2. Institutions and Trade Policy: A Review

There is an extensive literature on AD, its history and development. For a comprehensive review see for instance Nelson (2006) or Bloningen and Prusa (2003). Relevant characteristics of AD as well stylised facts will also be addressed in Chaps. 3 and 4 when focusing on the application of theoretical arguments to AD.

This has several obvious reasons: first, AD had only been used by a small group of countries for a long time, which obviously limited possibilities for cross-country analysis. This only changed relatively recently with AD use spreading and data on use being made available as a consequence. Second, scholars who study institutions and those who study AD often come from different disciplines, hence the development of separate literatures and the lack of connection between institutions, notably democratic rule, as an independent and AD as the dependent variable.

The respective legal definition refers to dumping as “(…) a product is to be considered as being dumped, i.e. introduced into the commerce of another country at less than its normal value, if the export price of the product exported from one country to another is less than the comparable price, in the ordinary course of trade, for the like product when destined for consumption in the exporting country.” See the Agreement on the implementation of Art.6 of the General Agreement on Tariffs and Trade 1994. Available at http://​www.​wto.​org/​english/​docs_​e/​legal_​e/​19-adp_​01_​e.​htm

Notably, delegation is an important issue in the literature on American trade politics. Members of the legislature represent narrower constituencies making them more prone for capture by special interests pushing for protection. Given their narrow constituency focus policymakers may engage in logrolling activities, i.e. they agree to barriers for other products in return for favours for their own constituency, resulting in higher trade barriers across the board—as illustrated by the infamous Smoot-Hawley tariffs (Oatley 2008). “Theories of Presidential liberalism” claim that the president tends to be more pro-trade because he represent a nationwide constituency and can better balance different interests. Hence, delegating powers to the president should help o foster openness. Looking beyond the US, the delegation of trade policy competences to the European Commission at an early stage of integration also reflects a similar logic.

Haggard and Webb (1993) similarly suggest that any autocratic advantage in terms of pursuing reforms essentially depends on leadership.

See Haggard and Webb (1993) and Haggard and Kaufmann (1995) for a more extensive discussion.

Scholars point out other factors that contributed to the success of liberalizing reforms for particular countries, e.g. American influence in the case of East Asia (Young 1996) or suggest counter examples such as Chile which also enacted tariff cuts under democratic rule (Henisz and Mansfield 2006) or Germany in the 1930s where autocracy went hand in hand with favouring autarky (Verdier 1998).

Ability, effectiveness and persistence of trade policy reforms then follow as a second step and require thorough theorising and empirical examination. See for instance the analyses by Giavazzi and Tabellini (2005) or Frye and Mansfield (2003).

Note that there is a debate in the literature as to what extent factor or sector-based models perform better to explain trade policy preferences. Factor mobility is important in this respect, with higher mobility of production factors across sectors, factor-based models provide a more accurate view and when mobility is low, sector-based models are better suited. Alternatively, the two approaches can be interpreted as short and long-term views. Apart from that, analyses of survey data also reveal that other noneconomic factors such as gender, social status or nationalism play a role in the formation of individual preferences about trade policy (Mayda and Rodrik 2005). Also, Mansfield and Mutz (2009) challenge the view that trade policy preferences are mainly based on income effects and stress the importance “societal effects” based on perceptions of how trade affects the entire country for the formation of preferences.

On the link between trade and growth see for example Sachs and Warner (1995), Dollar and Kraay (2001), and Wacziarg and Welch (2003). Wintrobe (1998) suggests that democracy helps to reduce rent-seeking. Verdier (1998) focuses on trade policy in industrialized countries in the nineteenth century and similarly argues that democratic leaders are less likely to use trade barriers to extract economic rents.

See Kono (2006) for an example.

While Aidt and Gassebner focus on reductions in bureaucratic inefficiencies, better protection of property rights—a characteristic typically associated with democracies—is another example with possible positive repercussions on trade. Better protection of property rights is conducive to FDI (Jensen 2003), which can help encourage trade as recent research suggests that both are complementary (Busse and Hefeker 2006).

Rodrik (1994, p. 69) for example makes a general claim based on historic observations: “Indeed, historically sharp changes in trade policy have almost always been preceded (or accompanied) by changes in the political regime. (…) Not all political transformation result in trade reform, but sharp changes in trade policy are typically the result of such transformations.” Verdier (1998) also starts from a historical approach examining trade liberalization in nineteenth century Europe.

Economic liberalisation is not limited to trade policy in this case. Rather, the authors start from a broader concept of reforms and then assess as part of their analysis to what extent policy reforms translate into changes in outcomes such as trade volumes.

The actual number of countries varies from 14 to 156 for the different regressions and increases over time.

See Milner and Kubota (2005, p. 120).

Gravity models which include geographic distance in their estimation solve this problem only to some extent because distance does not equal accessibility which may be promoted or hindered by geographical conditions (e.g. coastline, high mountains).

Kono is careful to demonstrate that higher NTB’s in democracies are not the result of policymakers being simply more responsive to public concerns, e.g. for higher product safety standards. He backs his argument by examining how the level of NTB’s is affected by several proxies measuring consumer demand for high quality regulation as well as interest group pressure and concludes that it is not consumer demand proxies but rather interest group pressure that explain NTB’s.

Additional evidence for this logic is provided by examining whether high tariffs prompt frequent challenges to protectionist policies in the domestic political arena. Kono finds that high tariffs are associated with parties more frequently denouncing protectionism in election manifestos.

For example, Bliss and Russett (1998) use a dummy to distinguish democratic dyads from all other categories and therefore differences between pairs of autocracies and mixed dyads cannot be discerned. Dixon and Moon (1993) exclusively look at trade relationships of the US with other countries and hence no inference about autocratic pairs is possible.

See McGillivray and Smith (2008, p. 121): […] “but if all else is equal on this dimension, the trader strongly prefers to trade with another large coalition system because of the stability of cooperative relations between large coalition systems. Traders in large coalition systems disproportionally trade with other traders from large coalition systems. This diverts trade by merchants in large coalition systems away from small coalition systems toward other large coalition systems.”

See however Dai (2002) for a critique of the model.

While this is a monadic argument, the authors argue that the probability to join a trade agreement for A is independent of B and therefore the probability of two democracies to cooperate is highest.

Tomz (2002) provides evidence for this logic examining the case of Argentina and the link between popular attitudes, debt repayment and electoral success.

Henisz and Mansfield (2006) similarly suggest that trade openness in democracies is more responsive to shocks and present empirical evidence for this but their argument is more general and not limited to the use of specific flexibility provisions.

It could be argued that trade disputes are “hybrids” in the sense that they include both elements of cooperation and conflict. On the one hand, they involve disagreement between nations. On the other, disagreement is addressed within a cooperative forum and subject to adjudication.

Given some recent evidence of China becoming a more active player in dispute settlement, this may change somewhat in the future.

See also Bown (2009a, p. 94): “Governments that feel the need to implement new import protection will try to do so against trading partners through the use of policies with the lowest costs of implementing such protection.”

As Smith (1999, p. 1256) notes: “Strategic choice is the explicit study of counterfactuals.”

While there are approaches to take particular types of interdependence into account, e.g. for cross-sectional time series data (Beck et al. 1998), the type of interdependence in strategic models is different as it directly arises from the structure of interaction between the two players.

While non-linearities can be modelled to some extent with interaction effects or squared terms, a simple logit or probit with an interaction effect would still not properly account for the interaction between two players, which is at the heart of a strategic model.

The game-theoretic solution concept with agent error is the Quantal Response Equilibrium (McKelvey and Palfrey 1995, 1998). Here, the distribution over outcomes basically comes from wrong decisions. With private information, players are observationally equivalent but hold preferences that are only privately known (Kenkel and Signorino 2014). For a discussion of the different sources of uncertainty also see Signorino (2003).

Another type of strategic interaction would be suggested by (anti)dumping models that focus on the activities of firms to engage in dumping practices. However, this is not the primary focus of this analysis which focuses on the decision-making processes associated with AD that involve government (re)actions.

Two factors that affect retaliatory capacity are the industry’s export share to the trading partner and whether the target country has an AD mechanism.

The authors use interaction terms to construct variables for retaliatory capacity. For the first stage, they use product specific export shares of industry to the potential target country and interact this with a dummy variable for having an AD law. The scope of the export component is broadened from product to sector level as part of robustness checks to capture the possibility that firms export multiple products. For the second stage, Bloningen and Bown interact US export shares to the target with several measures to proxy legal capacity (whether a country has brought a case against the US before, whether it has ever been a plaintiff and whether it is a GATT/WTO member).

Busch, Reinhardt and Raciborski examine US AD cases from 1978 to 2001 using a Heckman probit selection model. Note that both Blonigen and Bown (2003) and Busch, Reinhardt and Raciborski rely on hypothetic control cases, i.e. they try to identify a set of suppliers that could have been named in investigations but where no cases were ultimately initiated.

Their total sample includes 1,321 cases with 17 investigating countries and 33 targets. The sample for the second set consists of 849 measures.

They construct a measure to capture bureaucratic quality and litigation capacity based on survey data. Not combined with export shares.

For direct retaliatory capacity, Bown considers the value of US export shares at industry level to the target. The higher they are, the larger the target’s capability for retaliation.

Busch et al. (2009) propose the argument in reverse as they argue that dispute settlement and “the rule of law” can (better) deter protection against members lacking the ability to retaliate via other means.