TRADEPOWER Project
February 11th, 2022


This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 724107). We are very grateful to the many respondents for their support for this research.


INTRODUCTION

Between spring 2021 and spring 2022, as part of the TRADEPOWER project, we carried out a survey of legislators on their attitudes on trade policy. The key aim of the TRADEPOWER project is to shed light on how global value chains impact on trade policy and trade politics. Next to the survey, we use trade data, stock market data, and case studies to do so. This survey allows us to get insights about one of the most relevant actors’ motivations in terms of trade policy: legislators.

Throughout the survey, by “trade agreement” we understand agreements signed by countries to facilitate trade and investments across borders. Important recent examples for such agreements are the Comprehensive Economic and Trade Agreement (CETA) between the European Union and Canada, the United States-Mexico-Canada Agreement (USMCA), or the China-Australia Free Trade Agreement (ChAFTA).

In this report, we first offer descriptive information on the respondents to the survey and then present a series of key findings.


THE RESPONDENTS

We decided to survey legislators because a survey allows us to understand the views of various legislators in different contexts. To this end, we invited legislators in national parliaments in democracies across the world to participate in a short, 5-minute long survey. We did not apply any selection criterion in terms of interest in trade policy or similar. Rather, we allowed all legislators from a country for which we could find contact information to participate. Until now, we received 1028 responses from legislators in the following 47 countries:

The average respondent is 53 years old. The figure below shows the distribution of legislators’ age. Our data thus contains legislators across all age groups.

33 percent of responses are from female legislators. This distribution, which is visible in the figure below, reflects the lower representation of women in the parliaments that we studied.

The sample also includes variation in terms of the political ideology of respondents as measured by the economic position of their parties on a scale that ranges from 0 to 10 (provided by the Global Party Survey 2020). That we have nearly the same number of respondents left and right of the center indicates that our sampling procedure or non-responses do not introduce a bias in terms of ideological predisposition of the respondents.


THE IMPORTANCE OF TRADE POLICY

Going beyond these respondent characteristics, we were first interested in understanding how important trade policy is for legislators relative to other policy fields. Thus we asked them In your function as member [of your chamber of parliament], how important is trade policy for you in comparison with other topics on the agenda of parliament? Responses were recorded on a scale from 0 = “Not important at all” to 10 = “Extremely important”. Our survey responses suggest that — overall — legislators view trade policy as an important part of their work in parliament. The figure below shows the distribution of responses.


WHOM DO LEGISLATORS REPRESENT IN TRADE POLICY?

Our survey included a total of four experiments to better understand whom legislators represent in trade policy. For these experiments, we randomly selected specific question wordings to see whether and how legislators’ trade attitudes respond to different frames. From this, we can learn which constituencies are foremost in legislators’ minds when thinking about trade policy.


Are legislators more responsive to business associations or the population?

The first experiment scrutinises whether legislators are more concerned with business associations or the population when forming their attitudes towards trade agreements. Specifically, the experiment has two different treatment conditions. Either legislators heard that business associations supported the agreement but there is public opposition or that the public is supportive but business associations oppose the agreement. These treatments were embedded in the following question (here and below in square brackets the parts of the question wording that vary depending on country or treatment): Imagine [country]/[the European Union]1 negotiates a new trade agreement with an important trading partner. [Business associations support this agreement, but there is opposition to it within the population.]/[Business associations oppose this agreement, but there is support for it within the population.] How much would you support or oppose this trade agreement? 2 The figure below shows the distribution by treatment. We observe that legislators are substantially less supportive of trade agreements if the population is sceptical despite business support (blue bars). The mean support in this treatment group is 5.1. This is in sharp contrast to the other treatment group. If the population supports the agreement (orange bars), the average support on the 11-point scale is 6.5. A student t-test confirms the significance of this difference.3


How to respond to adverse effects of trade policy?

Moving on to the second experiment, we were interested in legislators’ choice of policy instruments to address the potentially adverse effects of trade liberalisation, e.g. job losses in some sectors or firms going bankrupt. We particularly investigate whether the policy choice differs depending on who loses, namely businesses or workers. Hence, we asked the following question: Trade agreements often lead to losses for [businesses]/[workers] via increased international competition. To help these [businesses]/[workers], please tell us how much you would support or oppose the following measures.

We specifically asked legislators about their support for the following four policy instruments:

  • Direct []/[wage]4 subsidies
  • Higher tariffs on imports
  • Tax cuts
  • Infrastructure projects

For this question, we used a five-point response scale from “Would completely oppose (0)” to “Would completely support (4)”.

Starting with subsidies, overall support for this policy instrument is fairly high. 54 percent of respondents would partly or completely support it. As regards the experimental design, the figure below suggests that the treatment – namely whether workers or businesses are mentioned as losers – only induced mild differences in legislators’ support for subsidies as a compensation of trade losses. A student t-test confirms the visual inspection and suggests that both groups do not differ in a statistically significant way.5

A second potential policy instrument are tariffs, which would allow for the protection of industries harmed by trade liberalisation and the associated foreign competition. The overall support for this policy instrument is mixed. Only 43 percent of respondents would partly or completely support it. Similar to the instrument on subsidies, the figure below suggests that whether workers or businesses lose hardly matters for legislators’ support for tariffs. A student t-test shows that indeed the difference between both treatment groups is not statistically significant.6

The third potential policy instrument are tax cuts. Such a policy measure would allow legislators to ease the pressure for industries and make them more competitive facing international competition. The overall support for this policy instrument is high. 64 percent of respondents would partly or completely support this measure. Again, however, we find only mild differences in legislators’ support for tax cuts when we either tell them that businesses or workers lose. These differences are not statistically significant at conventional levels.7

The final policy response outlined in our survey are infrastructure investments. Such a policy measure would allow legislators to ease the pressure for industries and create employment in at least some sectors of the economy. The overall support for this policy instrument is very high. 86 percent of respondents would partly or completely support this measure, and there is only little opposition to this policy instrument. Again, we find no statistically significant differences across the two treatment groups.8

To sum up, this question suggests that different policy instruments to deal with potential adverse effects of trade liberalization receive different levels of support. The treatments that directed legislators to think about businesses or workers, however, hardly induced any differences.


Does it matter who wins and who loses from trade?

In the next experiment, we scrutinise whether support for trade agreements depends on legislators’ perceptions of who gains and who loses from trade. Hence, we asked legislators about their support for a trade agreement with an important trading partner after telling them that either workers or firms would benefit or lose from this agreement. Specifically, either 70% (30%) of workers/firms would benefit (lose).9 Again, we recorded responses on an 11-point scale.

The results suggest strong treatment effects. Legislators are unlikely to support agreements that are framed in terms of losses. Not fewer than 81 percent of legislators oppose a trade agreement if workers lose. Similarly, 61 percent oppose trade agreements in case that firms lose. In contrast, legislators are supportive of trade agreements if workers benefit (85 percent support it). We observe similarly large effects for firms, with 86 percent of respondents supporting an agreement if firms benefit. Overall, this evidence indicates that legislators are slightly more sensitive to losses from trade agreements for workers.

In a final experiment, we referred to both winners and losers in the same treatment, to have even more direct insights into whether legislators respond differently to gains or losses of specific actors. Concretely, we asked respondents: Trade agreements tend to create economic winners and losers. Concretely, a trade agreement with India10 could [increase profits of firms exporting to India but lead to lower wages for workers in [country]]/[lead to higher wages for workers in [country] but reduce profits of firms from [country] that compete with imports from India.]11 We then captured support for this agreement on an 11-point scale from “strongly oppose” (0) to “strongly support” (10).

The findings shown in the figure above suggest that the treatment induces strong differences. By and large, respondents are substantially more likely to support agreements that benefit workers even if they harm firms, than vice-versa. While 59 percent oppose an agreement when workers lose and firms gain (and only 25 percent support it), a staggering 59 percent support the agreement if workers are said to gain and firms to lose (with 18 percent opposing it). These findings suggest that legislators tend to consider workers’ wages substantially more than firms’ profits.


LEGISLATORS AS DELEGATES OR TRUSTEES IN TRADE POLICY?

We concluded the survey by asking legislators about their views on democratic representation to find out whether they see themselves mainly as delegates (namely bound by the desires of their constituents) or trustees (namely with an emphasis on their free mandate, not necessarily following the explicit desires of their constituents) in trade policy. Specifically, we asked respondents to state their agreement or disagreement with the following six items:

The responses indicate that legislators predominantly see themselves as trustees at least as far as trade policy-making is concerned (see the table below). Illustratively, 86 percent of legislators agree or strongly agree with the statement that parliamentarians should act in the country’s interest even if unpopular. Similarly, a clear majority (53%) disagree or strongly disagree with the statement that legislators should act in accordance with public opinion, even if they themselves disagree. Clearly, this demonstrates that respondents in our sample see a strong notion of guardianship that legislators should follow.

This does not mean that they object to the public voicing its view via demonstrations or letter writing campaigns. In fact, most respondents (549, that is 65 percent) believe that demonstrations and letter writing campaigns are legitimate ways to influence parliamentarians on trade policy. Interestingly, legislators are more split when it come to the question of whether trade agreements are too technical for citizens or not. While 49 percent (strongly) agree, 35 (strongly) disagree with this statement. Legislators are also concerned about the (negative) influence of populists and economic elites on trade policy. Overall, however, it seems fair to conclude that the trustee model of representation dominates over the delegate model among our respondents.

Question Agree Disagree
Demonstrations and letter writing campaigns are legitimate ways to influence the decisions of parliamentarians. 65 12
The influence of economic elites on trade negotiations is too big. 68 12
Populists exploit the lack of knowledge of ordinary citizens about trade agreements. 74 14
In a representative democracy, parliamentarians should act according to public opinion even it is at odds with their own opinion. 24 53
Trade agreements are so technical in nature that ordinary citizens cannot comprehend them. 49 35
In a representative democracy, parliamentarians should act in the interest of the country even if that means making unpopular decisions. 86 6


SUMMARY

The survey at hand provided novel insights into how legislators think about trade policy and trade agreements. Our respondents consider trade to be fairly important. The experiments included in the survey demonstrate that legislators tend to assign more weight to citizens and workers when forming an opinion on trade agreements than to business interests. Finally, we showed that the legislators that responded to our survey mainly view themselves as trustees rather than as delegates in trade policy. These findings are interesting and important as they allow us to get a better understanding of how legislators form trade attitudes at a time when trade policy is increasingly contested across many countries of the world.



  1. For EU member states, the European Unions negotiates all trade agreements. Thus we substituted the country’s name with the European Union for EU member states.↩︎

  2. The outcome was recorded on a scale from 0 = “Would strongly oppose” to 10 = “Would strongly support”.↩︎

  3. t(884) = 8.5, p < 0.001.↩︎

  4. Wages were only asked in the workers’ treatment,↩︎

  5. t(924) = -0.6, p = 0.57.↩︎

  6. t(932) = 1.5, p = 0.14.↩︎

  7. t(932) = -1.8, p = 0.08.↩︎

  8. t(932) = -0.8, p = 0.42.↩︎

  9. One concrete wording was: Imagine that [country]/[the European Union] negotiates a trade agreement with an important trading partner. 70 percent of workers in [country] would benefit from this agreement via higher wages. How much would you support or oppose this trade agreement?↩︎

  10. For India, Bangladesh and Malaysia, we substituted India with the European Union as partner in the trade agreement.↩︎

  11. Additionally, we reordered the treatments to start with losses rather than gains. For this report, we collapse both groups since the information is the same.↩︎