Presidential Position-Taking, Presidential Success, and Interest Group Activity

Abstract

Do lobby groups help the American president achieve policy objectives? Existing research seldom evaluates interest groups and the president in conjunction, and as a result we have little systematic knowledge about how groups respond to presidential actions or whether they assist in realizing the president’s policy agenda. Building on existing data obtained through interviews with 776 lobbyists, combined with variables we generate describing issue salience, congressional attention, the political context, and policy adoption, we show that interest groups adjusted their lobbying activity to better reflect the president’s voiced preferences. Despite this strategy, we find that lobby groups had no significant marginal effect on policy adoption when controlling for the overwhelming influence of the president. The strong association between policy adoption and position-taking by the president withstands the inclusion of five alternative variables found in previous studies to condition the influence of the president over policy adoption.

Notes

Notes

1 Baumgartner et al. (2009) found that the support of “high-level government allies,” including the president, was the strongest predictor of policy success for interest group lobbyists working toward the officials’ policy goal.

2 A third rationale for greater lobbying is that of LaPira and Thomas (2017), who show that more experienced lobbyists, especially those who formerly worked in Congress, are hired by clients looking to minimize possible harm from government activity. Their argument relates to the conditions under which groups engage different kinds of lobbyists, and it is not relevant to this study’s focus on the relationships between interest group activity and presidential activity, nor which actor is a more powerful predictor of policy adoption.

3 See Salisbury et al. (1987), appendix 2, for more details on the sampling procedure.

4 For example, “FDA, FTC, and USDA announce a joint initiative in food labeling, including declaration of ingredients, and extended open-date and nutritional labeling in standard food items.” A nearly complete list of the proposals can be found in Heinz et al. (1993) (several proposals were part of the data collection but dropped for the book).

5 Specifically, we used partisan polarization, defined as the distance between the Poole and Rosenthal (1997) estimates of member ideology of each Congress, averaged between the two chambers (correlated with divided government at 0.97 and the location of the median member, also using Poole and Rosenthal (1997) first-dimension estimates (correlated with divided government at 0.99).

6 Including in the model all the control variables that appear in Table 1 also fails to produce a significant relationship between salience and presidential position taking.

7 Clustering errors by the proposal yields greater errors than clustering by the lobbyist or the lobbyist’s employer, making proposal-clustered errors the most conservative choice.

8 Our coding of whether each issue was complex may not be entirely consistent with Canes-Wrone and de Marchi’s (2002) definition of complex issues as those that are regulatory-related rather than social-oriented. We did estimate whether each issue was regulatory or social in its effects, as well as a simpler coding that energy and agriculture issues were technical, while health and labor issues were social. Neither of these codings confirmed their findings.

9 To test the effects of divided government on presidential success, we interacted the president’s position with an indicator for unified government and use this and the constituent terms, along with the control variables (salience, conflict, agency-initiated, and congressional attention) to predict the policy outcome. In this model unified government was not a significant predictor of policy adoption (0.36 at p=.535), and neither was the interaction between the president’s position and unified government (–2.25 at p=.101), whereas the president’s position was still significant (3.01 at p=.009).