… is from page 218 of the 2015 Fourth Edition of Douglas A. Irwin’s important volume, Free Trade Under Fire ; Doug is here discussing Japan’s Ministry of International Trade and Industry  (“MITI “) (footnote deleted; link added):
But the actual evidence of MITI’s contributions to Japan’s [economic] success is weak. The two industries that achieved the most notable success on world markets – automobiles and consumer electronics – did not benefit from extensive government support, unlike some other heavy industries such as chemicals and steel. MITI also had notable failures in promoting its biotechnology and computer industries. In fact, one statistical study of Japanese industrial targeting  found that a disproportionate amount of support went to low-growth sectors and sectors with decreasing returns to scale. This study failed to find evidence that productivity was enhanced as a result of industrial policy measures.
DBx: Industrial policy – that is, governmental attempts to ‘pick’ or promote industrial winners – has a poor historical track record of fueling economic growth and raising living standards compared to the trial-and-error discovery process of competitive, entrepreneurial markets. (Perhaps the best single-volume explanation of why remains my late colleague Don Lavoie’s remarkable 1985 book National Economic Planning: What Is Left? ) Nevertheless, myths of such successes remain alive, in large part because such myths serve the goal of crony capitalists to get tariffs, subsidies, and other special privileges from the state.
A couple of notes on the quotation above from Doug. First (for most non-economist readers): Do not infer from Doug’s quotation that there is something inherently wrong with, or inefficient about, sectors or industries with decreasing returns to scale. There isn’t. Doug’s point is rather that it is a myth to believe that industrial policy is either especially effective at enabling industries or sectors to achieve increasing returns to scale, or is generally directed toward industries and sectors that have as-yet-untapped economies of scale to capture.
Second, in a footnote to the above quotation Doug cites this paper  (published in a 2005 issue of the Journal of International Economics) by Hiroshi Ohashi which, as described by Doug in the footnote, “finds that there were few intra-industry knowledge spillovers in the case of Japan’s steel industry and that export subsidies did not help the industry’s growth.”