Sunday 30 October 2011

Uploads

The following files are available for download:

CSV file containing data scraped from Beating Bowel Cancer's map site (the data there come from UKCIS).
A spreadsheet for drawing funnel plots
An alternative version of the spreadsheet, converted by Chris for use in older versions of Excel
A spreadsheet which runs a simulation to find the typical range of bowel cancer mortality rates across UK districts

Saturday 29 October 2011

Heartwarming nerd

Guardian readers welcome.  If you want to try playing with funnel plots you could try this spreadsheet.

Friday 28 October 2011

Sovereign CDS

The Eurozone deal is done.  Judging by the reaction of the equity markets, even the 50% 'haircut' on Greek government debt, when taken together with the rest of the package, is a good deal for the banks. One interesting aspect is that the 'haircut' being taken by private banks has been structured as a 'voluntary' exchange of bonds, with the intention that it should not be a default event triggering CDS payoffs.  ISDA, the body responsible for ruling on CDS, seems willing to go along with this, even while acknowledging that "there has been a lot of arm-twisting".

The effect is that anyone who has insured against loss on Greek government debt by buying sovereign CDS will find that their insurance is worthless and they just have to take the loss.  Conversely, anyone who has blithely written Greek sovereign CDS will make a tidy profit.  It seems strange that we should be encouraging banks to accept downside risk and discouraging them from insuring against losses.  Not that I am going to spend any time feeling sorry for investment bankers - if you trade sovereign debt derivatives you have to be aware that governments can manipulate the market.

I suppose that this will destroy the sovereign CDS market in Europe, especially if the threatened ban on naked buying of sovereign CDS is implement, since that would leave no buyers.  Instead, the EFSF (European Financial Stability Facility) will be offering insurance.  I doubt that the bond markets will be in a hurry to trust the EFSF actually to be there with the cash should the need arise.

Wednesday 26 October 2011

Synecdoche

The answer to the challenge is that "Short" is a verb in the imperative mood, and the rest is a synecdoche for "the shares of companies with powerful CEOs".

Probably you've worked this out already if you care at all, but I don't expect to get another chance soon to write a post whose title is both a figure of speech and part of the title of a favourably reviewed film.

Monday 24 October 2011

Short Powerful CEOs

I've stolen this title from another blog.  The challenge is to guess what the post there is about.

Eurozone bailout

I apologize to Ed Miliband, who must be chortling to himself this morning.  Despite what I wrote, David Cameron, who I suppose has detected political danger in allowing the Eurozone bailout to be finalized in his absence, is going back to Brussels on Wednesday after all.  I don't suppose he'll enjoy himself much sitting there while Merkel and Sarkozy ignore anything he says.  Sarkozy is unlikely to be charming about it: "We are sick of you criticising us and telling us what to do."

It seems that the bailout deal will include a Greek government default on its debts, in the form of a bond 'haircut'.   The eurozone is no more interested in my advice than it is in Cameron's, but I say that abandoning the absurdity of pretending that Greece can pay its debts is a start.

Alcohol limits

The Royal College of Physicians is advising us to drink no more than 21 (for men) or 14 (for women) units of alcohol a week, and to have two or three alcohol-free days a week.  Ian Gilmore, the RCP's former president and currently the UK government's special adviser on alcohol says:
In addition to quantity, safe alcohol limits must also take into account frequency.
There is an increased risk of liver disease for those who drink daily or near-daily, compared with those who drink periodically or intermittently.
We recommend a safe alcohol consumption limit of between 0 and 21 units a week for men and 0 and 14 units a week for women provided the total amount is not drunk in one or two bouts, and that there are two to three alcohol-free days a week.
At these levels, most individuals are unlikely to come to harm.
So I can drink 21 units a week, provided that I spread my consumption over three, four, or perhaps five days a week (a UK unit is a centilitre, so there are 9 units in a 75cl bottle of wine containing 12% alcohol by volume: nowadays most wine is a little stronger than that).  It seems strange that somewhat uneven consumption should be safer than drinking the same amount every day: what's the evidence?

The BBC report arises (indirectly) from a written report by the RCP to a House of Commons committee.  The most relevent section of the report says:
Current Department of Health sensible drinking guidelines state that regular consumption of between three and four units a day for men, and between two and three units a day by women will not accrue significant health risk. Regular is defined as ‘drinking every day or most days of the week'.

26. This suggestion that daily drinking is low risk runs against evidence which suggests that frequency of drinking is a significant risk factor for the development of alcohol dependency, and the development of alcoholic liver disease.[7]
The report proposes that:
A very simple addition would remedy this problem namely a recommendation that to remain within safe limits of alcohol consumption that people have three alcohol-free days a week.
Reference [7] is this paper, which reports most pertinently that:
A large North American study of over 22,000 also showed that daily drinking carried more than twice the risk of liver damage compared with intermittent drinking once or twice per week [28].
Reference [28] therein is this paper, which investigates the health effects of frequent "risk drinking":
Risk drinking was defined as consuming the equivalent of 5+ standard drinks in a day for men and the equivalent of 4+ standard drinks in a day for women.
It doesn't define a "standard drink", but it seems that in the USA a standard drink contains about 14g of pure alcohol.  The density of alcohol (ethanol) is 0.789g/cm^3, so that's about one and three quarter UK units.  4 standard drinks would be 7 units, 5 standard drinks would be 8.75 units.

This table in the paper quantifies the risk of a list of adverse consequences at various frequencies of risk drinking.  It shows that drinking a bottle of wine nearly every day gives about twice the risk of liver disease of drinking a bottle of wine three or four times a week.  I'm not surprised - that's about twice as much alcohol consumed.

My summary is that the papers cited do not support the RCP's conclusion.  There is no evidence that drinking 3 units of alcohol every day is worse for you than drinking the same 21 units a week spread over only four days.

I suspect that they've made this recommendation because they think if you drink less often, you'll drink less in total.  They may well be right.  But I'd recommend a more honest approach, along the lines of "most people who drink alcohol several times a week consume more than they think they do.  One good way to reduce consumption is to have three days a week on which you drink no alcohol."

Sunday 23 October 2011

Real Estate

There's a half page in The Times Saturday opinion section often occupied to amusing effect by Giles Coren (in his spare time he writes humourless tosh for the Daily Mail: I suppose he needs the money).  This week the space is taken up by the thoughts of Roger Scruton on home ownership in England.  Suddenly, he says, we are being encouraged to take out second mortgages on our homes and gamble the proceeds.  Perhaps this is a piece he wrote ten years ago, which a sub-editor pulled out of a drawer at the last minute when Coren's effort got pulled because he'd written something the lawyers wouldn't risk publishing.  [The Times is behind a paywall, so there's no link here to the article.]

I mention Scruton's dull space-filler mainly in order to marvel at this under-informed comment: "...real ownership does not mean leasehold or rental but freehold - what the brokers call "real estate" on the understanding that nothing else is really real."

No.  Real estate is a legal term meaning fixed property, as distinct from personal property.  It is perfectly possible to have a leasehold on real estate.  (It's from the Latin res, a thing.  Some etymologies claim instead that it's from regalis, meaning royal as in "real tennis", but there's no evidence to support them.)

I wonder why Scruton didn't check this with a fellow don: a lawyer or historian perhaps.  I think the problem may be that the universities he has studied and taught at are "mere teaching machines rather than a community of scholars that is a humble part of a larger organism."

Saturday 22 October 2011

Myob

Ed Miliband has told off David Cameron "for not seeking a place at next week's crucial eurozone summit".  Ed, the reason why Cameron isn't going is that it's none of our business.  The UK isn't in the eurozone, thanks in part to the Gordon Brown's "five tests", carefully crafted to allow Brown to keep us out of the Euro without embarrassing Tony Blair.

I suppose the eurozone countries would be delighted to have us along if we offer to chip in a few billion for the latest "defence fund".  Ed?

But it's the weekend.  If you want to think about financial obligations, you could do worse than read this novella, which may be the source of the acronym I've used for the title of this post.  You might even buy the book.

Thursday 20 October 2011

Sunlight and bowel cancer incidence

I mentioned a fortnight ago that I was going to investigate a possible relationship between the regional distribution of bowel cancer incidence in the UK and of sunlight.  And so I have.

I've chosen to work with incidence rather than mortality, because the incidence numbers are larger and hence less affected by random variation, and because it seems to me that sunlight is more likely to affect incidence than outcome.

The hypothesis is that sunlight is linked to bowel cancer incidence because exposure to ultraviolet light promotes synthesis of vitamin D, which has a protective effect.  It's the UVB (higher energy UV) component of the light that's important for vitamin D synthesis, and the process is not linear.  However, for a first pass at least, I've used data I could quite easily get online, which is sunlight incidence from this site, intended for use in solar energy calculations (I told it I'd use horizontal panels).  For each district I queried for one or more population centres, averaging the results if necessary.  The solar incidences I found varied from 2.14kWhr/m^2, averaged over the whole year, in Shetland, to  3.00 in Kingsbridge, Devon.


Here's a scatter plot of the data.  I've used a non-zero intercept on the sunlight axis.  I've resisted the temptation to let Excel draw a regression line.

It looks as if there might be something there.  I calculated a population-weighted correlation: the result is -0.300 .  To check for significance I tried shuffling the sunlight data so that the numbers were applied at random to the districts, recalculating the correlation each time.  In 1000 tries none of the correlations where anywhere near that large: there may be a lot of perturbations that would give that much correlation, but they're a tiny fraction of the 380! available.

However, this is an in-sample test - I've used the same data to form the hypothesis and to test it.  To avoid that, I tried ignoring the eleven labelled points I looked at when I came up with the notion.  The population-weight correlation was reduced to -0.256 .  This is still larger than anything I got in 1000 shuffles of the data (the largest therein being -0.199).  I conclude that there is a genuine negative correlation in the UK between bowel cancer incidence and sunlight.

Correlation does not prove causation - I suppose I would have got a similar result if I'd looked at latitude instead of sunlight, and one could attribute that to lifestyle differences between different parts of the country.  , But more rigorous analyses in the USA and Japan have found a negative correlation of bowel-cancer mortality with sunlight even after correcting for dietary differences.

Population-weighted linear regression gives a regression slope of -15.34, i.e. the best guess is that an increase in solar incidence of 1kWhr/m^2 reduces bowel cancer incidence by about 15 per 100,000 people.  To give an indication of the potential effect, I worked out the effect of applying an adjustment in each district as if every district had the solar incidence of the sunniest.  That would reduce bowel cancer incidence by 4298 cases per year.  If the ratio of mortality to incidence were maintained (research suggests sunlight might in fact improve survival) that would save 1742 lives per year.

I am not proposing that we promote climate change to allow more sunlight through the atmosphere.  But it might be possible to encourage people to expose themselves to sunlight more (not of course neglecting its dangers: there were 2067 deaths from malignant melanoma in the UK in 2008), or dietary vitamin D might have a similar effect (the data are suggestive but unclear).

Tuesday 18 October 2011

Baby Grand

Dinner last night at De Belhamel in Amsterdam, which I commend.  The menu includes "Petit grand dessert".  I've been trying to work that out: there seems to be a not very well defined term "Grand Dessert" for an elaborate dessert comprising assorted smaller desserts.  This would be a nouvelle version of it.

Incidentally, the restaurant is proud of its "art nouveau" decor (it doesn't call it "nieuwe kunst").  It occurs to me that the adjective is relatively accurate, as compared with its use by New College.

I'm more confident about Little Big Horn.   The battle was fought on the banks of the Little Bighorn river.  Which is a tributary of the Bighorn river.  The Bighorn river was named by fur trader François Larocque for the bighorn sheep he saw there.  And the bighorn sheep is so called for its big horns.

Friday 14 October 2011

Like twins

The Daily Mail has the shocking allegation that Liam Fox and his close friend Adam Werrity sometimes dress similarly.  It illustrates its case with a picture from Fox's wedding, where Werrity was best man.

It's a pity they never served in the Army together.

Thursday 13 October 2011

How to stop rogue-trading losses

I'm now better informed than I was about how Kweku Adoboli managed to lose so much money for UBS.  I needn't reveal any confidences: this is a fair outline of what I have been told went on.

Whenever losses like this are uncovered, the bank concerned is censured, and rightly so, for whatever lax practices allowed the risk management system to be fooled.  But there is one simple procedure that would have caught each of the notorious rogue traders - Nick Leeson at Barings, Jérôme Kerviel at SocGen, or Kweku Adoboli at UBS.  And that is to check positions giving rise to any funding requirement out of proportion to the risk being reported.

All three traders were supposed to be carrying out low-risk trading, in which equity futures hedged the risk on other, quite simple instruments.  For Leeson, it was similar futures on another exchange.   For Kerviel it was a basket of shares or warrants, and for Adoboli it was ETFs.  Because the positions were supposed to be low risk, they were authorized to trade in very large size.  But in each case, the offsetting trades were faked so the actual risk was enormous, as were the losses when markets turned sharply against the positions.

When you hold an equity futures position, you pay or receive daily "variation margin" as market prices move. In each of these rogue-trading cases, the banks will have paid out most of the losses in variation margin even before the losses were discovered (the balance of the losses being due to market moves caused by trading out of such large positions after their discovery).  Barings ought to have been receiving offsetting variation margin on the offsetting futures trades, but weren't.  In the other two cases the purported offsetting trades would not have been expected to generate margin.  But large profits were being reported on the offsetting trades, slightly greater than the large real losses.  It would have taken only a brief investigation to discover that many of the offsetting trades had been neither confirmed nor settled by their counterparties.  At which point a thorough investigation would have been instigated.

I'm not saying that this sort of check would catch all rogue traders.  But it would detect sizeable losses from rogue trading before they become huge losses.

This sort of investigation might well pay for itself even if there were no rogue traders.  This is because arbitrage businesses exploiting small margins should be accounting carefully for funding costs (which have become much larger since the 2008 crisis) and for credit risk.  Investigation might show that the apparent profits may not really exist in businesses that seek to exploit narrow arbitrage opportunities by trading in high volume.

Another thing that should be monitored is an explosion of trading volume in any one instrument - in Adoboli's case that would be forward-settled ETFs.  The trading desk concerned should have to demonstrate that the market demand is there and that it's genuinely profitable to satisfy it.

There's one more thing that banks might consider doing: appoint a senior banker with a small team whose sole remit is to prevent large trading losses, whether by rogue trading or by concentration of risk in illiquid instruments.  This team should be paid mostly in salary: any bonuses should be paid in cash (not shares) deferred for three years and forfeited if major trading losses are uncovered during the deferral period.  Team members should not be allowed to transfer to trading jobs.

Monday 10 October 2011

Some say Theresa May

Theresa May, the British Home Secretary, has attracted much adverse comment for some of her remarks on the Human Rights Act, not least from her own side and its partners.  But let's look at what she actually said (full text here):
We all know the stories about the Human Rights Act. The violent drug dealer who cannot be sent home because his daughter – for whom he pays no maintenance – lives here. The robber who cannot be removed because he has a girlfriend. The illegal immigrant who cannot be deported because – and I am not making this up – he had a pet cat.
"We all know the stories about the Human Rights Act" - That's unlikely to be exactly true, but if you're looking for an audience that reads the Daily Mail and follows the speeches of Nigel Farage, where better to go than the Conservative Party Conference?  (I assume that the UKIP conference cannot be a better place.)  I think Ms May can be excused this rhetorical flourish.

"The violent drug dealer..."  This story is true.  (Ms May omits the detail that the drug dealer had been in this country since he was four, so it would seem rather unfair on Trinidad to deport him there.  I suspect the Tory Party Conference is not much concerned with fairness to Trinidad.)

"The robber who cannot be removed...".  This story appears to be true.

"The illegal immigrant"  The cat-owner was not an illegal immigrant, he was a Bolivian student who'd overstayed his visa.  But let's not get picky.

"I'm not making this up".  This is true.  It was the Daily Mail who invented the story.  Ms May didn't even make up the line about not making it up.

"because".  This is a whopper.  But it's not Ms May's whopper.  And it's only one word.  And she said it was a story: stories don't have to be true.

"he had a pet cat".  Yes he did.

In further support of Ms May, I might mention that:
i) She wears better shoes than does Ken Clarke
ii) Ken Clarke has made a lot of money peddling carcinogenic drugs in developing counties.
iii) Chris Huhne's ex-wife alleges that he lied about a speeding offence
iv) The Labour Party's most recent Immigration Minister, Phil Woolas, was a proven liar.

On the other hand, Ms May cited these examples in support of her view that "the Human Rights Act needs to go".  This is just populist pandering.  The Human Rights Act simply requires British courts and public bodies to act in accordance with the European Convention on Human Rights  so long as British law permits them to do so.  Repealing it would remove enforcement of the Convention to the European Court of Human Rights.  How is it desirable to have the law work more slowly, and with the final decision in the hands of foreigners?  If I read the feeling of the meeting correctly, the Conservative Party Conference is against them.

Nick Clegg on the Euro

Here's Nick Clegg in January 2009 advocating that Britain should join the Euro:
Mr Clegg says he is not pushing for immediate entry and admits that the past housing bubble might have been even worse had Britain been tied to eurozone interest rates.
But he says the “page has been turned” in economic policymaking, highlighting the need for an “anchor” against the “incredibly vulnerable exposure to international financial markets”. Refusing to discuss the euro is a “failure of political leadership”, he says.
and here he is being interviewed at his party conference recently:
"I think, clearly, with the benefit of hindsight, you can say [joining the Euro] would have been a huge, huge error,"
"I don't think anyone could have predicted at the time the euro was created that the rules which were supposed to be in place to ensure that everybody looked after their own financial affairs properly would be so spectacularly ignored and broken.
"I think history will judge the then French and German governments very, very unkindly who, some years ago, basically signalled that the rules could be relaxed because that then sent a signal out to everybody else - oh well, we don't need to keep our house in order.
I don't think it would have been very difficult to predict in 2009 that the rules would be ignored. Here's an official EU report from 2006.  Table A.2.1 shows that Belgium (1995-2007), Germany (2002-2007), Greece (1997-2007), France (2003-2007), Italy (1995-2007), Austria (1995-2007), Portugal (2005-2007) were all in breach of the 60% limit imposed by the Stability and Growth Pact on debt to GDP ratio, and that Greece and Italy were the worst offenders, having had a ratio over 100% ever since 1995.

Eddie Izzard for Mayor?

There's an interview with Eddie Izzard in Saturday's Times, in which he says he's going to run for Mayor of London in 2020.  He's a comedian, but two years ago he ran 1100 miles round the UK in 52 days.  So he's got my respect and attention.

There's one paragraph of political analysis:
He thinks of the Conservative Party as the Dementors from Harry Potter ("It cannot be proved scientifically, but we're pretty sure that the Tory party wil suck your face off"), and is against the cuts. "Why the hell do we have to pay everything off so bloody quickly?" he asks me rhetorically, before delivering a small lecture on Britain's Second World War debt.
"It was paid off in 2007. That says it all to me..."
That might be a good joke about the Dementors - being funny is not my area of expertise. But I thinks he's serious about the cuts. It bothers me that as an aspiring politician he should be so disconnected from the facts on the one political subject mentioned in his interview. The fact is that we've not paying off a penny of government debt, and plan to continue to be net borrowers until 2016.  The cuts are about reducing the government deficit - how much it's overspending by - not about paying back debt.

The point about the War debt is misguided too.  We did pay back the money the USA lent us at the end of the war.  But we didn't do it by running a surplus.  There was a schedule of payments attached to the loan, which we met mostly by borrowing more money - the UK government has made net debt repayments in only 14 out of 55 years since the Second World War ended.

It's not that I expect better of politicians.  But I expect better of Eddie Izzard.

Wednesday 5 October 2011

Funnel plot of UK bowel cancer incidence

If you've already read more than you want to about colorectal cancer, please look away now.
This chart is a funnel plot of incidence (i.e. new cases) rather than mortality.  It shows that there is significant variation from one district to another - there are several points outside the dashed lines.

It would be a mistake to deduce from this and the previous analysis that incidence depends on where you are but mortality doesn't.  We didn't prove the null hypothesis for mortality, we just failed to disprove it (outside Glasgow).  But it's easier to get statistically significant results for incidence, because the numbers are larger.  Strangely, the best evidence that mortality varies with district may be in the incidence numbers not the mortality numbers.

However, the data are for diagnosed incidence, not for the (unknown) total incidence.  Starting a screening programme will cause a jump in diagnoses, which will mostly but not entirely fall off with time - in the absence of screening, some people with bowel cancer may live out their lives without being diagnosed, and die of something else.  So for example the high incidence of diagnoses in Wirral in 2008 may be more to do with the start of a screening programme in 2007 than with a high total incidence.

I note that all the points well above the upper dashed line lie in the north of the UK, and all but one of the points well below the lower dashed line lie in the south of the UK.  I asked an oncologist, who told me that bowel cancer is known to be negatively correlated with sunlight, more so than any other cancer.  I'll try to show to what extent this factor is explanatory.  Meanwhile, here's a US map, generated on this site, of the sort that was first used to generate the sunlight hypothesis.

Tuesday 4 October 2011

Age Standardization

I accused Beating Bowel Cancer of doing its age standardization wrong: I withdraw that.  What they seem to have done is to standardize against European norms, without saying so - I realised this when I had another look at the Cancer Research UK spreadsheet, which gives a very similar standardization.

Actual bowel cancer mortality across the UK is substantially higher than the numbers from Beating Bowel Cancer tell us, which are for a hypothetical population with fewer old people in it.