The persistent electoral benefits of redistribution: Evidence from Italy 1946-92

Bruno Caprettini, Lorenzo Casaburi, Miriam Venturini 01 April 2021

a

A

The global rise in inequality creates social conflict and fertile ground for populism (Guriev and Papaioannou 2020, see also the Vox debate on populism). Many even wonder whether the current levels of wealth concentration are sustainable and whether governments should intervene to correct them (Alvaredo et al. 2017, Chetty et al. 2017 and Bozio et al. 2020). However, the political feasibility of redistribution interventions depends, among other things, on the electoral benefits politicians accrue from implementing them. Does redistribution lead to political gains? Are these long-lasting? Why?

While the idea that successful redistribution translates into lasting electoral gains sounds reasonable, surprisingly little empirical evidence supports it. If anything, empirical work suggests that the electoral impact of redistributive public policies may be short-lived (Bechtel and Hainmueller 2011).

The Italian 1951 land reform

In a recent paper (Caprettini et al. 2021), we study the long-run electoral effect of a major redistribution policy in the context of post-WWII Italy. In the years following the war, the ruling Christian Democratic (DC) government passed several ambitious reforms in an effort to promote growth, cool social conflict, and avoid the formation of a Communist government. We focus on the flagship reform of the first DC government, the 1951 land reform. Between 1951 and 1955 the DC government expropriated 800,000 hectares of land owned by wealthy landowners and redistributed them to more than 120,000 poor rural workers. To date, this is still the single largest episode of wealth redistribution policy in Italian history. While the 1951 reform is unique in Italian history, land reforms are a common form of wealth redistribution and have been implemented by the governments of many agrarian economies (e.g. Russia, China, South Africa, among others). 

Our approach

Besides its scope and significance, the 1951 land reform is interesting also because it lends itself to clean policy evaluation. In 1950 DC was worried that a full-fledged land reform would take years to approve and decided to ‘hurry up’ by redistributing land only in clearly defined areas of the country. Figure 1a shows these areas – in the end, only large, unproductive estates located in these areas were expropriated and redistributed. 

This setting allows the comparison of changes in electoral outcomes in towns on the two sides of the reform borders to identify the causal effects of the reform (a ‘panel spatial regression discontinuity’). Archival research and formal econometric tests suggest that, before the reform, two sides of the border were similar in the Center-North, while important differences already existed in the South. Therefore, our main analysis focuses on the Center-North (Figure 1b).

Figure 1

a) Reform areas

b) Study area

Land reform and DC vote: 40 years of persistence

Did the reform help DC at the polls? Figure 2 gives the answer. For every Parliamentary election between 1946 and 1992, the figure plots the differential DC vote share in reform towns compared to towns that did not redistribute on the other side of the border. In the first two elections (both before the land reform) towns on the two sides of the border show very similar voting trends. Starting from 1953 (the first elections after the reform), DC gains around 4% in reform towns, from a baseline of 36%. The effect is very large, but it would not have surprised contemporaries. One of the DC leaders observed that “in the reform areas, the Scudo Crociato [the DC symbol, a red cross on white shield] shines, while the hammer and sickle rust”.

What may have surprised 1950s DC leaders is how long these effects lasted. Figure 2 illustrates that the initial gains persist almost unaffected for over 40 years. In 1992 DC still receives more than 5% in reform towns. This result stands in stark contrast with evidence of short-lived electoral effects of public policies in Germany (Bechtel and Hainmueller 2011) or  Brazil (Zucco Jr 2013).

Figure 2 Impact of land reform on DC vote share 

The roots of persistence

How did DC manage to capitalise on the reform? During the actual redistribution, favoritism in land allocation arguably helped DC to increase consensus in reform towns. We analysed several original land applications and found that political sympathies often determined who received land. Figure 3 shows, for example, the application of one rural worker who did not receive land because they were singled out as “rosso” (“red”, or Communist) on the top left corner. 

Figure 3 Example land application

Once all land was redistributed, however, other mechanisms must have allowed DC to maintain the initial gains. One plausible explanation is long-term electoral memory. Voters may have remembered some of the benefits of the reform and continued to be grateful and support DC for a long time. While we can not rule out that memory plays a role, we note that the electoral effects persist despite profound economic and social changes (including the deaths of many of the original beneficiaries). For this reason, we consider alternative mechanisms that may have allowed DC to maintain political support after the initial effects of land transfers started to fade. We provide evidence for two such mechanisms. 

First, land redistribution strengthened in reform towns the major farmers’ association, Coldiretti. Coldiretti benefitted from the reform because land beneficiaries had to join one local cooperative, and Coldiretti coordinated a vast network of these cooperatives (Primavera 2018). Coldiretti was also very active politically – throughout the period its leaders were MPs in the ranks of DC, and the organisation explicitly supported DC at the polls. We measure the strength of Coldiretti with new data from the elections of farmers’ health centres boards (Casse Mutue). Figure 4 illustrates the results – it shows on the horizontal axis the distance to the reform border (normalised to 0) and on the vetical axis the log votes for Coldiretti per agricultural worker. The figure shows that, at the border, reform towns that redistributed (on the right) vote significantly more for Coldiretti than towns that did not redistribute (on the left). This result supports the idea that the reform strengthened Coldiretti and helped DC build grassroots support in reform areas.

Figure 4 Coldiretti votes per agricultural voter

Second, persistent political support in reform towns may motivate DC politicians to reward these towns with additional policies several years after the reform. We examine this possibility with two additional variables. In Figure 5a we look at municipal transfers per capita and find that towns on the two sides of the borders received similar transfers in the years of the reform (1952-55). As soon as land redistribution was completed, however, we observe larger municipal transfers going to reform towns (1959). Similarly, Figure 5b shows that reform towns employ a significantly higher share of workers in the public sector (public sector employment is a common form of patronage in Italy). Together, these results suggest that in the years after the land reform, when the initial political effects may have started to fade, DC politicians found ways to transfer additional resources to reform towns.

Figure 5 

a) Municipal transfers per capita

b) Public sector employment share

The economic impact of the reform

One last plausible mechanism of persistence is economic. The reform may have changed voting through its impact on the economic structure of reform towns. For instance, land redistribution may have changed production, expanding social classes more likely to vote for DC. Alternatively, the reform may have made voters wealthier, turning them into petit-bourgeois voting for conservative parties (cf. Di Tella et al. 2007). We do not find evidence for any of these explanations. First, in the 50 years after the reform, towns at the border have similar levels of income and wealth (proxied by home ownership). Second, the economic and demographic structure of the population is similar across the border for the entire period. Third, economic activity and sectoral composition were not affected by the reform. Finally, when in 1994 Berlusconi’s Forza Italia replaced DC as the major Italian conservative party, we do not observe greater support for Berlusconi’s Party in reform areas. This result speaks against reform towns becoming more conservative in general, and is consistent with stories that are able to explain support for DC only. Taken together, these results suggest that the economic effects of the reform are unlikely to explain its persistent electoral effects.

Lessons

The 1951 Italian land reform illustrates how large redistribution policies can generate lasting electoral gains and create the conditions for ruling governments to cultivate political support for many years. A comparison between our results and other studies (Bechtel and Hainmueller 2011, Zucco Jr 2013) suggests that redistribution of (productive) assets can have more enduring effects than cash transfers. Our results also suggest complementarity between the initial large-scale investment (the land reform) and subsequent investments (brokerage of grassroots organisation, fiscal transfers, and public sector employment). Because grassroots organisations and favouritism in the allocation of public resources are common around the world, our results may offer general insights into how these systems come about and persist over time.

References

Alvaredo, F, C Chancel, T Piketty, E Saez, and G Zucman (2017), "Global inequality dynamics: New findings from WID.world", American Economic Review, 107 (5), 404-09.

Bechtel, M M, and J Hainmueller (2011), "How lasting is voter gratitude? An analysis of the short‐and long‐term electoral returns to beneficial policy", American Journal of Political Science, 55 (4), 852-868. 

Bozio, A, B Garbinti, J Goupille-Lebret, M Guillot and T Piketty (2020), “Pre-distribution versus redistribution: Evidence from France and the US”, VoxEU.org, 18 November.

Caprettini, B, L Casaburi, and M Venturini (2021), "The Electoral Impact of Wealth Redistribution: Evidence from the Italian Land Reform", available at SSRN 3767181.

Chetty, R, D Grusky, M Hell, N Hendren, R Manduca, and J Narang (2017), “The fading American dream: Trends in absolute income mobility since 1940” VoxEU.org, 5 May, and Science 356 (6336), 398-406.

Di Tella, R, S Galiant, and E Schargrodsky (2007), "The formation of beliefs: evidence from the allocation of land titles to squatters", The Quarterly Journal of Economics 122 (1), 209-241.

Guriev, S, and E Papaioannou (forthcoming), "The political economy of populism", Journal of Economic Literature.

Primavera, N (2018), La gente dei campi e il sogno di Bonomi La Coldiretti dalla fondazione alla Riforma agrarian, Roma-Bari: Laurana Editore.

Zucco Jr, C (2013), "When payouts pay off: Conditional cash transfers and voting behavior in Brazil 2002–10", American Journal of Political Science, 57 (4), 810-822.

a

A

Topics:  Economic history Politics and economics

Tags:  land reform, redistribution, electoral benefits, Italy, Christian Democratic Party, populism

SNF Ambizione Fellow, Department of Economics, University of Zurich

Associate Professor, Department of Economics, University of Zurich

Ph.D. Candidate in Economics, University of Zurich

Events

CEPR Policy Research