Moving beyond the political trust crisis debate: Residual analyses to understand trends in political trust

For decades, political scientists have hotly debated longitudinal trends in political trust rates. An important undercurrent in the debate is that any decline in political trust might signal a legitimacy crisis. Yet, descriptive figures are unable to distinguish between two interpretations of these downward trends: (i) declines that can reasonably be expected as a reflection of declining political trustworthiness (i.e., procedural or output performance) and thereby reflect critical citizens who monitor their democratic institutions; and (ii) downward trends that are not warranted by democratic performance and thereby suggest a more fundamental disconnect between citizens and their democratic institutions.

This research note argues that residuals to multilevel models of political trust allow us to distinguish between these two types, and thereby provide a better understanding of trends in political trust. These residuals do not only reveal short-term aberrations to the explanatory model (often reflecting short-lived, country-specific events), but also the extent to which a country's trust rate systematically underperforms in the middle- to long-term. To the extent that declining trust rates are lower than explanatory models predict, the residuals express excessive distrust. To the extent that declining trust rates are in line with the explanatory models, the residuals reflect critical, monitoring citizenship.

We outline the approach of residual analyses as a tool to better understand trends in political trust. We illustrate the use of these residual analyses on a cross-national, longitudinal data set (the Eurobarometer), covering 15 Western and Southern European countries between 1999 and 2019. While political trust rates fluctuate in all these countries, we only find evidence for a structural decline in two of these countries. In France and Spain political trust failed to recover in line with improving economic and institutional performance after the Great Recession. We then test the versatility of the tool to different conditions, including retests on an alternative set of countries (11 Central and Eastern European countries between 2004 and 2019) and an alternative dataset with different measures and time points (the European Social Survey).

Finally, we elaborate on the two main conditions under which residual analyses offer a useful tool to the trend debate in political trust research: (1) a firm understanding of the object-driven determinants of political trust, and (2) a detailed coverage of country-wave combinations to separate structural trends from short-term fluctuations.