A quick rebuttal

Watching this week’s episode of This Week (a programme I greatly enjoy watching for its off-beat humour and incisive political analysis), I couldn’t help feeling that their discussion on the economic crisis in Europe and the eurozone was somewhat lacking a dose of common sense.

The first bone of contention for me was when Andrew Neil asked one of his guests whether Greece defaulting would mean a simultaneous exit from the euro. She immediately responded ‘yes’. My response would be an instantaneous “why???”.

There’s absolutely no connection between Greece defaulting and departing from the Euro. One is almost inevitable (in my view) and the other virtually impossible. Greek sovereign debt is simply a contract between the Greek government and a private individual. Greek membership of the eurozone is a much larger and intractable issue, as it is not one contract but a universal standard amongst millions of individuals.

(A quick explanatory analogy: Greece defaulting would be as complicated as asking your friend, or a group of friends, to change a previously agreed meeting time. Greece leaving the euro would be like requiring them to alter their timezone, and everyone else’s, to change the actual time of the meeting)

To clarify and underline: I believe Greek default is a certainty and it will happen within the next 12 months. Or rather, it will happen as soon as the non-interest Greek budget moves into surplus, which could be a lot sooner than 12 months away, going by reports. The Greek government would announce it would no honour its debts, institute a long bank holiday to prevent capital flight and would go through the motions of calling in the IMF and the other European leaders in to hold some sort of crisis talks. Then it would default, go back to business as usual and return to the markets within two years. The full details are written in this excellent 2002 article “How to default”.

Membership of the Eurozone is an entirely different affair. I have not seen a credible explanation as to how this would happen without causing massive currency flight, financial collapse and the destruction of the Greek economy. A report by UBS puts the potential damage to Greece at as high as 50% of GDP in one year. I do not think any government led by anyone with a sound mind would permit this to happen. Even the most ruthless of dictatorships remembered to pay their army.

Just think Greek exit through: Greece is leaving because it wants to devalue its currency. But the problem is that the currency is not going to default by a modest amount, say 20%. It is going to go through the floor and be worth say around 15% of the euro. Firstly that makes all foreign imports worth 6 times more than they did the day before. It means that every single person in Greece is going to want to make sure their savings in euros don’t turn into drachmas, meaning every Greek bank suffers a bank run large enough to bankrupt it immediately. It means Greece literally drains of money as capital flees abroad, and due to EU restrictions I suspect Greece could not impose currency controls. Greece also holds lots of debt in euros, so leaving the euro to join the drachma would increase the debt by a factor of six also.

People talk of Greece leaving the euro as if it were a possible policy choice. I do not share this view. I believe that leaving the euro is an impossible action, not least because of the suicidal implications it would have for any economy attempting to do so at a time of crisis.

So where do we go from here? Well, the European government leaders are still doing their usual dithering routine so I don’t expect them to come up with any great package to avert crisis. But here is my plan, assuming magic-wand powers:

1. Greece defaults in an ‘orderly’ way, having moved into a primary surplus.
2. The ECB engages in a widespread interest rate targeting regime on government debt to prevent interest rates in at-risk countries spiralling out of control in response to the Greek default. For example, the ECB would ensure Italian bond yields do not rise above 6%. (They are currently at 6.5%)
3. The ECB does what it should have done 12 months ago and moves interest rates to close to 0%.
4. A long-term solution to internal balance-of-payments imbalances within the eurozone is brought about. This could be greater fiscal integration, either explicitly through a central European treasury (highly improbable), or more likely through the back door via some sort of debt relief package.

The key thing here really is the ECB. The ECB lacks the full powers of a traditional central bank, so acting in the way described above is not in its DNA. However I am optimistic that under its new management it might have a fair stab at saving the eurozone.

Change of scene

Seeing as though I would actually like to do quite well in my degree, and hopefully to pass some useful knowledge on to others, from now on I’m going to be posting my notes from lectures up on to my blog.

There’ll be four main strands for now,  with a fifth later to follow:

• Microeconomics
• Macroeconomics
• Econometrics
• International Economics
• (Cost-benefit analysis)

Whether any of this remains comprehensible to anyone but myself remains to be seen.

Revenge of the dodgy mortgage bond?

You know in many newspapers they have those boring legal notices? You know, the ones which have to be published by law in prominent newspapers so they’re visible but are normally formatted to look really dull – so most people glaze over them?

Well, I found one in the FT yesterday, and strangely enough … it wasn’t very dull at all.

Holland Homes – Notice:

Notice is hereby given to the [holders of Holland Homes debt]  that Fitch Ratings, following an update of its criteria for assessing mortgage backed securities, placed on ‘negative rating watch’ several securities related to Holland Homes.

Further to this criteria update, the existing debt no longer satisfies the new Fitch criteria for an AAA rating. A downgrade of the current ratings assigned to the debt to BBB or a similar level is expected if no action is taken.

(edited for clarity, FT 18/08/2011 p17). Which immediately made me go ‘wtf’? You’re jumping from ‘never lose money’ AAA to ‘quite speculative’ BBB? I thought we’d left the silliness around mortgages-which-were-supposedly-AAA-but-not-really all back in 2008. Obviously not.

The most interesting thing about this is that these securities were issued in 2007 (they expire in 2039), so it’s not like Fitch hasn’t had very long to correct their sums. They’ve had 4 years, most of which were filled up with a financial crisis caused  in part by incorrectly rated securities.

The detail of the debt isn’t too complicated. Holland Homes has two types of debt, firstly “class A notes”, which get paid back first if things go belly up and secondly “class B notes” which lose out first if things go bad. Obviously the class B notes have a higher rate of interest to compensate for this higher rate of risk. There are €577m of Class A and €2.9m of Class B debt. But the problem is that it is the Class A debt that is being downgraded.

The proposed solution by Holland Homes is not encouraging. It seems to consist of more financial wizardry:

1. A mandatory issue of a new class of debt (“the class S notes”), which will be subordinated to the class A notes and will be senior to class B notes in terms of payment of principal and interest,
2. The reduction of the amount outstanding on the class A notes to equal the amount of the new class S notes,
3. The undertaking of reasonable efforts to obtain a new rating for the class A notes from a second rating agency, subject to the condition that the class A notes will receive the highest credit rating.

(same source, edited for clarity) Which means that if you’re a Holland Homes investor, basically they’re diluting the Class A debt you hold into Class S debt and then playing the system so than can try and maintain a AAA rating on the Class A stuff. Which will undoubtedly work in the short term, but misses the point: the house-building industry is not a AAA industry. The risk doesn’t go away. Even the most highly stacked, highly classified debt is subject to the factors like the bankruptcy of the company.

This is where the fundamental disconnect happens, which led in part to the collapse of big banks like Lehman Brothers. The banks pile into the mortgage industry, and take the Class A debt, which is AAA rated. Translation: you’ll never lose money. Only a fool would miss out on this bargain of a century!

The problem is of course, that the wider view is not taken. You don’t have to be a genius to see that if Holland Homes losses in any year are greater than the €2.9m, then the class A debt gets affected. Lehman and other banks had taken positions such that only a tiny shred of the class A debt would have to be affected, such as 5%, before Lehman made huge, huge losses.

Consider this: 99.5% of all Holland Homes debt is class A. The shield of the class B debt was paper thin – 0.5%. But because class B debt existed at all, class A debt was supposedly invulnerable – so naive investors would look at its AAA rating and miss the bigger picture.

A 6% decline in the housing industry – which is really, really not very much, meant that Lehman would lose an enormous amount of money very quickly.

Abstract of the Month

People often like to take pot-shots at statisticians, psychologists and other data-driven professionals for seeing correlations and relationships where none really exists. Especially if it involves an over-complicated model.

But rarely does such criticism come from inside those professions. And even more infrequently does it come through the medium of satire. But out of the blue, a paper by “Uchen Benzimeni” (apparently serb-croat for ‘Unnamed Scholar’) does just that.

Age is often found to be associated with a plenitude of socioeconomic, politico-administrative, biological and thanatological variables. Much less attention has been paid by scholars, however, to explaining `age’. In this paper we address this unfortunate scientific lacuna by developing a model of `age’ as a function of several factors suggested by (post)rational choice and social constructionist theories. Using state-of-the-art multilevel statistical techniques, our analysis allows the determinants of age to vary with the institutional characteristics of European countries. Our findings convincingly show that generalized trust in strangers, support for incumbent extremist political parties in provincial elections held in the month of January, and the percentage of overqualified women in the cafeterias of national parliaments are all statistically significant explanations of `age’. Our findings have obvious implications for conspiracy theorists, organizational advisors, spin doctors and ordinary charlatans.

The only thing I’m left thinking about is how the hell it got past the editors.

Quote of the Day

They’re scum, it’s just important to realise that they’re not all reactionary scum. A lot of anarchists have been trying to get this [clean-up] going as well.

~Donnacha DeLong, President of the National Union of Journalists on the twitter cleanup efforts, in response to the following question:

Al Mikey: [What about] the reactionary part of these clean-ups like in Clapham Junction that have overtly classist twists to them. What’s your take on them?

Source. PS A minor battle of wills occurs in the comments thread below as to the precise nature of what Donnacha was actually getting at, see below.

Update: For you pleasure, here are some more gems of wisdom from Donnacha Delong’s twitter feed:

• Britain is a failed state – 2 real longterm choices – bolster it with massive repression #fascism or abolish it #anarchism (link)
• The #Labour Party’focus on police rather than youth services cuts shows they’ve learnt nothing from defeat (link)
• Not quite, repression is the primary response of a self-perpetuating hierarchical system that’s under threat. (link)

Which reminds me very much of this:

Deals, deals, deals

While the Senate is yet to vote on it, it looks like a deal has finally been reached between the 4 congressional leaders and President Obama to alter fiscal policy and raise the statutory debt limit.

I have naturally been following events intensely, partly because I am a sad strange little man keenly interested in US politics and partly because I seem to have lots of time on my hands in the summer holiday.

Column inches are pouring in all over on the internet on who’s up, who’s down, who gained and who lost. The picture building up is somewhat confusing.

For example, compare these disparate analyses. The New York Times offers up Paul Krugman, the left-wing economist, who writes an op-ed entitled “The President Surrenders” – in it he calls the deal a ‘disaster’. Ross Douhat, also at the NYT but more centrist, writes another called “The Diminished President“. The paper itself has written an editorial called “A Terrible Deal“.

The New Republic, the left-wing magazine which I think is the US equivalent of The New Statesman, also possess a healthy barometer of conventional left-wing thinking. “Kill The Bill!” shouts John B. Judis. “This is not leadership” pens power-blogger Jonathan Cohn (it’s worth noting the original title of that piece was “The Debt Deal is not as awful as advertised, but still awful”).

You can get even more zany headlines if you move from the print to the online froth-party that is left-wing blogging in the US.

So, DISASTER FOR OBAMA is the order of the day? Well, not so fast. It seems a lot of those opinions (especially the NYT ones) were in reaction to the headlines, rather than the detail, of the deals. Those who have taken closer looks are not so adverse.

The Economist’s WW asks Democrats to “relax!”, noting in his view that we “didn’t really cut a thing”.  His colleague Ryan Avent notes the fiscal outcome of the deal is mild (although he doesn’t think there’s anything to be happy about). Also supporting this theory of mine is the analysis of Nate Silver, who points out that really the detail is very favourable to Democrats (he argues the large defence spending cuts are a valuable weapon against the Republicans). Liberal Lamp Post goes so far as to declare “3 Big Victories for Obama“.

So, as is often the case, we seem to have two parallel political universes in operation: those who take care in reading the details, the actions of politicians, and those who pay attention to the theatrics, words and ‘spin’ of the political game.

If you follow the former, the outcome of this messy process is that the US will raise the debt ceiling (amazing that this is an achievement!), and reduce spending below the baseline over the next 10 years. Spending will rise, and the largest proportion of cuts comes from Defence than anything else.

If you follow the latter, Obama has folded, caved even, at a time when impoverished school-children in Somalia Minnesota are having their livelihoods swept underneath them by a nefarious conspiracy led by Big Oil/Pharma or Wall Street, seeking to profit from their downtrodden plight. Or something.

Certainly following the latter is a lot more interesting and exciting, if misguided. It’s incredibly boring to read the explanatory notes to Bills and reading Bills themselves can be a fate worse than death. Understandable why most let someone else do the digging, often after they’ve committed thoughts to print.

I remember reading Jonathan Powell’s book “The new Machiavelli: how to wield power in the modern age” for its opposition to greater Freedom of Information – the argument being that policy-making, like sausage-manufacturing, was a process best left unseen. While I’m not sure about weakening FoI (that it’s pissing off the executive is a sign of success, not failure in my book), certainly policy-making seems to involve lots of unsightly stuff such as arguing and shifting positions rapidly.

I guess we’ll know the true impact of the deal upon Obama’s Presidency in about a week’s time, when the dust has settled and even the hottest of tempers have calmed or moved onto another enraging subject.

But I would add three caveats to this fairly bland summary of other people’s analyses.

The first is that there is an odd desire upon the part of many to see Obama be a bit more, well, recalcitrant. To invoke fire in the bellies of his supporters, and be truly radical, rather than all things to all people. But equally this is probably a desire by some for politicians to be in permanent campaign mode.

The second is that while a deal is all nice and lovely and I’m sure many Republicans are patting each other on the back for this wonderful deal they’ve extorted out of the Democrats (even if the Tea Party might be unhappy), I think the Republicans have made a strategic error in using the Debt Limit as a negotiating tactic. For it is now easily foreseeable that Democrats will take their revenge upon Republicans at a later date, or that Democrats will campaign to remove the Debt Limit totally. Regardless, I have the feeling that the Republicans are playing with fire here and will pay a price for it.

Finally, in recent days I have grown noticeably less patient with the views of some on the left of the american left-wing. Stacie Borello of Liberal Lamp Post deems this group the “Bath Water Party” (she claims “No Perspective + Lack of Context + Assumptions = The Bath Water Party”). This insufferable group – a small lefty closed-loop, roughly equivalent to the Guardianistas of the UK (yes those who were of such hand-wringing purity that at the most recent election decided to abandon the main left-wing party when it was desperate for votes), has recently been turning up the volume.

A prime example of BWP nuttery was this article in Salon by Steve Almond: “What we wish Obama had said“. It only takes a few paragraphs before the smug, self-righteous attitude becomes suffocating, and you gag for oxygen as you reach for the ‘close-tab’ button at the top of your screen (or at least, I do). It’s a shame Almond can’t see his Michael Moore-style whiny liberalism is a season ticket to a Republican Presidency. I can’t decide whether it was the bit where he denounced all those concerned about illegal immigration as “fat, white, racist pigs” or that anyone opposed to greater financial regulation was “on the take” was more offensive.

I bring this final, seemingly unrelated point up because back over here in the UK, the Labour party is increasingly playing host to people of a similar persuasion. The Tory government is pursuing a similar course to the compromise thrashed out today – lowering spending in the medium turn. Meanwhile they receive criticism ranging from the technical or sensible to that they’re nothing more than the fiscal version of Harry Potter’s Dementors. And all this means is that the important – yes – details of the debate get completely obscured.

It will be interesting to see how these shrill criticisms evolve on both sides of the Atlantic.

Why I’m voting yes to AV

This is a long post, and I make no apologies for its length. If you want it summed up in one sentence, here it is: I support AV because it allows voters more choice in deciding who should be their MP, and ensures the MP elected has the support of the majority of their voting constituents.

What I find most infuriating about the current debate on AV is the fact that most of it seems to focus on tangential issues to the matter in hand: questions of ‘do we really need it’, (as if democracy were a frivolous expense), endless & contradictory accounts of which parties will benefit and who would have won which previous election if held under AV.

The merits and demerits of AV harbour in the none of these areas. Whether one party will gain or lose is not a reflection of anything other than the particular methodology employed in that projection. If a party stands to lose under the switch to AV, it will simply adapt its habits of campaigning or its political strategy to the new electoral system.

A further irritating factor in the current debate is that a lot of people misunderstand what it means to vote in the AV system. Putting a #1 next to Party A and #2 next to Party B does not mean you have voted for Party A. Counting up the first preferences and saying “this many people voted for Party A” is wrong.

Under AV, a voter express a preference between the candidates. For instance, if I preferred Labour above Conservative, and I preferred Conservative above all the other candidates, I would put #1 for Labour, #2 for Conservatives and give none of the other candidates a numbering. This overall numbering is my vote. I have not voted for Labour, in the same way I have not voted for the Conservatives. What I have done is said I prefer Labour the most, Conservatives second and I am indifferent between the others.

This misconception between expressing a preference and voting is one that has fuelled a lot of the criticism. One of the silliest criticisms is the idea that AV can produce a winner who did not win the most first preferences – this is the whole point of using AV! Another is that the voters whose first choice drops out get votes worth more than voters whose first choice stays in the race.

An excellent way of illustrating why AV is a good idea is to explain how it works. So let’s hold a fictitious election. Remember, to win election, you have to secure more than 50% of the vote. (And no, this is not arbitrary, it means that most people prefer you than anyone else).

Imagine there are six parties contesting an election: the Roman Party (RP), the People’s Front of Judea (PeFJ), the Judean People’s Front (JPF), the Judean Popular People’s Front (JPPF), the Campaign for a Free Galilee (CFG) and the Popular Front of Judea (PoFJ).

Election to the seat of Jerusalem Central
Name	1st Round	2nd Round	3rd Round	4th Round	5th Round
RP	32%	32%	33%	36%	43%
PeFJ	23%	23%	27%	35%	57%
JPF	21%	21%	24%	29%	-
JPPF	12%	12%	16%	-	-
CFG	11%	11%	-	-	-
PoFJ	1%	-	-	-	-
RP	8054	8055	8157	8411	8934
PeFJ	5789	5894	6662	8255	11838
JPF	5285	5295	5751	6648	-
JPPF	3020	3035	3787	-	-
CFG	2768	2834	-	-	-
PoFJ	252	-	-	-	-
Total	25168	25113	24357	23314	20771

So to sum up this table: the unified Roman Party holding a strong minority of popular support is finally beaten into second place by the People’s Front of Judea as the AV system redistributes votes according to preferences among the disparate Israeli Resistance groups.

If first preferences were the same as First Past The Post votes – which they are not – then the Romans would triumph always in the election, because although the majority disapproves of them, they are too fragmented to unify against the Romans. However under FPTP, because people vote tactically, it is conceivable that the support for smaller parties would be less than it is in the first round of elections here, to the benefit of the larger Israeli parties.

The single defining benefit of AV is clear – under AV people can’t win unless they have the broad support of their constituency – over 50% of the people who expressed a preference.

In Australia, where the law states that you must express preferences amongst all candidates, this means that the winner has 50% of the support of all who voted. In the proposed UK system, this is not the case. You will notice that the total number of voters in each round declines as some people’s preferences are exhausted and thus are thrown away. This means that the winning party secured 47% of all votes. However, all those who expressed a choice between the candidates in each round still had their vote counted – if they didn’t, they were indifferent to the outcome in the rounds after their preferences were exhausted.

Hopefully the criticism that some people have more votes than others has already become demonstrably false. To think that it is true is to simply misunderstand how the system works. You can see quite clearly that the election consists of five sub-elections, due to the fact that nobody secures 50% of the vote until the final round.

If nobody wins on the first round, everyone gets to vote again, with one less candidate. The AV system ensures that we don’t have to physically go to the ballot box each time, so if the candidate you voted for last time is still in the race, then it is assumed you vote for them again. If your candidate is knocked out, then your vote is moved to one of the remaining candidates depending on your second preference. If you have no second preference, your vote is discarded because you have decided to opt out of the remaining contest! To sum up then: nobody gets to vote for the same candidate more than once in any one round.

No electoral system is perfect, and AV like any other has its share of flaws. It is possible to ‘game’ the system, and contrive a situation where a candidate would be better off if less people gave him their first preference. Switching the ranking of your second and third preferences can affect the outcome of your first preference. AV can also produce situations where, if put in a head to head contest with every other candidate, the overall winner would lose some of those contests.

But to be perfectly honest most of the above are fairly academic considerations and hard to replicate in practice. AV has the advantage over FPTP of reducing the need for tactical voting, and ensuring that in tight races with several front-runners, a clear winner with the largest base of support is chosen.

If I were attacking AV (and defending FPTP), I would expect I would probably make three arguments: (1) the current system is fine, (2) AV is unnecessarily complex and (3) AV has the risk of producing wacky results.

I would counter-attack these three arguments by pointing out that the current system isn’t fine, that AV is a perfectly logical system that’s easy to use (hopefully I’ve shown this above) and the ‘wacky results’ really aren’t a cause for concern.

Firstly, FPTP isn’t fine. Yes, this fact is disguised by the many safe seats we have, where one party has a clear overall majority of support. But just look at the results of the Norwich South constituency:

Liberal Democrat		13,960	29.4%
Labour		13,650	28.7%
Conservative		10,902	22.9%
Green		7,095	14.9%
UKIP		1,145	2.4%
BNP		697	1.5%
Other		102	0.2%

Clearly in this fight, if run under AV, the Conservative and Green votes, if re-distributed, would have played a decisive outcome in the battle between the Lib Dems and Labour. The redistribution mechanism might have also meant that more Lib Dems or Labour supporters put their first preference for the Greens, who might have then beat the Conservatives into fourth place. We don’t know, but what we do know is that the tightness of this race, and others like it, should compel us to use a voting system that allows voters to cover all such eventualities – and that is what AV allows us to do, by giving voters the power to decide between all the candidates.

The second point my hypothetical critic made was that AV was unnecessarily complex. First of all, it’s not particularly complex, although many seem to wilfully misunderstand it. John Rentoul memorably described it as “number the candidates in order of preference; votes are then counted in the obvious way”. More seriously though, when you give voters the ability to be more sophisticated in the way they vote, you should expect to have to be more sophisticated in the way you count those votes.

But not only is AV not particularly complex, it is also not unnecessarily complex. In any two-party system, the best electoral system is FPTP. In fact with only two candidates, an AV election is a FPTP election. No further complexity is necessary, because the simple FPTP system is the best. That’s also why in America, unusual voting arrangements will never catch on while their two-party system continues to exist. When you get a three party system, or even a two and a half party system, the electoral system should be altered to reflect it, because otherwise FPTP cannot cope with the new nuances of the multidimensional elections that ensue. FPTP, as we saw above in Norwich South, allows candidates to win elections on less than 30% of the vote.

The third point of my hypothetical criticism was that AV had some fairly major flaws. I’m sure any supporter of AV would contest the significance of these flaws – I did so above – any quirk AV possess pales into the background when compared to the major flaw of FPTP – that candidates  can get elected with minorities of the vote amongst those who voted, and do so regularly. Less than one-third (216 of 650) of all MPs won election with a majority of those who voted. This is a disastrous situation, and needs remedying – the AV system is the best way of doing so.

So, to bring this lengthy discussion to an end. I support AV for two reasons, both on issues of principle. Firstly, I believe MPs should be elected with the support of the majority of people who voted. AV can produce situations where candidates don’t win election of 50% of those who went to the ballot box, but this is only because people actively chose to be indifferent to the end result. Secondly, I believe that a preferential ranking system – one where people number the candidates in order of preference – is fundamentally superior to one where people simply scribble in an ‘x’, because it allows the voters greater sophistication and eloquence in choosing their MP.

And that’s why I’m voting Yes to AV on May 5th.

Update: I forgot to mention in this thread that I support AV precisely because it is a ‘competitive’ (ie non-proportional) single-member electoral system. I think the system of winner-takes-all and one MP per constituency is a distinctive hallmark of British democracy which would bin at our peril. I would never support PR, for reasons which I may go into here on a different post, but needless to say AV is not a stepping stone to PR, as some people have claimed.

Update Update: I got into a discussion about AV on a different website’s comments thread. There I extended some of my arguments and rebutted the unthinking platitudes of a No to AV supporter.

Election to the seat of Jerusalem Central
Name	1st Round	2nd Round	3rd Round	4th Round	5th Round
RP	32%	32%	33%	36%	43%
PeFJ	23%	23%	27%	35%	57%
JPF	21%	21%	24%	29%	-
JPPF	12%	12%	16%	-	-
CFG	11%	11%	-	-	-
PoFJ	1%	-	-	-	-
RP	8054	8055	8157	8411	8934
PeFJ	5789	5894	6662	8255	11838
JPF	5285	5295	5751	6648	-
JPPF	3020	3035	3787	-	-
CFG	2768	2834	-	-	-
PoFJ	252	-	-	-	-
Total	25168	25113	24357	23314	20771

Why direct action is wrong

After the farce of yesterday’s TUC rally and the hooliganism that followed, a few apologists for anarchists have claimed that their actions were justified because either a) they were ‘driven’ to this by the wild actions of the government or b) peaceful protesting didn’t work for Iraq, so why should it work now.

Direct action, ie kicking a few shop windows in and claiming there’s a political rationale for it (similar to the IRA’s ‘political terrorism’), was wonderfully described by someone on twitter as “what football hooliganism is to football”.

Why is direct action so wrong? Well, imagine you have a generous grandmother who gives you a sizeable allowance of money.

One day, your grandmother tells you she is stopping your allowance because she needs the money to spend on looking after grandpa. Now, direct action would be the equivalent in this situation of going round to your grandparents house and beating her up until she gave you the money.

What could possibly justify that?

Ed Miliband’s latest speech

An interesting passage:

Millionaires with the right accountant pay nothing while pensioners pay VAT on fuel.

Offshore trusts get tax relief while homeowners pay VAT on insurance premiums.

Middle income taxpayers get stung, whilst perks and privileges at the top roll on unstopped.

And because the Government changed the rules, two million more people now pay the top rate of tax.

We will create a tax system that is fair which is related to ability to pay. Where the abuses end, the perks stop, and where ordinary families are not squeezed to pay for the privileged.

Another on tax and spending:

They have brought us injustice and division but these have not been the price of economic efficiency.
Because tax is also up – 800 pounds a year extra for the average family.

Spending is up and growth over the last 15 years is down.

And look at what they wasted on the way.
Billions of pounds gifted by Nature, the God-given blessing of North Sea oil, billions we could have invested in our future.
Billions they squandered.

One hundred and eighteen billion pounds – five thousand pounds for every family in this country – gone, wasted, vanished.

And to hide the truth of the nation’s problems they have sold our nation’s capital assets, built up over many years and used the proceeds not to invest, but to cover current spending. Seventy billion pounds gone forever.

Finally, one on principles:

Let us say what we mean and mean what we say.

Not just what we are against.
But what we are for.

No more ditching.
No more dumping.

Stop saying what we don’t mean.
And start saying what we do mean, what we stand by, what we stand for.

Whoops, that was Tony Blair in 1994. Compare that centrist approach – one where the leader reaches out to middle income families, where he attacks higher rates of tax and unnecessary spending – to Ed Miliband’s slovenly pandering to every passing public sector worker or trade unionist.

At least Blair had the guts to say what he would actually do, rather than just attack the Tories. And by the way, this was his first speech to conference – one he delivered 3 months into his leadership, and a good 2 and a half years before the next general election.

Boat Race predictions – a statistical approach

Update: Well the results are in – a double win for Oxford. The model successfully predicted the win for Isis, but its supreme overconfidence in a Cambridge victory backfired. As a statistician, I would put this to the model putting too strong a weight on the height differences between the crew.  The more underlying problem remains the poor quality of available data. As a rower, I would say it was Cambridge’s race to lose, and they underperformed. Regardless, the extra data point will be of use for next year’s race!

As an avid rowing fan (previously of King’s Chester RC, then Agecroft RC and now Newcastle University BC), I obviously take a keen interest in the Boat Race.

Standard discussions before the race often encompass whether being heavier, taller, or on one of the particular sides has any measurable advantage.

Doing a statistics related course, I thought why not use the tools I’ve being taught to come up with something a bit more concrete.

Before I go any further, I might as well give you the predictions, and then tell you how I got them. So, my computer says: Cambridge to win the Boat Race, Isis to win the Reserve Race.

I started out by assembling a large data set of what was available to me – every boat race going back to 2001. I compiled the average weights, heights, age, what side the crew was rowing on, how many blues they had each, what their BMI was, even the size of the squad. This was far too much data, so I pared it down to the most important few – weight, height, blues and which side the boat was on.

The computer then takes the data values of these and performs what is known as a multinomial logistic regression. This means it works out how the values of each piece of information affects the overall probability of either crew winning. So to be clear, the model does not bother predicting times or by how far either crew will win, instead it predicts simply who will cross the finish line first. For all the computer cares, they could be racing on a 2K lake.

It’s worthwhile pointing out at this point that the results the model came up with have to be taken with a pinch of salt. There really isn’t enough data out there to work out who won what with anything like the accuracy statisticians would normally expect. Statisticians like to use data sets in the thousands, not the dozens, and there really haven’t been that many Boat Races since 1829!

A final thing before I put the data table up – the computer pronounced on the old Surrey vs Middlesex debate. It said that, all other things equal, being on Middlesex increases the chances of that crew winning by 22%.

So here is the table documenting the relative height difference between Cambridge/Oxford and Isis/Goldie. If a number’s positive, it means that the named crew has an advantage that big over the other crew; a negative number is the disadvantage that crew is at compared to its rival. Side and Injury are what statisticians call “dummy variables”,  they only have values 1 or 0. For side, 0 means Surrey and 1 means Middlesex.

	Weight (kg)	Height (cm)	Blues	Side	Injury
Cam	1.62	0.04500	3.75	1	0
Goldie	5.56	0.00625	-0.75	1	0

So the computer then multiplies the values by a coefficient it has calculated.

	Weight	Height	Blues	Side	Injury
Cam	-0.44	4.52	0.65	0.97	0
Goldie	-1.51	0.63	-0.13	0.97	0

The values in the second table are added up to produce the z value. z is the sum of all the coefficient weighted values. p is the probability that crew will win, and is calculated by 1/(1+exp(-z)). The probabilities are also provided as odds.

	z	p	Odds
Cam	5.40	99.5%	1/99
Goldie	-0.34	42%	7/5

As you can see, the computer is very confident that Cambridge is going to win. I’m not so sure. But it makes for an interesting perspective.

As for Goldie/Isis, the race is much closer – to put things in perspective, my computer thought Cambridge last year had a 40% chance of winning.

Overall, I’ve tested the model against previous results – and it seems to get 3 out of 4 results right. Not perfect, but an interesting comparison. Whether it gets this boat race right or not – we’ll have to wait till 5pm!