While I haven't found anything that can directly answer my question on whether the incidence of taxation at the very top of the pyramid makes a difference on growth rates, it has led me to reading some excellent papers on taxes.
A couple of these are available without journal access. The main one I want to discuss is the OECD paper "Taxes and Economic Growth" which can be downloaded from here.
The first thing I want to mention is that this paper does claim that highly progressive tax brackets are associated with lower rates of entrepreneurship. However, a quick look at Wikipedia (sufficiently accurate for my purposes here) showed that many states have their top tax bracket kick in well before the very highest income brackets are reached (though I could not easily find data on income by decile for many countries, this is representative of my time commitment rather than lack of data I'm sure). This is consistent with my earlier hypothesis, that entrepreneurship is concentrated in the top fifth of the distribution but the top 1% (or some subset thereof) is likely acting differently. Of course, this does weigh against that hypothesis but I don't have data specifically analyzing changing rates at the top 1% so it also does not falsify it. Perhaps interestingly enough, consulting Saez for the income among the top earners in the US does reveal we are one of the few with an additional bracket (at $372,951) not far off from the $382,600 Saez identifies as the top 1% (these are different years, but I'm lazy and I doubt there has been much change).
There are a few other takeaways from this paper. First of all, there are strong labor promoting effects (with some costs) for high personal exemptions. I can personally speak to this, I had worked in both Canada and the U.S. out of my undergrad in fairly low wage jobs and I really noticed the difference in the basic personal amount (or standard deduction in the U.S.). Here in the U.S. we could gain substantial efficiencies from raising this amount, as well as addressing equity concerns. It also claims that after tax benefits are far more efficient than using the tax code for this purpose, consistent with other things I've read.
The paper also discusses some other important to consider impacts of income taxes such as reductions in human capital investment since higher income is taxed at a higher rate. Again though, I think there is a possibility that this result has more to do with taxing the top decile than it does with taxes on the top 1% or so.
The second thing is how much more efficient taxing property is. The biggest thing here is how negative housing incentives are to growth, it distorts investment and spending in a large number of ways and is very costly to government. Definitely the single biggest problem problem in most tax schemes. Very well known but it's nice to have the links spelled out in a format where it can be compared to other taxes.
The second observation here is how efficient estate taxes are, especially if applied at the individual rather than the estate level. This is also well known so I won't linger.
The third issue related to property taxes is how well capital gains taxes hold up to other means of taxing business holdings. While there are disadvantages to taxing capital gains relative to other options it looks pretty good. Ultimately, capital gains are more of a property tax than they are a disincentive to saving.
The paper also points out that corporate taxes are very bad for growth, though considerations involving foreign investment (which the paper mostly considers attracting FDI, for the US though I'm more worried about taxes leaking out, our market is large enough that I'm more concerned about lost income than I am about the need for FDI) and the possibility of people trying to conceal personal income as business income also come into play here. Still, a simple, low corporate tax rate is worth it for the efficiency gains as long as they are not radically different from other taxes.
Also, consumption taxes are good especially when they are achieving other social ends, such as taxes on pollution and drugs.
My ending thoughts on this? They haven't changed much, though this paper has made me think a bit more about the negatives of income taxation as well as the potential negatives of too big of a difference between income and corporate taxes. My ideal tax reform remains eliminating the vast majority of deductions and exemptions, lowering rates overall especially the corporate tax, raising property taxes particularly on housing and capital gains, adding a VAT, adding in more excise taxes particularly on additional drugs such as cannabis as well as on pollution and perhaps soft drinks, instituting a much larger personal exemption, and imposing a higher tax on the top 1%, ideally a smooth upward progressive tax rate (these guys have graphing calculators and can hire accountants so I'm not too worried about the math involved for this particular segment). Entrepreneurship concerns* are increasingly leading me to think that the progressivity of taxes after the personal exemption can be reduced until the very peak but I wouldn't eliminate it entirely.
*Asset levels are the single largest factor determining new business start ups. Historical cases lead me to think that new, disruptive (creative destruction) businesses tend to be concentrated in those just below the very top of the income pyramid meaning these individuals should be a focus of particular concern. However, helping those lower down the scale build up assets is also critical to smaller scale entrepreneurship. Benefits are probably a bigger deal here. Providing health care and a greater level of income supports for failed entrepreneurs will help here as well as getting rid of a lot of means testing requirements to help those on the lower rungs who want to take a chance on something like starting a restaurant to do this without also risking their health and benefits as they save up for the initial investment.