Reading Brad DeLong's latest word on the subject, I find that we seem to agree on the economics, but perhaps not on the history. We both like Roosevelt's policies of going off the gold standard and providing deposit insurance. Where I complain that monetary and fiscal policy were not expansionary enough, Brad says yes, but Roosevelt removed the contractionary bias of the Hoover Administration. Point taken.
But, to borrow a locution from Daniel Gross, "One would be very hard-pressed to find a serious professional historian--I mean a serious historian, not a think-tank wanker, not an economist" who would argue that the New Deal consisted primarily of deposit insurance, going off the gold standard, and taking the foot off the monetary and fiscal brakes. I think that the serious historian would feature the NRA, the SEC, the TVA, Glass-Steagall, the Agricultural Adjustment Act, the Wagner Act, Social Security, ...
Incidentally, Gross promises that his forthcoming book
contains a chapter that argues that the financial infrastructure laid down by the government in the New Deal helped pave the way for the immense growth seen in America's capital markets, banking system, credit industry, and housing markets over the past 70 years.
Robert Higgs has a chapter in his book, Depression, War, and Cold War that takes Bryan Caplan's view, that Roosevelt scared the daylights out of private investors. While 'PGL' jeers, "Might we remind Bryan that investment demand (in 2000$) rose from $11.5 billion in 1932 to $91.1 billion in 1937?", Higgs writes,
both real GDP and real GPI [gross private investment] plunged from 1929 to a trough in either 1932 or 1933, the former by 29 percent and the latter by 84 percent. Both variables recovered rapidly after 1933: by 1937, real national product has regained 96 percent of its loss...and investment had recouped 64 percent of its loss. The "Roosevelt recession" of 1937-38 cut short the recovery...gross investment fell by 34 percent...even in 1941, when stimulus from the defense mobilization had become substantial, real GPI had not quite regained its 1929 level.
So the interpretation of investment data is subject to choice-of-endpoint decisions. Yes, it was higher in 1937 than in 1932, but it was lower in 1939 than in 1930 (at least in the chart in the Higgs chapter).
I have yet to read an economic history of the 1930's that I trust as definitive. I wish I knew what really happened.
UPDATE: Alex Tabarrok weighed in on the anti-New-Deal side of the issue.