Tag Archives: Globalisation



Sheve and Slaughter

Chapter 3: Cleavages in Public Preferences about Globalisation

 A Summary

In a Nutshell

The authors seek to predict who might oppose measures to open up trade based on the assumptions that individuals are self-interested, that individual welfare is dependent on income, and that the individual knows how trade policy affects his income. They expect that those whose income suffers as a result of more open trade will oppose freer trade policy. They largely focus on the Heckscher-Ohlin (HO) framework of factor mobility where workers can move costlessly from sector to sector and across industries. Thus wages differ not by industry but by type of worker. They posit that this model better reflect long term horizons as compared with the Ricardo-Viner (RV) framework whereby workers cannot move across all industries, so wage differences are due to type of work rather than type of worker.

 They run empirical models to test the hypotheses they derive from the model, which is that those less skilled workers will be more opposed to trade liberalization than highly-skilled workers. Their data analysis confirms this. It also seems to suggest that the type of industry that employs a worker is not particularly relevant in predicting if he will be anti-free trade measures. Other important variables prove to be race (in certain circumstances), membership of a union and political awareness. However, these variables do not significantly alter the coefficients of the skills variable. Surprisingly political affiliation is not significant in predicting if a person will be against free trade policy.


  • Trade policy affects factor incomes by changing a country’s product prices.
  • An industry enjoying a rise in product price post liberalisation will earn +ve profits, and if product price is falling the reverse is true. Optimising firms earning +ve profits increase production so economy wide demand for different types of labour changes i.e. demand for labour in expanding sectors will rise relatively. For equilibrium to be restored the wages of the two industries must adjust. 
  • The HO model says that trade protection is extended to sectors that employ factors of production in which the country is relatively poorly endowed as these are the workers who lose income in the move from autarky to free trade. So the types of workers who are abundant support free trade, and those that are scarce support protectionism. E.g. US tariffs were higher in less skill intensive industries in 70s and 80s (as they have a greater proportion of skilled workers). 
  • The RV model says income is linked to sector of employment not the type of worker. So sectors whose prices fall post free trade (ones with comparative disadvantage) realize income loss for their workers, and thus they are opposed to free trade. i.e. those workers who work in protected industries should oppose trade liberalisation. 
  • Trade policy affects regional economic performance, and so if it acts to shrink industries with less comparative advantage, then areas with a lot of such industries will see demand for housing fall and thus housing prices fall. So in areas with a greater concentration of activity in sectors with weak comparative advantage, homeowners will tend to oppose free trade.


Data Analysis

They do statistical analysis of National Election Studies surveys of samples of the US population. I will not describe the data nor the regressions, just summarise the findings reported.

 There is little statistical evidence to suggest that the person’s industry of employment influences policy preferences. This counters the RV model.

  • However, the coefficients on the skills variable (measured in terms of years of education) is significant and means that if you take a high school dropout and give them a full education they are 25% more likely to be in favour of free trade. So, in support of the HO model, it is the individual’s skill level that matters most for probability  of supporting free trade measures. This indicates that intersectoral labour mobility is high in the US over the time horizons in which people are evaluating trade policy.
  • They perform a robustness check by including other variables, such as political party identification, ideology, race, gender etc. and find that the coefficients in the skills variable is not significantly affected. However the check reveals that there is another important variables as trade union membership appears to be a determinant of opinions regarding free trade. Also, a small effect is found in those supporting environmental causes in being against free trade.
  • They also conclude that political awareness is a factor. Those who are more attentive an knowledgeable about politics are less likely to support new trade restrictions.
  • The confirm their hypothesis regarding home ownership.