Thank God Matt Yglesias is not a Rawlsian

I always thought the Rawlsian minimax principle was rather odd.  Recall that John Rawls once argued that public policies should be implemented if and only if they improved the welfare of those who are worst off in society.  I came up with all sorts of bizarre counterexamples, like what if a public policy massively improved the welfare of the top 99%, but slightly reduced it for the bottom one percent.  Say it made them $2 worse off.  At the time, I never thought my silly thought experiment would ever show up in the data.  Until today.  I found this graph at Matt Yglesias’ blog:

Yglesias argues that this graph shows that the British public did better under the Labour party than under the Conservatives.  He bases that claim on the fact that most of those in the bottom half did better.  I have no problem with that argument; it’s based on solid utilitarian reasoning.

As an aside, I still think the Conservatives did more to improve Britain.  They inherited a country going down the tubes, and made some very painful decisions to shut down a lot of uncompetitive manufacturing and mining.  The made the economy more efficient.  They ended double digit inflation.  These reforms hurt various sectors of the public, but were needed in the long run.  In contrast, Labour inherited an economy in very good shape, and left a fiscal train wreck when they left office in 2010.  And a bad recession.  Notice the data only goes up to 2008—let’s see how it looks in 2 years when we have the full data showing the Labour government’s entire term in office.

But I digress.  My main argument here is that Yglesias is quite rightly ignoring Rawls’ silly maximin principle.  The poorest of the poor didn’t do well under the Conservatives (losing about 0.2%), but they did even worse under Labour, losing 1.1%.  Rawls would clearly vote Conservative, but for the wrong reason.

If I thought this graph actually captured all the effects of government policy, I’d probably vote Labour.  But as I said, my hunch is that Labour was dealt a somewhat better hand.  And I think their record will look worse when extended up to 2010.

Tags: ,


19 Responses to “Thank God Matt Yglesias is not a Rawlsian”

  1. Gravatar of Morgan Warstler Morgan Warstler
    30. September 2010 at 07:29


    Rawls concerned himself with outcomes, as such, the application of theory is NOT focused on a single society, as though there is an island of people who know only of themselves, using Rawls’ ideas to become as just as possible.

    Instead, you are to approach it as a Very Poor Person thinking rationally from a veil of ignorance.

    You are then looking at All the possible societies, and seeing in which model you would have the best life at the bottom. That society then is the most just. And it does NOT follow that Rawls would say, “OK, let’s take the most society thats system has produced the most justice, and screw around with what is working!”

    Instead he would concern himself with what good cause would force all the other societies to become like the most just one…. which is of course OURS.

    This is why disparity means so very little. This is why Rawls is considered an apologist for capitalism.

  2. Gravatar of Indy Indy
    30. September 2010 at 07:33

    And, of course, it’s even worse reasoning than that, because it’s all about what that chart would look like *relative* to the unknowable counter-factual of a change in policy or party in political control.

    And a less-ideally-Rawlsian curve that was nevertheless higher at every point would clearly be preferable from a utilitarian perspective, if not necessarily a progressive political one.

    At ant rate, “minimax” is a much better description of what Rawls’ system looks like. It’s completely Orwellian not to create a new precise coinage for the concept and to instead choose to commandeer the ancient concept of “Justice”.

  3. Gravatar of ssumner ssumner
    30. September 2010 at 08:00

    Morgan, I’m guessing Rawls wouldn’t agree with your interpretation of his views.

    Indy, Yes, that’s a good point, we need to consider the unobserved counterfactuals. I wasn’t suggesting the graph proved anything, just that if people were going to take it seriously, Rawlsian values would imply an odd interpretation of the outcome.

  4. Gravatar of Wonks Anonymous Wonks Anonymous
    30. September 2010 at 08:23

    If you know you are a Very Poor Person, you aren’t behind a complete Veil of Ignorance. I thought Rawls’ argument was that if you didn’t know how rich or poor you were, you would be most concerned about the worst possible outcome.

    Rawls was anti-capitalist:

  5. Gravatar of Benjamin Cole Benjamin Cole
    30. September 2010 at 08:43

    Speaking to Scott Sumner’s example of 99 percent of the country benefitting from a hypothetical action, I sometimes ponder what would happen if the US seized the assets of the top 400 richest Americans, which I think is about $1 trillion. Yes, it would be wrong; a violation of property rights.

    On the other hand, the rest of the country could enjoy a tax holiday for several months, due to a microscopic number of Americans being wronged. Sheesh, we could leave each billionaire with $1 million.

    Wrong? Well, who dies in our wars? That’s wrong too, and the youngsters getting killed (or maimed) pay with a lot more than just money.

  6. Gravatar of StatsGuy StatsGuy
    30. September 2010 at 08:58


    “But as I said, my hunch is that Labour was dealt a somewhat better hand. And I think their record will look worse when extended up to 2010.”

    I predict that there will soon be a cottage industry in taking time series regressions from pre 2007 and re-running them to include that last decade. I suspect lots of things we ‘know’ will turn out to be historically contingent.

  7. Gravatar of Blackadder Blackadder
    30. September 2010 at 09:25

    I sometimes ponder what would happen if the US seized the assets of the top 400 richest Americans, which I think is about $1 trillion. Yes, it would be wrong; a violation of property rights.

    On the other hand, the rest of the country could enjoy a tax holiday for several months, due to a microscopic number of Americans being wronged.

    My guess is that the long term cost of doing this would be a lot greater than $1 trillion.

  8. Gravatar of Morgan Warstler Morgan Warstler
    30. September 2010 at 09:36

    Yeah Scott, you’d be wrong.

    “People must recognize that they cannot make up for failing to regulate their numbers or to care for their land by conquest in war, or by migrating into another people’s territory without their consent.”

    Rawls concerned himself greatly with getting to his realistic utopia, where we have negative rights, primary goods, and capitalism (even with great disparity).

    And HOW we get there? It was based on the above. Take multiple societies and compare them. The best one is where the poor have the most of 1 and 2, and every other society has good cause to alter their own approach.

  9. Gravatar of Morgan Warstler Morgan Warstler
    30. September 2010 at 09:45


    no, Rawls was a tricky bastard, he used the demand to aid the poor first as justification for disparity / capitalism.

    His VOI argument is: since you might end up poor, you will rationally only look at the long-term bottom (how much of 1 and 2) of any given system to determine which guarantees you the best lifestyle.

    this is accurate to my recollection, and I asked him about it at harvard in 1991.

  10. Gravatar of scott sumner scott sumner
    30. September 2010 at 17:58

    Wonks Anonymous. That’s also what I thought. He may have favored a mixed economy.

    Benjamin, Unfortunately, it would do more harm than good in the long run. Short run it would be nice, however.

    Statsguy, Good point.

    Blackadder, I agree.

    Morgan. That’s what I did here. The best society is Conservative (according to Rawls) because the bottom 1% are better off.

  11. Gravatar of Peter Whiteford Peter Whiteford
    30. September 2010 at 18:08

    When one looks at the original Institute of Fiscal Studies report that this chart comes from they point out that for the people with the lowest incomes in the surveys, cash incomes are not accurate measures of living standards: “To be precise, the roughly 1% of children living in households with incomes below £50 a week have average living standards comparable to those with incomes of £250 to £500 a week. The lowest average living standards are to be found amongst children living in households with equivalised incomes of £100 to £200 a week, which represents about 11% of all children and corresponds to roughly 30-50% of median income.” And these families became better-off under Labour.

  12. Gravatar of Maxim Maxim
    30. September 2010 at 19:13

    Scott — thank you for your posts like these. I’m glad there’s a blogger out there who’s explicitly utilitarian. 🙂

    What bothers me most about Rawls is the way he attempts to use the veil of ignorance to reach the “maximin” principle — his claim that, as Morgan accurately describes, “since you might end up poor, you will rationally only look at the long-term bottom of any system…” isn’t it obviously absurd to only consider your worst possible outcome? If I’m considering whether or not to, say, start a business, is the only factor I pay attention to the worst possible outcome? I don’t think so.

    It’d be better if Rawls just stated that he preferred lots of wealth redistribution, instead of trying to argue that the “veil of ignorance” concept (which works nicely with a utilitarian framework) leads to a principle that it in fact does not.

  13. Gravatar of Morgan Warstler Morgan Warstler
    1. October 2010 at 07:43

    Maxim, HE DIDN’T.

    Thats the whole point of Rawls. He REALLY WAS a tricky fucker sticking it in the eye of liberals.

    He said, “ho, ho, ho… watch me start with Help the Poorest and the Weakest… and end up with Capitalist Democracy Rocks.”

    The libertarian response to him was the one he found most interesting…

    “whoa there Nelly, it’s nice you reach the correct form of Economy, but if we start with your first principals, and things don’t turn out right on the back-end, we’ll have granted far too much.”

    It annoys me to no end, the way that neo-liberals misuse and mis-mention Rawls.

    As a young guy I was fascinated by him, he’s easily my favorite philosopher, because he’s so RADICALLY sure of the positive outcome of markets, that he’s willing to GRANT a first principal, that most libertarians won’t even mutter – that helping the poorest and the weakest is the most important thing.

    I’ve always felt Rawls actually informed The End of History argument (arguments for capitalism masquerading as arguments for democracy), which principally is still true today: we have an evolutionary morality, today is better than yesterday, brought about largely by property-rights economics.

    But we shouldn’t convince ourselves EVER that Rawls would when pushed have favored a anti-capitalist agenda in the name of “democracy.”

    And that’s why Scott’s notion of being a “utilitarian” is suspect… ultimately the greatest “good” is really the greatest “goods” – it isn’t about hedons and dolors, it is about TOILETS.

    Only technology raises up people, only productivity gains create progress, only honest capital formation can be loaned to risk takers (who seek to bring productivity gains).


    Wages must be liquidated because our workers are NOT worth as much in the global market. This is a basic fact. It cannot be out maneuvered.

    We cannot inflate money (past the fall in V) to outflank sticky wages.

    Policy MUST focus on bashing sticky wages in the head… fortunately the only sticky wages we still have are:

    1. public employees.
    2. professors.

    And both are willing to make less money, after all, what else are they going to do?

    Note: it is nice that we can finally see why Scott and DeKrugman are REALLY in the same boat – they are the ones most likely to believe in “sticky wages” – because their profession hasn’t yet experienced how stupid that concept is.

    Very soon, Economics professors will say, “of course there is no sticky wage problem, my salary just got cut by 33%”

  14. Gravatar of scott sumner scott sumner
    1. October 2010 at 17:49

    Peter Whiteford, Thanks for that info. That makes it one of the most poorly labeled graphs I have ever seen.

    Also interesting that in Britain the kids with the lowest livings standards in the entire country are those in households with the median income. That corresponds to a point I have made many times. The people in the various income quintiles are not those whom they are assumed to be. Many poor people are in the middle of the pack, and many middle class are right at the bottom. I’ve been there myself.

    Thanks Maxim,

    Morgan, You said;

    “Very soon, Economics professors will say, “of course there is no sticky wage problem, my salary just got cut by 33%””

    That’s precisely what sticky wages mean–it takes way too long to cut the pay of professors.

  15. Gravatar of Morgan Warstler Morgan Warstler
    3. October 2010 at 20:35

    Scott, THAT’S THE POINT, only professors and public employees still have sticky wages.

    You will soon say, “wtf was I talking about.”

  16. Gravatar of ssumner ssumner
    4. October 2010 at 06:43

    Morgan, Lots of people have sticky wages–look at industries like health care. That’s 16% of GDP, equal to manufacturing.

  17. Gravatar of Morgan Warstler Morgan Warstler
    4. October 2010 at 09:14

    Scott, you keep pointing to UNIONS – they are MEANINGLESS to this discussion.

    They are 5% of private workforce, and are VERY BENDY (see the UAW screwing the newbies).

    Healthcare will see serious wage deflation when we have PRICE EXPOSURE – HSAs for everyone and the ability to find the cheapest provider because you want to keep your cash in your pocket.

    You seem to live in fringes… in NORMAL PRIVATE EMPLOYMENT wage stickiness is a JOKE.

    I can take my top 20 friends, and not one of them has “sticky wages” – we bounce all over the place.

    Stop trying to keep the future from happening. Stop trying to protect the fat from the brutal forces of productivity gains. Stop running up the white flag.

  18. Gravatar of scott sumner scott sumner
    5. October 2010 at 06:20

    Morgan, The data suggests you are wrong–why haven’t wages fallen enough to restore full employment?

  19. Gravatar of Morgan Warstler Morgan Warstler
    5. October 2010 at 08:18

    My god man!

    Because we haven’t tightened the SCREWS!

    The data suggests I’m right…. LOOK AT UAW, when they suddenly own the damn thing, and if they don’t take pay cuts, they go bankrupt, they cuts wages faster than ever. Poof! Wage fall.

    What the hell do you think I’m cheering for? I’m staring at the final delicious act of the play, finally the hero will win, and the villains will lose, and you are trying to SCREW IT UP.

    The important data question is: what have been the circumstances where large changes in wages headed south have occurred?

    Now we simply want to replicate those circumstances. WE WANT CHANGES IN POLICY! Why are you such a surrender monkey?

    If changes never occurred, you’d be right, but since they happen frequently, it means you are wrong, it means you are ACTUALLY fighting changes in policy.

Leave a Reply