The literature on the economic origins of political radicalism revolves largely around the ‘China shock’. Yet, trade not only causes shifts in the labour market but can also affect living standards via the expenditure channel. This column shows that the relative consumption loss of the European middle classes, who benefit less than low-income consumers from cheap imports from China, and the associated perceived risk of societal decline can explain the large success of right-wing political radicalism in the last three decades, above and beyond standard effects on the supply side.
The last three decades have witnessed a substantial rightward shift in political preferences up to the fringes of the political spectrum, including the success of national far-right parties in numerous European countries. The growth of empirical scholarship on the origins of political radicalism is impressive, but one argument that remains central to much of the available literature is the idea that trade globalisation in the form of a ‘China shock’ or ‘China syndrome’ has prompted affected workers to hide behind the protectionist electoral platforms of far-right political parties (e.g. Colantone and Stanig 2017, Autor et al 2020). At the same time, it is generally recognised that trade globalisation, and in particular the emergence of cheap producers, has increased the availability of goods as well as consumers’ purchasing power, especially at the bottom of the income distribution (Khandelwal and Fajgelbaum 2015). While this benefit is unlikely to redress the loss in income from the displacement of low-skilled jobs (see also Dorn and Levell 2022), it may well trigger knock-on effects elsewhere, a mechanism that is able to reconcile the fact that the pool of far-right voters is much larger than that of trade-displaced manufacturing workers.
The China demand shock
In a new paper (Marzinotto 2025), I start off by conceptualising trade globalisation as a demand-side rather than a supply-side shock. This follows influential work on the pro-poor bias in international trade through its impact on expenditures. While low-income consumers are the main beneficiaries because they spend a significant share of their income on cheap tradables, middle-income consumers stand to lose from the process of trade liberalisation (Fajgelbaum and Khandelwal 2016, Heins 2023). Seen this way, trade exposure, especially in the form of a China shock, is relevant through a mechanism that is complementary to standard supply-side effects from import competition yet at the same time more cross-cutting.
Figure 1 correlates percentage changes in consumption by income quintile with the change in the share of imports from China over total imports for a large sample of European countries from the early 2000s to 2014. The association between trade with China and consumption gains is more clearly positive for the first quintile (a correlation coefficient of 0.35, p<0.05) and loses significance when moving up to higher income groups. Overall, the evidence delivers a very intuitive message and one that is fully consisted with the sophisticated trade models in Fajgelbaum and Khandelwal (2016) and Heins (2023): China has been flooding international markets with products that cater to low-income groups, whose consumption grew by more than that of other groups in the same society.
Figure 1 China-induced shifts in consumption by income group
Note: Own elaboration based on Eurostat and Household Budget Survey (HBS). The sample includes: Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Netherlands, Poland, Slovenia, Slovakia, Spain, Sweden, and the United Kingdom.
The argument
The relative consumption loss of middle-income agents (i.e. those in the third quintile of the income distribution) becomes politically consequential because they are generally more concerned with social status than any other income group, as acknowledged in the literature on social comparison and relative deprivation (e.g. Kim et al. 2018). Far-right political preferences should thus mainly concern middle- or low-to-middle-income groups, who see their consumption behaviour approximating that of poorer households while departing from that of high-income agents, which comes with a threat of societal decline.
The expectation is that the consumption misfortunes of the middle class are better captured by the electoral platforms of the far-right rather than the far-left, if the latter is believed to represent the interests of the same socioeconomic groups the middle class feels threatened by.
Treating trade globalisation as a demand-side shock comes with two important implications. First, the distributive effects from trade exposure do not vary from industry to industry but from one income group to the other (i.e. within-country heterogeneity). Second, gains from trade also vary from country to country as a function of relative technology endowment, which is an additional source of variation I exploit in the data (i.e. cross-country heterogeneity). Trade exposure has probably been politically consequential mostly in the rich Western European economies of the euro area because trade increased the availability of affordable products rather than of sophisticated varieties as happened instead in the less affluent Central and Eastern European countries, a stylised fact that is well described in Heins (2023).
The evidence
To account for political preferences, I use pooled European Social Survey (ESS) data registering individual-level attitudes in up to 18 European countries from 2002 (ESS1) to 2014 (ESS7). I consider whether respondents “feel closest” to parties that are classified in the available political-science literature as radical (or populist) rightist configurations (e.g. Burgoon et al. 2019). Because none of the ESS rounds contains information on individuals’ consumption behaviour let alone on its evolution over time, I measure changes in consumption in the five years before each survey round using data from the Household Budget Survey (HBS), which include mean consumption expenditures by income quintile in a given country-time, and then match these aggregate measures to the ESS data.
To make sure that consumption behaviour is not simply income-driven but stems from an exogenous variation in the costs of living (i.e. so as to isolate relative changes in consumption that are fundamentally price- rather than income-induced), I devise a simple empirical strategy. I calculate, for each quintile, the average propensity to consume, which is the share of income that is devoted to consumption, as well as its change in the five years preceding each survey round. I further hypothesise that a fall in the cost of living that is trade-induced would result in a lower propensity to consume mainly at the bottom of the income distribution because poorer agents can devote a smaller share of their income to (cheaper) consumption, all else equal. The same shall not be true of agents belonging to the third quintile, who feel threatened in their social status if, paradoxically, their propensity to consume grows (falls) faster (less) than that of the first quintile because of their differential exposure to low-quality imported varieties.
Hence, the key determinant of political preferences is country-level time-varying changes in the propensity to consume of the third relatively to the first income quintile, where my speculation is that higher values indicate that lower income agents receive a consumer price benefit than does not equally accrue to the third income quintile.
Through simple logistic regressions and a battery of individual-level controls and robustness tests, I find that the consumption of the middle class relative to the first quintile is an important driver of far-right voting preferences at the country-time level, net of other individual-level determinants of political preferences (age, gender, education, employment status, foreign-born status, urban status, religiosity, subjective income). It is much more so in the affluent Western European economies than in Central and Eastern European countries because of the former’s larger pro-poor bias (and hence the more unequal gains from trade). The results are non-random but relate to selected parties’ substantial electoral platforms. In countries where the middle class loses in consumption terms relative to poorest households, individuals prefer political parties, whose agendas score high in terms of net nationalistic autarky.
Lessons
My results provide important insights. First, it is desirable to distinguish between consumption-related and income-related misfortunes, as they lead to differential preference formation both in ideological and substantive terms. Those that suffer from consumption deprivation might still be wealthier than those that gain in consumption terms, which is an assumption that is both plausible and at the core of the reference trade model. Second, it is the consumption deprivation of specific income categories that drives my overall results and, more generally, shapes political preferences at the country level, especially if the relevant group coincides with the median voter. Third, the reason why the received wisdom is that political radicalism in Central and Eastern Europe is unaffected by economic factors should not be interpreted as evidence that culture is a superior explanation of political preferences, but might simply indicate that the effects from trade are here much less unequal than elsewhere.
To wit, recent political shifts are part of the revolt against trade globalisation, as now in the received wisdom, but the backlash is coming from groups that have not been traditionally identified as the main losers from trade globalization.
Source : VOXeu
