This figure, which is from a recently revised and submitted paper of mine, shows how the multiplier varies across the income distribution. It shows that equivalently sized tax changes for lower income groups have larger macroeconomic impacts on employment growth and that the effects of tax changes diminish as they go to higher income groups. Finally, the effects for the highest income groups are quite close to zero and are substantially smaller (in absolute value) than the impacts of tax changes for lower income groups. This graph, which shows the effects of a large one percent of GDP increase in taxes, suggests that progressive tax cuts targeted to lower income groups tend to be much more effective at stimulating job creation at a business cycle frequency than tax cuts for high income taxpayers. Put another way, if we actually want to give tax cuts to job creators, we shouldn’t provide tax relief for the highest income groups. High income tax cuts are substantially less effective at stimulating growth.
Finally, note that these are the estimated effects from modestly sized tax changes, so the magnitudes of these effects for very large tax changes could be different.