Let me tell you a story – was how a talk this lunchtime by Prof. Steven Levitt started as he gave a talk about economics and some of the weird but true facts he has come across in his research.
He is noted for being the co-author, along with Stephen Dubner of best selling book, Freakonomics and is in the UK promoting the suitably titled sequel, Superfreakonomics.
A quick run though of some of the more interesting or plain funny anecdotes started with the tale of John Szilagyi.
Mr Szilagyi, who worked at the US tax office wondered why American children were being given increasingly weird names and had a suspicion. After some effort, the tax returns form was changed to require a social security number to be placed alongside each child being reported by parents when claiming their $250 child’s tax credit.
In 1986, some 7 million children “vanished” from the USA, including as it later transpired, a non-existent brother and sister to Prof. Levitt. The moral was that if you create an incentive, then people will find a way to abuse it.
The purpose of the talk, and the book, is to expand on how simple insights into information and human behaviour can lead to surprising outcomes.
For example, the oft-cited, but not acted upon finding from studying the data that special seats for children in cars make no difference to fatality rates in car accidents when the child is more than 2 years old.
Unfortunately, the politicians felt the facts disagreed with their political policy and not only disregarded the findings, but actually attacked them. Although the talk was being given in a House of Commons Committee Room, no mention was made of a more recent episode of politicians and facts conflicting – this time the recent Nutt sacking over drugs policy.
A more recent bit of work he has been engaged in lead to the title of this blog post – and that was the analysis of bank account activity (with necessary privacy issues dealt with). Not the big stuff that governments like to worry about, and pass silly laws on – such as money laundering etc.
How about monitoring when a person uses a cashpoint machine? If they never ever use one on Friday evenings, is there a chance that they are Jewish or Muslim? What if they lack life insurance or other long term financial plans?
It’s a lot more sophisticated than that – but out of 10 million accounts enough data existed to identify around 30 accounts that had particularly unusual behaviour.
He said that any suicide bomber seeking to evade detection prior to their crime should simply take out a life assurance policy. You can bet the security services have a few other key items they are watching that haven’t been put into books though. Indeed, they might be watching for a spike in people who never use ATM’s on Friday evenings suddenly taking out life assurance policies!
During the Q&A session, much time was spent dealing with their fairly controversial views on climate change – but if you get away from the tabloid media hype, their views are not actually that weird and indeed make a lot of sense.
In essence, it will cost a fortune to reduce CO2 emissions, and whatever we do will take 50+ years to have an effect. Or, we could spend a comparatively tiny amount on geo-engineering which can have an effect almost immediately and if it doesn’t work, we haven’t really wasted any time.
He is arguing that we should look at mitigating CO2 in the short term, and spend the money saved on developing the low-carbon economy that we will need in the future.
Hardly controversial, but the way some in the green-lobby have reacted you would think he was one of the climate change deniers. He isn’t – he just thinks that spending $1 trillion per year on reducing CO2, when alternatives costing maybe $200 million are available seems a bit odd.
His metaphor was that of an approaching asteroid.
One team of scientists come up with a plan that costs $1 trillion, takes 50 years to work and requires every human on the planet to be involved.
Another team comes up with a plan costing $200 million, can be tested in just one year and involves a few thousand people.
Lacking a Bruce Willis to save us, which option would you try first?
If you want to know more, then the duo are giving another talk, this time tomorrow at the RSA, and they will be broadcasting it over the internet.
Tune in, it will be entertaining (especially if he includes the potty training story), and quite possibly enlightening.
Thanks to the Henry Jackson Society for the invite to the event.
I went to their discussion at the LSE on Monday night and it covered similar ground.
I was taken aback by the barely concealed rage at their position on Climate Change. I’d estimate that almost half of the questions from the audience were about it and, although nobody was overtly aggressive, it was clear that many people had decided they were dangerously subversive.
To my mind it was just further evidence that, for many, climate change is a convenient figleaf under which various campaigners can achieve the ends they would be seeking regardless. Which is a real shame because such behaviour has a counter-productive and will result in total loss of public support for *any* efforts to fix the problem.
It’s a pity very little of this guy’s work actually stands up to any scrutiny, and a pity that you’ve apparently taken it in entirely uncritically.
A pity you didn’t offer any concrete examples.
He did say at the talk I was at that they are would be delighted if people didn’t just have knee-jerk reactions, but studied the actual data to see if they had made mistakes in their calculations. He accepts that he might have made mistakes – but no one else wants to look at the data, so how can he be proved to be wrong?
What was mentioned as annoying was how few people/politicians etc are willing to look at the data before announcing policies on issues.
I can assure you I have critically studied some of their stuff and, while that doesn’t come across in a short blog posting, my posting wasn’t meant as a 16 page thesis either.
I agree with Graham. We are writing blog comments, not school papers, so don’t expect a thesis or for me to crunch the numbers, although sometimes, like you, I will if I read something that doesn’t make sense to me. But, some of Freakonomics claims are just plain silly and you don’t need to be a statistician to see that. You can be persuaded against them by looking at the arguments they make. Others have demonstrated the fallacy of their approach and you can find them on the web. One I just read before I read yours was on Fr. . ics’ drunk driving essay, which showed how Fr…ics made unwarranted assumptions, compared apples and oranges, in order to come to an exciting conclusion. I sure hope the next time any reader of their book has a choice between drunk walking or driving, they will walk. Drunk walking is bad enough, as it can be fatal too, but at least they will not take anyone with them. Perhaps they could show statistically that few people die walking drunk on train tracks, but would you believe that this was more advisable than walking on the street. Bet they could sell it. Huff and Geis’s How to Lie with Statistics published in 1954 and republished in the 80s is a much more honest guide for people who want to understand how to avoid being taken in by statistics. They do not von Danikenize statistics, but show the art that Freakonomics tries to pass as science. There are other similar books out there which I would recommend than Fr…ics. I don’t need to read every book to find out whether it’s worth it or not.
But, all that said, I like your blog. You have something to say.