Utility: Disparity’s Despair?

I’ve been hearing the complaint that only the top 10% of income earners are benefitting from productivity growth because middle-class personal incomes are stagnating, with the implication living standards are only rising for the rich, but I think (outside of monopolies) for the large majority of goods and services it’s generally not possible for productivity improvements to only benefit the top 10%. Over time, competition ensures prices approach costs, so all productivity benefits eventually flow to the consumer — with the exception, of course, of that small portion of products and services consumed only by the top 10%. But new technologies tend to be aimed at the mass markets — for instance, 20 years ago no one could use Google or Facebook, few had email, and telephone services were on average considerably more expensive — precisely because they have the most consumers and thus the most potential for profit.
I’ve seen this claim justified on the basis that productivity improvements tracked wage growth until about 20 years ago, with the notion being that by improving productivity employers could offer higher wages, but I don’t think that reflects reality. As a greedy self-interested rationally acting factory owner, I don’t want to pay my workers more just because they’re more productive. In fact, of course, I would prefer to pay them minimum wage, or better yet nothing at all — and the only reason I don’t is that it’s hard to get people to work for free. So I’m not sure there was ever a cause and effect relationship there, and if there was it must have been because more productive workers, being more specialized, were harder for employers to find. The reason the correlation to median wages has dropped off in recent years is likely that as the range of worker productivity has expanded the expansion has increasingly happened in the upper quintiles — the highly specialized are getting more specialized, the relatively unspecialized less so.
There’s also a fairly well-known phenomenon in studies of income distribution, which I’ve seen referenced in some papers from Brookings (sorry, couldn’t find the link), that the lower income quintiles have seen living standards improve considerably more than is measured by incomes in real dollars. At least some of this effect is ascribed to the fact CPI ignores supplier-consumer efficiencies such as those created by WalMart, as well as qualitative improvements (MIPS per dollar, cell phone capability, etc).
It’s really a shame we don’t have a standard measure of median utility per capita (as fanciful as that notion is, given that everyone defines their personal utility a bit differently). When you realize that the marginal utility curve is logarithmic for dollars (i.e. the first $1000/yr of good and services keeps you alive so it’s almost infinitely precious, while the 49,0000 – 50,000th dollars mean relatively little, the thousand between 99K and 100K even less, and so forth), then you also realize the utility disparity has been falling even as the income disparities have grown (esp since the mass markets are the most attractive for new utility-improving ideas), making it hard to justify the notion income inequalities are a major societal problem as opposed to an inevitable result of the increasing specialization necessary to continually grow overall consumer utility.