A recent Sowell column characteristically argues for extending the Bush tax cuts, claiming that lower tax rates for the richest increase tax revenues and to think otherwise is 'sheer hogwash.' He cites data (probably from some Heritage Foundation report) from the 1920s.
Some things he doesn't tell us, however, tell us a lot. For instance, after Reagan's income tax cuts took effect in 1982, here's what happened over the next six years (in $millions):
Income Taxes/ Cut
Year Taxes GDP GDP Loss
1982 346,951 3,253,200 0.107
1983 325,960 3,534,600 0.092 51,002
1984 355,308 3,930,900 0.090 63,919
1985 395,862 4,217,500 0.094 53,931
1986 412,102 4,460,100 0.092 63,564
1987 476,483 4,736,400 0.101 28,650
1988 495,689 5,100,400 0.097 48,264
Taxes in the absolute rose for the simple reason the economy (GDP) grew. In other words, removing the effect of the rising economy, taxes would have fallen from the 1982 base by at least one whole percentage point in five of the six years (that's over $300 billion over the six years).
Maybe you could try and argue that the tax cuts for the rich were a cause of the GDP growth, but you need more evidence than Sowell's claim that if you press into the wealthiest's income with too high tax rates (he, of course, never specifies the tax rate range over which this works--how about zero?), they will simply shift some portion of their their income into some unspecified non-taxable shelters (presumably, then, entailing some loss of productive resources to society).
There are other periods, likewise, that don't support the stuff-the-fed-coffers-by-lowering-tax-rates-on-the-rich scheme. Like the Clinton 1993 tax increases with resulting increased revenues. There are good reasons Sowell's claim is counter-intuitive, if not 'sheer hogwash.'