The goals of this chapter are twofold. The first is to describe the empirical research design strategy that we employ to evaluate the two testable hypotheses and the causal claims that lead to these hypotheses. The second goal is to statistically test hypothesis 1 produced by the theory in Chapter 2. Recall that this hypothesis posits that higher geographic concentration of domestic private SMEs in autocracies increases the probability that these firms will form a national-level business association. We test this hypothesis on a comprehensive time-series cross-sectional (TSCS) dataset of authoritarian countries and report the results obtained from this sample.

The rest of this chapter is organized as follows. We first discuss the empirical research design employed in this book. We then describe the sample, the dependent variable, and the statistical models used for testing hypothesis 1. This is followed by discussing the procedure employed to operationalize the independent and control variables. We then report the results obtained from testing hypothesis 1 and the robustness checks of these results. The chapter concludes with a discussion of the implications of our main findings.

Empirical Research Design

We adopt a multi-methodological approach to test hypotheses 1 and 2, the corollary to hypothesis 2, and the causal claims that produce these two hypotheses. This “multi-method” approach combines large-N statistical analysis, in-depth study of three cases selected using a combination of quasi-experimental and most-similar design, evaluation of within-country survey response datasets, and time-series data on corruption. We adopt this multi-method approach, as marrying large-N analysis to in-depth case studies permits us to exploit the advantages of both approaches. As stated by Fearon and Laitin (2008), “qualitative and quantitative tools can be used jointly to strengthen causal inference” since large-N analysis evaluates “whether and what sort of patterns or associations appear in the data” while “case studies” are “extremely useful…for assessing whether arguments proposed to explain empirical regularities are plausible.” Gerring (2007: 85) similarly suggests that large-N analysis enhances the utility and validity of case analyses by noting that “the case study is, by definition, a study of some phenomenon broader than the unit under investigation. The more one knows about this broader population of cases, the easier it will be to choose cases and to understand their significance.”