I basically concur with Tyler Cowen's take on Greg Mankiw's recent column--you can certainly disagree with the analytical choices he made in arriving at the government's cut of his children's inheritance, but the basic point is sound: we tax income many times as it is earned, which is especially hard on capital formation. The means testing of various benefits, from Medicaid all the way to 401(k) contributions, makes this effect even more pronounced.
I don't know how strong the bequest motive is in the average affluent person, but from talking to financial planners, I know that in some cases, it is very pronounced indeed, and influences decisions not just about estate planning, but about retirement and business expansion. The case of parents who spend their whole lifetime building up a small fortune to make sure that their (only) developmentally disabled child will be well taken care of, rather than retiring, is not necessarily typical, but it certainly occurs. Or else financial planners just make these stories up to make me feel sorry for their rich clients, which seems possible, but not all that likely.
How we should think about the effect of taxation on deferred consumption seems to me to be an interesting philisophical debate, as well as a thorny empirical question. But I didn't find much enlightenment on those points from those who disagreed with Mankiw; what I found was a bunch of people who seemed to be angry that rich people are allowed to voice opinions on the optimal level of taxation for people who make a lot of money. Interestingly, no one thought it was odd that they should have opinions on Greg Mankiw's taxes; apparently, it is only opinions on your own taxes that are in bad taste.
There was also, especially from the professors, a lot of weird argumentation that Mankiw essentially must not have any marginal value for money if he writes for the New York Times, because writing for the New York Times does not pay much, relative to Mankiw's existing income. Now, I've known a lot of rich skinflints who would walk a mile to save a quarter, so I am simply inherently skeptical of attempts to derive peoples' time value of money from their income. But more to the point, this seems to assume that there is a single activity--"writing"--for which each individual has a single supply curve.
But this is not even close to being true. I get paid for speaking sometimes. You cannot assume that because I speak to my alumni groups for free, I must therefore be willing to speak whether or not there is money involved. Other groups have to pay me, because I don't have so much free time, and I have no sentimental attachment to the Greater Omaha Chamber of Commerce; if they want to hear me speak, they're going to need to pay me well enough for me to give up some fairly scarce leisure time. Likewise, even if Mankiw is willing to write for the Times, there might be many other outlets that he would write for if the pay were greater. If it were not for the wedge introduced between their costs, and his income, by taxation, we might have more Mankiw pieces.
I think that would be a good thing; I sense that many of the most vehement critics disagree. But it's also worth extrapolating this case study to folks who aren't Greg Mankiw, but who face similar choices. Professors, and to a lesser extent journalists, compete for non-monetary status much more than most jobs. The guy who owns six McDonalds, and doesn't want to open a seventh because that's just one more place he's got to be willing to go in and scrub the toilets when his assistant manager calls in sick . . . those people tend to be more sensitive to tax incentives than Greg Mankiw or Brad DeLong, whose friends will snicker if they buy a Porsche and a McMansion.