Winner of the New Statesman SPERI Prize in Political Economy 2016

Tuesday 19 April 2016

In defence of George Osborne over Brexit

This by Fraser Nelson in the Spectator (HT Tim Harford) starts well: “Sometimes, George Osborne’s dishonesty is simply breathtaking.” Who could disagree with that? Except that the statement Nelson objects to is the following:

Britain would be permanently poorer if we left the European Union, to the tune of £4,300 for every household in the county. That’s a fact everyone should think about as they consider how to vote.”

Nelson does not object to the economics behind the number, set out clearly in a Treasury study released yesterday. (For an excellent review of the study, which makes both of the points I make below, see Chris Giles here.) Instead he has two objections:

  1. With economic growth we would not be poorer under Brexit, just less richer than we would have been if we had remained in the EU

  2. The household figure is derived by dividing the GDP ‘loss’ by the number of households in the UK.

Nelson makes the point that household after tax income is only about two thirds of GDP/households. So implicitly he is saying is that we shouldn’t count lower taxes (and therefore government spending) and investment (future incomes) when assessing whether people would be poorer. But that seems silly. We all benefit from total government spending and investment, so we would feel less well off if we lost some of that. [1]

Indeed I have done exactly as the Chancellor has done when assessing the impact of 2010 austerity. I calculated, using OBR figures, that austerity cost each UK household at least £4000. The two figures appear comparable, but in fact they are not. My figure is a total one-off cost, on the (admitted very optimistic) assumption that the UK economy had completely recovered from 2010 austerity by 2013. The Brexit cost is a continuing loss each year.

Alas for Fraser Nelson dividing any GDP loss by the number of households is standard practice among economists (see John Van Reenen here for example), and we do it to make our analysis more relevant to those who do not commonly think in terms of GDP. It is also common practice to think about counterfactuals: if we did X (Leave) rather than Y (Remain) how much better/worse off would we be. It is just much clearer to do things that way. Who knows how much richer we will be by 2030: that would be a pretty unreliable forecast, because it depends on pretty well everything. In contrast we can be much surer (although still uncertain) about what the impact of just one change (leaving the EU) will have.

Now if you wanted to avoid any ambiguity, you could rewrite the first sentence of the Chancellor’s statement as follows:

““Britain would be worse off if we left the European Union compared to if we stayed in, to the tune of £4,300 for every household in the county by 2030, and for each and every year after that.”

If this is honest, does that make the original version dishonest. I do not think so.

If this report illustrates anything, it is that it would have been much more effective to have launched a 2003 joining the Euro type exercise immediately after the 2015 election victory, consulting widely among outside experts, and perhaps getting some of them to write key parts of the report. That would have produced a wider range of numbers for the cost, and the Chancellor could have then chosen the higher one, simply inserting ‘up to’ in front of it. But here I am, defending George Osborne and advising him on spin. I think I better go and lie down for a while.

[1] Slightly pedantic economic point: we should really use GNP rather than GDP, but it makes little difference here.  


  1. I'm no supporter of Brexit, in fact I'm most likely to vote for the EU, but IMHO arguments such as Osborne's do economics no favours, and supporting his ideas does economists no favours.

    You can argue all you like about how the number is calculated, but it rests on the assumption that GDP will fall. Can you or anybody argue that with any certainty? [BTW, I believe it would, but I wouldn't want to put a number on it] Especially given that mainstream modelling has such a poor record in predicting the future [as far as 2030!!].

    The Brexit case will be that the GDP prediction is false, or that any impact on GDP will be mitigated by policies that they will be released to follow - they are, after all, promising to end austerity, save the NHS and give free apple pie to everybody.

    Osborne has always been a politician rather than an economist, and he wanted a big number, and he's been given one. But this kind of "forecasting" is just disreputable. Formulae can give you precision, but that precision is useless if the inputs and priors are junk.

  2. If employers started employing and training up local labour through apprentice schemes rather than rely on inflows in hundreds of thousands of cheap labour from Eastern Europe and this gave rise to a reduction in long term unemployment and 'bottom up' and more balanced growth how does this effect the 4300 figure? Are these potential effects embedded in the model?

  3. The 'less richer' rather than 'poorer', point is quite important. The BBC were reported that the economy would shrink by 6%, which seems a shockingly biased presentation.

  4. Have you seen Osborne in the Guardian:

    "George Osborne has said the British government would lose £36bn in net tax receipts, equivalent to 8p on the basic rate of income tax or 7p on VAT, if the UK leaves the EU and negotiates a bilateral trade agreement with the bloc."

    The whole thing is a total fabrication to introduce further public spending cuts. It is blame shifting from Osborne.

    government investment => taxation + increase in private savings

    Each time, every time for any positive set of tax rates.

    How much the economy grows or shrinks is mostly down to the level of government spending undertaken by the government to support the businesses of the UK. Similarly the BoE and Osborne have full control over 'borrowing costs' as the BoE can just buy any bond that falls below par and cancel it QE style.

    The only way you can 'lose' that much taxation is if you spend £36bn less on public services.

    And that is actually Osborne's intention.

    If we leave the EU we just need to elect a government capable of using the new freedoms from the finance restrictions (Article 123) in the EU treaty and off we go.

    At the moment:

    Government spending from cash buffer/intraday overdraft => taxation plus private savings => purchase newly created government bonds => QE by the Bank of England of government bonds.

    And we *only* have to do the double shuffle because the EU treaty prevents us from simply running an overdraft at the BoE.

    Besides that I am far less pessimistic on Brexit than you Simon. The main (and very large) issue I see is with David Cameron doing the negotiation.

    Remember the Eurozone economy is the one in depression and needs ‘export-led’ growth.

    Wouldn’t the Irish simply veto any restrictions on the UK that are likely to result in retaliation from the UK?

    The woodford report is pretty good reading. I don’t agree with it all because it is in neo-classical language:

    Here’s the crunch paragraph.

    “In addition, falling tariffs, the decline in manufacturing and Europe’s diminishing importance in the global economy mean we doubt that even the absence of a trade deal with the European Union would hurt the United Kingdom’s overall exports materially. The benefits of being in the European Union are smaller than they were a few decades ago, when a Brexit would have been a far bigger deal. However, the effects will vary across sectors. Brexit would give Britain a crucial opportunity by allowing it to broker its own trade deals with non-European Union countries; indeed Britain could even have a unilateral free trade policy. Non-European Union countries may find negotiating with Britain easier and quicker than dealing with the European Union’s bureaucratic machine, as Switzerland has shown.”

    Similarly David McWilliams over in Ireland doesn’t see much of a problem:

    It is in no-one’s interest to upset the 5th largest economy in the world. Particularly when there is no demand from anywhere else!

  5. If your head has cleared after your wee lie down could you address the point of why the treasury report was not on the more standard GDP per capita measure rather than households. Further, does per capita make any difference.

  6. Of course your gloss - and the Treasury's document - rests on many suppressed premises. For example, 1) we couldn't, in any imaginable world, trade as freely with EU countries as we do now. Free trade needs the present bureaucracy. Nothing would promote the end as well as the present means. The arrangements of other countries exhaust the possible models for the world's fifth largest economy. 2) Trade with the rest of the world wouldn't be better as a consequence of leaving. The sclerotic EU bureaucracy that can't seem to manage anything, in hock to the obstructionist whims of any of 28, is a leaping panther of decisiveness next to anything we might do for ourselves. 3) There are no peculiar externalities attached to being a member. Anything that would damage the UK as a member of the EU would be equally (or more) damaging to the UK as an independent country trading with the EU. 4) There are no economic consequences that follow from the free movement of persons, or they're all positive, or we'd have to have such free movement regardless of whether we are members.

    I could go on. a) The answer that comes out of your model depends on what you put in. b) There are Rumsfeldian unknown unknowns. c) Just like forecasting the present value of shares using expectations of future earnings growth, the models are exquisitely sensitive to the assumptions you make.

    What weight, for example, do you and the Treasury attach to the (I'm sure) four standard deviation event of Britain being a member - with no new relationships with the rest of the world which we might have had - when there's a complete meltdown in the Eurozone in 2030? Might there be two worlds, one where we're either £4,300 up or destroyed, and another where we're fine(ish) whatever happens? Is this included in the model?

  7. "In contrast we can be much surer (although still uncertain) about what the impact of just one change (leaving the EU) will have."

    Now you are of course right here in that you can be more certain about the impact of one change rather than many but it's still has an air of sophistry about it.

    The "still be uncertain" qualification is still a huge uncertainty and, in my view, renders all these exercises as nugatory at the end of the day. The dynamic effects will outweigh anything else and are simply unknowable.

    You must be perfectly well aware that the record of such august bodies as the IMF and OECD at short term forecasting is very poor (for very good and understandable reasons) so why one should dignify an exercise such as that undertaken by the Chancellor in trying to forecast what the world will be like in 2030 as being anything other than totally speculative escapes me.

    The methodology may be defensible but the overall principle is not.

  8. I think we can mark both yours and George's efforts with a D grade: 'must try harder'.
    If you take a one-eyed view of the world you will come up with some peculiar perspectives.

    1. Indeed 'U' do StuartP. Or should that be StUartP...

  9. > Alas for Fraser Nelson dividing any GDP loss by the number of households(...)

    This does seem flimsy though, given the disparity in wealth between households, the majority of households are far below the mean exposure to GDP.

    So to say that Bob in his council flat is going to get the same 4k hit as George in his mansion is a little disingenuous.

  10. This Telegraph Tory is - I'm assuming - also making no sense, despite the fact that he seems to be making perfect sense.


Unfortunately because of spam with embedded links (which then flag up warnings about the whole site on some browsers), I have to personally moderate all comments. As a result, your comment may not appear for some time. In addition, I cannot publish comments with links to websites because it takes too much time to check whether these sites are legitimate.