Well, I have to say that pointing to that one infamous quote - which I have seen numerous times now - and saying - "See how wrong that is!" is 20-20 hindsight, cherry-picking, and pretty shallow. If Samuelson was generally faulty, there should be many quotes to pick apart. I hear crickets.
Since Roberts is going after Krugman, he ought to take a look at his predictive abilities, rather than dragging up a 70-year idea enunciated by someone else. That is changing the subject.
Re: demobilization, I just did a quick google search to make sure what I thought I know wasn't off base.
I came up with these.
At any rate, it's awfully hard to put a robust economy story together with the real GDP data in a way that makes sense.
1941-01-01 1371.5 17.1%
1942-01-01 1623.5 18.4%
1943-01-01 1887.9 16.3%
1944-01-01 2040.2 8.1%
1945-01-01 2016.6 -1.2%
1946-01-01 1798.2 -10.8%
1947-01-01 1784.8 -0.7%
1948-01-01 1864.8 4.5%
1949-01-01 1854.2 -0.6%
1950-01-01 2016.5 8.8%
Negative growth 4 years out of five, starting in '45.
I'd love to see data on the 7.5 million new jobs.
I checked FRED again, and returned.
Also from FRED, PAYEMS series. (All Employees: Total nonfarm)
Jan 1, 1946 39839
July 1, 1947 43742
OK 3.9 million. This would be more impressive, though, if the previous 18 months hadn't seen the loss of about 2 million jobs.
July 1, 1944 41904.
The job additions from Jan, 1946 finally totaled about 7.5 million in Jan, 1951 - way on the far side of the 1948-9 recession.
Roberts started with "the release of 10 million people into the labor market with demobilization," which became 7.5 million NEW jobs over 18 months, which actual data showed to be 3.9 million over 18 month, and 7 million over 5 years. And they weren't all new jobs. Two million jobs had been lost in the previous year and a half. Roberts acknowledges the effects of the GI bill and women leaving the work force with little more than a nod. Those niggling little details don't fit the Libertarian narrative particularly well.
With a few seconds on the Google, I found this:
Unemployment almost disappeared, as most men were drafted and sent off to war. The government reclassified 55% of their jobs, allowing women and blacks to fill them. First, single women were actively recruited to the workforce. In 1943, with virtually all the single women employed, married women were allowed to work.
Isn't that something. Seems there was a bit of a labor shortage. They even allowed women and blacks to work -- freaquing amazing! But further, as the war ended, unemployment went up
From the same source we get these population figures (Civilian noninstitutional population)
So, from 1945 to 1946, the civilian population increased by 8.98 million. Of these, 3.66 million entered the work force, while 5.3 million did not. From '46 to '47, civilian population increased by another 2.9 million, the workforce by 2.65 million, and the non-workforce by 300 thousand.
OK - the Truman administration did a pretty reasonable job of redeploying GI's into the work force. Credit it where it's due. But this data in no way refutes the original Krugman statement Roberts was ridiculing: "Pay no attention to those who invoke the confidence fairy, claiming that tough action on the budget will reassure businesses and consumers, leading them to spend more. It doesn’t work that way, a fact confirmed by many studies of the historical record."
In fact, the post WW II "experiment" is essentially irrelevant to Krugman's point. It is, as I said above, Roberts changing the subject. And if this fact is not immediately obvious, I suggest you ponder it at your leisure.
Well, I learned a few things about the post WW II period during this exercise, and that's pretty valuable on its own. I didn't learn anything new abut Libertarians, though. With them, it's the same old same old, all the time.