Tobias Nowacki: The effect of electoral systems and rules on candidate success and representation

I am grateful to The Europe Center for a generous grant to support my research on the effect of electoral systems and rules on candidate success and representation more broadly. The grant enabled me to collect rich data on over 10,000 Polish elected officeholders’ financial assets, property and income from mandatory annual declarations. Combining these data with electoral records and modern causal inference methods, I evaluate whether a change of electoral rules from plurality rule to proportional representation not only affects the number of candidates running, but also produces wealthier or higher-earning winners on average. While the findings from this study are still preliminary, early evidence suggests that, empirically, local election winners in either electoral system report similar levels of personal wealth on a number of dimensions (monetary assets, property, liabilities). This is a surprising result with notable implications for political representation.

 

This project opens a new avenue in research on electoral systems. Canonical research on the topic of electoral systems studied the effects of plurality and proportional representation (PR) on system-level or election-level outcomes such as the number of parties or candidates (e.g., Lijphart 1994), government survival (Cheibub 2002), or the overall levels of redistribution (Iversen and Soskice 2006). The study of the effect of electoral systems on election winners’ individual-level attributes is more nascent; among the more-covered questions is whether PR aids in the election of more women (Salmond 2006). Research on electoral systems’ effects on winners’ wealth, however, is still scarce – in no small part because of the lack of extensive, high-quality data sources on their assets and holdings.

 

This fact sits at odds with the question’s importance. In the discussion on electoral systems’ (dis)advantages, we should care a great deal about effects on descriptive representation across various dimensions: if our elected officeholders are, on average, vastly wealthier than the electorate, we may worry that policy-making and decision-making outcomes are skewed towards the interests of those demographics. In a study of Norwegian local politics, for example, Folke et al (2021) find that politicians are more likely to live in wealthier areas and, in turn, are also more likely to shield their neighbourhoods from local public bads. We should therefore strive to understand whether (and how) different electoral systems can address the problem of distorted representation with respect to wealth and give policy-makers more information when deciding how to reform and design elections with the objective of improving representation and democratic quality in mind.

 

The data collection made possible by the Europe Center’s funding takes us one step closer towards answering this important research question. I take advantage of a law in Poland that requires all elected officeholders in the country to submit an annual (usually handwritten) statement of assets, business interests and income. These statements are then published in a diffused manner across a wide number of institutional websites – ranging from municipalities (for local officeholders) to the website of the Sejm, the lower house of the national legislature. Thanks to this grant, I have been able to hire research assistants who collected, digitised and cleaned these statements for more than 10,000 candidates from over 500 towns and combined them into a well-organised database.

 

My research design combines these newly collected data with already existing records on electoral results and then utilises a sharp population threshold across municipalities to compare otherwise similar elections that differ in their electoral system (regression discontinuity). More precisely, Polish law specifies that all municipalities with less than 20,000 population should use plurality in their local elections to the city council, while those exceeding the threshold should run their elections with proportional representation (PR). A regression discontinuity design compares election winners in places just under the threshold to those just over the threshold; under certain assumptions (which are likely met in this case), we can identify the effect of the change in the electoral system on candidate attributes.

 

The preliminary results from this analysis suggest that towns on either side of the threshold elect candidates with similar levels of wealth: I find no statistically significant evidence for a discontinuity in overall wealth, property wealth, income or gross liabilities across the population threshold at which the electoral system changes. In further work, I intend to extend these analyses to compare the full distribution of wealth across officeholders in any one city (rather than just summary statistics such as mean or median per town).

 

These results are surprising given theoretical expectations and warrant further research into the mechanisms of candidate selection and the relationship between wealth and political careers. We may have expected that plurality elections, on average, produce less wealthy winners as candidates run in more heterogeneous single-member districts – hence increasing the representation of poorer electoral districts. Such a scenario would strengthen the argument that some opponents of electoral reform make that single-member plurality elections guarantee more equitable representation and ensure that no area (and, by extension, no poorer part of town) is left without an elected officeholder. Alternatively, we may have expected PR elections to produce wealthier winners if we think that winners in plurality contests need to be of higher quality overall (in order to win first place), whereas elected officeholders in PR can ‘shirk’ their way into office by occupying a lower list rank and avoid putting in effort (see Cox et al 2021). The fact that there is no significant observable difference in wealth in either direction suggests that these theoretical mechanisms either cancel each other out or are absent altogether. Moreover, they speak to an important dimension of representation for policymakers and electoral reformers: there is no evidence that either form of electoral system produces a more representative set of office holders (with respect to wealth – they may still differ on other candidate attributes such as gender or age).

 

Finally, I intend to use the data collected thanks to the grant in future and related work. In a follow-up project, I will examine the distribution of officeholder wealth across parties and independent (non-partisan) candidates, as well as across geography: are those who represent more socio-economically disadvantaged districts or neighbourhoods also less wealthy themselves? All told, I am grateful to the Europe Center for supporting my data collection and research agenda on electoral systems

 

References

 

Cheibub, J.A., 2002. Minority governments, deadlock situations, and the survival of presidential democracies. Comparative Political Studies, 35(3), pp.284-312.

 

Cox, G.W., Fiva, J.H., Smith, D.M. and Sørensen, R.J., 2021. Moral hazard in electoral teams: List rank and campaign effort. Journal of Public Economics, 200, p.104457.

 

Folke, O., Martén, L., Rickne, J. and Dahlberg, M., 2021. Politicians’ neighbourhoods: Where do they live and does it matter?. Working Paper.

 

Lijphart, A., 1994. Electoral systems and party systems: A study of twenty-seven democracies, 1945-1990.

 

Iversen, T. and Soskice, D., 2006. Electoral institutions and the politics of coalitions: Why some democracies redistribute more than others. American Political Science Review, 100(2), pp.165-181.