number is 500 or more, I owe him $50.

Over this period, the total number of deaths in France from terrorism was 265.  (Wikipedia's numbers come from the canonical Global Terrorism Database).  I cannot find any official statistics on rioting deaths.  But even the most famous riots during this period appear to have a cumulative body count of zero.  I have therefore won our bet - and by a large margin.

Prof. Franck, to be fair, disputes this determination.  With his permission, I'll soon post his alternative count in his own words.  The key disputes, as far as I can tell:

1. Franck counts the Germanwings Flight 9525 incident as terrorism, even though almost no informed observer concurs.

2. Franck counts almost all people killed by French police during this period as "rioters."

Needless to say, I maintain that my victory is beyond reasonable doubt - though I do regret my failure to pre-specify an official data source and arbiter to avoid this impasse.  I should have taken Tetlock even more seriously than I already do.

Still, my current track record now stands at 18 wins, 0 losses.  And out of all my bets I've won so far, I think we learn the most from this one. 

To see why, imagine that Franck and I had publicly disagreed about whether French terrorism and riots would be a "severe problem in the next decade."  Looking back, virtually everyone would agree that Franck was right and I was wrong.  Every incident in France would be treated as further proof of my Pollyanna blindness to the horrors of the world. 

Since we settled on a specific number, however, fair-minded spectators will now see a very different story.  Franck - a vocal pessimist on these matters - wouldn't have bet if he didn't expect the standard number to be worse than 500 - probably markedly worse.  Hence, contrary to casual observers, the pessimists were not "vindicated by events."  Far from it.

The flip side, of course, is that I would not have agreed to 500 if I didn't expect a lower number.  What actually happened stands at roughly the 80th percentile of what I expected based on long-run base rates.  But statistically speaking, what occurred remains a tiny problem.  Over a thousand people are murdered on Earth on a typical day.  From 2008-2015 (the last available year), France's murder rate ranged from 1.2 to 1.6 per 100,000.  That comes to roughly 800 murders per year in France, about 3% of them committed by terrorists. 

If ethnic Frenchmen from Breton or Normandy were responsible for a crime wave of this magnitude, would anyone outside of France even take notice?  It's hard to imagine.  Would anyone in or out of France use it to justify any form of collective punishment?  Again, it's hard to imagine.  But collective punishment is what terrorism-inspired immigration "reforms" around the world amount to.  There are many polite ways to say, "Most terrorist attacks are perpetrated by Muslims, so let's keep Muslims out," but good manners don't change the fact that countries are punishing millions for the villainy of a small minority.

As someone who's had a public conversation with Nassim Taleb, I know that terrorism has a big right tail.  But so do a great many risks that we calmly endure.  Every bite of food you take could be poisoned.  Every car ride you take could be your last.  Every baby born could be the next Hitler.  The sober response in all these cases is to exert moderate caution, not assume the worst.

Last question: Would I take the same bet again?  Yes, though only with a pre-specified data source and arbiter.  While things turned out worse than I expected, I also think that ISIS was a once-in-a-generation disaster.  Email me for specifics if you're interested.

(10 COMMENTS)

Hear Rob Wiblin, by Bryan Caplan

If you haven't already heard of Rob Wiblin of 80,000 Hours, you should.  He's brilliant, charismatic, inspiring, sobering, enthusiastic, realistic, knowledgeable, open-minded, clear-headed, engaging, curious, funny, earnest, forthright, fair, and pleasantly quixotic. 

Perhaps bashful as well. 

When he's not crusading for Effective Altruism, Rob's conducting some of the most in-depth and well-prepared interviews in the world of ideas.  Besides his recent interview with me, EconLog readers will appreciate...


#15
- Prof Tetlock on how chimps beat Berkeley undergrads and when it's wise
to defer to the wise#25
- Prof Robin Hanson on why we have to lie to ourselves about why we do
what we do#7
- Julia Galef on making humanity more rational, what EA does wrong, and
why Twitter isn't all bad#17
- Prof Will MacAskill on moral uncertainty, utilitarianism & how to
avoid being a moral monster

Enjoy.

(0 COMMENTS)