The Effect of Far-Right Parties on the Location Choice of - - PowerPoint PPT Presentation

1

The Effect of Far-Right Parties on the Location Choice of Immigrants: Evidence from Lega- Nord Mayors

Emanuele Bracco, Maria De Paola, Colin Green, Vincenzo Scoppa Lancaster University, Università della Calabria

2

Motivation

• Immigration is an area of increasing political focus
• Rise of parties where immigration is a core platform
• There is increasing evidence that the presence of

immigrants affects the electoral success of these parties

• We explore the opposite dynamic:
• Do Immigrants react to these parties? Are their choices

affected by the presence of an anti-immigration mayor?

• We focus on the case of municipal elections in Northern
• Italy. This region has seen the ascent of a party with a

core anti-immigration political platform, Lega Nord

3

Existing Literature

• Existing literature has analyzed how the presence of immigrants

affects political preferences:

• Mayda (2005): attitudes towards immigration depend on one’s

level of skills, and the skills of immigrants

• Russo (2010): theoretical model of voting on immigration policy
• Friebel et al. (2013): emigration decision affected by xenophobic

attacks in destination country.

• Barone et al (2014): immigration increases votes to centre-right

(but not in big cities)

• Halla et al. (2014), Otto and Steinhardt (2014), Sekeris and

Vasilakis (2016): immigration increases votes to anti-immigration parties

• It is hard to disentangle the effect of electing a mayor

supported by an anti immigration party on the location decisions of immigrants

• Strong correlation between immigration and influx of

immigrants

• Causality goes in both directions
• We know how areas with more immigration see a rise in anti-

immigration parties (Barone et al. 2016)

• To address this we use Regression Discontinuity Design: look at

municipalities where the Lega-Nord-supported mayor won/lost by a small margin

• Assumption is that this generates a quasi-experimental setup

in which treatment (being subjected to a Lega Nord mayor) can be considered as quasi-random

4

Identification strategy: RDD

5

Background: Lega Nord

• Anti-immigration, pro-federalism (at times pro-secession),

anti-establishment (anti-Rome!) party born in the Eighties

• After the collapse of “old” political system in early Eighties,

gains power (elects the Mayor of Milan in 1993)

• Subsequent unstable alliance with centre-right (Berlusconi)

parties

• Gains votes in Northern Italy, esp. Lombardy and Veneto.
• Between 5 and 15% of (national) votes, but first party in many

areas (esp. Veneto)

6

Background: Municipal Elections

• Directly elected mayors, by plurality (<15k inhabits.) or runoff

(>15k inhabits)

• Mayors are partisan
• City council elected together with mayor through a open-list

proportional representation

• Electoral system is such that mayor always has a majority in the

city council

• Balance of power Mayor-Council is strongly tipped in favour of

the Mayor

7

Data

• Data from 4,000 municipalities in Northern Italy (Piemonte,

Lombardia, Liguria, Veneto, Friuli-Venezia Giulia, Emilia- Romagna [no Lega Nord presence in Valle d’Aosta or Trentino- Alto Adige]).

• From Interior Ministry (elections from 1997):
• Data on mayoral elections including parties linked to each

mayoral candidate

• From ISTAT: data on population and immigration, data on

geographic and economic characteristics

• From Finance Ministry: average taxable income
• Data from local councils’ balance sheets on expenditures in

social services

8

Identifying Lega-Nord mayors

• Lega: any mayoral candidate who is supported by Lega Nord,

as from the electoral data

• (Lega2): Lega+ any mayoral candidate who is supported by a

centre-right coalition (“CEN-DES”) in small (plurality, single- list) municipalities

• (Lega3): Lega Nord only if the LN member was the mayoral

candidate or deputy mayor.

• We probably lose some “Lista Civica” mayors who are actually

Lega Nord (downward bias in our estimate)

• We cannot distinguish easily the intensity of Lega Nord

presence (noise in our estimates?)

9

Margin of Victory

• Votes to the mayoral winning candidate − votes to the runner-

up

• Second-round votes used when there is a second round
• First-round votes used otherwise
• Results materially unaffected if we discard municipalities that

go to a second round

• Too few municipalities go to a second round to do the
• pposite.

10

Controls

• Population
• # employed people in agriculture/industry/trade/transport as a

share of employed people

• Area of municipality, dummy for urban munic., above-sea level
• Education (avg years of education of adults)
• Average taxable income

11

Three maps: regions where Lega Nord runs

Three maps: municipalities where Lega Nord runs

12

Three maps: municipalities where Lega- Nord rules

13

Immigration Data

For each year/municipality we know:

# of newly registered foreigners coming from other municipalities # of cancelled foreigners going to other municipalities # of newly registered foreigners coming from abroad # of cancelled foreigners going abroad Create variables for net inflow (registered-cancelled), and divide these by the population. “Inflow from abroad” may hide previously illegal immigrants getting their papers

14

15

Regression Discontinuity Design

and in its non-parametric version:

16

Regression sample and summary statistics

• A standard concern with this type of identification strategy

is that other relevant characteristics may also vary discontinuously with respect to the margin of victory.

• We focus on observed characteristics and test whether a

discontinuity is present in any of these variables when a Lega Nord Mayor is elected.

17

18

19

Histogram of Margin of Victory

20

Regression: Total Net Flows, Parametric

21

Results: Net Flow — Non-Parametric (CCT) Optimal Bandwidth à la CCT: ±15.25%

22

Scatterplot by LegaNord

23

Inflows and Outflows

Robustness check: Alternative Lega Nord Measures

24

Immediate and lagged effects

26

Falsification check

27

New mayors and incumbents

28

Social expenditures

Do immigrants run away from more selective public good provision / stricter policing? Is LegaNord-mayor just an instrument for attitudes towards immigrants?

29

Evidence that immigrants’ location choice is affected by Lega- Nord Mayors They don’t flee, but they may choose other locations (mostly through fewer inflows from other municipalities) Where do they go instead? Do they just move locally? Is there any (sound) way to disentangle these two mechanisms?

Conclusion