← More or Less
More or Less
Government waste
00:27:36
In More or Less this week: Government waste, a logic puzzle, the statistics of spying,...
Loading summary
Transcript86 lines
- [00:00]
A
Thanks for downloading the More or less podcast from BBC Radio 4. Here's Tim Harford.
- [00:06]
B
Hello and welcome to the last in this series of More or Less, the programme which deals in numbers. This week we'll be interviewing Alex Allen, the chairman of the Joint Intelligence Committee, about the use of statistics in fighting terrorism and outwitting foreign spies. We'll also be asking whether hosting the Olympics is a good deal for Britain. And we'll work out the chances of this happening.
- [00:28]
C
Introduced to Scrabble when I was 19, at a social event, I drew to go first. Staring at my seven blind picked letters, I realised they could make a whole word and put down janitor to general consternation.
- [00:42]
B
But first, let's talk about government waste. Wouldn't it be nice if we could have less of it? Here's the communities and Local Government Minister Bob Neill, speaking in the House of Commons on Monday.
- [00:53]
D
The estimates are that savings of some 20% could be made on £50 billion
- [00:59]
E
at local authorities spend with third parties.
- [01:02]
D
That's about £452 a year. So to that end, the Government, as I say, is working with the Local Government group on behalf of the sector to identify short and longer term savings through the local Productivity programme.
- [01:15]
B
So useless, local Authorities are wasting 10 billion pounds a year through poor spending decisions. That's a huge potential saving, almost as much as the Government raised by lifting VAT to 20%. And much more than labour raised by introducing the 50p rate of tax. Or at least it is if the claim stacks up. When we asked the Department for Communities and Local Government about the figures, they referred us to a press release they published in June attributing precisely the numbers the ministers used to a report by a private consultancy firm. Sadly, however, the Guardian columnist and bad science author Ben Goldacre thinks the report is less than impressive.
- [01:50]
D
It's possibly one of the most ridiculous figures I've ever seen quoted in parliament. The 20% figure comes not from a department of competent statisticians in the civil service with access to all of the local government data. It comes from this incredibly flimsy advertising leaflet from a private management consultancy firm touting for business a private management consultancy firm, which I hope we're sadistically not going to mention by name, because they were obviously only doing it for the publicity. And the short story of how they derived this 20% figure is they said, oh, we reckon we can save you 20% on your mobile phone bills, therefore we can save you 20% on the entirety of local government expenditure.
- [02:35]
B
Aren't mobile phones really the most famously complicated and difficult to procure product in history.
- [02:42]
D
Everybody knows that the mobile phone industry kind of makes money out of having really annoyingly complicated tariffs that are so fiddly that it would take you more time to work out how to get the best deal than it would to earn the money to get work around not having the best deal. Now, it's great if somebody says that they want to make sure that everybody in local government has the best deal on their mobile phones, but the problem is that only reflects £600,000 in their sample of all of the local government expenditure. That's the tiniest, teeniest corner. So what they've done is they've said, we can get you a massive 20% saving in procurement costs on this teeny, tiny corner, the teeniest, tiniest corner of local government procurement expenditure where you are most able to get a good saving. And then they've extrapolated that they could therefore get you a 20% saving on the entirety of local government procurement costs. To see it being wheeled out in Parliament three months, four months after the figure was shown definitively to be absurd. Is it not amazing? You know, I mean, I just.
- [03:46]
B
Although I'm amazed by your capacity to continue to be amazed, because it just,
- [03:50]
D
it happens, you know, I'm not always like this about things. Honestly, genuinely, sincerely.
- [03:56]
B
So, Ben, we did our BBC duty. We asked the Minister whether he'd like to come on and talk to us, or even better, talk to you, and his office said he wasn't available, but they did refer us to the original press release, which you destroyed in the Guardian some time ago, so wonder what you thought of that.
- [04:12]
D
It's obviously ridiculous, isn't it? I mean, you can't brandish around ridiculous figures like this one and then not expect to have to defend the methods by which they were created. So it's not good enough to say, go and read the PDF. I've read the PDF and it's nonsense. And that's why we want you here on the Naughty Step, Minister, to defend it.
- [04:31]
B
We should point out that after we recorded all that, the Department for Communities and Local Government got in touch to point us to another report making a similar claim. We'll have a look at that one in the next series. But Ben Goldacre was aghast that a report he'd already debunked was. Was still being cited by the government. Sadly, we're used to that sort of thing here on more or less, we call dodgy numbers that refuse to go away. Zombie statistics. Doesn't matter how often you shoot them down, they never die. And while we were disappointed that the Minister wouldn't come on the program to talk about the numbers he used, we did manage to secure this statement from a zombie. It's quite easy for governments to promise cost cutting through greater efficiency. It's harder to deliver on those promises. The Cabinet Office has been getting in on the act too, declaring at the beginning of June that thanks to a new efficiency drive, government departments would no longer be spending £2,000 on a laptop which elsewhere in government had cost a mere £350. That's a striking disparity if it's true. I asked Patrick Casey from the fact checking organisation and website fullfact.org what he made of it.
- [05:41]
F
They were quite different laptops with quite different specifications. Some of it included services and additional charges, not just the hardware. And even the hardware was different.
- [05:50]
B
Well, we spoke to the Cabinet Office and one of the things they said was, well, it wasn't really about whether the two identical laptops cost a very different amount of money. It was really about, did the person with a 2000 pound laptop really need the 2000 pound laptop? Maybe he was just doing the same job.
- [06:05]
F
The press release said, and I quote, the move will end poor value for money contracts such as those where government departments and agencies pay between £350 and and £2,000 for the same laptop.
- [06:16]
B
Okay, so they basically did not tell the truth in that press release.
- [06:20]
F
The information they gave there was not accurate.
- [06:23]
B
Patrick Casey. In fairness, the Cabinet Office did give a second example in its press release, which turned out to be accurate. And if half the examples were correct, well, that's better than some press releases we've seen. You're listening to More or Less, with me, Tim Harford, in association with the Open University. Now, a couple of logic puzzles emerged in the news this week. The Justice Secretary, Ken Clark, said in Monday's Guardian that almost three quarters of people charged with offences from the recent riots have previous convictions. And that prompted a number of you to email us at more or lessbc.co.uk with the following. Does that mean that most of the rioters had previous convictions, as Ken Clark seems to be suggesting? Or were the police simply more likely to catch and charge looters who were already known to them because they had previous convictions? Well, Charlotte MacDonald's here. Charlotte, what do you think?
- [07:15]
A
Well, I think that seems like a reasonable theory. If you take a look at London, for example, the day after three days of rioting, the Metropolitan Police reported that around 685 people had been arrested. However, now a month on, more than 2,300 people have been arrested in the Capitol.
- [07:32]
B
OK, so most of the people were arrested only after a delay. And I suppose that gives weight to our listeners suspicions, doesn't it? It suggests the police made most of arrests after reviewing CCTV footage. And it must be a reasonably safe assumption that police are more likely to recognise people who've been convicted before.
- [07:48]
A
Exactly. And I've spoken to a few academics and people in the Met who agree that a lot of rioters will have been identified that way. But in fairness to Mr. Clarke, he was trying to make a wider point about the problem of reoffending. And if you look at the Ministry of justice figures, around 80% of adults who are convicted in Crown Court have a previous conviction. In the Magistrates Court, where less serious crimes are heard, the figure is even higher. Around 93% of people convicted of an offence there have a previous conviction.
- [08:17]
B
Gosh, that's really high.
- [08:19]
A
Yep, for certain. And it's partly because we're talking here only about what are called recordable offences, which exclude things such as speeding tickets and failure to pay your TV license, etc.
- [08:29]
B
OK, so most people convicted are already carrying a previous conviction. I suppose that the same point that our listeners have been making holds true writ large. It must be easier for the police to catch people already known to them. And when they're investigating crimes, known criminals are presumably one of the first avenues the police would explore. I mean, still, it seems, if Anything, that the 75% number cited by Ken Clark is actually lower than the average for other crimes. Now tell us about the other logic puzzle also spotted by several listeners.
- [08:57]
A
Well, this comes from a report published by Alcohol Concern which said that the number of off licenses in an area is statistically linked to the number of underage drinkers admitted to hospital for alcohol related to related problems. In other words, the more off licenses there are in an area, the more under 18 year olds are admitted to hospital because of drinking in that area.
- [09:17]
F
Hmm.
- [09:18]
B
Oh dear. The first rule of more or less
- [09:20]
A
is correlation is not causation.
- [09:22]
B
Exactly. And this is a great example. If you plan to open an off licence, wouldn't you want to do it in an area where there's a demand for booze? I mean, we don't know which way the causation runs here and indeed we don't know if there's a causal link at all.
- [09:34]
A
Well, exactly. We called Alcohol Concern and they said they accept they haven't shown A causal link.
- [09:40]
B
Although that didn't stop their chief executive from saying the numbers of off licenses in any one area has an impact on under 18's drinking and ending up in hospital. Charlotte, thank you. Now, in the 10 years since the terrorist attacks of September 11, Britain, like many countries, has devoted considerable effort and money to tackling the threat from terrorism. And it turns out that the chairman of the Joint Intelligence Committee, which advises the Cabinet and oversees aspects of the British intelligence services, he's trained as a statistician. His name is Alex Allen. And I went to speak to him in the Cabinet Office. On the way to this interview, I passed a poster for the movie Tinker Tailor, Soldier Spy. And I couldn't help noticing that the whole poster was printed on a background of tiny, tiny letters and numbers. We have this vague idea that somehow statistical analysis or numerical analysis must have something to do with intelligence gathering. Even if we don't know how does it?
- [10:39]
E
I mean, in terms of the work we do here in the Joint Intelligence Organization Joint Intelligence Committee, I mean, we do a lot that's based around what would be recognisably sort of statistical approaches in that, you know, our work is a lot about decision taking under uncertainty, trying to piece together fragmentary information, trying to draw conclusions based on uncertainty. So to that extent, a statistical background, statistical training is quite helpful. One of the issues, for example, is how we actually describe our judgments about what may happen. You know, when we're saying something is likely, what sort of probability are we assigning to that? There was a study in the U.S. i think, by the CIA that showed indicators. There was one term where people used a serious possibility in a document and two different people. One interpreted a serious possibility as meaning 20% probability, another interpreted as serious possibility as meaning 80% probability. So you have to be very careful and there are sort of guidelines in the use of language. I mean, it may seem a trivial point, but, you know, actually you can convey completely the wrong impression through that.
- [11:54]
B
Do you think that statistical reasoning training as a statistician, even if you're not explicitly using statistics, is helpful in intelligence gathering and intelligence analysis?
- [12:05]
E
Well, I certainly think it's helpful in intelligence analysis because one of the issues, for example, is trying to test different hypotheses, which is very much at the core of statistical analysis, even if we're not using specific statistical techniques. But I mean, there are great dangers in intelligence analysis of, for example, groupthink, where everybody starts accepting something as the truth and interpreting all new information in that light. The infamous assessment of Weapons of Mass destruction in Iraq, for example, I think there was a general conviction that Saddam Hussein did have weapons of mass destruction or was trying to develop them. And so the intelligence was interpreted in that light and not properly challenged. And so one of the things we do is use various techniques for investigating alternative hypotheses, sometimes using technique called red teaming, where you get another group who are specifically tasked with arguing a contrarian point of view. And indeed the whole purpose really of the Joint Intelligence Committee is to get together a group of people with a wide range of interests to actually debate and really test out whether the hypotheses you're presenting are soundly based.
- [13:23]
B
I remember meeting the author of Freakonomics, Steve Levitt, and he was telling me, this is six years ago, he was about to use his clever statistical techniques to identify terrorists, identify money launderers. I suppose the problem with this kind of thing is the problem of false positives. Most people aren't terrorists, most people aren't spies. And so any statistical techniques you use are going to throw up a lot of suspicious looking activity that in fact is perfectly innocent.
- [13:54]
E
I mean that's clearly a risk. And certainly MI5, the security service do analysis of what's happening online. I mean, I don't really think I should go into the details of that, but clearly there are civil liberties issues about, as you say, suddenly going and investigating people on the basis of what may be very flimsy correlations from online activities.
- [14:18]
B
What about your own academic qualifications?
- [14:21]
E
Well, I did an undergraduate degree in maths and then I did a postgraduate MSc in Statistics and in fact I did my thesis on dice games. So quite interested in some of the sort of game theory issues that can arise in intelligence analysis where you've got somebody you're looking at maybe deliberately trying to mislead you. What can you learn from that? How do you get into the business of if they're trying to mislead you and point you in one direction, you know, perhaps you should be looking in another direction. And you do get into some quite classic sort of game theory analysis at times.
- [14:54]
B
I'm a game theorist myself, that's my background as well. And I was struck by the fact that game theory seemed to get very, very big in the, in the 1960s, used to analyze the potential nuclear war, nuclear deterrence. The film Dr. Strangelove was made and then somehow it seemed to slightly fade into the background. Is game theory still formally used or is it more an informal thing that you keep it in the back of your mind?
- [15:17]
E
I'd say it's probably more informal. I mean clearly, as you say, there are other particularly were during the Cold War issues on the nuclear doctrine which were very much played out in, in game theory terms. I'm not conscious of it being used as explicitly nowadays, but certainly it's there in the background and I find it very relevant to quite a lot of the work.
- [15:39]
B
Alex Allen, chairman of the Joint Intelligence Committee, you're listening to More or less. I'm Tim Harford and our website is BBC.co.uk more or less. Now, as we all know, the International Olympic Committee has the honor of announcing that the games of the 30th Olympiad in 2012 are awarded to the city of London. Several loyal listeners, perhaps inspired by the fact that the One Year to Go to the Olympics stories have been running throughout the summer, have been curious to know whether there's any evidence that the Olympics have any financial value. Do they make a profit in their own right? What about the wider economic benefits such as tourism or urban regeneration? And if the Olympics do help with urban regeneration, are they really the best way to do the job? Oh, and one other thing. Does hosting the Olympics inspire a nation to take up sport, as is sometimes claimed? Let's begin with the economic questions. And who better to speak to than Stefan Szymansky, until recently a professor of economics at Cass Business School in London, now professor of sport management at the University of Michigan. I began by asking him whether Olympic Games ever make a profit.
- [16:51]
G
The most famous one for making money is probably the 1984 Los Angeles Olympics. And the reason they made money was because they really didn't spend any money in the first place on big stadiums and so forth. They had all the infrastructure already in place, so they just hosted the games and the operating costs were covered by the ticket revenues and broadcasting income. Once again, in London and in most Olympics, the operating costs are covered by the broadcast income and ticket income. But it's really the promises to build large scale infrastructure that cause all the financial problems. Which means that the 9.3 billion that is commonly talked about in the UK, that number really comes from building the Olympic site so that it can be handed over to just to run the Games. The legacy company which is currently tasked with finding a use for the facilities after the event, they're generating some income from selling it, but not very much because really the value is quite limited.
- [17:47]
B
I think of the Sydney Olympics, I think of all of those shots of the harbour and the Opera House and really putting Sydney in particular on the map. Won't these Olympics help to boost tourism and help to boost the international image of London. Is there some benefit to that?
- [18:00]
G
Well, I mean, in the case of London, it would be interesting to find a map which London isn't on already. So in that sense it doesn't need a boost of that kind. But also there's been some very interesting research done which shows that none of the Olympic sites over the last 20, 30 years have produced a significant benefit in tourism. So the tourist story is really a bit unlikely as being a major contributor to any economic benefit.
- [18:24]
B
But what about the wider legacy of the Games? What about the regeneration of East London?
- [18:28]
G
Much of the development that's taking place on the Olympic site was planned to take place anyway because it was part of the redevelopment associated with building the high speed rail link, which was another huge public expenditure. So in that sense there's very little what you'd call additionality in the Olympic Games. What there is is really the remediation of a very dirty industrial site and that's cost a lot of money and it will be a very nice area. The question is, will that then be a catalyst for further developments in the rest of East London?
- [19:01]
B
So I guess the question is, is all this stuff worth £9.3 billion?
- [19:06]
G
Well, for an economist, that's a very good question. In terms of, when you say worth, if you mean in strictly economic terms, are the financial returns greater than the investment, then the probable answer is no, because the infrastructure that you build doesn't have a great value. After the event, for example, the Olympic stadium is going to be the host of West Ham football club, but for most of the week the facility is going to be empty. The utilisation you're getting out of that facility is quite limited. Whereas if you compare that to say, investing in a new bridge or a housing complex or a hospital or some other kind of facility that will be used on a regular basis, then you get a lot more benefit. In fact, the biggest catalyst for development in the area of the Olympics complex is in fact going to be the Westfield shopping centre, which again, Westfield have said that they would have built their shopping centre there anyway. That's going to employ more people in the long term, probably than any other employer in and around the Olympic site and will generate more economic development in all probability than the entire Olympic project that the government has funded.
- [20:15]
B
Stefan Shymansky of the University of Michigan. One of the selling points of the London Olympics was the idea that it would inspire a generation of young Britons to get off the sofa and take up sport. But what's the evidence for that. I spoke to Professor Mike Weed, director of the Centre for Sport, Physical Education and Activity Research, and began by asking him whether anything specific had been promised.
- [20:37]
H
Initially, there was a sort of vague promise by the organising committee in Singapore in 2005 at the final bid presentation, that London would inspire a new generation to take part in sport. Then a couple of years later, the previous government established a specific target of getting 1 million more people participating in sport three times a week for at least 30 minutes each time. So that's three times a week for 30 minutes of time. And they want to get a million more people participating at that level between 2007, 8 and 2012. 13. So over a five year period, getting a million more people participating at this level, this effect is something we call the demonstration effect. So it's the idea that people get inspired by elite sports, sports people and sports events to participate themselves. One of the problems about this is that using elite sport and elite sports people to try and inspire the population in general to participate is that you can get an effect called the competence gap. And here, this is where people who don't and never have participated in sport watch somebody like Kelly Holmes winning a double gold medal in Beijing and they think, well, if that's what sport is, it's so far beyond anything I could ever do that it's almost not worth me trying. So there can be a negative effect among people who haven't really played sport in the past in terms of putting them off sport because they think it's too difficult for them going back then to the 1 million more people participating in the five years from 2007. 8. Essentially, what this requires is 200,000 people a year for five years. There's a very big national survey called the Active People Survey that's run by Sport England. What we've seen halfway through that period is that rather than having got 500,000 people participating so far, we've got 109,000 people participating at the current rate. In order to hit that legacy target of a million more people will need another 20 years.
- [22:41]
B
Mike Weed of Coventry University. And by the way, the new government has quietly scrapped the target. I'm not a Scrabble player myself. I'm a bit of a purist and never quite trusted a game where the most concise expression of a legal move requires the deployment of a heavyweight dictionary. But we're a broad church here at More or Less and loyal listener Alan Kirby is a huge fan of the game. He had a question for us.
- [23:05]
C
Could someone Try to calculate the odds against the following. I can't believe it. Coincidence that happened, I swear to me. Introduced to Scrabble when I was 19 at a social event, I drew to go first. Staring at my seven blind picked letters, I realised they could make a whole word and put down janitor. To general consternation. I've played Scrabble many hundreds of times Since. I'm now 61 and can only remember a couple of occasions when I was able to make a seven letter word when going first. About two years ago, a newspaper gave away CDs of electronic Scrabble. I inserted the disc into my computer and opted to play first in my first game against the program. And, well, you've guessed it, the program dealt me the seven letters necessary to make the word janitor. My flabber has never been so gasted. So what are the odds of drawing those same seven letters out on the first go in one's first Scrabble game in two differing formats?
- [24:08]
B
Well, a software engineer called Michael Phelan has written a program which, among other things, carries out precisely this sort of calculation. I asked him how it worked.
- [24:17]
F
There are a certain number of ways that you can draw seven tiles out of a bag. And basically what the program does is it calculates the different ways that you can draw seven tiles and how many of them form a particular word. For example, there are 16 billion different ways to choose seven tiles out of 100.
- [24:35]
B
And there are 100 tiles in a Scrabble bag out there.
- [24:37]
F
Yeah, there are 100 tiles in the Scrabble bag. And it makes it a little tricky to calculate because I'm also counting the two blanks where a blank can be used as any letter.
- [24:45]
B
What are the chances of drawing the word janitor?
- [24:48]
F
So the chances are about one out of 20,000.
- [24:51]
B
Okay. And to draw it twice in a
- [24:53]
F
row, it's that number squared. So it's about 1 in 500 million.
- [24:57]
B
Okay, that's pretty unlikely now.
- [24:59]
F
Yeah, very unlikely. And that's one way of looking at it. You know, of course, if you were to predict in advance, I want to draw Janitor twice in a row, and then you do it. Of course that's extremely improbable. But what happened wasn't exactly that these occurrences happened about 40 years apart. And he said that he played many games of Scrabble in the interim. And I just would estimate maybe if he'd played 25 games of Scrabble per year over 40 years, maybe he's played a thousand games. And so we may want to consider what would be the probability of drawing Janitor twice in any two of those thousand games?
- [25:32]
B
You've done that particular calculation.
- [25:34]
F
The chances of drawing it twice in a thousand games are about one in 400.
- [25:39]
B
Okay, so it's not very likely, but it's not unimaginable anymore.
- [25:42]
F
Yeah, we're starting to get into the realm of from the extremely improbable into these sort of somewhat improbable.
- [25:48]
B
And of course, it doesn't necessarily have to be janitor. You know, if it was any seven letter word twice, that's the kind of thing he would remember.
- [25:55]
F
Exactly. Janitor is a common word and it also stands out because it has the unlikely J. And so if you draw the word janitor, you're likely to say, wow, I just drew Janitor. I can't believe it.
- [26:06]
B
And what's the least likely word that you could actually play?
- [26:09]
F
The least likely word that you could actually play is ziziva. And that's Z Y, Z, Z, Y, V A zizy.
- [26:19]
B
What does it mean?
- [26:19]
F
It's some kind of tropical weevil. I don't know exactly. It's like a bug.
- [26:23]
B
Okay.
- [26:24]
F
I don't know exactly what it is.
- [26:25]
B
Michael Phelan and his program is available via his website, which is of course zizaver.net and I'm afraid that's all we have time for this week and indeed this series. No more zombie statistics to slay, no more press releases to debunk. But we'll be back in December at the new time of 4:30 on Friday afternoons. We'll still be in the same Sunday evening slot and of course, podcast via BBC.co.uk more or less. Please keep your letters and comments coming in between now and December. We'll be reading them before the new series starts, and until then. Wait, what's that noise? Charlotte, take a look.
- [27:05]
D
Tim, it's zombies.
- [27:09]
B
Charlotte, we need to get equipped. Fetch me my calculator, a blunt pencil and my slide rule.
- [27:14]
D
Yes.
- [27:15]
B
No, that's my after dinner slide rule. Fetch me my fighting slide rule. Excellent. Ready?
- [27:20]
A
Yes.
- [27:21]
B
Let's go.
- [27:27]
A
More or Less was presented by Tim Harford, the Financial Times undercover economist. The producer was Richard Knight. The programme was made in association with the Open University.