Political economy research has contributed significantly to our understanding of the mechanisms of vertical transmission of individuals’ beliefs and preferences. For instance, Paola Giuliano, Antonio Spilimbergo and Alberto Alesina describe how individuals’ preference formation within families depend on historical family types or exposure to economic recessions (e.g. Giuliano and Spilimbergo 2008, Alesina and Giuliano 2011). Focusing on the historical persistence of attitudes, Acharya, Blackwell, and Sen 2016 show that contemporaneous voter preferences in the US are shaped by the long shadow of slavery; attitudes and values are passed down over time through processes of intergenerational socialization. While research in political science has focused substantially on vertical transmission mechanisms of ideology, a key controversy concerns how well we can actually identify this vertical transmission mechanism.
A nascent and growing interest within the historical political economy literature centers on the horizontal transmission of ideology. By horizontal-, as opposed to vertical transmission, I refer to processes through which political attitudes are acquired through interactions with peers. That is, the process through which individuals form political attitudes not only through exposure to specific events, but through the process of learning from peers via horizontal social and political networks. For example, a recent working paper by Giuliano and Tabellini (2021) shows that US citizens living in counties that were more exposed to historical European immigration are more likely to hold liberal ideology and have stronger preferences for redistribution today. The authors suggest that the key mechanisms behind this result are not economic channels but, instead, the horizontal transmission of ideology from immigrants to natives in the early 20th century. Nonetheless, the question of how to explain the persistence of the proposed horizontal mechanisms remains unclear.
The flourishing historical political economy literature has a lot to say and contribute to our understanding of the mechanisms behind the horizontal transmission of ideology. In this blog, for example, Vicky Fouka has already written about the need to better understand the oftentimes very rapid changes in social norms. Jared Rubin has written about the key role of network structures in understanding the spread of the Reformation. The opportunity to exploit historical data at very disaggregated levels provides many advantages when studying horizontal transmission mechanisms. Sometimes we are fortunate enough to be able to exploit historical registers at the individual-level. An example of such rare sources of data are the precious and unique individual voting roll calls (i.e., individual-level information on turnout) during the Spanish Second Republic (1931-1936). An example appears in Figure 1 below. Matching these individual voting roll calls with historical population censuses enables the opportunity to study the determinants of individual turnout decisions. By doing so we can focus directly on historical outcomes, exploring for example the determinants of individual-level turnout in the 1930s – thereby avoiding the need to make assumptions about the long-term persistence of effects.
Understanding Mobilization and Horizontal Transmission with Individual Voting Roll Calls
In our JoP article (2020), “From Political Mobilization to Electoral Participation: Turnout in Barcelona in the 1930s”, co-authored with Carles Boix, Jordi Muñoz and Toni Rodon, we examine the process of electoral mobilization that followed the extension of voting rights to low-income citizens exploiting those very unique individual voting roll calls. We show that individual electoral participation was driven by the direct mobilization strategies developed by political parties and by social organizations, such as trade unions. This was the case especially among unskilled workers and in areas with a high density of working-class voters. Of equal importance for our understanding of the processes of horizontal transmission, we also show that turnout was shaped by indirect mobilization channels – such as the social networks in which partisan ideas and organizations were embedded.
Our main focus of interest is the role of political organizations in mobilizing voters in 1936, especially unskilled workers. In order to estimate it, we measure the distance from each voter’s precinct centroid to the nearest organization (in log meters) and interact these measures with the unskilled worker’s dummy. To estimate local effects among low-skilled workers, we include the interaction between unskilled worker and log distance to the nearest republican/socialist association. Results appear in Figure 2 below. Ceteris paribus, distance away from a republican/socialist organization is significantly negatively related to voting in 1936, conditional on voting in 1934, and a number of individual and precinct-level controls. More precisely, when we move one standard deviation away from a republican/socialist organization, the probability of turning out to vote in the 1936 election is 3.3 percentage points lower among unskilled workers.
The organizational effects of republican and socialist associations were stronger in the neighborhoods of the city in which the density of unskilled workers was higher. We interpret this result as evidence of indirect political mobilization effects. As such, the article advances our understanding of both direct and indirect mobilization effects – arguably, the main pillars of horizontal ideological transmission – for which, surprisingly enough, there has been relatively weak and scant evidence in the existing literature so far and even less so for early democratization periods.
A stronger test of our argument relies on the strategic switch by the powerful anarchist trade union CNT. We exploit it to identify a very unique causal effect of organizational mobilization. Breaking with its traditional election boycott campaigns, in 1936 the anarchist leaders decided to stop campaigning actively for abstention. That short-term strategic change by the anarchist trade union provides us with a good opportunity to identify the impact of organizations. If the latter mattered for voters, we should expect two things: that anarchists effectively deterred voters from turning out in the 1934 elections; and, second, that they mobilized new voters or at least did not generate differential levels of abstention in the 1936 elections.
The empirical results are consistent with our expectations. See Figure 3 below. Distance to anarchist organizations within the city of Barcelona is significantly more vote-deterring in the 1936 election than in 1934. Looking at the figure below we can see that whereas the anarchist ateneus had a vote-deterring effect (turnout increased with distance) in 1934, they had a mobilizing effect in 1936 – turnout decreased with distance. In other words, the effect basically switches. Substantively, the empirical results show that in 1934 voters close to a CNT center were 4.5 percentage points less likely to participate than those 1km further away, while in 1936 the same voters close to anarchist centers were 5 percentage points more likely to go to the polls.
In a second and more recent working paper, that focuses more directly on women’s horizontal ideological activation, we argue that past political experiences – before the extension of the franchise – might be key to understand women’s electoral participation once they got the right to vote. We analyze this question by exploiting a wave of labor unrest during the 1915-1920 period in the textile sector, a heavily feminized sector, prior to the introduction of women’s suffrage. We exploit detailed information on labor conflicts and strikes – including the number of women taking part in strikes – to investigate the ideological activation of women and its effects on electoral participation later on. Put differently, we analyze the variation in the magnitude of the gender gap upon the political enfranchisement of women.
We focus again on the case of the Spanish Second Republic, which granted women the right to vote for the first time in 1933. We argue that the economic approach to understanding the gender gap, based on women’s outside options in the labor market, is not sufficient to comprehend differential gender turnout patterns. Instead, we argue that – once women’s suffrage was introduced – the gender turnout gap crucially varied as a function of the previous political experiences of women, such as exposure to episodes of labor unrest in feminized industries in the period prior to enfranchisement. In other words, women’s electoral mobilization depended on to what extent they had been politically activated, prior to the introduction of women’s suffrage.
The Road Ahead: Identifying Social Complementarities and Ideological Spillovers
Future research should aim to provide a deeper theoretical and empirical understanding of the social complementarities, defined as externalities among peers, which shape horizontal transmission of ideology. Horizontal ideological activation can be understood as a process in which two key ingredients interact: individual exposure to specific events and the social basis under which such individual exposure takes place. The social basis –political structure, network characteristics and the resources for mobilization– shape and characterize the social complementarities between participants’ actions. From this perspective, in order to understand horizontal transmission of ideology is necessary to identify both individual exposure to specific political events and the extent to which social complementarities are at work.
To put it differently, if individual exposure to a charged political event takes place under strong social complementarities, such individual exposure is likely to be associated with strong horizontal ideological activation. However, if individual exposure occurs under weak social complementarities, individual exposure to political events might not cause ideological activation. The obvious lesson is that we should observe greater horizontal transmission of ideology when there are significant ideological spillovers driven by strong social complementarities. The recent paper Heroes and Villains by Cagé, Dagorret, Grosjean and Jha (2020) investigates soldiers fighting in Verdun under Pétain during the IIWW and argue that diffusion of anti-democratic values occurred because of strong complementary hierarchical networks. On the other hand, the paper discussed before by Giuliano and Tabellini (2021) explores the horizontal social complementarities shaped by inter-group contact: intermarriage, residential integration and linguistic similarity.
The challenge ahead is therefore to identify the nature of the social complementarities that characterize processes of horizontal transmission of ideology – either amplifying or diluting individual exposure to specific politically charged events and other political experiences. The opportunity to reconstruct voters’ families and social networks – exploiting individual voting roll calls and population census such as the ones described for the Spanish case – are likely to offer a rich opportunity to explore the role of social complementarities in the horizontal transmission of ideology. The fundamental question we are dealing with is understanding how ideology spreads across social networks. Or more generally, understanding how the diffusion of ideology within peer groups depends on the strength of social complementarities. My intuition is that the seeds of ideology should work very differently depending on how fertile the soil in which they are planted is. We are just starting to develop historical political economy studies on that front and arguably the exploration of historical individual-level registers (individual voting roll calls, population census, etc.) is one of the most promising avenues for answering such questions.