Co-blogger Scott Sumner has written an excellent post this morning pointing out that the Keynesian model per se is not a big government model. He is right, for the reasons he gives.
When the late Ben Rogge came to give a talk at the University of Winnipeg in the winter of 1969, I asked him if there were any non-big-government Keynesians. He named one: David McCord Wright. Isn't it striking that he named one rather than saying, "Oh, yes, there are a lot of them."
Then, when I first learned the Keynesian model later that year in my one undergraduate course at the University of Winnipeg, I came to the same conclusion as Scott: If the government raises taxes during booms and lowers them during busts, or decreases government spending during booms and increases government spending during busts, there is no built-in growth of government from following the policy implications of the Keynesian model. Nevertheless, even then I noticed that the Keynesians who wrote my textbook--Paul Samuelson and Canadian economist Anthony Scott--seemed to be big-government people. When I started looking around, I noticed that most Keynesians were big-government people.
Why was that? I think it was because they were big-government people anyway and they could use the Keynesian model, quite consistently with no sleight-of-hand, to advocate tax increases during booms and government spending increases during busts. I hasten to add that I'm not accusing them of any kind of dishonesty. They honestly believed in big government and, with that belief, it came naturally to them to advocate as they did.
If you could convince me that it [the Keynesian model] worked in a technical sense, I'd immediately favor tax cuts in recessions and tax increases in booms.
My bottom line would be different from Scott's and it's informed by something Alan Reynolds told me in the mid-1970s:
If you could convince me that it worked in a technical sense, I'd immediately favor tax cuts in recessions and cuts in government spending in booms.
Alan pointed out that that would lead to a small-government bias in Keynesian policy.