Almost all economists ridicule the notion that - at its current margin - the U.S. government can raise revenue by cutting taxes. Well, let me put forward a promising new candidate for a free lunch tax cut: Much lower taxes for families who have additional children.
According to Arthur Brooks (of Gross National Happiness fame), Ronald Lee and Tim Miller estimate that from the point of view of the IRS, the net present discounted value of the average American child is $100,000. If $99,999 in tax incentives produce one more child, it's a good deal. In fact, it's a no-brainer.
What fraction of couples of child-bearing age would accept such an offer? It's got to be at least 25%, doesn't it? Of course, you'll dilute the incentive if e.g. pre-existing large families headed by parents in their 40s qualify, too. But even if natalist incentives were quite bluntly targeted, they could easily be the kind of free lunch that Laffer dreamed of.
Admittedly, unlike the usual Laffer curve story, you have to wait decades for natalist tax cuts to pay off. But since Lee and Miller's estimate is a present value, that's already accounted for. And as long as bond markets work reasonably well, the time pattern of payments makes little difference.
Of course, if you're really looking for a free lunch, you'd give extra-big natalist tax incentives to the rich, because their kids are especially likely to pay a lot more taxes than they consume in services. But hey, we should probably stick to one outrageous proposal at a time.