Prices, markups and trade reform

In our globalized economy, information about the costs, benefits, and distributional consequences of lowering trade barriers is essential to policymakers trying to decide if a particular agreement should be supported. This research fills an important gap in our knowledge concerning the effects of reducing trade barriers when firms have some degree of monopoly power.

The political debates over the Trans-Pacific Partnership (TPP), the Transatlantic Trade and Investment Partnership (TTIP), and issues involving international trade more generally have been contentious. Advocates cite the benefits of trade liberalization such as lower prices of goods to consumers, increased competition, better quality, and an increase in the variety of goods available in the marketplace. Opponents counter that there are reasons why opening markets may not be desirable. In the presence of market failures and economies of scale, the costs of trade liberalization may exceed the benefits depending on the institutional context.

Does lowering trade barriers benefit a country? To answer this question, we need a better understanding of the costs and benefits of reducing trade barriers in the presence of factors such as market power that can have a large impact on standard cost-benefit calculations. Our paper “Prices, Markups and Trade Reform” takes a step in this direction, and demonstrates that the presence of market power can affect the degree to which consumers benefit from opening markets to international trade.

The gains from reducing trade barriers typically come about through lower costs that are passed through to consumers in the form of lower prices. While this mechanism is unquestionably important, it ignores another channel featured prominently in the public debate. When firms operate in markets that are less than fully competitive, they are able to raise the price of goods above the cost of producing them, and this markup above costs allows the firms to earn excess profits. Reducing trade barriers exposes these firms to more competition forcing them to lower their markups and ultimately their prices. This second mechanism is the so-called “pro-competitive” effect of trade. It is completely absent in traditional models of trade that assume either perfect competition (in which case the price of a good equals its cost) or constant markups that are invariant to policy (the latter arise as a result of the standard modeling assumptions in trade models).

Markups actually increased as a result of the tariff reductions

A recent set of theoretical and quantitative papers have developed models with variable markups that allow for trade to have pro-competitive effects. However, empirical evidence on such effects is scant. Our paper fills this gap by examining how prices and markups in India responded to the large trade liberalization India implemented in the early 1990s. The answer is surprising: we find that markups actually increased as a result of the tariff reductions.

Prices still declined due to the cost reduction effect, but the prices facing consumers declined by much less than one would have expected in the absence of market power. These findings do not provide evidence against the pro-competitive effects of trade. Rather, they demonstrate that the pro-competitive effects are dominated by a second effect that arises in settings where firms have market power and their markups are variable: the incomplete pass-through of cost reductions to prices.

To understand these surprising findings, it is important to interpret them in the context of the particular policy under study. As in many developing countries, the Indian trade liberalization had two conceptually distinct impacts: First, it reduced trade barriers (primarily tariffs) on final products imported from other countries. Second, it reduced trade barriers on intermediate goods used by domestic firms as inputs in the production of final goods sold to Indian consumers. The primary effect of the reductions of final goods tariffs (output tariffs) was to intensify the competition faced by domestic firms as imports became cheaper. This represents the “pro-competitive” effect of trade.

In contrast, the effect of the reductions of tariffs on intermediate products (input tariffs) was to reduce the cost of producing domestic products. This reduction consists of a direct effect (lower cost of imported intermediate goods) as well as an indirect effect (spillover effects on the prices of domestically sourced intermediate products).

Importantly, the “pro-competitive” effect and the “cost-reduction” effects have different implications for markups and prices. In the first case, we would expect markups to decline as competition intensifies, leading to lower prices. In the second case however, it is unclear what happens to markups. Depending on the market fundamentals (demand, market structure, and firm behavior), markups could stay the same, go up, or go down. Prices in turn would only decrease in full proportion to the cost reductions if markups did not increase. But if markups increased, prices would decline by less than costs. This is exactly what we find: the cost reductions associated with the reductions in the tariffs of intermediate imported inputs are not completely passed through to prices. In fact, this effect dominates the pro-competitive effect of output tariff reductions so that on net, markups go up and price declines are modest.

Before discussing the interpretation and policy implications of these findings, a few words on the methodology used to arrive at these conclusions are in order. Calculating the markup requires data on both prices and costs. Prices are observed. However, firm costs are not fully observed since surveys used to construct the data do not report the cost of capital. This difficulty is well known from the Industrial Organization literature. We therefore develop a unified framework for obtaining marginal costs and markups that is based on estimating production functions. This approach allows us to estimate markups and marginal costs at the firm-product level. These are then related to the tariff reductions observed in India. Since India’s tariffs were passed quickly as part of a “shock therapy” with little debate and no time for the opponents to express their objections, our correlations have a causal interpretation.

These results lend support to the view that firms used the higher profits generated by the trade liberalization to finance innovation

Our estimates indicate that the trade liberalization led to significant cost decreases. However, cost pass-through is incomplete. We find that on average only a small proportion of the cost variation is passed through to prices. As a result, the price declines are small relative to the cost reductions. This does not mean that pro-competitive effects were absent. If costs are held constant, declines in output tariffs lead to a reduction in markups, and this decline is most pronounced for high-markup products. This implies that while markups did increase due to the cost reduction effect, they would have increased by even more if it had not been for the pro-competitive effects of trade.

What explains the incomplete pass-through? Our methodology does not pin down the mechanism(s) generating markup variability and incomplete pass-through. One way to think about our approach is that it reveals what happened as a result of the trade reforms based on relatively few assumptions that could inadvertently influence the results. But in order to shed light on the exact mechanisms behind the effects we document, our work needs to be complemented with the case-study approach used in the Industrial Organization literature. That is left for future research.

Importantly, however, the markup variability that is associated with incomplete pass-through does not necessarily indicate a change of firm behavior (for example the adoption of more collusive or anti-competitive practices). Markup variability can arise simply as a result of the market fundamentals (e.g., non-constant price elasticity of demand). Hence, the increase in markups does not necessarily translate to a call for a more stringent anti-trust policy.

The markup increases, which benefit firms, do not necessarily imply that consumers did not also benefit from the trade reforms. There are at least two channels through which consumers may have benefited in addition to the modest decreases in prices. First, it is likely that the reforms led to an increase in the quality of the products produced by Indian firms. That said, quality is costly, and we find that the marginal costs of production significantly declined as a result of the trade liberalization. But it is possible that the cost declines would have been even larger if quality had not improved. Second, firms that experienced the highest markup increases were also the most likely to introduce new products. Though this evidence is only suggestive, it lends support to the view that firms used the higher profits generated by the trade liberalization to finance innovation. To the extent that this is true, the total benefits to consumers from the trade reforms would exceed the benefits captured in our study.

Changes in tariffs or other trade policy instruments do not necessarily translate into proportional changes in prices as typically assumed in the literature

At any rate, the important insight for policy is that changes in tariffs or other trade policy instruments do not necessarily translate into proportional changes in prices as typically assumed. In the presence of less than fully competitive markets and variable markups, the response of prices is substantially more complex. This insight has implications for the aggregate gains from trade, their distribution across consumers and producers, and the channels through which consumers benefit from trade reforms.

Further reading

This columns summarizes “Prices, Markups and Trade Reform,” by Jan De Loecker (Princeton), Penny Goldberg (Yale), Amit Khandelwal (Columbia) and Nina Pavcnik (Columbia). The article is forthcoming in Econometrica.