Archive for May, 2003

94942846

May 27, 2003

Struck gold!

Junius, among others, spotted the same article as me in today’s Guardian about the Enron episode, and a rather good and thoughtful one it is too (although frankly, the reference to Wittgenstein is a bit forced) But the real gold strike is the webpage of the bloke who wrote it, Donald Mackenzie, who is (it sez ‘ere) a Professor of Sociology at Edinburgh University. You wouldn’t think that an article entitled “An Equation and its Worlds: Bricolage, Exemplars, Disunity and Performativity in Financial Economics” would be any good, but actually it is. Also the best account of the LTCM affair I’ve seen, plus an extremely good paper on the sociology of financial markets which demonstrates the massive quality gap between the “behavioural finance” crowd in economics and proper sociologists who do fieldwork and everything. Science studies at its best, applied to economics, and he’s clearly actually looked at the financial economics literature before taking it on (fans of Andrei Shleifer’s Clarendon Lectures “Inefficient Markets” will be very interested).

A clearer candidate for a Houblon-Norman Fellowship I could not think of, by the way.

Health Warning: If you read all those papers at one sitting, they get a bit repetitive …

Advertisements

94695799

May 21, 2003

The tribute that vice pays to virtue

What with my acid remarks about “leftist hypocrisy” below and my occasional expressions of sympathy for social libertarians, some people in comments (and, as ever, in vast volumes of fictitious email correspondence all of which agrees with me, except some terrible, abusive fictitious letter-writers who everyone who disagrees with me should be ashamed of) have started suspecting me of harbouring Conservative tendencies. Oh yeh, and I said I was going to vote Tory at the next election. That also might have set people thinking.

To an extent, they’re right. I’ve just received my P60 and an itemised bill telling me precisely how much Camden Council proposes to charge me for 2003. Added to that I increasingly dispair of the likelihood of the ECB ever democratising, I’m having trouble distinguishing one pudding-faced Geordie fixer on the Labour frontbench from another and the Home Office still haven’t done anything about that bastard running a pirate radio station on the LBC frequency and ruining the Charlie Gregory show for me, and it’s perhaps not surprising that I feel nostalgic for a government that would combine xenophobic authoritarianism with a slightly lower higher rate of income tax. Feh.

Anyway, the Tory vote is non-negotiable. Blair lied to me about Iraq, I wanna lash out at him, Iit’s stupid and irrational but I’m going to do it anyway. But the question of “hypocrisy” bears a bit more explanation because it does appear to go to the heart of a lot of people’s emotional politics.

Think about it this way. In my post below, I suggested that the difference between the progressivity of the tax systems students suggested for income versus for their own grades “might serve as a useful index of the hypocrisy of leftist students”. When I use the word “hypocrisy” here, what do we actually mean? Well, the combination of the following two qualities:

1. A moral belief that (some loosely defined concept of) equality is (an actual or instrumental) good.
2. A personal desire to accumulate more, even at the expense of others.

The first is simply a baseline definition of what it means to have left wing politics. The second … well put it this way, Buddhist monks spend twenty years living ascetically and meditating for hours at a time before they presume to believe that they have conquered all selfish desires. If you’re talking about “leftist hypocrisy”, you’re just talking about “leftists who have not been able to transcend history, biology and socialisation in order to develop an unparticularised love for all sentient things”. In other words, you’re just talking about “leftists who happen to be humans”.

Contrast with rightwing politics. As I’ve posted earlier, the single most sensible thing said in political philosophy in the twentieth century was JK Galbraith’s aphorism that the quest of conservative thought throughout the ages has been “the search for a higher moral justification for selfishness”. Some rightwingers are not hypocrites because they admit that their basic moral principle is “what I have, I keep”. Some rightwingers are hypocrites because they pretend that “what I have, I keep” is always and everywhere the best way to express a general unparticularised love for all sentient things. Then there are the tricky cases where the rightwingers happen to be on the right side because we haven’t yet discovered a better form of social organisation than private property for solving several important classes of optimisation problem. But at base, the test of someone’s politics is simple; if their political aim is to advance all of humanity, they’re on our side, while if they have an overriding constraint that the current owners of property must always be satisfied first, they’re playing for the opposition. Hypocrisy doesn’t really enter into the equation with rightwing politics; you don’t (or shouldn’t) get any extra points for being sincere about being selfish.

So where does that leave our students? Well, they’re young. They’re most likely insecure. They don’t actually have a lot, and it’s hardly surprising that they’re a bit precious about what they have (a close runner for the most sensible thing said in political philosophy in the twentieth century was Michael Oakeshott’s remark that “a conservative is a man with something to lose”, and the genius of this remark is its ambiguity). One shouldn’t blame them for not being Boddhisattvas.

In general, one of the biggest problems with the psychological politics of left and right is the need that people feel to think of themselves as not just having made what looks like on balance the best decision given the things they regard as important, but as morally good people themselves. People in general seem to be horribly uncomfortable with the idea that, by the standards they use to judge political situations, they themselves don’t come out as moral heroes. At base, this is a fairly childish and decidedly illiberal attitude; childish because it demands a sort of moral perfection which everyone intellectually knows can’t exist outside fairy stories (unless you count the way that parents appear to their children) and illiberal because it suggests that you’re only prepared to have normal social interactions with people who pass your own personal moral examination (a rather prominent political philosopher has told me to my face on a couple of occasions that he regards me as morally beyond the pale because of the job I do; I’ve nonetheless been made to feel very welcome at his house).

So anyway, hypocrisy in people is not a vice, particularly when the alternative is to be sincerely horrible. In political parties, it’s much worse; people who presume to take control of the state’s monopoly on the use of ppowerhave to be held to a much higher standard of honesty, because they are explicitly asking for us to trust them on matters important to our lives. A double standard? Perhaps. But I just told you that I don’t care about hypocrisy. Perhaps I should have termed my imaginary measure above an “index of political self-righteousness”. On which score, it seems fairly clear, the political science professor himself would outscore most of his students.

94672819

May 20, 2003

Thankyou Mr Mankiw

I have no post to go with this headline, it’s just that some day the current Head of the Council of Economic Advisers might do something really good and I want to prove I thought of this rhyme first.

Update: I’ve just realised I have no real idea whether his name is pronounced the way it looks or not. If everyone is reading this as “Thankyou Mr Moernkiv”, please forgive me.

Update: He’s another one with a Robert Barro grin, by the way.

94443709

May 16, 2003

Ooooh, Mr Wilson, Oooooh Mr Heath

This guy has a sweet little thought experiment designed to sharpen up a few people’s intuitions about egalitarianism (like all we cool blogger people, he has dispensed with the outmoded concept of permalinks; scroll down, you lazy swines). The experiment:

At the beginning of class I ask my students how many of them are in favor of progressive redistribution — taking from those who have a little more and giving it to those who have a little less. About half to 60% of the class stands up (I make them commit to their position by standing up.) I then tell them what I actually was thinking about was the progressive distribution of their grades, taking a few grade points from those who are above the median grade and distributing those extra points to those below the median. (I also propose less severe redistributions that would not put everyone at the median, but still would have the effect of collapsing the grades to the class median.) The immediate reaction is that almost all the students sit down, only one or two students actually remain standing or stand up. Assuming that most of them thought I was originally referring to income (or wealth, not the same thing), I then ask them to explain why they were in favor of income/wealth redistribution but not grade distribution.

And what do his class do? Well, start rationalising away like fury, apparently, rather like everyone with a weblog who’s commented on it. Grades aren’t like money because they’re meant to be about people’s real ability (bollocks) they can’t be exchanged for things in the economy (yes they can). Of course, this apparent inconsistency between people’s views about what ought to be done with other people’s money versus what they say when it might affect them, is taken as great evidence of leftist hypocrisy. And to be fair, to a pretty significant extent, it is.

But at base, I still think it’s a trick question. Here’s why:

1. Try another thought experiment at the end of term. Get all the ‘B+” students to stand up. Tell them that at the end of the year, three students have bad enough results that you will have to throw them off the course. Then offer them the deal that if the “B+” students standing agree to be marked down to a flat “B”, you will let the flunked students stay on. I’m interested in the finer grains of this prejudice against redistribution of grades and suspect that the B+s will agree to a bit of fudging in order to save people from really desperate straits.

2. But the real trick is in the ambiguous phrase “progressive redistribution”. Is it a pleonasm1? Not really. Progressivity and redistribution are not the same thing when one applies them to a tax system. If I get it into the tax code that every year we have to take a $100 bill from Bill Gates and burn it then that’s progressive (the incidence of the tax increases as one goes up the wealth distribution) but it’s not redistributive. Progressivity is quite a specific feature of tax economics, and it’s not just as simple as “taking from the rich and giving to the poor”. The point of view that taxation should enforce a smoothing out of the income distribution is actually quite a stronger point of view that simple support for progressivity in the tax code. One might believe what William Petty did, that there were a number of necessary expenses to the running of the state, and that the incidence of paying for those expenses should fall for the most part on those “with the greatest interest in the public peace”.

Since the “grade tax” proposed above is just a pure redistribution rather than a financing of necessary expenditures (you can’t use grade points for feeding the poor or defending your class against its enemies), the analogy is testing a stronger view than the one the students would really support. I must say that myself I can’t really see the point in a progressive tax policy just for the sake of evening out the counters, but I’m quite passionately in favour of the principle that people ought to be given a fair go of things, that “a fair go” is a much more substantial concept that most people’s view of “equality of opportunity”, and that the rich should pay more for this important social good. After all, if there was a revolution tomorrow then my binman would presumably still be in his old job, but Warren Buffet would be looking at a seriously reduced standard of living. So Warren Buffet ought to pay more than my binman for the cause of keeping the proles contented. But that doesn’t mean that I’d necessarily support the principle of taking Warren’s wallet and sharing it out pro rata.

3. Possibly, therefore, one could try the following experiment, again at the end of term rather than the beginning. Say “the department has informed me that I need to reduce the GPA of this class until it fits on a curve. Sorry, not my fault and all that, but as an aggregate, you lot are going to have to be docked some points. Now how do we decide who to dock them from?”.

I suspect that the class would agree to some sort of progressivity in the redistribution schedule, but not as much as they might suggest for the marginal rate of income tax. The difference between the answers would be a useful quantitative measure of exactly how hypocritical left-wing students are, a question that I suspect a lot of us would be mildly interested in.


1Look it up, look it up!

Endnote: I only posted this here because I couldn’t get his comments section to work. I saw the link onJunius

94286914

May 13, 2003

Log Books – an Impudent Suggestion

Important caveat: It’s just struck me that anything you put up on the internet has a chance of being used for mischievous purposes. So I’d just like to clarify that I make repeated references to John Lott in this essay for two reasons a) his paper happened to be the subject of the debate in the comments section that set me off on this line of thought and b) I think my lay readers will be more familiar with “More Guns, Less Crime” than most of the rest of the econometrics literature. I have no special knowledge of how John Lott goes about his work, and no evidence of any kind which might support claims about whether he is more or less scrupulous in his use of iterated stepwise regressions than any other econometrician (I’d note that the problem is endemic, so it would hardly be a criticism of Lott if he had all the faults described in this article; I certainly do). I happen to dislike the guy’s work, and I personally don’t trust him because he’s attached to the American Enterprise Institute, but this is far more likely to reflect my personal political prejudices than anything else. As far as I can tell, in the current edited version of this essay, there is only one reference to Lott’s work in a context where it might be reasonably suspected that I’m referring to him pejoratively (a joke about the Stanford Law Review that I’m leaving in because it’s funny), and this reference would apply equally to his critics. Everyone should make up their own mind about Lott’s work based on the evidence and on detailed analysis of his work, and this essay is neither. If John Lott or his representatives want to complain about particular references, I’d be grateful if they used the comments link below for first point of contact, as I have a curious phobia about posting my email address on this site.

Anyway, on with the show.


What with things like this, plus the whole John Lott fiasco, econometrics is having a pretty hard time of things at the moment … and probably, deservedly so. Basically, read on for a story about one of the dirty little secrets of econometric analysis and my suggestions for something that could be done about it. I call this an “impudent suggestion” because, naturally, my personal contribution to the published econometrics literature is pretty precisely fuck-all. But anyway …

A statistician whose name I forget once said of the Ordinary Least Squares regression model that it was the internal combustion engine of econometrics; something so convenient that everyone tended to use it for trivial tasks where it wasn’t really needed, without regard to the polution it caused. This was a bit precious, perhaps, but nowadays, computer power has reached the point at which everybody can have the SPSS regression package (I believe that’s the standard in universities; I’m a TSP man myself) on their desktop, along with a CD containing 5,000 different economic data series for whatever they want, plus the Internet to download most any specific data series collected for whatever question they might be interested in. And this is the point at which the internal combustion analogy really does become valid; it is not at all healthy for the general advancement of economic knowledge to have the power to do OLS regressions so democratically distributed.

To pursue the analogy with cars a bit further, we all know that it’s simply not possible for everyone in China and India to drive a Ford Galaxy. The world just doesn’t have enough oil. And OLS regressions are also pretty intensive in their consumption of a scarce resource – a scarce resource called “uncontaminated data”. This is the point, by the way, at which matters start to get both a bit technical (an inevitable result of the territory, which I am trying to mitigate) and more than a bit inaccurate (both a result of my trying to simplify things for the layman, and my characteristic error-laden style). I happen to believe that it’s massively, monumentally important that the educated plain man should be able to understand this sort of thing, and that if he can’t it’s the fault of professionals for not explaining it better. I also believe that anything that can be understood, can be explained to someone with no specialist background, without making material oversimplifications. If I don’t achieve either aim, I hope someone will pick me up in the comments.

Anyway, the issue of data contamination is this. When you’re building a model of something, here’s one of the ways you might choose to go about it. (NB that first, this is a very general and oversimplified characterisation, which will be sharpened up considerably below, and second that there are a number of other ways you might set about this task, all of which share this problem, but not always in as transparent or obvious a way).

  1. Decide on the structure of your model (take an easy example; we’re going to assume that consumption equals X times income plus Y times wealth).
  2. Estimate your model (in this case we take the US consumption, income and wealth numbers from, say 1945 to 1971 and use SPSS to pick the values of X and Y which minimise the “in sample prediction error”
  3. Now, you might be a bit worried that your model had fitted itself to random features of your data rather than to the underlying structural relationship you assumed was there.
  4. So, you can now use the data for income and wealth for 1972-2002 to generate thirty years’ worth of “predicted” consumption data and compare it to the actual. If your “out of sample prediction error” isn’t too bad, then you’ve got a model that might be valid, and you’ve learned something about the relationship between consumption, income and wealth.

Fine and fun. But what happens if your out of sample prediction error is laughably awful, as it most likely would be in such a model? Well, the obvious thing to do is to go back to step 1 and alter your model a bit; say, try to model the logarithm of consumption, or changes in consumption, or add lagged values of consumption as an explanatory factor, or something. But now you run up against a really serious problem, which I shall illustrate with a Christmas cracker joke.

Q: How many hamburgers can William “The Refrigerator” Perry eat on an empty stomach?
A: Only one. After that, his stomach isn’t empty any more.

Similarly, how many times can you test a model against out-of-sample data? Only once; after you’ve done that, the data isn’t out-of-sample any more.

Unscrupulous readers will have noticed immediately that if the referee didn’t see it, it didn’t happen (a fortiori, if you’re publishing in a non-refereed journal like the Stanford Law Review, it certainly didn’t happen), and that there is nothing to stop you from doing the statistical equivalent of cheating at patience and “resetting the clock”, trying out lots of models on the 1945-71 data until you get one that has an acceptable fit with the 1972-02 data as well. You can even, I am ashamed to say, buy computer programs which carry out this process (“step-wise regression”) for you; you just feed in the data series that you want to find a relationship for and they try out all sorts of models until they come up with one that has decent performance. Or, if you don’t own one of these programs, you can do what economics graduate students get up to on those rare evenings when they lack hot dates, and sit in a computer lab, mindlessly pressing the button on SPSS over and over again, looking at the goodness-of-fit statistics and trying to ignore your undeniable and growing resemblance to the rodent half of one of BF Skinner’s operant conditioning experiments.

Obviously, it wasn’t possible to do this back in the days when linear regressions had to be carried out by armies of men with slide rules, or even in the days when you had to book time on the faculty mainframe, but technology has advanced to the point at which SPSS can be an acceptable substitute for Minesweeper for a certain kind of econometrician. As a social pastime, obviously, endlessly re-estimating permutations of regression models has all the disadvantages of masturbating into a sock with few of the advantages, but is it really all that harmful? Well yes. I’m not just being a luddite here.

The problem is that at the end of an evening spent in this fashion, our economics graduate student will have finally arrived at a model which fits the data very well, but of which it is almost impossible to judge the significance. Why?

Well (digressing for the layman here), let’s take a fairly standard significance test. Specifically, let’s take the absolutely standard t-test for significance in a multivariate linear regression (a test of this sort lies at the heart of John Lott’s “More Guns, Less Crime” paper, for example). The idea behind this test is pretty simple. Along with the regression coefficient (in the model above, X and Y were the regression coefficients), the software calculates something called the “standard error”. The standard error in this context is a measure of the variability of the coefficient; the extent to which that particular coefficient would have needed to be bigger or smaller to explain the various data points in the sample. And, under the usual assumptions, the estimated value of the coefficient can be treated as a draw from a probability distribution with a mean at the “true” value of the coefficient and a dispersion around that mean defined by the standard error. So what you do to “test significance” of a coefficient is to take your estimated coefficient, divide it by its standard error to “standardise” it, and then look up in a book the critical values of the relevant distribution (it’s called the “t-distribution”, or for those who are showing off, “Student’s1 t-distribution”. Oversimplifying mightily, the book tells you, for a given “t-ratio” (the ratio of the coefficient to its standard error”, what the probability is that a random draw from a t-distribution with mean zero would be no greater in magnitude than the t-ratio you’re looking up. So if I get a coefficient of 6 (an estimate of 6 for X in my equation above) and a standard error of 2, then my t-ratio is 3. I look this up in the book and (say) find that a random draw from a t-distribution with mean zero will be greater than 3 or less than -3 no more than 2% of the time. This is great news for me, as it means that I can say that (speaking loosely, or as a Bayesian) I can be 98% confident that my coefficient is not zero, or that (speaking rigorously, or as a frequentist), in an arbitrarily large number of similar studies, I would only expect to see results like mine 2% of the time if the true value of the coefficient was zero. In other words, it’s highly unlikely (speaking loosely again) that there is no relationship between the variable associated with my coefficient X (income) and the dependent variable (consumption). Or in other words, this variable is “significant”2.

Or is it? As you notice above, I’ve stressed on a number of occasions that the book of t-distributions tells you what to expect if you make a random draw from a t-distribution. But, if you’ve been up all night pressing the fire button on your regression program over and over again until you get a sufficiently high t-ratio on your coefficient of interest, then that t-ratio isn’t a random draw from a t-distribution. It’s a number that could only in principle be about 1.9 at the lowest, because you’ve iterated through the regressions until you found one. There is a very big problem if you pretend that a number created in such a way is a t-ratio randomly drawn from a t-distribution and start pretending to do statistical tests on it as if it was. If you publish an article based on this sort of methodology in a journal (or worse, submit it as a policy paper), you’re actually polluting the information space, because your results are indistinguishable from genuine statistical analysis, but they are damn near to being statistically meaningless. Not only that, but anyone who reads your article is contaminated; if they ever do any work of their own on the dataset that you did, they start from a basis of knowing how your model fit the data. Laymen reading this, if you take nothing else away from this article (and if you take nothing else away, sorry for wasting 1904 words of your time), remember every time you see a piece of statistical analysis which looks definitive, you normally have no guarantee that it wasn’t produced in this way.

This is the issue of “data mining”, and we had a bit of a go at it in the comments section a couple of weeks ago. In actual fact, there are all sorts of ways in which one might go data mining, and some of them are far less pernicious than others. In particular, I have to confess a soft spot for the “LSE Econometrics” approach of David Hendry3 and his mates. Under this approach, you don’t start by choosing a model at all; you use a highly general model to begin with (Hendry refers to his method as “general to specific”), one which is structured so that it has a lot of redundancy and explains the data (another slogan of the LSE school is that one should “let the data determine the model”, and I’ll explain below why this is a more controversial slogan than you might think). You then start placing restrictions on the general model (for example, you might delete a coefficient by restricting it to be zero) and re-estimate, then perform tests to see whether the restriction in question is one that impairs the model’s ability to explain the data. If the restriction doesn’t reduce the explanatory power, then the restricted model is said to “encompass” the more general model; it does the same amount of work with fewer assumptions. Using a very well-thought-out algorithm for searching through the potentially massive space of possible restricted models (an algorithm implemented in the program PCGets, whose authors it is fair to say are a bit sensitive about criticisms of data mining, you finally arrive at a model which, like the Reality of Divinity, “encompasseth all created things”. This is your “local data generating process”, your best estimate of the structural model which generated the dataset that you started estimating your general model on. Now your job is to interpret that LGDP; to see what theoretical model corresponds the best to the data. Contrast this with what one might call a Popperian approach to econometrics – the more usual method of setting up a model which corresponds quite closely to your a priori theoretical view of how you think the world works, and then setting it up against a dataset to see whether the data falsify your theory or not. In general, people who are sympathetic to the LSE approach tend to be a bit more interested in philosophy of science (and thus less inclined to be knee-jerk Popperians) than econometricians of the more orthodox tendency.

On the other hand, the LSE approach isn’t a panacea. It has a number of problems:

  • You can’t be sure that you’re going to come up with a theoretical interpretation of the LDGP that makes sense. Unless your interest was in forecasting a process rather than understanding it to begin with, that leaves you no better off than when you started.
  • Although encompassing and general-to-specific methodology is a way of data-mining which doesn’t give flat out deceitful results like the caricature of stepwise regression based on t-ratios I described above, it’s still pretty difficult to interpret the results of PCGets’ output. You don’t really have a feel for the sensitivity of your estimates. It’s much better for macroeconometric work (where you’re usually either forecasting, or attempting to test the validity of an entire theoretical approach such as rational expectations) than for microeconometrics (where you’re often more interested in the specific coefficients).
  • Related to the above, there is a considerable danger of developing a “theory of a single dataset”. Although the problems of statistical validity of data-mining are addressed by the LSE approach, the issue of overfitting and poor out-of-sample performance persist. If you develop entirely different LDGPs for, say, the UK, the USA and Japan, then no matter what your forecasting performance, how credible are your results? Not very, I’d say, although there are people who do a lot of semi-interesting work about “model uncertainty”.

So there is still, in my opinion, a significant role for non-LSE-style, Popperian econometrics, to answer questions like “If we pass concealed-carry laws, will there be less crime?”. And there is no real equivalent of PCGets for this kind of econometrics; I disagree with the commenter a couple of weeks ago who thought that data-mining in this area wasn’t always and invariably pernicious (the comments are attached to the story “Readers’ Digest” and very good they are too. If you aren’t absolutely 100% confident of the good faith of the people involved (and on a politically sensitive issue, how could you be?), then there’s a very good chance you’re being fed sawdust rather than steak.

Which is where my proposal comes in. In an ideal world where I was the wealthiest man alive, I would run the world’s most prestigious econometric journal. It would be called the Journal of Definitive Results, and it would have a submission procedure as follows:

  1. The editorial board would decide on a problem to be tackled
  2. Having decided on this, researchers would get grants to assemble datasets. These researchers would be barred from taking any further part in the project
  3. Thirty different teams from different universities around the world would be invited to contribute to the Spring issue of the JDR. This would be entirely theoretical, as everyone contributed a paper on how they felt the problem should be modelled. Nobody would be allowed to see the data at all until the Spring issue was published, at which point, half the dataset would be released to the econometricians from each team.
  4. The Summer issue would have two sections. In the first section, the theorists would respond to each others papers from the Spring issue, while in the second section, the econometricians would write methodology papers on the subject of the correct approach to be taken to modelling the data.
  5. The Autumn issue would have the econometricians responses to each others’ papers. On publication of the Autumn issue, the second half of the dataset would be released.
  6. Each team would contribute one econometric study on the first half of the dataset and a second study, using the same methodology on the second half, to the Winter issue. The Winter issue would be introduced with an editorial summary of the theoretical debate, and would conclude with a metastudy of the thirty econometric papers by an eminent econometrician not connected to any of the teams.

Obviously, this isn’t going to happen any time soon, and like all pipe-dreams, it would presumably be impossible to make it work at an interpersonal level. But a weaker proposal based on the same ideas might have a chance.

One of the things that they do in some natural sciences is to keep lab books. This is a book in which you write down every experiment you perform. Partly to keep track of what you’re doing, partly to help in future patent lawsuits and partly to avoid duplication of effort. It strikes me, though, that the introduction of this practice to the social sciences could help to stamp out a lot of the more mindless “data dredging” (like data mining, with the difference that a miner occasionally strikes gold but a dredger just stirs up shit) that takes place. If you had to write down what you were doing every time you pressed “estimate”, and if you knew that you would have to submit your log book along with your results to whatever journal you intended to publish in, and that people would tend to look askance at your “significant at the 99% level” results if they found out you’d estimated 500 versions of the same model before you got them, then you might be a little less trigger happy with SPSS. And this is the real source of most of the data dredging that goes on; not calculated deception but just the fact that a lot of economists have experienced something close to operant conditioning at the hands of their regression software. There will always be people who intentionally iterate through models in order to reach a politically convenient conclusion, but if we can find them guilty of the heinous crime of “forging laboratory notebooks”, it will be a lot easier to chuck them out of the profession. Hopefully


1“Student” was the pen name of the bloke (William Gosset) at the Guinness factory who invented this incredibly useful distribution in 1908. I have no idea why the Guinness factory was employing statisticians but they were, and they also had a policy of not allowing them to publish journal articles under their own names (presumably Guinness Breweries was not subject to Research Assessment Exercises). So “Student” it was.
2Statistically significant that is, not necessarily practically significant. In all honesty, “statistical significance” is really just a measure of how many data points you have, which is something you already know or ought to. But everyone treats the two meanings as equivalent and I am not inclined to kick against the pricks.
3“LSE Econometrics” is a bit of a misnomer; Hendry is actually at Oxford and so are most of his school. But the tradition of this approach apparently goes back to LSE.

93887335

May 6, 2003

Tactless voting1

Apparently a few people in the comments section (and vast volumes of fictitious email correspondence) want to know what the merry hell I’m playing at by saying I’m going to vote Conservative next time round. So I suppose I’d better clarify the matter. I probably made a mistake in the way I phrased the issue below, because it implied that I was only going to vote Tory if I could be sure that they weren’t going to get in. That’s not really true; I’m going to vote Tory and I really don’t care whether they get in or not. This is an explicit “moral equivalence” position; I simply don’t think that it makes a difference whether the blue Tories or the red Tories get in. They’ve already done all the things that they were claiming the Tories would do (abolishing the right to silence, university tuition fees, introducing workfare, being complete bastards to asylum seekers), and their performance on that little lot means that they will hardly be convincing when they come round in 2006 trying to put the fear of God into me about the Health Service and the last vestiges of progressivity in the income tax. People have been treating this “they’re all the same”, “New Labour = Old Tory” thing as if it were a joke for too long. I’d even argue that there is, these days, far less ideological distance between our two main parties than there is between the Republicans and the Democrats. The centre of gravity in UK politics is a little bit further over toward the Big-State side of things, but even that’s not particularly anything to be proud of; both Blair and Duncan-Smith appear to be firmly of the belief that the main purpose of the State is to make people do things that they don’t want to do, unless they can afford not to.

Needless to say, this point of view scandalises my friends. Objections to it fall into two categories: 1) but the Tories are such bloody horrible people!, and 2) why don’t you just vote LibDem/Green/not vote? Taking them in order:

First, yes, it really, really goes against the grain to vote for the other side. Not all Tories are bad people, but a lot of really awful people are Tories, and voting for them gives direct succour to the worst kind of fox-hunting, race-baiting, queer-bashing arsehole. Given that I tend to favour theories of morality which have a lot more to say about the kind of people we are, than what we do, this matters to me. Added to that, good old-fashioned team spirit is a powerful force. Even now I know what kind of saccharined authoritarian nightmare Blair has been, I still can’t forget how great it felt to give the other lot a bloody nose back in 1997. Believe me, I wouldn’t be considering this point of view if I wasn’t really extremely angry.

Why so angry? Well, it’s basically the lying thing. In general, I can’t claim to be an absolutist on the subject of lying. On the whole, I’d actually probably say that for any given lie, I’m as likely as not to be in favour of it. But I have a complicated code of morality which helps me to distinguish between types of lie; the actual features which I distinguish probably lie in the area of Hayekian “tacit knowledge” rather than anything else, but I attempt to codify them below out of a belief that other people are likely to recognise some sort of truth in them.

  1. Lies told about matters of fact are worse than lies told about one’s own opinions.
  2. Promising to do something and then not doing it is not as bad as promising not to do something and then doing it.
  3. Lying to someone before they’ve made their mind up on what to do, is not as bad as telling a lie calculated to change someone’s mind.
  4. And most importantly, lies told by people who have the monopoly on the legitimate use of force to people who don’t are per se worse than any other relevantly similar lies

This to me suggests why I am more outraged over the cocktail of falsehoods that have been thrown our way over the recent war, than about the general run of fibs that all politicians of any stripe tell us. Recognising that I can’t prove that any particular member of the government was actually lying (rather than passing on in good faith some terribly untrue information gained from sources they regarded as trustworthy), I have concluded that if these statements were lies, they were very bad lies indeed. And since I’ve always regarded “innocent until proven guilty” as a wonderful rule of procedure for putting a man on trial for his liberty but a bloody awful way of selecting a people’s representative, I’m operating a somewhat more relaxed standard of evidence for making my personal assessment of who to vote for. In any case, I’d note that since the Crichel Down affair, it’s been a general principle of British government (albeit one honoured in the breach rather than the observance, and one pretty badly damaged by Michael Howard) that a Minister ought to resign when it is shown that he has misled the House, even if he did so in good faith.

Which brings me onto the second question; that of why the Tories rather than some other form of protest vote? Well, for one thing, I can’t just stay at home as a protest because I’m already doing that as a protest against a) the number of meaningless elections, local elections, European elections, London mayoral elections, devolution referenda and other such calls on my time, and b) against the Labour party’s continued failure to make good on its promises to the drinking man.

(This last one was a big issue for me in 2001, and directly contributed to canvassers not bothering with my house anymore. A sample conversation on my doorstep went as follows:

New Labour: Hello sir, I see you’re down on my list as a Labour supporter. Can we count on your support?
D^2: I’m sorry, I don’t wear one. ([c] Spike Milligan, 1922)
New Labour: (insincere laugh) But seriously, will you be voting Labour?
D^2: Depends … what is New Labour offering the drinking man?
New Labour: Oh plenty! We intend to lower the rate of alcohol duty to European levels as soon as is practical ie never, and even better than that, we have a manifesto commitment to open the pubs 24 hours a day!
D^2 Oh. Yeh. Right. Sure. And I suppose that while we’re in the 24 hour open pubs, we’ll be drinking out of the oversize “fair measure” pint glasses that you promised us last time you wanted to get elected? My memory hasn’t gone that much, you bastard. I’ll be staying at home on polling day, drinking the last remnants of my Calais trip.
New Labour: I’ll put you down as a “maybe”.
D^2: While you’re here, can I ask what your party’s policy is on telling fucking lies2 in order to start a fucking war? That’s the sort of thing which often influences my vote.

Sorry, I appear to have digressed)

But anyway, I’m already on a stay-at-home protest. And the LibDem/Green Axis of Ineffectuality doesn’t really appeal. Mainly because both those parties are extremely rich in a) the kind of meddling, everything-would-be-so-much-better-if-only-everyone-listened-to-my-precise-200-point-plan type I really can’t stand (the kind of person who spends evenings on design of voting systems is always going to make me nervous) and b) the kind of careerist I-want-a-career-in-politics-but-couldn’t-quite-make-it-in-a-proper-party-so-I’m-going-to-stand-in-the-West-Country type, that I can’t stand. Plus, ironically, lots of people I really, really, like. But there you go.

But also for sound economic reasons. As anyone who’s studied the economic theory of signalling (for which Spence won his Nobel gong) will know, there’s no point in doing something to “send a message”, unless it’s something that it costs you to do. Voting for a more leftwing party would just send a sign to Blair etc that Davies has dropped off the left edge of British politics into the swamp of People Who Can Safely Be Ignored Forever. What I want to do is to show Blair that faced with a choice between him and the Tories, I’m prepared to select the option that is both lesser and evil – the choice which is genuinely repugnant to me – in order to send a message to him about being a lot more careful about the bloody “intelligence reports” with which he pollutes the public discourse during a time when a nation is trying to make up its mind whether it wants to go to war. In order to do that, I have to do something I don’t want to do.

I don’t know whether, when shite comes to bust, I’ll be able to actually go through with it. I’m more likely to sit at home, particularly if there’s something good on the telly. But I know this; if any canvasser comes to my house, I am for certain going to tell them that I’m voting Tory because Blair didn’t tell the truth about Iraq. If enough people do this, it might come through in the focus groups.

I would never presume to tell anyone else how to cast their vote in a secret ballot. But telling little lies to politicians strikes me as the smallest step toward a world in which politicians don’t tell big lies to us.

Update: Note that it is perfectly consistent to take one view about whether we should have fought a war in Iraq, but another view altogether about whether Blair should have misled Parliament about Iraq.


1Thanks to the chap in the comments section who suggested this name.

2Oh all right then, “material falsehoods”, in case anyone feels like waiving their Parliamentary privilege in order to sue me.

93662550

May 2, 2003

Big with Nuts

Oh dear. As a political junkie, I think I’ve hit rock bottom. Last night, Mrs Digest came up the stairs late at night wondering why I wasn’t in bed and found me watching coverage of the (UK) local elections. I honestly think I’d have been less ashamed if I’d been caught watching a porn channel. I have never remotely been tempted to vote in these elections (or indeed to find out whether I could have voted; I’m not sure whether they had them in London), but I’m just addicted to the coverage. I love the fancy logos (the BBC had a particularly good “Vote ’03” one which frankly deserved a better election), the silly computer graphics of debating chambers, the “results coming in” from places I couldn’t point to on a map, the sheer self-importance of it all. And, of course, the sadist in me loves to see how badly most politicians cope with staying up late at night. (Thinking about it, I seem to remember posting a while ago about how the appeal of boring sports like snooker was that they basically turned on the psychological destruction of one of the contestants. I was also watching the snooker from the Crucible last night, and it sort of segued into the elections. Same appeal).

But although I’m an addict, I’m not so far gone as to believe that these elections matter. In fact, I would support the view that local council elections matter a negative amount; the better a political party does in them, the worse a state it is in. To explain this view, let us consider the makeup of the electorate for these elections. We know (or at least, credibly guesstimate) that of the voting population of the UK:

a) 2% are fascists
b) about 4% are Welsh or Scottish nationalists.
c) 10% are civic-minded souls who will turn up and vote just because there’s an election.
d) A thick 10% are single-issue nutters of one kind or another
e) about 1% are more or less fictional characters because they are the result of ballot fraud.

All of these groups will be voting in the local elections for sure; the fascists and nationalists have to turn out because it’s their only chance of getting anyone elected at all, and the other three groups have to vote by definition. Already, with these groups, all five of which are by any reasonable standard weirdos, we’re up to 27% of the electorate. A further 5% of the population belong to political parties, and about four-fifths of these can be expected to get out and vote. We have now reached 31% of the electorate and turnout was only 36% in the local elections.

The weirdoes have usually made up their mind how they’re going to vote on issues of no relevance to a general election campaign (I include the “civics” in this category, because they make up their mind based on their assessment of the issues and thus are swamped in a proper campaign by a mass of voters who just treat it like a football game). The party hacks vote the party line. Therefore, when we’re “taking a barometer reading of the nation” in these elections, we are looking at a sample which is made up of fourteen percent regular Joes who just happened to be walking past the polling station, plus eighty six percent cranks, ideologues and hacks. If you’re doing well with that crowd, you’re in trouble.

The problem is that the single-issue cranks are the swing voters in local elections. If something’s on the cover of (say) the Daily Mail every day, then you can guess that about eighty per cent of the readership don’t give a fuck, and just turn to the features about Catherine Zeta Jones getting a bit fat and the Fred Basset cartoon. But the more gullible fifth will be foaming at the mouth with rage as directed, and if it’s the Mail, that’s four hundred thousand voters out of a turnout less than 15 million. One party ends up being the beneficiary of this 2.5% swing, and believes that there’s been a massive endorsement of their entire platform. Since one tabloid or another (in fact, usually more than one) is always in the self-righteous rage game, this swing is completely random, and since usually the tabloids are catering to a niche audience (people who need to be told they are right all the time because at some level they realise their views are bizarre), there is a come-uppance at the general election when the population as a whole make up their minds. There are real consequences to this, unfortunately; hundreds of people have quite likely been tortured and killed because the political class of the UK got it into their head that being “soft on asylum-seekers” was an important political issue on the basis of local election results.

Bottom line: “Really popular with nutters” isn’t an election winning strategy, but it can get you a good showing in UK local elections. All of which makes me even more determined to vote Tory next time. I really want to punish Blair for lying to me, and I now know that I can do so without taking any material risk that the buggers will get in. And the frightening thing for Duncan Smith is that my kind of support is probably quite a material proportion of his entire franchise.