The Piketty Affair

Summary

Podcast Summary: More or Less – "The Piketty Affair" (May 30, 2014)
Host: Tim Harford | BBC Radio 4

---

Episode Overview

This episode dives deep into "The Piketty Affair," a major debate in economics sparked by claims of data errors in Thomas Piketty’s bestselling book, "Capital in the 21st Century." The episode also explores the veracity of commonly repeated housing-and-golf statistics, examines media claims around video games and tragedy, and assesses reports of rising racism in Britain. With clear-eyed statistical analysis, Tim Harford and guests debunk, explain, and put numbers in context.

---

Section 1: The Piketty Affair – Debate Over Wealth Inequality Data

Key Points

• Piketty’s Thesis: "Capital in the 21st Century" argues that wealth and capital are becoming concentrated among a smaller group, reversing the more egalitarian post-war period.
  • “We’re potentially heading back to a world with a sharp divide between those who live off capital income and the rest of us mugs who have to work for a living.” (Tim Harford, 01:33)
• Data Controversy: Chris Giles (Financial Times Economics Editor) challenged Piketty’s data on UK wealth inequality, finding large discrepancies he couldn’t replicate.
  • “I couldn't replicate Thomas Piketty's numbers for the UK. … I found difficulty replicating any of his other numbers from his own source materials.” (Chris Giles, 03:50)
• Types of Critique:
  • Mistakes: Errors in spreadsheet data entry.
  • Judgment calls: Handling of missing data/choice of sources, accused of "cherry-picking" data to support conclusions.
  • “Lots of missing data, where data seemed to just essentially come out of thin air…cherry picking data to suit his conclusions.” (Chris Giles, 04:41)
• Piketty Responds: Denies substantive errors; argues methodology is sound, and minor corrections don’t change the overall trend.
  • “There is no error or mistake in my series... the little corrections made by the FT…have little impact on the overall patterns.” (Thomas Piketty, 06:22)
• Ongoing Debate: Academics and commentators weigh in.
  • Justin Wolfers argues small flaws don’t undermine the bigger story.
    • “Even if you took everything that Chris said at face value, I don't think it fundamentally changes the message about wealth inequality through time…” (Justin Wolfers, 08:52)
  • Calls to remember that all economic data have uncertainties, and trending evidence still points to rising inequality.
• Conclusion: The statistical debate continues, but the broader message about rising wealth inequality seems robust.

Notable Quotes and Timestamps

• “[Piketty] is the rock star French economist whose book…has combined lots of data with a grand theory about the economy…” (Tim Harford, 00:47)
• “He [Chris Giles] tweaked the data. … cherry picking data to suit his conclusions.” (Chris Giles, 04:41)
• “Let me tell you right away that there is no error or mistake in my series… It is quite unfortunate that the FT ignores this.” (Thomas Piketty, 06:22)
• “All economic data have errors in them, and a little bit of noise upward on some data points and a little bit of noise downward on other data points isn't going to change the underlying trends.” (Justin Wolfers, 08:52)
• “The first law of numbers is that they go up and they go down.” (Michael Blasland, 26:26)

Important Segment Timestamps

• [00:08–03:26]: Set up of the Piketty debate
• [03:26–05:45]: Chris Giles details concerns with Piketty’s data
• [05:45–08:29]: Piketty and Giles exchange, with further academic commentary
• [08:52–11:13]: Justin Wolfers analyzes implications of errors and the broader evidence

---

Section 2: Golf vs. Housing – Myths About Land Use

Key Points

• Urban Legend: Claims that as much or more land in England/Surrey is devoted to golf courses than to housing.
• Origin: Statistic popularized by charity Shelter and Professor Paul Cheshire (LSE).
  • Both use back-of-the-envelope calculations based on estimates of golf course numbers and house footprints.
• Fact-Checking: Severin Waites (LSE) uses mapping software and Open Street Map data for accuracy.
• Findings:
  • In Surrey: True; 2.8% of land is golf, 2.1% housing.
  • For England: False; just 0.59% golf courses versus 1.13% housing.
  • Including gardens and related land, housing takes up much more land — up to 9 times golf courses.
    • “England 0.59% golf course, but 5.38% houses and gardens—nine times as much.” (Laura Gray, 17:15)
• Conclusion: The catchy stats about golf courses crowding out housing are misleading.
• Broader Context: Most suitable housing land isn’t occupied by golf courses; England not running out of space compared to European neighbors.

Notable Quotes and Timestamps

• “Looks like Shelter found a bunker with that one.” (Tim Harford, 17:50)
• “If you take a broad view about the land necessary for houses and gardens…then it’s wrong by a factor of nine.” (Laura Gray, 17:40)

Important Segment Timestamps

• [12:10–17:50]: Data deep-dive and mythbusting on land use

---

Section 3: Do Video Games Cause Tragedy?

Key Points

• Trigger: Media links between the video game "Call of Duty" and youth suicides/murders.
• Expert Perspective: Dr. Andrew Przybelsky explains the "base rate fallacy."
  • Since virtually all teens play video games, any tragic event among youth is statistically likely to involve a gamer by chance.
    • “It would be a remarkable thing if a young person… didn’t play video games.” (Andrew Przybelsky, 20:09)
• Host Clarification: About 97% of teens play video games; almost half play violent games. Presence of gaming in tragedy doesn’t prove causation.
  • “It's not surprising at all to hear of a young person who killed himself and who also played video games…” (Tim Harford, 21:19)

Notable Quotes and Timestamps

• “This is something called the base rate fallacy. So you have to be very, very skeptical of this…” (Andrew Przybelsky, 20:09)
• “We ignore the general information that gives us context and focus on the specific. It's only human to think in this way, but it doesn't help us figure out the truth.” (Tim Harford, 21:47)

Important Segment Timestamps

• [19:29–23:20]: Breakdown of video game causation debate

---

Section 4: Butter, Margarine & Divorce – Listener Feedback

Key Points

• Humorous listener speculation about a graph correlating margarine consumption and divorce rates.
• Listeners suggest marital harmony restored when butter is reintroduced.
  • “I begin to consider escaping from this butterless hell…The Lurpak reappears. I begin to smile again and marital harmony is re-established.” (Listener letter, 23:58)

Timestamps

• [23:39–24:40]: Light segment on correlation versus causation

---

Section 5: Is Racism on the Rise in Britain?

Key Points

• Guardian Headline Analysis: "Racism on the rise in Britain" based on NATSEN survey of self-declared prejudice.
• The Data:
  • 2013: 30% admitted some prejudice (down from 2011 and 2008).
  • All long-term trends: admissions of racial prejudice have been mostly declining since 1983.
• Statistical Lesson: Trends fluctuate; interpreting a single year out of context can mislead.
• First Law of Numbers: "They go up and they go down." Don’t cherry-pick endpoints; look at overall trends.

Notable Quotes and Timestamps

• “Michael's point is that these numbers are variable and they do bounce around a bit. It's often more informative to try to pick out a general trend…” (Tim Harford, 26:26)
• “Let's hope this particular number continues to go down.” (Tim Harford, 27:18)

Important Segment Timestamps

• [24:46–27:18]: Examination of racism statistics and interpretation

---

Memorable Moments & Quotes

• “[Piketty] is the rock star French economist…” — Tim Harford (00:47)
• “All economic data have errors in them…isn't going to change the underlying trends.” — Justin Wolfers (08:52)
• “Shelter found a bunker with that one.” — Tim Harford (17:50)
• “First law of numbers is that they go up and they go down.” — Michael Blasland (26:26)

---

Episode Takeaways

• Data Scrutiny: Big policy arguments rest on complex datasets. Mistakes and judgment calls are inevitable, but don’t always change the big picture.
• Statistical Soundbites: Sensational statistics (golf vs. housing, video games and violence) often crumble under scrutiny. Context is everything.
• Trends vs. Headlines: Don’t be swayed by year-to-year fluctuations; look for long-term patterns and be wary of catchy but misleading statistics.

For more or to contact the show:
More or Less, BBC Radio 4 | Email: [e-mail] | Website: bbc.co.uk/moreorless

Transcript

[Tim Harford]

Thank you for downloading More or less from the BBC. This is the version first broadcast on Radio 4. Here's Tim Harford.

Hello and welcome to More or Less, your weekly hot date with Data. This week, are we building too few houses because the country is covered with golf courses? Should video games get the blame for murders and suicides? Racism is on the rise. Or is it? And More or Less founder Michael Blasland gave gives us the first law of numbers. The first law of numbers is. But first, this week, the world of economics has been abuzz with claims that one of their own has got his numbers wrong. And it's no ordinary economist. He's the rock star French economist whose book capital in the 21st century. Reviewers have called it a bulldozer of a book magisterial, seminal, definitive, a watershed. The man who's managed to unite the words rock star and economist in the same sentence is Thomas Piketty. He's a French professor whose vast tome capital in the 21st century has combined lots of data with a grand theory about the economy, and yet somehow has been the number one seller on Amazon in the United States. What's Piketty's big idea? You're probably familiar with the claim that inequality is rising. We've heard about the Occupy campaigns and talk of the 1%, people whose incomes and wealth outstrip everyone else. Piketty and his colleague Emmanuel Saez have for years been supplying detailed data on income inequality, especially for the richest. It's been rising sharply, with the rich earning more and more relative to others. Without Piketty in size, I don't think we'd even have the phrase the 1%. But Piketty's book adds a new that capital is accumulating in the hands of fewer and fewer people and the post war years in which inequality was low are a historical anomaly. We're potentially heading back to a world with a sharp divide between those who live off capital income and the rest of us mugs who have to work for a living.

[Thomas Piketty]

In a way, what my findings show is that most importantly, you have a huge rise of capital and the total value of capital has never been as large as what it is today.

[Tim Harford]

That's Piketty talking to the BBC earlier this year. His claim is that income from capital is going to become more and more important relative to income from working for a living, that this is an inherent feature of capitalism and there are political implications.

[Thomas Piketty]

This can be good news, assuming that we find the proper way to organize and to redistribute the gains so that everybody can get a Decent share of the total income and output coming out of this. You cannot have one group that is rising three, four times faster than the size of the economy forever. And I think everybody should pay careful attention to numbers here.

[Tim Harford]

We're all in favour of that here at more or less. The Big Theory is exciting and controversial and furious. Debate rages over Piketty's policy prescriptions, but almost everyone agreed that his careful data gathering was fantastic. Until last week.

[Chris Giles]

It was quite odd. And it's one of these things that is just pure luck.

[Tim Harford]

Chris Giles is a colleague of mine in my other job as a columnist at the Financial Times. Chris is the paper's economics editor. He was writing about the recent Office for National Statistics report on wealth in the UK and he wanted an international comparison. So he reached for the data set. Everyone's talking about Thomas Piketty's book.

[Chris Giles]

When I looked at the amount of wealth that he said the top 10% richest UK people held, his number for the UK in 2010 was 71% and the Office for National Statistics number was 44%. Now, that's not just a little difference. I was quite happy to say there would be definitional problems and you'd get some small difference and I would have accepted that, but that was so enormous. And then I thought the next day, well, I'll just go back and have a look at both data sets and see why are we getting such enormous differences. And so I went to the source material and I just found I couldn't replicate Thomas Piketty's numbers for the uk. And the more I looked at all these other numbers as well, I found difficulty replicating any of his other numbers from his own source materials.

[Tim Harford]

The Financial Times published a front page story and a blog post detailing these concerns.

[Chris Giles]

We found some slips in the spreadsheets of Thomas Piketty's data. We found he tweaked the data. We found the way he averaged between different countries didn't seem to be particularly reasonable. We found lots of missing data, where data seemed to just essentially come out of thin air. And we found him also to be what seemed like cherry picking data to suit his conclusions.

[Tim Harford]

So there are two different things going on. Mistakes, such as copying the wrong numbers into spreadsheets and then judgment calls, adjustments to make up for incompatible or missing data, and choosing, or as Chris Giles calls it, cherry picking, which data you use to look at wealth inequality.

[Chris Giles]

So in the us, he moves from an estate tax type of data which gave one result to a survey data which gave another result and that gave an increase in inequality. And in the UK he did the same sort of thing and particularly went to in his final year 2010, a form of data from the HMRC. And that's what drove the finding that UK wealth inequality was rising when all other UK data doesn't show that at all.

[Tim Harford]

Now, just to be clear, Chris Giles isn't criticizing the long standing finding that income inequality is on the rise. The argument instead is over rising inequality in wealth, which has always been harder to measure. Your income is for instance your salary from a job, while your wealth is your assets, your house or your trust fund if you have one. It's worth noting that the FT was only able to raise these questions because Thomas Piketty's put his data online. So top marks for transparency. So what does Thomas Piketty say in response? Well, he wasn't available for interview, but here's what he told us by email.

[Thomas Piketty]

Let me tell you right away that there is no error or mistake in my series. Unlike what the FT pretends the little corrections made by the FT, which I disagree with have little impact on the overall patterns for the us. The most recent research in fact reinforces my finding. If anything I underestimate the rise in wealth concentration for the uk. It is pretty clear that self reported wealth surveys are not the right data sources to detect the rise in top wealth. It is quite unfortunate that the FT ignores this.

[Tim Harford]

He was more forthright in a response to Agence France Press calling the Financial Times ridiculous and dishonest. Eventually he posted a more detailed response on his website on Thursday. It's about 4,000 words long, so let's give you a flavour. In the case of France, Chris Giles said there was no explanation for Piketty's apparently ad hoc tweaks to the numbers.

[Thomas Piketty]

Thomas Piketty this is a surprising statement because all necessary explanations are actually given in the technical research paper on which these series are based.

[Tim Harford]

In the case of the United States, Piketty agrees that there are very large uncertainties in the data and he could have explained some of those uncertainties more clearly. But he also points to research done by other academics since his book was published that backs him up. And what about the UK? Thomas Piketty says the top 10% of households own 71% of the wealth. Chris Giles points to Office for National Statistics Research that puts the number at 44%.

[Thomas Piketty]

Piketty writes, I think my estimate is more reliable and rests on better methodological choices. I also believe that this large gap reflects major uncertainties and limitations in our collective ability to measure recent evolution of wealth inequality and he adds a 44% wealth share for the top 10% would mean that Britain is currently one of the most egalitarian countries in history. History in terms of wealth distribution.

[Tim Harford]

Piketty doesn't find that plausible. I wonder how much of our view depends on what house prices are doing. Well, while waiting for Thomas Piketty's response, economists, journalists and bloggers had been piling in to give their opinion. One of them is Justin Wolfers, a fellow of the Brookings Institution and a professor of economics at the University of Michigan.

[Justin Wolfers]

It led me to admire Chris Giles more as a journalist for digging in and really trying to understand the data. But then I took a step back, a big step back. And while I think he's uncovered a bunch of small flaws, even if you took everything that Chris said at face value, I don't think it fundamentally changes the message of about wealth inequality through time or the broader message of the book. All economic data have errors in them, and a little bit of noise upward on some data points and a little bit of noise downward on other data points isn't going to change the underlying trends.

[Tim Harford]

Copying numbers wrongly into spreadsheets might be noise, but what about the suggestion that Piketty's cherry picked his data sources?

[Justin Wolfers]

Chris criticized Piketty for making specific choices about how you can put very different data sets together. Yet when he compiles his own time series of wealth inequality, he's also had to try and compare apples and oranges. And he's done so in a way that I think it looks like expert opinion is coalescing around the view that his proposal isn't rock solid and may even be substantially weaker than Piketty's.

[Tim Harford]

Keen listeners and economic buffs may be reminded of a controversy we covered last year. Two eminent economists, Carmen Reinhart and Ken Rogoff, had concluded that when government debt rises beyond a certain level, economic growth drops off a cliff. But then a graduate student found some fairly basic errors in the spreadsheets they'd been using.

[Justin Wolfers]

And both of these, I think, came into debates that were hotly politically contested. And so with the Reinhart Rogoff furore, the left was very unhappy that conservatives had used this as an argument for austerity. And so they took some evidence of a spreadsheet error as being a broader battering ram for trying to knock down the claims around austerity. And likewise, you're seeing the right today, looking at the simple transcription errors and other errors as a broader battering ram for knocking down something the left have really taken to, which is Piketty's views about rising wealth inequality and the inevitability of that. In each case, even if you knock over the prize graph or the price chart, there's still the broader constellation of evidence that needs to be reckoned with.

[Tim Harford]

Justin Wolfers and before him, the words of Thomas Piketty and the voice of Chris Giles of the Financial Times. Something tells me this one will run and run. A final point from Chris Giles original blog post on the FT website. He writes, while this post is clear about what is wrong with Piketty's charts, it is much less certain about the truth. Well, perhaps the truth will become clearer over time and perhaps the historical data on wealth just aren't good enough for us ever to know. You're listening to More or Less, and I'm Tim Harford. Every controversial topic needs a catchy statistic, something that packages everything that seems wrong about an issue into a soundbite easily repeated at dinner parties. And here's a couple that have been doing the rounds about Britain's housing shortage. There's more land for golf courses than houses in Surrey. We use as much land for golf

courses in England as we do for homes.

You get the idea. Isn't it terrible that there's more land for wealthy people in plus fours and loud sweaters to knock little white balls around than there is for a basic human need such as housing? But do we really have more of this than this? Hello.

[Presenter/Host]

Hi.

[Tim Harford]

Hello, Laura. Come in. Laura Gray's just popped over to explain. Laura, where does this number come from?

[Laura Gray]

The main claim that we use as much land for golf courses in England as we do for homes comes from the housing charity shelter. Last year they took a rough estimate of the number of 18 holed golf courses in England, multiplied it by a rough estimate of their size, and came up with the estimate that 1.1% of England's surface area is covered by golf courses. They compared this with land use data from Defraud, which shows that 1.1% of England's land is covered by domestic buildings.

[Tim Harford]

So 1.1% golf courses, 1.1% domestic buildings. So it's a draw.

[Laura Gray]

That's right. But add in nine hole golf courses and driving ranges and voila, more golf than houses.

[Tim Harford]

Well, fair enough. What about the claim for Surrey that

[Laura Gray]

came out in spring this year? Professor Paul Cheshire, an economist at the London School of Economics, wrote a paper about planning policies in which he worked out that more of Surrey is under golf courses about 2.65% than has houses on it. And he used roughly the same approach

[Tim Harford]

I'm hearing the word rough a lot. Is there any chance we can put these statistics right in the middle of the fairway of rigorous calculation?

[Laura Gray]

For that, we'd need a statistical superhero.

[Severin Waites]

I'm Severin waites. I'm a PhD student in economic Geography at the London School of Economics.

[Laura Gray]

Severin is here to bring his geography smarts and some serious computing power to working this out.

[Severin Waites]

The easy part really is to find how much land is given over to housing, because there's quite a lot of available data on that. The more difficult part is to find how much land is used by golf courses, because there's no available data on that. Particularly, I used Open Street Map, which is similar to Google Maps, to download polygon data for golf courses. Polygon data is some sort of essentially geographical map data which allows you to plot out the area of golf courses. So I loaded the Polygon data from OpenStreetMap into my mapping software and use the software to calculate the land share in Surrey, England and Great Britain of golf courses.

[Laura Gray]

So some pretty techy stuff there, but basically Sevrin is getting his software to work out the actual area of actual golf courses, rather than estimating.

[Tim Harford]

I like it. I'm all in favour of the back of the envelope, but here Severin's built a massive, super accurate electronic envelope for this calculation to be done on the back of. Let's get some answers. Let's start small. Is it true that there is more land taken up by golf courses than by houses in Surrey?

[Severin Waites]

Yes, that is actually true. I found that 2.8% of the total land in Surrey is used for golf courses, whereas about 2.1% is used for housing.

[Tim Harford]

So more golf than houses. So that's Paul Cheshire's claim. What about Shelter's assertion that golf courses take up as much space in England as houses do?

[Severin Waites]

The claim that been made for England isn't true. More land is used for housing than for golf. For Great Britain, the claim isn't true either.

[Tim Harford]

In fact, 1.13% of England has houses on it. But only 0.59% of England has golf courses on it. Twice as much house as golf course. But the claim about Surrey was true.

[Laura Gray]

Laura, in a sense, it all depends what you mean by given over to housing. Imagine that someone builds 40 houses on a one hectare plot. What area is given over to housing? Only the actual footprint of the houses or the gardens and the roads and pavements that connect them.

[Tim Harford]

Well, does it make a big difference?

[Laura Gray]

Yes, it makes a very big difference. Add in just the gardens for England, and it's almost five times as much land as the houses alone.

[Tim Harford]

And so what have we been talking about so far, then?

[Laura Gray]

Every calculation so far has just looked at the footprint of the house, the ground directly under the roof, if you like. Shelter's numbers, Paul Cheshire's numbers and Severin's numbers all take that quite compressed view.

[Tim Harford]

Well, fair enough, but if we actually wanted to put a few houses on the local golf course, we'd at least want some gardens. So let's recompute. Severin's figures show that Surrey is 2.75% golf course, but 14% houses and gardens

[Laura Gray]

five times as much, and England 0.59% golf course, but 5.38% houses and gardens nine times as much. And you could tell a similar story if you included Wales and Scotland, although there would be more land that was neither house nor golf course.

[Tim Harford]

So the claim that in England there's as much land given over to golf courses as there is to houses is wrong. It's wrong by a factor of two.

[Laura Gray]

And if you take a broad view about the land necessary for houses and gardens, let alone the roads that connect them, then it's wrong by a factor of nine.

[Tim Harford]

Looks like Shelter found a bunker with that one. Well, thank you, Laura. So this catchy statistic turns out to be wrong, but it is it even a sensible comparison? Do golf courses prevent us building the houses we need? Cathranson is president of the Royal Town Planning Institute.

[Cathranson]

I would say not, because the houses that we need are to serve an increasingly urban society in the uk, and that we need housing in close proximity to jobs, transport hubs. And these aren't necessarily the places where golf courses are developed. I think we're actually at about 13% developed at the last count. So there's a lot of space that's not developed. Not golf course, and may be more suitable.

[Tim Harford]

So although Shelter's statistic was wrong and bringing golf into things just confuses the search for solutions, the broader point might be right, trying to counter the suggestion in the press that we're somehow running out of space to build anything new or for new people to settle. Are we any more crowded than our European neighbours?

[Laura Gray]

According to the World bank, the UK as a whole has more than twice the population density of France and a bit more than Germany. But Belgium has a higher population density and the Netherlands has almost twice the UK's density.

[Tim Harford]

Well, it sounds like there's room yet for both homes and golf courses, then.

[Laura Gray]

Yes, although the population density of the Southeast is much greater, nearly at the level of the Netherlands. And that figure doesn't include London itself.

[Tim Harford]

Well, no wonder people keep arguing about building houses. Thank you, Laura.

[Presenter/Host]

Six minutes to seven. A coroner has raised concerns about the video game Call of Duty after investigating

[Tim Harford]

the death of a teenager on Wednesday morning. The Today programme addressed a sad and serious subject. A 16 year old boy died in his own bedroom after recording a verdict of suicide. The coroner pointed the finger at the combat video game Call of Duty. The coroner added that it concerned him greatly that after three or four inquests into the deaths of teens, the Call of Duty game seems to be figuring in recent activity before death. Well, there was a clash of cultures on the Today programme where the conversation quickly moved from suicide to mass murder. The guest, Dr. Andrew Schebelsky of Oxford University, wanted to talk about the statistics.

[Andrew Przybelsky]

Well, of course I'm concerned whenever death and young people are involved, but I actually approach the reports with a healthy dose of skepticism. This is a perfect example of where correlation and causation kind of get intermingled. It would be a remarkable thing if a young person who was involved in a shooting event actually didn't play video games. This is something called the base rate fallacy. So you have to be very, very skeptical of this if it's something that everyone does. There are other correlated factors, I'm sure as well, like wearing blue jeans.

[Tim Harford]

So what is the base rate fallacy, you might ask. And so did we. But Michelle Hussein had a different line of questioning in mind.

[Presenter/Host]

You can't rule out that there might have been a connection in the cases of these teenagers. Describe this game Call of Duty. I don't know if you play it yourself, but it's pretty graphic, isn't it, in terms of the violence. Kill or be killed is the kind of scenario it has been linked. Has it? I mean, we know for a fact that Anders Breivik, where he claimed to have trained himself to kill his victims by playing the game, Mohammed Merar, the Toulouse gunman, also played it before he went on his rampage. Would you let your 14, 15, 16 year old play it though, because it should be an over 18 game?

[Tim Harford]

Well, we think the statistics are important here and we're not sure everyone will have understood what Andrew Przybelsky was trying to say. A bit of context. Almost all teenagers play video games. The estimate is 97% in the United States and it seems to be similar in the uk. And nearly half of gamers play first person combat games such as Call of Duty, Call of Duty itself is one of the most popular video games around. In short, almost half of teenagers play violent video games. Now, every year in England and Wales, about 100 boys and young men between the ages of five and 19 die from what the Office for National Statistics calls suicide and injury or poisoning of undetermined intent. Assume for a moment there's no correlation at all between these tragedies and the playing of video games. In that case, about half of these boys, 40 to 50 would be playing violent video games. It's not surprising at all then to hear of a young person who killed himself and who also played video games, any more than we would be surprised to hear that a young person who died of cancer played video games. Andrew Przybelsky referred to the base rate fallacy. That's when we hear two pieces of information. Generally that half of young men play violent video games and specifically that Mass Murderer and Oz Breivik played violent video games. And we ignore the general information that gives us context and focus on the specific. It's only human to think in this way, but it doesn't help us figure out the truth. And if we can't figure out the truth, we can't do anything to prevent such deaths in future. You're listening to More or less. Our email is more or lessbc.co.uk and our website is BBC.co.uk more or less. Last week we looked at graphs that appear to show striking correlations between seemingly unconnected items. One which caught our attention was a graph showing clearly showing that margarine consumption and divorce rates were falling at the same rate. We invited listeners to suggest their theories for why this might be. Charlotte MacDonald's here. Hello, Charlotte. You've been going through the mountains of replies.

[Laura Gray]

Yes, Tim. There was one explanation which cropped up a lot. For example, in this email from Colin,

[Thomas Piketty]

just look at the wedding vows to have and to hold from this day forward for butter for worse.

[Laura Gray]

And a health conscious butter ban imposed by his wife prompted Michael to wax lyrical.

[Tim Harford]

Daily I plead in vain for a butter resumption. I begin to go into a decline. I begin to consider escaping from this butterless hell. Fortunately, my wife cannot, in the face of my misery, continue with her campaign. The Lurpak reappears. The I begin to smile again and marital harmony is re established. Good to hear. Thank you, Charlotte. And if you want to contact us about anything you've heard in this week's programme, please email more or lessbc.co.uk.

[Laura Gray]

Now is the time. Now is the time for your

[Cathranson]

The

[Tim Harford]

Guardian had bad news for us this week.

[Laura Gray]

Racism on the Rise in Britain.

[Tim Harford]

That was the headline on what the Guardian called an exclusive story based on data available from NATSEN, the National Centre for Social Research. For the past 30 years, Nat Sen has been conducting surveys which ask people whether they'd describe themselves as prejudiced against people of other races. In the Most recent year, 2013, 3% of people admitted to being very prejudiced and 27% to being a little prejudiced, for a total of 30%. James Fletcher's here. Hello, James.

Hi, Tim.

James, racism is apparently on the rise. 30% of people admitted to racial prejudice in 2013. How does that number compare to two years earlier?

Racism is down since 2011, not up.

OK, well, what if we go back five years to 2008?

Racism is down since 2008.

Lets go back 10 years to 2003.

Racism is the same as 2003. And since 1993, no numbers from 1993. But racism is down since 1994 and racism is down since 1983 too.

I see. So, James, in what sense is racism up?

Well, it was higher in 2013 than it was in 2012, so in that sense it's up. And as the Guardian reports, it's risen since 2001. But if you look at the five year moving averages which Natsen provides but the Guardian doesn't report, racism was slowly rising for several years after 2001. But more recently it's been falling.

This reminds me of the first Law of Numbers formulated by the creator of More or Less, Michael Blasland. If only Michael Blastland's disembodied voice could appear in this studio and remind us of the first law of numbers. The first law of numbers is that they go up and they go down. You may want to write that one down. Michael's point is that these numbers are variable and they do bounce around a bit. It's often more informative to try to pick out a general trend, assuming that trends in this particular number make sense at all. After all, what people mean when they admit to racial prejudice now may well be very different from from what they meant 30 years ago. Thank you, James, and thank you the spirit of Michael Blastland. Let's hope this particular number continues to go down. That's all we have time for this week. As always, Our address is BBC.co.uk more or less. Our email address is more or lessbc.co.uk. we'll be back next week, but until then, goodbye.

More or Less was presented by Tim Harford the Financial Times undercover economist. The producer was me, James Fletcher with Charlotte MacDonald and Laura Gray and the programs made in association with the Open University. You can download many more programs for free from Radio 4. You can find them at BBC.co.uk radio4.