Ireland’s Power in Europe: Guest Post by Macartan Humphreys on the Lisbon Treaty

We are pleased to carry this very insightful article by Macartan Humphreys of Columbia University (and a graduate of TCD!): you can download this article here.

17 replies on “Ireland’s Power in Europe: Guest Post by Macartan Humphreys on the Lisbon Treaty”

This is academically interesting, but fails to deal with the actuality in countries like Ireland, where there is consensus between the government and main “opposition” parties and their MEPs on most things EU-related, and the viewpoint of citizens is diminished or ignored — this is highlighted in the case of the Lisbon Treaty (and locally, NAMA, where a deeply unpopular coalition government ignores the popular will and the advice of leading economists).

It is also worth pointing out that Ireland only represents about 1 percent of the population of Europe, but according to this paper (which is very interesting) it controls 2 percent of the decisions. I am not sure if this means that Ireland is in fact overrepresented in Europe. My feeling is that Ireland is in fact overrepresented in Europe, both in voting and in practical terms through the presence of its representatives.

There is not much sense in complaining that Europe is undemocratic on the one hand and then complaining that Ireland’s influence is being diminished on the other.

There are obviously other issues to deal with regarding democracy in Europe that don’t have anything to do with voting strength.

An excellent article, crystal clear, hard facts, unemotional.
Surprising for a man named after St. Macartan who was consecrated by St. Patrick near Monaghan. Five generations in the US does improve the attitude of Celts.

Humphreys writes about voting in general and the Banzhaf index in particular. He does not apply it to the actual voting rules under Nice and Lisbon, however. His point “in general more decisions could be made more easily under Lisbon than under Nice” is therefore unfounded.

It is also untrue. See
Barr and Passarelli (2009), Mathematical Social Sciences
Widgren (2009), CESifo Economic Studies
Napel and Widgren (unpubl):

Double majority voting means that it will become easier to block a decision. Under Lisbon, it will also be easier for the European Parliament to block decisions, and for national parliaments to slow-down decision making.

The claim that Lisbon will make it easier to make decisions is patently untrue.

(I happen to think that that is a good thing as “Brussels” too often makes bad decisions.)


Many thanks for posting this article. Couldn’t access it yesterday because there was some kind of techie glitch prohibitiing access via internet explorer.Whether he’s right or wrong on some of the technical details, the article serves to debunk the ‘no’ campaign’s myth that Ireland, under Lisbon, is losing voting power within the EU.

Your point about the European parliament and national parliaments is well taken.

Unfortunately I could only access the last of the three papers you cite which use game theory to examine decision making by the Council of Ministers.

Having said that, the third paper, to which I have access, specifically states that it is concerned with comparing QMV under Nice with QMV under Lisbon. It should be borne in mind that under Lisbon a significant number of areas move from unanimity to QMV in the first place.

I would not be an expert in the Banzhaf and other voting indexes. Something that quickly occured to me though is that the Banzhaf voting index essentially adopts a binomial random coin tossing approach to sampling voting configurations. Such a procedure will introduce strong biases of its own. Being a binomial distribution there will be a central limit theorem like strong bias/concentration towards the centre. Voting configurations with roughly equal numbers of yes and no votes will predominate. The other complications of double majority voting will of course complicate this trend but random sampling should nevertheless impose a strong bias. The question is would real world voting patterns reflect this? A brief look around some of the literature has confirmed this unease of mine about this general approach. A paper by Gelman, Katz and Bafumi and a paper by Mulligan and Hunter demonstrate that actual real world voting patterns often lead to very different conclusions on voting strength. Theoretical work such as the Banzhaf index would indicate voting strength is in general roughly proportional to the square root of population size. This would appear to indicate that the influence of smaller countries can be much greater than actual voting power indicates. But these papers belie this and would indicate that in real world situations it appears voting influence is often just simply proportional to one’s voting weight. There can be quite subtle biases involved with the above type of theoretical analysis. It would therefore appear to me that seemingly simpler arguments based on just counting voting weights are probably just as valid as more sophisticated analysis by simulations or voting indexes.

Very true. What really matters, of course, is the probability that a decision would be made that is bad for Ireland. As long as Ireland’s interests are aligned with those of the real power brokers, all is dandy.

One of the Irish Members of the previous European Parliament consistently argued against the interests of Ireland and against the expressed preferences of the Irish electorate, at least in the areas of energy and the environment. Formal power (as measured by Banzhaf) is not the only thing that matters.

Correction to above: Banzhaf indicates voting strength of an individual (rather than country) is proportional to square root of country voting strength, so am incorrect in above and Banzhaff theory would imply larger country over-representation as opposed to smaller country over-representation, but wouldn’t affect my main point about the need to be wary of random sample biases with this type of analysis.

Thanks for the various interesting responses. I agree with the thrust of these.

For Richard: in fact the note is based on my calculations using the actual voting rules, not the general theory. If you want to examine the calculations they are done in R and the code is here: As to whether the decisions are easier or harder under Lisbon, the claims are based on the number of “decisive coalitions”; under Nice only about 2% of all possible coalitions could implement a measure, compared to about 13% under Lisbon. The Nice rules are surprisingly complex. So that’s the basis for that claim. Other non-treaty information can and should be factored in if you want to make statements about which or other of these coalitions are more *likely* to arise and some of the pieces you point to do that.

Here Finbar’s comments are right on target I think. What really matters for actual voting power is what coalitions are actually likely, and that should be central to the discussion. Interestingly the literature on aggregation that uses data on actual pivotality on the low level units (rather than assuming all coalitions are equally likely) suggests that the voting weight of countries ought to be proportionate to their size. That literature is conceptually tricky but would suggest that Ireland is *overrepresented* under both sets of rules (ie it should have one percent but in fact has two percent).

So I agree (and I make the argument in the original piece) that you cannot capture voting power properly by looking only at the rules; but the point is that many of the arguments informing the discussion, and in particular the arguments on voting weight given by Sinn Fein, Corr, and others *do* base their analysis on the rules, but only look at one part of it: they look at relative population sizes (see for a recent discussion) rather than factoring in the other components. So perhaps the best summary is: calculating relative power based on voting weights as given by the treaty only gets a part of the story. One should instead look at what is actually likely to happen with a Yes or No vote. But *if* you are going to make arguments based on the rules alone you shouldn’t do it by looking only at population weights, which greatly exaggerates the adverse effects for Ireland and which, sadly, is what most of the discussion is doing right now.

Thanks for the thought provoking article! It helped open my eyes as to the complexities of comparing voting systems. And yes, a simple relative comparison of populational voting weights cannot be the full story! One has several other factors thrown into the mix. The fall in the populational threshold in the double majority from 74% to 65% might tend to make blocking proposals harder but then again the raising of the country double majority requirement from 12 to 14 would tend to make blocking votes easier (probably increasing to a degree the voting power of small countries). And on top of this the new minimum four country requirement to block a vote would curb the voting power of larger countries (at least to an extent). A complex business! 🙂

Correction again!: in the above should read “raising of the double majority requirement from 14 to 15” instead of “from 12 to 14”.

If I read your code correctly, you have a single qualified majority rule for Lisbon, while there would a double qualified majority rule: X% of the votes (which you include) plus Y% of the population (which you omit)

Excellent article. On the wider issue of Lisbon and democratic accountability…see these comments from Ambrose Evans-Pritchard of the Telegraph where he identifies correctly (in my opinion) the wider issues about accountability and the EU.

His article is here:

The relevant bit is:

‘Critics call the Lisbon Treaty a “federalist blueprint”. That muddies the issue. It in fact concentrates power in a unitary state, giving the European Court jurisdiction for the first time over the whole gamut of EU affairs (all three pillars, in EU jargon – the Community pillar, the common foreign and security policy pillar and the pillar devoted to police and judicial cooperation in criminal matters). It will adjudicate over the Charter of Rights. Euro-judges will have power to reshape British society by court ruling if they so wish, just as the activist Warren Court reshaped America.

‘By creating a full-time EU president and by giving Euro-MPs the power of the purse, it mimics nationhood. Yet it should be obvious that Europe cannot ape the institutions of the historic nation states in this way. Shifting power from London, Madrid, or Copenhagen to the EU core does not transfer democratic accountability: IT BREAKS THE LINES OF ACCOUNTABILITY (my caps..).

The main problem I have with the whole project is that its a kind of supra-corporatism – whereby the Council or Member States are the basic voting units for key decisions: hiring commissioners, appointing judges. They then don’t so much delegate as abandon authority to legislate / interfere these appointees. e.g. the NGOs consulted by the EC when proposing new directives usually comprise organisations that are 100% funded by the EC itself!! A sort of off balance-sheet advocacy..

Comments are closed.