We develop competing hypotheses about the relationship between high performance work systems (HPWS) with employee and customer satisfaction. Drawing on eight years of employee and customer survey data from a financial services firm, we employed a recently developed empirical technique -covariate balanced propensity score (CBPS) weighting -to examine if the proposed relationships between HPWS and satisfaction outcomes can be explained by reverse causality, selection effects, or commonly omitted variables such as leadership behavior. The results provide support for leader behaviors as a primary driver of customer satisfaction, rather than HPWS, and also suggest that the problem of reverse causality requires additional attention in future HR systems research. Model comparisons suggest that the estimates and conclusions vary across CBPS, meta-analytic, cross-sectional, and time-lagged models (with and without a lagged dependent variable as a control). We highlight the theoretical and methodological implications of the findings for HR systems research.Keywords: human resource management, leadership, employee satisfaction, customer satisfaction, endogeneity, longitudinal, covariate balanced propensity score ACCOUNTING FOR SELECTION BIAS 2 Scholars began to report robust relationships between HR systems and firm performance outcomes in the 1990s (e.g., Arthur, 1994;Huselid, 1995;MacDuffie, 1995), which led to a surge of empirical research in strategic HR management over the next two decades (Paauwe, 2009). Much of the research suggests meaningful associations between HR systems and a variety of employee and organizational outcomes (see Jackson, Schuler, & Jiang, 2014;Subramony, 2009); a comprehensive meta-analysis showed that "high performance" HR management systems influenced firm financial outcomes via human capital, employee motivation, voluntary turnover, and operational performance (Jiang, Lepak, Hu, & Baer, 2012).Despite the large body of literature supporting the efficacy of high performance work systems (HPWS), some have questioned the methods employed in strategic HR studies. For instance, Wright and colleagues (2005) highlighted that many HR-performance studies up to that point were either cross-sectional or even post-predictive (i.e., HR practices were measured after performance), demonstrating the potential for reverse causality. Most research designs now ensure that measures are collected in the appropriate temporal sequence; however, past performance is still often found to be correlated with future HR practices, making it difficult to ascertain causal direction (Paauwe, 2009). Moreover, while HR scholars usually attempt to mitigate bias associated with common method variance and measurement error in research designs, we argue that reverse causality, omitted variables, and selection effects often go unchecked, or even unmentioned, in most studies.Concerns about bias -elsewhere referred to as endogeneity -in non-experimental research are certainly not new (Cook & Campbell, 1979;Rubin, 1974). More ...