Statistical significance

It would take a better statistician than I am to parse this from Ritchie. But purely by eyeballing it looks like the variance is so huge that I\’m not sure there\’s any statistical significance to the trend.

BTW, if there\’s an economics student out there, someone who can do the graph stuff with Excel etc, I\’ve a little project that might amuse. Perhaps a day\’s worth of work (at very most….but it would take me a week I\’m sure).

Take the position on the economic freedom index, strip out the two of the ten factors that are related purely to size of government (ie, we\’re now measuring how \”neo-liberal\” a place is, Denmark being number 1 I think, not how socially democratic a place is) and then plot that against GDP per capita. Scott Sumner has I thnk done something like this but I\’ve not been able to find that paper/blog post.

This is also amusing:

The Tea Party is a threat to just about everything of real value. Even democracy itself.

A democratic party in a democratic election, putting out its wares for the populace to make their democratic decision, is a threat to democracy.

Ho hum.

13 thoughts on “Statistical significance”

  1. Oh dear, why have you taken up the absurd American abuse of “parse”? Didn’t they teach you anything at St Ignoramus?

  2. …..A democratic party in a democratic election, putting out its wares for the populace to make their democratic decision, is a threat to democracy……

    Tim, not that it applies to the Tea Party, but have you not heard of the concept:
    One Man, One Vote, One Time?

    After all Hugo Chavez was voted into power, a development that has been undoubtedly a negative one for Venezuelan democracy.

  3. Hasn’t he got the x and y axes the wrong way round? He hasn’t even bothered to label the x axis, but I presume he wants it to be GDP per capita.

    But surely the x axis should be public spending and the y axis GDP per capita.

  4. This is RM’s reply to his critics, (those that he lets post at all), and it is just jaw dropping:

    “But let me tell you something really important as an accountant and one time auditor that you as a non-accountant need to know. That is that data whether it be accounts, stats or anything else, and any model, whether of economies or anything else, is a remarkably poor substitute for sound judgement. And sound judgement says that the association obviously exists. And sound judgement would also say that variations between countries and economies are also inevitable. But that this does not alter the very obvious conclusion I came to.

    Of course you may exercise unsound judgement if you so wish. That’s your right.

    But it won’t change the reality of the conclusion I’ve drawn.”

    I think Mr Flibble has better reasoning skills.

  5. “A democratic party in a democratic election, putting out its wares for the populace to make their democratic decision, is a threat to democracy.”

    Well, hitler did kind of did that, didnt he. Your leap of logic that democratic choices can’t be a threat to democracy is rather unfortunate.

  6. Interesting; there seem to be about 40 more countries on the CIA Factbook list (his main source) than in his Excel chart.

    I’ve not gone through the whole list to see which ones are missing, but a few obvious ones are:

    Jersey
    Guernsey
    Isle of Man
    Gibraltar
    Liechtenstein
    Monaco
    Andorra
    San Marino
    Cayman
    Bermuda
    British Virgin Islands
    Bahamas

    Have we spotted a connection yet? All low-tax, high GDP per capita.

    I’m not saying that he’s deliberately removed them to make his graph work; some data sets don’t include them, depending on their definition of a ‘country’, so it may just be that he didn’t have the data.

    But when comparing tax levels with GDP per capita it doesn’t make sense to exclude the very places that have built a very successful economy on low taxes.

  7. But the wider point is that if (as we believe) tax affects GDP per head, then it does so over time.

    Therefore comparing one year of tax revenues to a level of GDP that has built up over decades doesn’t really tell you much (which Richard Murphy concedes in his blog post).

    More relevant is the comparison I did of government spending vs job creation – because that shows how taxes are affecting movements in the economy now. And there the evidence is unambiguous – more tax means fewer jobs.

    See here:
    http://blog.iea.org.uk/?p=3159
    (Sorry, don’t know how to do the linky thing)

  8. This is exactly how Nazi Germany started – people going around demanding lower taxes and less State power.

    I guarantee that if these people win elections then the very next day the Jews will start being rounded up.

    Which should please a Socialist, I would have thought

Leave a Reply

Your email address will not be published. Required fields are marked *