How Deadly Is Ebola?

Summary

---

More or Less – "How Deadly Is Ebola?"

BBC Radio 4 | Hosted by Tim Harford | First aired: August 29, 2014

Episode Overview

In this episode, Tim Harford and the More or Less team critically examine headline-grabbing statistics and their underlying realities—with a primary focus on the 2014 West African Ebola outbreak. The team dissects the reported fatality rate of Ebola, explores what "up to 90%" actually means, and discusses the true risk of a global Ebola pandemic. The episode also touches on statistical distortions in stories about guide dogs, deaths after benefit assessments, and prison populations, as well as an analysis of football transfer fees.

---

Key Discussions and Insights

1. Ebola Outbreak—Understanding the Numbers

[00:35–08:44]

a) Is Ebola "Up to 90% Fatal"?

• The popular statistic that Ebola "kills up to 90% of those infected" is misleading.
• Ruth Alexander: Explains what "case fatality rate" (CFR) means with a fictional example.
  “If 100 people were diagnosed with [a disease] and 30 of them died, then the fatality rate ... would be 30%.” ([02:21])
• Dr. Maimouna Majumdar (MIT):
  “That 90% figure actually comes from one outbreak of Ebola ... in the Congo between 2002 and 2003. ... The cumulative average ... is actually closer to about 60 to 65%." ([02:54])
• Case fatality rate during the 2014 outbreak had dropped to around 52% ([03:41]).

b) Variation and Context

• Rates vary between countries (e.g., 42% in Sierra Leone, 66% in Guinea).
• Survival depends not just on the disease, but also on healthcare and available resources. "A CFR of 90% could be considered the upper limit ... the current outbreak ... has a CFR of about 54%, though it’s subject to change as the outbreak wanes.” (Majumdar, [02:54]) "This is the first time that we've seen Ebola in West Africa, so it kind of caught everybody by surprise." (Majumdar, [04:01])

c) WHO Estimates and Media Misinterpretation

• Reports that actual Ebola cases could be "four times higher" than reported are exaggerated. “They said there could be up to four times as many ... in some areas ... So there may be more than the recognised 3,000 cases, but probably hundreds more rather than many thousands.” (Tim Harford, [05:02])

d) How Contagious is Ebola, Really?

• Ebola is not highly infectious compared to airborne viruses like flu or measles.
• Transmission requires direct contact with bodily fluids—making widespread outbreaks in developed countries unlikely. “Ebola, you don't become infectious until you’re symptomatic. Contrast this with something like flu ... where people are walking about without symptoms and spreading the infection.” (Dr. Adam Kucharski, [06:51])
• Spreading in the UK or US would require a series of improbable events. “For an outbreak of Ebola to occur in the United States, a very large number of improbable events would have to occur.” (Majumdar, [07:30])

e) Focus on Africa, Not Western Spread

• Health experts are most concerned about further spread in poor nations with weaker health systems. “Right now, the focus is on dealing with the outbreak on the ground in Africa rather than the possibility that it might spread to the UK.” (Harford, [08:31])

---

2. Debunking Other Statistical Stories

A) Guide Dogs and Cyclists ([08:44–11:42])

• BBC headline claimed 25% of guide dogs in London had been hit by a cyclist—statistic was based on a misleading, unrepresentative Facebook survey.
• Only 16 collisions among 320 guide dogs are known (about 5%). “Of course, any guide dogs being hit is terrible, but this doesn't seem quite as bad as the 25% figure.” (Tom Chivers, [09:21])

B) Deaths After Benefit Assessments ([11:42–15:04])

• Widely-shared claim: 10,600 people died within six weeks of being declared 'fit for work' by Atos.
• The statistic misinterpreted official data—deaths occurred within six weeks either side of the benefit claim ending, not necessarily after losing benefits. “10,600 claimants didn't die six weeks after their claim ended. 10,600 claimants died within the same six week period as their claim ended. Not the same thing at all.” ([15:04])

---

3. Why Are Prison Numbers Rising?

[15:49–20:38]

• Justice Secretary Chris Grayling links the spike to an increase in sex offenders, especially historic cases.
• Carol Hedderman (Institute of Criminology, Cambridge): Only 80 of the 700 additional sex offenders are new entrants—most of the increase is fewer releases and longer sentences. "How much water you've got in the bath depends on ... how fast you've turned the taps on and how firmly you've stuck the plug in." (Hedderman, [17:38])
• The hike is due to policy keeping prisoners longer, not just new convictions.
• Political speeches and policy can have unintended effects on prison populations. "When Michael Howard made the comment about prison works ... sentences sitting around the country were thinking, oh, well, ... perhaps we can take the brakes off.” (Hedderman, [19:48])

---

4. Football Transfer Fees—Fact vs. Model

[20:58–26:29]

• Transfer fees often diverge from modelled "true" value due to club desperation, contract variables, and off-field issues.
• Dr. Raffaelli Poli (CIES Football Observatory):
  “For Di Maria ... this is more related to the very bad start of the [United's] season, the bad season they did last year and ... Real Madrid didn't ... need to transfer the player.” ([22:33])
• By Poli's model:
  • United overpaid by £42m on three big transfers this window.
  • Mario Balotelli was valued at £28m but cost Liverpool £16m—a bargain due to off-pitch concerns ([24:56]).
• “We’re not dealing in commodities, we're dealing with humans.” (Football agent Seb Ewan, [25:53])

---

Notable Quotes and Memorable Moments

• On headlines:
  "I wouldn't be totally sold by a retailer who offered 'up to 90% lower prices.' The 'up to'—it's a bit weaselly. Well, the stakes are a lot higher here." (Tim Harford, [01:35])

• On Ebola transmission:
  "Ebola requires intense contact with body fluids. It's not airborne in the way that something like SARS or flu is." (Dr. Adam Kucharski, [06:51])

• On statistical pitfalls:
  "So the moral of this story: It's always worth asking what a statistic is really counting, rather than assuming we know." (Host, [15:04])

---

Timestamps of Important Segments

• Ebola headlines and listener question: [00:05–02:21]
• CFR explained, expert input (Majumdar): [02:21–04:23]
• Fatality rates by country, outbreak scale: [03:41–05:02]
• How contagious is Ebola, risks of Western spread: [05:36–08:44]
• Guide dog-cyclist statistic debunked: [08:44–11:42]
• Benefit assessment deaths claim debunked: [11:42–15:04]
• Explaining prison population rise: [15:49–20:38]
• Football transfer fee analysis: [20:58–26:29]

---

Summary Verdict

This episode of More or Less is a compelling reminder to look past sensational headlines and critically interrogate the statistics that drive public fear and policy debate. Whether it’s how deadly Ebola truly is, how many guide dogs are at risk, or the price tag on a footballer, the More or Less team dissects the data with clarity, wit, and a healthy dose of skepticism.

Transcript

[Tim Harford]

Thank you for downloading More or Less from BBC Radio 4. Here's Tim Harford.

[Presenter/Host (possibly Tim Harford or co-host)]

Hello and welcome to More or Less, the programme which considers itself the sequined feather boa of statistics. And who are we to argue? In this week's programme, we have cyclists running over guide dogs and prisons full of sex offenders. But first, if you've been listening to the news, you could be forgiven for feeling very worried. Doctors at the Royal Free Hospital in North West London have begun treating a British nurse who caught Ebola in Sierra Leone.

[Tim Harford]

The head of the World Health Organization

[Presenter/Host (possibly Tim Harford or co-host)]

has warned that the outbreak of the

[Tim Harford]

deadly Ebola virus in Western Africa is

[Presenter/Host (possibly Tim Harford or co-host)]

running out of control. Ebola is probably the most contagious and most deadly virus in the world. The strain can kill up to 90% of its victims. Most of those who get it will die through uncontrollable bleeding. Ebola, a viral disease, has spread across West Africa, affecting Guinea, Liberia, Sierra Leone and Nigeria. Another separate outbreak has just been confirmed in Central Africa. Now it's come to the uk. William Pooley was a volunteer nurse who contracted the virus while treating patients in Sierra Leone. He was flown to London for treatment. The latest figures from the World Health Organization showed that over 1500 people have already died, making this the largest ever outbreak. Even worse, the WHO says that its scale has been vastly underestimated. It seems like the stuff of a Hollywood pandemic, but more or less. Listener Max Cousens is wondering whether Ebola is quite as terrifying as it seems. A widely reported statistic is that Ebola kills up to 90% of those infected. I've been wondering what this statement actually means. Is it that 90% die in some areas and less than 90% in others? If so, to what extent and why do mortality rates vary? Yes, I wouldn't be totally sold by a retailer who offered up to 90% lower prices. The up to it's a bit weaselly. Well, the stakes are a lot higher here. So what does up to 90% fatal actually mean? Ruth Alexander's here. Hello, Ruth.

[Tim Harford]

Hi, Tim. What we're talking about here is the case fatality rate, which is the ratio of people diagnosed with an illness to the people who die. So let's take a fictional disease, Dragonpox, from Harry Potter. If 100 people were diagnosed with that and 30 of them died, then the fatality rate, the CFR, would be 30%.

[Presenter/Host (possibly Tim Harford or co-host)]

We asked Maimouna Mazumda, a biostatistician and epidemiologist at MIT in the United States, what the WHO means when it talks about a fatality rate of up to 90%.

[Maimouna Majumdar (Biostatistician and Epidemiologist)]

That 90% figure actually comes from one outbreak of Ebola, and which took place in the Congo between 2002 and 2003. So it's the highest rate we've ever seen. We've also seen CFRs much lower than that in previous outbreaks, and I think that's something important to keep in mind. A CFR of 90% could be considered the upper limit of outbreaks caused by Zarebola virus. The cumulative average, considering all Ebola cases and deaths from 1976 to to present, is actually closer to about 60 to 65%. And the current outbreak thus far has a cfr of about 54%, though it's subject to change as the outbreak wanes on Mainmuna Majumda.

[Presenter/Host (possibly Tim Harford or co-host)]

And since we spoke to her, the fatality rate has dropped to 52%.

[Tim Harford]

And that's the average for the whole outbreak. But the fatality rate varies for each individual country. It's ranging at the moment from 42% in Sierra Leone to 66% in Guinea.

[Presenter/Host (possibly Tim Harford or co-host)]

Partly that's just chance, but a lot of the variation is no accident. Here's Maimona Majumda.

[Maimouna Majumdar (Biostatistician and Epidemiologist)]

Again, the two biggest factors that are influencing case fatality rate are first, familiarity with the disease. This is the first time that we've seen Ebola in West Africa, so it kind of caught everybody by surprise. And the second is availability of resources to contain, prevent and treat the people that are sick.

[Presenter/Host (possibly Tim Harford or co-host)]

This is important when we hear that the case fatality rate for ebola has been 60 to 65% across history. Well, that sounds like a fact about Ebola, but it's also a fact about healthcare systems and about how well they treat the symptoms.

[Tim Harford]

And yet, if you catch Ebola, it is still quite likely to kill you. Now, the World Health Organization says that the scale of this outbreak has been vastly underestimated. And that also sounds alarming, but there

[Presenter/Host (possibly Tim Harford or co-host)]

have been around 3,000 recognised cases so far, and it's been reported that the World Health Organization says there could actually be four times that number. In reality, that would make 12,000 cases.

[Tim Harford]

But that's actually not what they said at all. They said there could be up to four times as many people in some areas in areas of intensive transmission, so urban centres, basically, where it's hard to keep up with the numbers. So there may be more than the recognised 3,000 cases, but probably hundreds more, rather than many thousands. And the WHO said that the total number of cases could reach 20,000 in the next six to nine months. Again, that's been widely reported, but that would be only if the outbreak wasn't controlled. And of course, they're working hard to do just that. Control it.

[Presenter/Host (possibly Tim Harford or co-host)]

OK, so we know that Ebola isn't 90% fatal, but it's still very dangerous if you get it. So how likely are you to get it?

[Tim Harford]

Which is a key question, because the deadliest disease in the world isn't much of a worry if it's impossible to catch it. And a highly contagious disease can kill a lot of people, even if most people who catch it pull through. Now, thankfully, a Ebola is nowhere near as contagious as things such as flu or measles, for example. They both spread easily through the air, whereas Ebola is transmitted through contact with bodily fluids, which makes it much harder to get.

[Presenter/Host (possibly Tim Harford or co-host)]

There's still been a lot of concern that the virus could spread beyond West Africa. Airlines such as Emirates and British Airways have stopped or reduced flights into the region. But a British nurse with Ebola was flown from Sierra Leone to the uk. So was that a risk?

[Tim Harford]

American billionaire Donald Trump, displaying great self confidence in his medical credentials, tweeted that infected medical staff shouldn't be allowed to return to the US and that the US must immediately stop all flights from Ebola infected countries or the plague will start and spread inside our borders.

[Presenter/Host (possibly Tim Harford or co-host)]

Well, far be it from us to disagree with Mr. Trump. We're not epidemiologists either, but Dr. Adam Kaszarski is.

[Dr. Adam Kaszarski (Epidemiologist) / Seb Ewan (Football Agent)]

In theory, Ebola is a controllable disease, so we're very unlikely to see a major outbreak in the UK for example. The reason being that Ebola, you don't become infectious until you're symptomatic. Contrast this with something like flu, where a lot of the infection occurs when people are walking about without symptoms and spreading the infection. Likewise, Ebola requires intense contact with body fluids. It's not airborne in the way that something like SARS or flu is.

[Presenter/Host (possibly Tim Harford or co-host)]

Adam Kusharski of the London School of Hygiene and Tropical Medicine. Of course there is a risk that Ebola might spread around the world, reaching the UK or the US But Maimuna Majumda from MIT explains that that would be pret difficult.

[Maimouna Majumdar (Biostatistician and Epidemiologist)]

For an outbreak of Ebola to occur in the United States, a very large number of improbable events would have to occur. First and foremost, you would have to have a person that has Ebola get on a plane in West Africa en route to the United States. The likelihood of this is actually quite small. You would also need the person to not be symptomatic, because if they were symptomatic, they would be feeling so ill that they probably wouldn't be flying. So let's assume that does actually happen and they land in New York City, for instance. Now, by this time, if they're not showing symptoms yet and they go about on their own merry way, then there might be a potential for sustained transmission. What's more likely is that by the time that they arrive in the United States, they may not be feeling so well. If they're not and they turn themselves into a hospital, because the hospitals in New York City are very aware that this is a possibility, containment would occur right away.

[Tim Harford]

Health experts are far more concerned with the spread of Ebola in poor countries. Right now, the focus is on dealing with the outbreak on the ground in Africa rather than the possibility that it might spread to the uk.

[Presenter/Host (possibly Tim Harford or co-host)]

Thank you, Ruth. And thank you to Max for raising the question. Loyal listener Josh cast his eagle eyes over the BBC news website this week, and then he emailed us at more or lessbc code. Hi, just seen this article on the BBC website.

[Tim Harford]

London cyclists collide with 25% of guide dogs working in the capital.

[Presenter/Host (possibly Tim Harford or co-host)]

We saw that one too. Josh, what did you make of it

[Tom Chivers (Daily Telegraph Assistant Comment Editor)]

crunching the numbers down? 320 guide dogs, fifth of those were surveyed and a quarter of those had collisions. I make it that 16 dogs were in collisions.

[Presenter/Host (possibly Tim Harford or co-host)]

Of course, any guide dogs being hit is terrible, but this doesn't seem quite

[Tom Chivers (Daily Telegraph Assistant Comment Editor)]

as bad as the 25% figure.

[Presenter/Host (possibly Tim Harford or co-host)]

Good point. After all, the actual number of collisions reported in this survey isn't that large. So we'd want to know whether these are a few disturbing incidents or are representative of a much more widespread problem. A short time later, a note appeared on the BBC article. An earlier version of this story said a quarter of guide dogs working in London had been hit by a bike. Guide Dogs for the Blind association has since said the information it provided was incorrect. We've been looking into this and actually the story gets stranger and stranger. Guide Dogs advertised this survey on their Facebook page. Here's the thing, though. They said, if you are vision impaired and have a view on cyclists in London, please fill in our survey. Obviously that's fine if you want to gather case studies of people's experience, but you might well expect people who've been hit by a cyclist in London to have some very strong views about cyclists in London. So it's not even close to being a representative sample of visually impaired people. And we can't extrapolate from the Survey results As it happens, 33 guide dog owners from London were among the respondents and 13 of them had been hit by a cyclist. That's 42% according to Guide Dogs, although we make it 39%. In the end, the poor fellow at Guide Dogs admitted he wasn't great with numbers, and we're not going to disagree with him that. The 25% number doesn't refer to guide dog owners at all. No wonder the BBC withdrew that statistic, and we might ask why they ran it in the first place. Now, what might surprise some people is that according to Guide Dogs, there are actually only 320 people using guide dogs in London. It certainly surprised us there were two guide dogs here in our office and we know for sure that 13 of these 320 have been hit by cyclists, so that's already 1 in 25. So while we don't have any confidence in this survey or the numbers it produced, we are confident that for people who own guide dogs, being hit by a cyclist isn't a trivial risk. If you think you've seen a suspect statistic, we would love to hear about it. You can email us at more or lessbc.co.uk. a number of emails that have landed in the More or Less inbox over the last few weeks have shared a similar theme.

[Tim Harford]

Doing the rounds on Twitter, there's a picture titled A flower for every person who died within six weeks of Atos finding them fit for work. There are a lot of flowers. Can it really be true that so many people have died such a short time after their assessment by Atoz?

[Presenter/Host (possibly Tim Harford or co-host)]

How many people have died? Earlier this year, the trade union Unite. The government's own figures last year showed that 10,600 people died within six weeks of being declared fit for work by Atos. Attos refers to Atos Healthcare, the company which has the contract to carry out health assessments for the DWP, the Department for Work and Pensions. Here's the In 2011, existing incapacity, benefit and income support claims were replaced with something called Employment and Support Allowance. Claimants were made to undergo a work capability assessment to determine whether they were entitled to the new allowance and how much money they might get. Now some critics complain that these assessments are stacked against claimants. Seriously ill people are being dismissed as malingerers by Atos Healthcare and having their claims denied. And in the middle of this argument, up pops the truly shocking finding that 10,600 people have been cut off from this vital benefit and then died within Six weeks. Or at least it would be shocking if it was true. But is it? Tom Chivers is the Daily Telegraph's assistant comment editor and he detected the aroma of dodgy data.

[Tom Chivers (Daily Telegraph Assistant Comment Editor)]

In July 2012, there was a Freedom of Information request about how many people died within six weeks of their benefit claim, ending in this marvelously impenetrable response that came back. In total, between January 2011 and November 2011, some 10,600 claims ended and a date of death was recorded within six weeks of claims ended.

[Presenter/Host (possibly Tim Harford or co-host)]

Well, that seems that the statistic that everyone's cross about is slightly bureaucratically conveyed. But yeah, it's true. We've got ATOs bang to rights.

[Tom Chivers (Daily Telegraph Assistant Comment Editor)]

It does. It sounds exactly like that. When ordinary humans use the English language, when they say something happened within six weeks of something else happening, they mean that second thing happened after the first thing, you know. But I found this whole document so impenetrable, I felt I should at least give DWP a ring and see what they were talking about. And they said, no, actually, there is a rather weird, obtuse meaning of it, which they mean it was in six weeks either side of this thing, or at least there was a six week period either side of the death, and that was when the claim ended. Now, when you know, that suddenly makes you realize that a lot of these people would have died and then the claim ended shortly afterwards because they were dead.

[Presenter/Host (possibly Tim Harford or co-host)]

So hang on to what the DWP are doing here is they take a snapshot, they see a certain number of people are making a benefit claim and are alive, and then six weeks later they take another snapshot and they discover that these people are no longer making a benefit claim and these people are no longer alive. So do we know the truth? I mean, do we know.

[Tom Chivers (Daily Telegraph Assistant Comment Editor)]

The DWP says they don't keep records of the number of people who died after their benefits were cut off because that's irrelevant to them, that's no longer their problem. So we don't have the full figures. What this really comes down to, as far as I'm concerned, is just a dreadful piece of communication by the DWP that this fairly well, not simple but not complicated piece of information has been translated into 10,000 people dying within six weeks of being callously removed with their benefits.

[Presenter/Host (possibly Tim Harford or co-host)]

Tom Chivers from the Daily Telegraph. So 10,600 claimants didn't die six weeks after their claim ended. 10,600 claimants died within the same six week period as their claim ended. Not the same thing at all. So the Moral of this story. It's always worth asking what a statistic is really counting, rather than assuming. We know. And when you receive an incomprehensible answer from a bureaucrat, it may be worth checking what they meant. You're listening to More or Less with me, Tim Harford, the Justice Secretary. Chris Grayling has been having to defend the state of prisons in England and Wales. Part of the problem, apparently, is that there are more prisoners than he expected. Here he was on the Today programme.

[Chris Grayling (Justice Secretary)]

Recently, we've got challenges from an increased population that was not expected in the last 12 months. We're meeting those challenges, we're recruiting more staff. But I'm absolutely clear there is not a crisis in our prisons. There are pressures which we're facing, but there is not a crisis.

[Presenter/Host (possibly Tim Harford or co-host)]

Chris Grayling there you might have missed the almost subliminal message that there is definitely not a crisis. But why has the rise in prisoner numbers been so unexpected? Why weren't we prepared? The justice select committee asked Mr. Grayling to explain it to them.

[Chris Grayling (Justice Secretary)]

Last month got about 1,600 more people in prison than a year ago. The most obvious change has been there's been quite a big increase in the number of sex offenders coming into our prisons, mainly historic sex abuse cases. I think it's something like 700 new people in our prisons in that category. And happily supply the detailed figures to the committee.

[Presenter/Host (possibly Tim Harford or co-host)]

Clear enough, you might well believe that after the revelations of Jimmy Savile's crimes, people are more willing to come forward with reports of historic sexual abuse. You might also expect the authorities would be more likely to prosecute. And there you are, suddenly a clear explanation of why there are 700 new sex offenders in prison this year, a good chunk of the extra prison population. But does that story really make sense? There have been a lot of high profile cases in recent months, but 700 is a big number, especially since these sex offence cases often take a long time to conclude. I thought I'd run the more or less slide rule over Mr. Grayling's claims with a visiting scholar at the Institute of Criminology at Cambridge, Carol Hederman.

[Carol Hederman (Institute of Criminology, Cambridge)]

Chris Kraling's right to say that there are 700 more sex offenders in prison this year than last year. But only about 80 of those are new cases entering the prison population.

[Presenter/Host (possibly Tim Harford or co-host)]

So what explains the others?

[Carol Hederman (Institute of Criminology, Cambridge)]

Well, when you think about the prison population, it's kind of helpful to think about it as a big bath, where how much water you've got in the bath depends on whether there was some in there. Already. And in terms of the prison population, that's most of it. How fast you've turned the taps on and how firmly you've stuck the plug in. So at the moment, what we've got is a slight increase in the taps being turned on.

[Presenter/Host (possibly Tim Harford or co-host)]

So 80 more sex offenders coming in than in a comparable period in a previous year. Okay, so the flow rate's increased by

[Carol Hederman (Institute of Criminology, Cambridge)]

80, but releases have dropped dramatically in the same year.

[Presenter/Host (possibly Tim Harford or co-host)]

Plugs in more tightly.

[Carol Hederman (Institute of Criminology, Cambridge)]

Indeed.

[Presenter/Host (possibly Tim Harford or co-host)]

So it's right to say there are 700 more sex offenders in prison. It's not true to say there's been a sharp increase in new sex offenders being sent to prison. There has been a small increase, but far more significant is the fact that there's been a trend for years now to put sex offenders away for longer. So the rate at which they're released has been falling, and that fall is predictable. It's the result of decisions made years ago. There's also some recycling of the metaphorical bath water Carol told me. People let out on parole who breach the conditions are taken back in again. And there's been a trend in recent years to make them serve lengthier periods before they're let out. And it isn't necessarily just sex offenders who are being kept in prison longer. If you look at the overall numbers, fewer people were sent to prison in January to March this year than in the same period last year. About 300 fewer, but 2,000 fewer prisoners were released. So the prison population went up. It is hard to forecast prisoner numbers, Carroll says, not least because it's a very political arena.

[Carol Hederman (Institute of Criminology, Cambridge)]

I remember when Michael Howard made the comment about prison works in the early 1990s, he, along with a lot of other politicians, was enormously surprised to see the prison population go up immediately afterwards.

[Presenter/Host (possibly Tim Harford or co-host)]

And that wasn't because of a change of policy. Well, what was it?

[Carol Hederman (Institute of Criminology, Cambridge)]

Simply because a load of sentences sitting around the country were thinking, oh, well, if it works, we better all stop. You know, we've been told we shouldn't be using prison so much, but if the Home Secretary is saying that prison works, perhaps we can take the brakes off. And actually, you see, there's analyses done which show that those sorts of speeches do actually influence behaviour and that legislation that's intended to do certain things does not necessarily have the effect that people expect of it.

[Presenter/Host (possibly Tim Harford or co-host)]

The current Justice Secretary has been taken by surprise by the numbers, but Carroll says he shouldn't have been.

[Carol Hederman (Institute of Criminology, Cambridge)]

Chris Grayling's rather taken the brakes off. So I don't think anybody who understands anything about criminal Policy would find it particularly surprising that the prison population has started to go up again under his watch.

[Presenter/Host (possibly Tim Harford or co-host)]

Carol Hederman of Cambridge University. The football season is underway already in the top European leagues. And as if the media weren't busy enough reporting on the games, they're also saturated with rumours and reports of players moving from one club to another in Europe. The deadline to complete such transfers is fast approaching.

[Carol Hederman (Institute of Criminology, Cambridge)]

Signed for a British record fee, Manchester

[Presenter/Host (possibly Tim Harford or co-host)]

United secure winger Angle di Maria for 59.7 million pounds. Liverpool have finally confirmed the signing of

[Tim Harford]

AC Milan striker Mario Balotelli.

[Maimouna Majumdar (Biostatistician and Epidemiologist)]

The 24 year old is Liverpool's ninth

[Tim Harford]

summer signing and has cost the club 16 million pounds.

[Presenter/Host (possibly Tim Harford or co-host)]

The main thing that's reported when a player moves from one club to another is the transfer fee. That is effectively the compensation one team pays to the other to release the player from his contract. But how do the two clubs arrive at that value? I spoke to Dr. Raffaelli Poli, who co founded the CIES Football Observatory which is part funded by football's governing body FIFA and affiliated with the University of Neuchatel in Switzerland. They've created a statistical model of the underlying transfer value of every player in the top five leagues in Europe. That's Spain, Italy, Germany, France and of course England. This model is based on previous transfer fees and the variables it contains are the players age, the length of contract remaining, the position in the side, the league in which they play, the team they play for, matches played, goals, assists, factors such as the number of tackles, how many international matches they've played and for which team. One of the most eye catching transfers of the summer was Argentine international Angel Di Maria. Moving from Real Madrid to Manchester United for a British record transfer fee of £59.7 million. That's €75 million. I asked Dr. Poli if that was a fair price according to his analysis.

[Dr. Raffaelli Poli (CIES Football Observatory)]

I don't think so. Of course Manchester United started badly this season. This has probably pushed him to overpay. These players is double the value that we estimated according to our model. Of course Di Maria is an outdoor outstanding player, but I think this is more related to the very bad start of the season, the bad season they did last year and also the fact that the player himself didn't really want to leave Real Madrid and Real Madrid didn't really want or needed to transfer the players.

[Presenter/Host (possibly Tim Harford or co-host)]

But Manchester United have been desperate for a while. Angel Di Maria isn't the only player they bought. We asked Dr. Pohli to estimate how much they may have over overpaid just in this Transfer window. He reckons Anda Herrera was worth £16 million but £29 million was reportedly paid. Luke Shaw was worth £26 million but including various add ons, £31 million was reportedly paid. And then there's angel di Maria, worth 36 million pounds but apparently costing nearly 60 million pounds. That's a total of 52 million euros overpaid. 42 million pounds. That's not even including Roja who's come from the Portuguese league and so isn't included in Dr. Poli's modelling. Well what about another high profile transfer, Mario Balotelli's move from AC Milan to Liverpool. Balotelli is 24 years old, he's an Italian international and he's got a good goal scoring record. But Liverpool only paid £16 million or €20 million. Why? Well, Seb Ewan, a football agent, isn't surprised that the fee is modest.

[Dr. Adam Kaszarski (Epidemiologist) / Seb Ewan (Football Agent)]

You mean Balotelli, peripheral player, perennially in the press, not necessarily somebody who is generally regarded as being easily manageable. A lot of people within the game would consider him to be a risk playing for a club who has had better years financially than is currently ongoing. I think that's every reason why he's 16 million and admittedly still a risk at 16 million pounds but an extremely well calculated one I think.

[Presenter/Host (possibly Tim Harford or co-host)]

Dr. Poli though reckons Balotelli's worth 28 million pounds. That's 35 million euros. So Liverpool got a bargain.

[Dr. Raffaelli Poli (CIES Football Observatory)]

Yeah, for Balotelli it's a bit the opposite. As for Di Maria, it was underpaid because of non sporting related issues. Balotelli unfortunately wasn't able in the past to behave in the best way. I would say that outside the pitch and sometimes also on the pitch. But sportingly it's a good price, good thing for Liverpool to have been able to transfer these players to signing for only 20 million euro.

[Chris Grayling (Justice Secretary)]

Hmm.

[Presenter/Host (possibly Tim Harford or co-host)]

Well it seems every transfer we look at there's some sort of special case to explain why the actual transfer fee is different from the model. But I suppose we are looking at the most eye catching examples. Dr. Poli says that overall his statistical model accounts for 80% of the variability in transfer fees. Now even giving him the benefit of the doubt on that, it leaves 20% for the football agents to argue about. As Sebuwen says, it can be a messy business.

[Dr. Adam Kaszarski (Epidemiologist) / Seb Ewan (Football Agent)]

There are many things that you can't account for and because of that reason alone I think it's impossible to be able to sort of assess and quantify accurately via statistics. I think they will always give you a very, very good indication, but in many cases they can be very, very wide from the reality of it. We're not dealing in commodities, we're dealing with humans. You deal with emotions, you deal with families, you deal with, you know, many tragedies as well within these players. These aren't machines, they aren't robots who consistently perform and produce, because if they are, it wouldn't be fun.

[Presenter/Host (possibly Tim Harford or co-host)]

Football agent Seb Ewan, loyal listener Ian has been in touch. Inspired by Radio 4's In Business presenter, Peter Day's new vision of how data should presented on the radio. I enjoyed your illustration of Peter Day's suggestion of rendering graphs orally, but I feel you missed an obvious opportunity for improvement. Instead of merely humming, why not contact the offices of I'm Sorry, I Haven't A Clue and borrow their swanny whistle? I feel this would lend the BBC's financial reporting some much needed gravitas.

[Carol Hederman (Institute of Criminology, Cambridge)]

Hmm.

[Presenter/Host (possibly Tim Harford or co-host)]

Well, let's see now. Prisoner numbers are mainly because fewer prisoners are being released, The More or Less team will soon be huddling around the email inbox to see what you have to say about what we've had to say. So go on, send us an email. Comments, suggestions, questions, all are read and considered. Our email address is more or lessbc.co.uk and if you'd rather just subscribe to a free download of the program or. Or wallow in our many past editions like a pig in mud, go to our website, BBC.co.uk more or less until next week, goodbye.

[Tim Harford]

More or Less was presented by Tim Harford of the Financial Times. The producer was Ruth Alexander. You can Download many more Radio 4 programs at the Radio 4 website. BBC.co.uk radio4.