Competition in the Chilean Banking Sector: A Cross-Country Comparison

Ronald Fischer:

Meral Karasulu provides a good, clear description of the Chilean banking sector and its imperfections. The paper lays out some of the potential reasons to believe that the sector may be characterized by imperfect competition. Among the most relevant are the stamp tax on credit operations, the concentration of wealth in vast conglomerates, the ownership of pension funds by the bank holding companies, and the high capital requirements for entering the sector (which were recently lowered). Two additional factors are also important: first, no new banks were allowed to enter for a long period after the crisis of the early 1980s, because the banking regulator apparently feared that excessive competition would reduce bank profitability and thus lead to excessive risk taking; second, the lack of a centralized guarantee system means that it can be very expensive for small companies and individuals to switch banks. There are thus good reasons to suspect a lack of strong competition, and the sector's profitability, measured as return on assets or on equity, is indeed among the highest in Karasulu's sample. This is so even though banking concentration is not high and new banks have entered the market since the regulator changed its restrictive policy in the late 1990s.

Karasulu uses Panzar and Rosse's methodology to study competition in the Chilean market over time, as well as to compare it with a sample of other emerging countries.¹ As Karasulu mentions, the traditional competitive analysis based on concentration ratios suffers from the fact that markets with intense competition can be concentrated since only a few firms can survive, while others that are cartelized may attract entry.² The Panzar-Rosse approach, on the other hand, makes very few assumptions and is based on the observation that different industrial organizational structures respond differently to changes in input prices. Competitive markets transmit input price changes into output prices better than uncompetitive markets. Under perfect long-run competition, the sum of the elasticities of revenue to input prices (the H statistic) equals one. The sum of these elasticities is negative, however, for monopolies or cartelized industries.³ Finally, in the case of monopolistic competition, the values of H lie in the range (0,1), with higher values of H representing more competition.

The Panzar-Rosse methodology has been used extensively to model the airline industry, as well as financial markets, and numerous papers use this methodology to study competition in the banking sectors of different countries. It is fairly easy to construct the data required to perform the analysis in the banking sector. The main drivers of bank costs are deposit costs (that is, the interest rate), labor costs, and the cost of fixed capital (which contributes marginally in most cases). Revenues, in turn, can be represented by either interest income or total revenue.

Karasulu's cross-country regressions indicate that most countries in the sample operate under monopolistic competition, with the exception of Australia, Greece, Hungary, Ireland, Israel, and the Netherlands, which have perfectly competitive markets. According to this regression, competition is imperfect in Chile, although it is stronger than in a few countries in her sample. The author ran a pooled regression with a dummy for Chile to determine whether the Chilean market was significantly less competitive than the average country in the sample; her results confirmed this hypothesis. Finally, Kara-sulu shows that the net interest margin in Chile is about 1.50 to 2.25 points higher than in other countries, which is consistent with the low level of the competition index (although it could mask, at least in part, unobservable differences in credit risk). The author has some suggestions as to what causes the lack of competition, with an emphasis on the industry's high capitalization requirements. As mentioned above, other options include the switching costs associated with the stamp tax and the lack of a centralized guarantee system.⁴

Marcela Meléndez:

This paper examines the conduct of Chilean banks using microeconomic data for the period 1995-2004. The author provides a rich characterization of the workings of the Chilean banking sector and includes a detailed discussion of the factors...