Defenders of inequality are doing a desperately poor job.
And Greg Mankiw's claim that the rich deserve their fortune barely counts as an effort. Even if we concede - heroically - that high salaries are due to high marginal productivity, questions remain such as: what about Rawls' claim that talents are arbitrary from a moral point of view? Do people really deserve to have been born at a time when their marginal productivity is high: if Wayne Rooney had been born 80 years ago he'd have earned £20 per week, not £300,000? And even if people do deserve credit for cultivating their own ability, they hardly deserve credit for the fact that others' inability to do so renders their ability scarce. Problems such as these mean that the more intelligent rightists, such as Hayek or Nozick, have acknowledged that the link between merit and income is weak.
So, is it possible to defend inequality better?
The issue here is not one of free markets: it's perfectly possible for a free marketeer to worry about inequality precisely because so much of it arises from market imperfections or state interventions such as bank subsidies, QE, crony capitalism or copyright laws.
Subject to that caveat, here's how I would try to defend inequality:
1. Bosses and bankers have epic levels of narcissism and high pay is the least bad way of gratifying this. If they were paid less, they would seek other ways of satisfying their self-regard such as pursuing non-pecuniary status symbols such as corporate jets or by plundering the firms' assets. There's a reason why the word "efficiency" appears in the phrase "efficiency wage models". It's because it might be efficient to bribe workers not to exploit their power.
Income inequality was low in the 1970s, but industrial relations then were poisonous, productivity low and management poor. There might be a link here.
2. Some policies to curb high salaries would merely shift incomes between the rich. Limiting footballers' pay, for example, would make John Terry worse off but Roman Abramovich better off; ticket prices would stay high because demand outstrips supply. Much the same is true of CEOs were paid less; shareholders would benefit, not customers or workers.
3. One of the best ways to increase equality would be to create countervailing power to the rich - either by strengthening trades unions or by encouraging worker control. But there is no demand for such policies.
4. Even if we accept Wilkinson and Pickett's evidence that inequality has several social costs, it doesn't automatically follow that there's a case for reducing the incomes of the top 1%. They measure inequality by the ratio of the top 20% to bottom 20% of incomes. And it's plausible that it is middling inequalities that hurt us: it's our neighbour's new car or our colleague's pay rise that upsets us from day to day, not footballers' million-pound salaries.
5. The question of whether inequality is bad for growth (pdf) or a cause of economic crises (pdf) is a tricky (pdf) one; much depends upon what causes inequality, and the form that redistribution takes. One big fact, however, warns us not to expect a growth bonanza from ergalitarian policies. It's that long-term economic growth doesn't vary much across countries or time. This suggests that variations in inequality might not have big effects upon growth.
FWIW, I'm not sure about these arguments; I've listed them in rough order of what I think is their plausibility. But I suspect they are better than wibble about desert.