Ronald Dworkin, answered from the texts and cited to the page.
Equality of resources is the right metric of distributive justice, not equality of welfare.12 The argument runs through several stages. First, equality of welfare cannot handle expensive tastes: if someone cultivates a taste for claret and plovers' eggs while another is content with beer and beans, equality of welfare requires that the claret-drinker receive more resources to reach the same welfare level, which is the wrong result.34
Second, the apparatus of an idealized desert-island auction shows what equality of resources comes to: each person bids for the bundle of resources he prefers, and the distribution is egalitarian when it passes the envy test — no one would prefer to trade his bundle for anyone else's.5 Third, the hypothetical insurance market extends the apparatus to handle differences in personal endowments (talent, skill, handicap): we ask what insurance against bad endowment luck a person of average risk-aversion would have bought from behind a thin veil of ignorance, and we redistribute to approximate that insurance.
Fourth, the brute luck and option luck distinction does the moral work: bad luck for which a person bears no responsibility (genetic handicap, the parents one is born to) is brute luck, and society owes compensation through the hypothetical insurance scheme; bad luck a person assumed by his own choice (gambling losses, risky career choices) is option luck, and society does not owe compensation.
G. A. Cohen's expensive-taste objection — that the resources/welfare distinction does not track the responsibility line because some tastes (cultivated identifications, religious commitments) are not chosen — names a real difficulty, but the response in Sovereign Virtue and the Burley reply (2007) is that responsibility runs through endorsement rather than choice: a person is responsible for the tastes she now endorses, even if she did not choose them, and the resources framework remains correct.
He therefore needs some argument why relative success is nevertheless the right metric of distributive justice; why justice demands transfers in the name of welfare equality when those who gain think they are already overall better off than those who lose. His difficulty underscores the danger in his initial assumption that he can defend equality welfare without confronting the question of what welfare is.Replies to Critics in Burley, p. 10
Indeed, I argued that equality of resources is to be preferred to other theories of distributive justice because it better fits a more attractive account of what well-being really is. In particular I emphasized, in my original account of the crucial phenomenon of insurance, that people buy insurance not to improve their resources but to improve their welfare.Replies to Critics in Burley, p. 13
The information left to an independent political level under equality of welfare is therefore brought to the initial level of individual choice under equality of resources. The elements of luck in the auction just described are in fact pieces of information of a crucial sort; information that is acquired and used in that process of choice.Sovereign Virtue - The Theory and Practice of Equality, p. 70
Equality of resources, however, offers no similar reason for correcting for the contingencies that determine how expensive or frustrating someone's preferences turn out to be.'^ Under equality of welfare, people are meant to decide what sorts of lives they want independently of information relevant to determining how much their choices will reduce or enhance the ability of others to have what they want.Sovereign Virtue - The Theory and Practice of Equality, p. 69
Nor is the choice of sets of bundles arbitrary. Many people will be able to imagine a different set of bundles meeting the no-envy test that might have been established, but the actual set of bundles has the merit that each person played, through his purchases against an initially equal stock of counters, an equal role in determining the set of bundles actually chosen.Sovereign Virtue - The Theory and Practice of Equality, pp. 68–69
Equality of resources is one of two major contemporary egalitarian metrics, with G. A. Cohen's access to advantage / midfare and Richard Arneson's equal opportunity for welfare as the principal alternatives within luck egalitarianism, and Amartya Sen's capability approach as the alternative metric from outside the luck-egalitarian framework. The Burley volume (Chs. 1-5) contains the most sustained recent engagement: Cohen's 'Expensive Taste Rides Again,' Christofidis on talent and slavery and envy, Van Parijs on undominated diversity, Otsuka on liberty and abstraction, Arneson on the foundations of liberal equality. Elizabeth Anderson's 1999 Ethics article 'What Is the Point of Equality?' argues that luck egalitarianism gets its motivation wrong (pity and resentment rather than the relational equality of free citizens); the response in Justice for Hedgehogs Ch. 16 treats Anderson's challenge as a misreading of what the luck-egalitarian project is doing. The family disagreement among luck egalitarians (Dworkin, Cohen, Arneson, Otsuka, Van Parijs, Casal, Williams) is one of the most productive ongoing debates in contemporary political philosophy.