I occasionally listen to EconTalk, hosted by Russ Roberts. Recently, I listened to an episode where Russ Roberts was even more off-the-mark than usual. Sometimes EconTalk is like a broken pencil... pointless. Russ can go on for an hour not getting close to understanding Keynesian economics, in general, and Paul Krugman, in particular. I want to discuss something quite general about Keynesianism, but I wanted to cite someone who makes the mistaken point that Keynesians are basically just advocates of big government, which is wrong. I hear it a lot and it annoys me each time I do. Here's a quote from Russ Roberts, which was quoted by Matt Yglesias at ThinkProgress.org.
The evidence for the Keynesian worldview is very mixed. Most economists come down in favor or against it because of their prior ideological beliefs. .
I have a huge problem with his statement because he is quite clearly misrepresenting (and maybe misunderstanding) what Keynesian economics is all about. Keynesian economics is not about a social model of the way society should be (like having a big government), but is an economic model about how the government should respond to macroeconomic problems. A lot of the time, Keynesians are very pleased to advocate the use of monetary rather than fiscal policy to stabilize the macroeconomy because they believe (correctly, in my view) that cutting interest rates can induce additional investment than would otherwise be forthcoming and increasing interest rates can reduce the amount of investment in the economy. However, when the macroeconomy is seriously depressed and interest rates are as low as they can go (the zero lower bound, zlb), Keynesians are usually at the forefront advocating aggressive fiscal policy to stimulate the economy.
Mark Thoma of the University of Oregon sums it up in the Fiscal Times this way:
Myth 3: Keynesians are advocates of big government: This is probably the biggest and most common confusion about Keynesian economics. Keynesian stabilization policy involves increasing government spending or lowering taxes to stimulate the economy in recessions, and then reversing those policies when the economy improves. Thus, under Keynesian policy the change in spending or taxes is temporary, e.g. spending goes up in recessions and then goes back down after the recovery, and the average size of government does not change over time. If, however, politicians decide to increase taxes rather than cut spending after the economy recovers then the average size of government will increase. If politicians do the reverse, use tax cuts to stimulate the economy and then pay for it by cutting spending when things improve, the average size of government will fall. But when the changes are truly temporary as Keynesian policy requires, the average size of government does not change at all.
The fact is that Keynesian policies do not even need to increase the size of the government in order to work. In fact, Milton Friedman actually supported the idea that fiscal policy could be used to stabilize the economy in his book Economics and Liberty, and he suggested that a way to support the economy could be to use tax cuts instead of spending increases.
After saying that, I really should add that the most recent iterations of Keynesianism (New Keynesianism, or NeoWicksellianism) have taken the position that temporary spending increases or tax cuts should have less impact than permanent spending or tax changes. So permanently increasing the size of the government could be advocated, although if the history of the U.S. at the end of World War II is any evidence, it probably would not be necessary once full employment is restored. At the end of World War II there was a lot of concern that if the government reduced spending, the U.S. economy would fall back into depression, but it didn't happen. So I do not think there is any reason to believe that Keynesian policy would need to be permanent and I have yet to encounter any explicit endorsement of such a permanently increased size of government in the current U.S. depression.
 One really good example of this is this interview by Russ Roberts of Don Boudreaux, where they prattle on endlessly about how they fail to understand the most basic things about Keynesianism and especially the arguments related to the costs of public borrowing in a liquidity trap. Their argument gradually degrades into grunting and ceaseless repetition of their assumptions as further proof of their assumptions.
 I referred to this in my previous post.
 On a side note, government expenditures as a percent of GDP are projected to increase, but that has to do with the demographics of Social Security and projected increases in the cost of Medicare (which is now on a much more modest trajectory than in the past).