Schumpeter, Joseph A. "John Maynard Keynes: 1883-1946." American Economic Review. Vol. 36, No. 4 (September, 1946), pp. 495-518.
Explanation is not far to seek. The higher ranges of mathematical economics are in the nature of what is in all fields referred to as “pure science.” Results have little bearing – as yet, in any case – upon practical questions. And questions of policy all but monopolized Keynes’s brilliant abilities. He was much too cultivated and much too intelligent to despise logical niceties. To some extent he enjoyed them; to a still greater extent he bore with them; but beyond a boundary which it did not take him long to reach, he lost patience with them. L’art pour l’art was no part of his scientific creed. Whereverelse he may have been progressive, he was not a progressivein analytic method. We shall see that this also holds in other respects that are unconnected with the use of higher mathematics. If the purpose seemed to justify it, he had no objection to using arguments that were as crude as those of Sir Thomas Mun.
Historia de la Macroeconomía [pdf]