Against Kaldor-Hicks, or an ethical weight of zero on distributional concerns is still an ethical weight.

Definition of a Kaldor-Hicks improvement & efficiency 

The notion of a Kaldor-Hicks improvement is a way of evaluating policies. As we shall see, it is a very bad one.

A change is a Kaldor-Hicks improvement if and only if:

A) It is a Pareto improvement (at least one party wins and no parties lose) OR

B) The parties that win receive enough that they COULD completely compensate the losers , if they so choose, though it is not required that they actually do so.

Plus a tweak to deal with the Scitovsky paradox (1).

A situation is Kaldor-Hicks efficient if and only if no change is possible that would be a Kaldor-Hicks improvement.

Generally speaking, a policy will be a Kaldor Hicks improvement if it increases total wealth or income, completely irrespective of distribution. The concept of Kaldor-Hicks improvements is often incorporated into government decision making.

I aim to convince you that the notion of a Kaldor-Hicks improvement is is a bad, even repugnant way of evaluating policy. The arguments I canvass here have all been considered before elsewhere in the literature. If there is something original about this piece’s approach it is in the emotions I want you to feel. It is my aim to argue that the Kaldor-Hicks criterion is not simply ‘incomplete’ or even just ‘flawed’, it is morally disgraceful. You should be angry that economists and other public policy theorists use it.

The moral repugnancy of Kaldor-Hicks improvements as an ethical criteria

Two travelers pass each other in the desert. Traveler A steals many of travelers B’s things. Later traveler A sells traveler B’s possessions, achieving an extortionate price for them through his superior bargaining skills. Meanwhile traveler B hauls herself out of the desert, barely alive. Suppose that traveler A acquired enough resources from so oppressing traveler A that he could, if he wanted, compensate her. Of course, he won’t. This banditry is a Kaldor-Hicks improvement over the alternative of traveler A leaving traveler B alone.

Does an ethical criterion that recommends this sort of behaviour as an improvement sound like a suitable guide to policy to you? Does it sound just? Good? Right?  Of course not. It sounds insane. Whether you’re an egalitarian, a utilitarian or a Nozickean of some sort, what I’ve described will offend your conscience. Nonetheless, the Kaldor-Hicks criterion says go for it.

In this topsy-turvy moral universe any kind of wanton greed is justified, so long as the acquirer gains more than the loser loses. I would say this is not merely wrong, it is appalling.

The problem stated

You might be wondering why anyone would endorse Kaldor-Hicks efficiency in light of this.

The great allure of the Kaldor-Hicks criterion comes from the 2nd fundamental theorem of welfare economics according to which, under some strong assumptions about markets, all possible Pareto optimal outcomes can be achieved as competitive equilibrium if we allow lump sum costless transfers. Hence the argument goes, the Kaldor-Hicks criterion tells us how to grow the pie, and if we have distributive concerns we can simply change the cut and achieve a more desirable competitive equilibrium through a lump sum transfer. Leave the cutting of the pie to ethicists, or politicians, or the democratic public- anyone but economists.

The picture that emerges is a neat view (entirely too neat) in which economists can deal with “efficiency” whereas philosophers, politicians etc. deal with questions of distributive fairness. Thus the economist is freed from a specifically political responsibility. Think of it as a kind of alchemical process. The economist hopes to be able to create the gold of normative advice (policy prescriptions) from the lead of purely positive assumptions. Indeed the approach arose during a time when economists who wanted to say anything about social welfare were very much on the defensive, and it is the bitter fruit of a generation’s counter-productive efforts to expunge ethics from an applied science ( I often wonder what Kaldor-Hicks would look like in medicine).

The problem with relying on the second fundamental theorem of welfare economics to separate distribution from efficiency is that the theorem:

1. Requires a kind of absurdly perfect market which never and nowhere has existed. Further, the theory of the second best shows that it’s not good enough to say it is ‘approximated’.

Further, if the theorem is taken as demonstrating that distributive justice and efficiency can be separated, this is wrong because:

2. This line of reasoning ignores the role of power in society. Winners have more money with which to intervene to defend their new holdings against redistribution.If the winners are powerful enough to push for the Kaldor-Hicks “improvement”, they are quite likely also powerful enough to resist the compensatory redistribution.

3. Worst of all, the second theorem of welfare economics ignores the reality that costless transfers between people, especially if facilitated by taxation, don’t exist. At an absolute minimum there is the cost of tax-collection, but more worryingly there is the deadweight loss of taxation.

In other words, we can’t just make the pie as big as possible then cut it how we like. The size of the pie, who yields the knife, and how the pie is cut are all interrelated in subtle ways.

And there a host of lesser, but still important objections. For example, relative income effects mean that on certain parameter specifications, the wealthy getting wealthier might do more harm to the poor than it does good to the rich.

To conclude our criticism,  KH as a decision criteria is simply in practice a preference for maximising total wealth, irrespective of distribution. This not more ethically neutral or uncontroversial than its alternatives, rather it represents an extremely demanding philosophy of distributive justice that almost no one holds. The left will object to this philosophy on egalitarian grounds. The right will object because no premium is placed on deserts going to the productive. If you remember nothing else, make it this: A weight of zero on distributive concerns is still a weight.

On alternatives

At this point, the defence’s best line is probably like:

“Well, what’s your alternative then? Some kind of policy framework is necessary, but all existing options (utilitarianism, prioritarianism, the max-min approach, historical theories of justice) depend on very meaty normative assumptions. Economists aren’t paid to be moralists. The idea of doing efficiency recommendations through economics and distribution recommendations through economics may be a fiction, since in practice all recommendations have implications for both, but there is no alternative.”

My first response is to reiterate that continued use of the Kaldor-Hicks criteria is no less normative than any alternatives, and the notion of KH’s ethical neutrality is not sustainable even as a simplification or approximate truth. In trying to avoid answering difficult ethical questions, Kaldor & Hicks just ended up answering them in a way everyone hates, as opposed to traditional moral philosophies, which only some people hate. The idea of policy recommendations without normativity was folly to begin with.

There are many reasonable alternatives. I propose the closest thing to a fair alternative is utilitarianism. We are not all utilitarians- I myself am not a utilitarian- but utilitarianism is at least partially persuasive to most people, most of the time. Even if you reject utilitarianism, the fact that a policy is recommended by utilitarianism is usually a reason in favour of it, even if it is not decisive.

Moreover, utilitarianism can be seen as occupying a sort of natural middle ground between egalitarianism and priortarianism on one hand, and the historical theories of justice on the other. It denies the left by not including any premium on what they see as equality, it denies the right by not including any premium on what they see as just deserts.

The combination of these two points points to an argument in favour of using utilitarianism that mirrors the logic of utilitarianism itself. Using utilitarianism as our decision rule maximises aggregate ethical persuasiveness across all agents: the greatest academic, political and ethical interest for the greatest number.

Though there are theoretical concerns, the practical feasibility of utilitarian welfare analysis is, by this point, well attested, whether the focus is on subjective well-being, positive affective states or preference satisfaction .

Perhaps you don’t buy my pitch for utilitarianism, fair enough, but I want to be clear that even if utilitarianism isn’t the best solution, it is certainly a better solution than a Kaldor Hicks approach. So is Little’s criteria. So is prioritarianism. Heck, I’m not a Nozickean, but the logic is certainly more compelling than Kaldor-Hicks.

FOOTNOTES

(1) After the Scitovsky paradox showed that under some situations this can lead to an endless cycle because both a change and its reversal could count as an improvement under certain situations, the criterion was altered to exclude as Kaldor-Hicks improvements changes whose reversal would also be Kaldor-Hicks improvements. The details of this need not concern us here.