Unprotective tariffs, ineffective liberalization, and other mysteries: an investigation of the endogenous dimensions of trade policy formation in Australia.

• Author: Tombazos, Christis G.

1. Introduction

   Since its formation as a federation in 1901, Australia embraced a fairly comprehensive tariff schedule in relation to imported manufactures. (1) Tariff rates fluctuated considerably over the years. However, over the course of the century they remained sufficiently high to earn Australian manufacturing the reputation of one of the most heavily protected manufacturing sectors in the industrialized world. (2) In recent years, a number of liberalization waves have surfaced. The most notable of these took place in 1973, 1977, 1988, 1991, and, most recently, in 1994, with the signing of the "Uruguay round" of the GATT agreement. Still, tariffs remain a principal actor in Australia's menu of industrial policies. In a recent edition of the annual report of the Productivity Commission, (3) published in November 1999, it is noted that although tariffs have dropped considerably since 1994 "they remain an important form of assistance ... tariff assistance still accounts for around 90% of measured effective assistance to the manufacturing sector" (4) (p. 52).

   Given the pervasiveness of Australian tariff policy, the significant welfare implications of changes in that policy and the particular importance of international trade for this country, ensuing from its idiomorphic cultural, resource-endowment, and geographic characteristics, (5) considerable research effort has been invested in the examination of various dimensions of this issue. At the forefront of pertinent empirical research, two items have attracted considerable interest: (1) the host of determinants of the highly dispersed tariff concessions and (2) the impact of liberalization on imports. Research that falls in the category that examines the former, including Anderson (1980), Conybeare (1984), Aislabie (1988), and Feaver and Wilson (1998), is consistent, at varying degrees, with the theory of endogenous trade policy formation. (6) According to the foundations of this theory, tariff concessions may be viewed as "public outcomes" that facilitate a political equilibrium. The mechanics are rather elementa ry. When the profit loss that arises from import competition exceeds the transaction cost of a successful lobbying campaign for protection, private interests will undertake such lobbying activities. The higher the loss from foreign competition, the more fervent the lobbying effort. Hence, an increase in import penetration is likely to trigger greater protection (7) (Trefler 1993, p. 139). Unlike this line of inquiry, the critical causal relationship of interest to studies that fall in the category that investigates item (2), including Simmons and Smith (1994), relates to the effects of tariff changes on imports, rather than the reverse.

   The causal directions of the nature examined by studies that investigate issues (1) and (2) are, of course, not mutually exclusive. Higher import penetration increases tariffs, and higher tariffs, in turn, decrease import penetration. However, despite the simultaneity of these variables, studies of Australia's trade policies associated with research lines (1) and (2) followed separate paths by only considering unidirectional causal relationships. (8) And yet, these paths do converge. The meeting point: perplexing results! On the one hand, studies of endogenous protection, such as Anderson (1980) and Aislabie (1988), have consistently failed to establish the expected result that an increase in import penetration will lead to higher tariffs. On the other, studies of trade liberalization often find that tariff reductions have a surprisingly small effect on imports, with a study by Simmons and Smith (1994) concluding that, at least in relation to the examined industry,"... removal of the tariff would have no effe ct on the level of imports..." (p.57). It is important to note that analogous, and equally puzzling, findings pertaining to improbably small effects of liberalization on imports have been identified in the case of other countries, such as the United States, by a number of authors. In a recent article, Trefier (1993)argued that the small magnitudes of related estimates in the case of the United States derive from methodologies that ignore the political economy of trade policy formation. This author showed that, when the level of protection is modeled endogenously, the relevant estimates are altered considerably.

   In the context of the existing pertinent literature, the present article constitutes the first attempt to model the determination of imports and tariffs relevant to the Australian manufacturing sector simultaneously. This exercise has three distinct objectives: First, implementation of the proposed methodology will shed light on relevant dimensions of endogenously determined tariff concessions while independently accounting for important "feedback" effects that have been ignored by similar studies. Second, extending Trefler's (1993) analysis, my model will explore the degree to which the scale of misspecification inherent in previous studies that investigate research line (2) may have been responsible for their unsettling results. Last, and most important, the proposed framework of analysis will augment the scope of inquiry adopted by previous contributions in this area by investigating whether findings regarding the impact of import penetration on protection that are derived using single equation techniques, and which are incongruous with the predictions of endogenous trade theory, may be appropriately rehabilitated in a simultaneous equation setting.

   The remainder of the article is organized as follows. The model is examined in the next section. In section 3, issues of econometric implementation and the empirical results are discussed. Concluding remarks are reserved for section 4.

2. The Model

   Single-equation regression techniques "entangle" reciprocal causal flows that link trade barriers with trade volumes. To extricate distinct dimensions of this relationship, Ray (1981b) proposed a simultaneous equation approach, which was subsequendy adopted by Trefler (1993) and Lee and Swagel (1997). The first two studies explored endogenous trade policy formation in the United States, whereas the latter relied on cross-country observations. However, all three implementations of this analytical framework used limited-depended-variable formulations to investigate the joint determination of nontariff barriers (NTBs) and imports. (9)

   In the present article, I introduce a continuous-variable interpretation of the analytical framework pioneered by Ray (1981b) to investigate the simultaneous determination of import penetration and tariffs in Australia. The system of equations that are jointly estimated is given by

   T = [[alpha].sub.T] + [[beta].sub.M] X M + [c'.sub.T] X [X.sub.T] + [[epsilon].sub.T] (1)

   M = [[alpha].sub.M] + [[beta].sub.T] X T + [c'.sub.M] X [X.sub.M] + [[epsilon].sub.M]. (2)

   These equations represent a cross-sectional characterization of the Australian manufacturing sector. T denotes the nominal tariff that corresponds to the level of protection granted to an industry that faces a level of import penetration given by M; [X.sub.T] and [X.sub.M] represent vectors of industry characteristics that determine tariff concessions and the demand for imports, respectively; and a' = [[alpha].sub.T], [[alpha].sub.M]] and e' [[epsilon].sub.T], [[epsilon].sub.M] represent vectors of constants and residuals, respectively.

   Similar to other studies of the political economy of trade-barrier concessions, including the most seminal contributions in this area, such as Pincus (1975), Caves (1976), Anderson (1980), Ray (1981a,b, 1987), Trefler (1993), and Lee and Swagel (1997), my model does not incorporate the desired level of rigor often evident in behavioral functions that are explicitly derived from fully specified optimization frameworks. In an effort to, at least partly, ameliorate this deficiency inherent to the relevant area of study, the model is subjected to sensitivity analysis. The relevant details are discussed in section 3. The specification of Equations (1) and (2) is examined below.

   The Tariff Equation

   Discretionary Tariffs and the GAIT

   The dependent variable of the tariff equation represents the unweighted average nominal tariff rate (10) that prevailed during fiscal year 1990/199 1 in the case of each of the examined industries. (11) Clearly, the dispersion in cross-industry tariff concessions may be viewed as the product of political process only if policymakers have discretion over the relevant tariff rates. In this context, it should be noted that in 1991 tariffs were not constrained by the GATT in any significant way. Although, among the 23 original signatories to the 1948 GATT, Australia did not take part in the GATT negotiations of the 1950s and did not fully participate in the Kennedy Round of 1964-1967 or the Tokyo Round of 1974-4979. As Capling and Galligan (1992) explained, Australia's "in-principle" support for the GATT was "eventually tempered by domestic considerations" (12) (p. 106). Indeed, it is widely acknowledged that, whatever the reasons for joining the GATT in the first place, the ultimate determinant of Australia's in herently incongruous position of a nonparticipatory membership was a vociferous commitment to high levels of industrial protection. (13) "... The Australian government was, quite simply, not prepared to make the sacrifice of autonomy with respect to tariff making that goes with tariff commitments" (Snape, Gropp, and Luttrell 1998, p. 365). (14)

   Still, to the extent feasible by its political economy, Australia did make a handful of pre-1991 concessions to its trading partners under the GAIT. Table 1 outlines a host of rather generous measures of the coverage and incidence of these concessions. The first column reports the total value of imports "corresponding" to each of the industries examined in the present study, and columns 2-9 outline the portion of...