He started by debunking the idea that this isn’t as bad as the 82 recession because the US hasn’t lost as many jobs, first quipping "just wait a couple months" then noting something more important: the 82 recession was caused by the Fed. We knew what caused it and we knew how to end it. It was ugly, but it was something we could fix.
This downturn was caused by factors we still don’t entirely understand, the Fed can’t end it with normal monetary policy because rates can’t go below effective zero and while other policies might work, we don’t know for sure what they are or how to do them.
His fear is that deflation might set in permanently, as people’s expectations of lower prices and wages become set and as a result they keep spending and investing less and less. While we may not know exactly what will work, especially because of the financial component, the best tool in the toolbox to stop deflation remains Keynesian stimulus spending.
The best type of Keynesian stimulus spending is sustained (there must be an expectation that it will continue for some years) and on public goods like infrastructure, education, research and so on. This sort of public spending peaked under Eisenhower.
Since Reagan it’s been at about only 2%, and that has left a huge deficit in public spending, which has shown up in education statistics, collapsing bridges, declining research, lousy broadband and a transportation network which is falling apart. Krugman argued, therefore for multi-year spending on public goods as the primary stimulus method.