When union leaders signed on with liberals in supporting the National Labor Relations Act in 1935, they were well aware of the risk they were taking by giving up traditional organizing tactics in exchange for promises of government protection through the NLRB and the courts. They took that risk in part because they were having little or no success except for a few business sectors in which employers could not afford to bring in replacement workers for one or more of several reasons, including high skills levels (e.g., printing), geographic isolation (e.g., coal mining), and time-sensitivity (e.g., railroads) (Kimeldorf 2013). Furthermore, the usefulness of the original act for union organizers was not automatic. Its value depended on the protection and possible extension of the several specific statutory guarantees that were included in it. But the Taft-Hartley Act and many later decisions by both Congress and the NLRB narrowed or withdrew those guarantees in what turned into a class struggle at the legislative and regulatory levels.
Due to the strong corporate support for the Republican challenger, Roosevelt was outspent $8.8 million to $5.1 million, and most major newspapers endorsed his opponent. But he won 62.5% of the two-party vote, documenting once again that those with the biggest war chest don't always win and that the influence of the mass media can be overstated. The ultraconservatives' appeal to traditional values, their claims that the constitution was being shredded, and their insistence that the New Deal was socialism in liberal clothing (sound familiar?) fell on deaf ears. Organized labor's efforts seemed to make the difference in Ohio, Illinois, Indiana, and Pennsylvania, including "crucial local elections in the steel and coal towns of Pennsylvania and Ohio" (Dubofsky 2000, p. 157). The Democrats increased their already overwhelming margins in both the Senate and House, and also elected New Deal Democrats to the governorships in Pennsylvania and Michigan.
It's therefore much more likely that liberals and labor leaders were able to pass this legislation for very different reasons than what historical institutionalists or Marxists claim. First and foremost, the liberal-labor alliance was able to convince most moderate and conservative Democrats in Congress to vote for the act willingly by excluding agricultural and domestic labor from its purview. This purposeful exclusion meant that the great bulk of the southern workforce would not be covered, making it easier for Southern Democrats to support the legislation (cf., Farhang and Katznelson 2005). The exclusion of farm labor also made it easier for the Progressive Republicans of the Midwest to vote for the act. Translated into class terms, the exclusion of agricultural and domestic workers meant that the Southern segment of the ownership class did not have any direct stake in opposing the act, so Southern Democrats in Congress were free to support their Northern counterparts instead of voting with Northern Republicans, as they usually did on labor legislation. Contrary to historical institutionalists, then, the corporate leaders did not lose power in general despite the calamity of the depression. Instead, they lost on this issue because their key allies, the plantation owners, did not stick with them.