In my post about deficits paying for themselves, I decided to make the point generally, i.e. in a normal economy, and therefore discounted the notion that deficits are actually self-financing. But maybe I should have taken the time to discuss how the result might change in a severe recession and at the lower bound for interest rates. As Paul Krugman points out, in this case it's much easier to argue that such spending will, in fact, improve the long-run budget picture:
But to some degree the self-financing question steers us away from the main point. A more general interpretation of Krugman's point is that in an economy like we have today, the cost of deficit spending is far lower than during more normal times. That is, even if the spending is not self-financing, there are still substantial offsets that make it much cheaper than it would be closer to full employment.