South Korea

Passenger Cars

Buses

Trucks

Hyundai

39.1%

58.3%

50.3%

Kia

16.5%

15.9%

44.3%

Daewoo

36.8%

25.8%

4.6%

Samsung

7.3%

0.0%

0.5%

Others

0.0%

0.0%

0.2%

Imports

0.3%

0.0%

0.1%

Sum

(Total units sold in
Korea)

100.0%

(570,056)

100.0%

(75,185)

100.0%

(136,739)

Post-merger HHI

4,499

6,171

8,970

HHI increase

1,290

1,854

4,457

Source: KFTC Decision 99-43, ‘Concerning Violation by Hyundai Automobile of Article on Combination of Enterprises’ (April 1999). The market shares are based on number of cars sold, not revenue.

the passenger car market and the truck market was virtually a merger to monopoly (with the combined Hyundai-Kia controlling close to 95 per cent of market share).

The KFTC determined that the merger was anti-competitive on three grounds. First, as shown above, the merger increased the concentration in the relevant markets to very high levels. Second, new entry was difficult because massive investments were required but the Korean car industry already suffered from excess capacity (with capacity utilisation of only 47.3 per cent in 1998). Third, actual foreign competition in the Korean market was minimal (with foreign firms controlling only 0.3 per cent for passenger cars in 1998).[25]

Nonetheless, the KFTC approved the merger with minimal conditions for five reasons. First, Kia was bankrupt and there was no willing buyer in the first two rounds of international auctions for Kia. As a result, Kia’s assets would exit the market unless Hyundai took them over. Second, Daewoo had made significant inroads since 1995 (with shares jumping from 20.5 per cent to 36.8 per cent from 1995 to 1998 in the passenger car markets) and as a result, Hyundai/Kia’s combined share had dropped significantly (from 74 per cent in 1995 to 56 per cent in 1998) in the passenger car market. According to the KFTC, the emergence of a strong domestic competitor had a significant potential to limit the exercise of market power by the combined Hyundai/Kia.[26] Third, imports from Japan were scheduled to be liberalised from June 1999, which were expected to further constrain Hyundai/Kia’s market power.[27] Fourth, foreign car markets were very competitive and Hyundai/Kia exported more than half of its cars. (The proportions of Korean automobiles exported were 68 per cent for passenger cars, 54 per cent for buses and 25 per cent for trucks). The KFTC claimed that the vibrant competition in the foreign markets would restrict price increases in the domestic markets.[28] Fifth, the combined Hyundai/Kia would have a total capacity of 2.6 million cars and thus would be in a position to realise economies of scale. Automobile industry participants often claimed that a firm needed at least a capacity of two million cars in order to reach a minimum efficient scale. Before the merger, Hyundai had five automobile platforms[29] and Kia had eight, but the KFTC Decision stated that the combined firm could integrate them and use common parts, streamlining both future model developments and part procurement processes.

The KFTC ruled that the efficiency effects from the merger outweighed the anti-competitive effects in the passenger car and bus markets.[30] However, given the almost monopoly position of the combined Hyundai/Kia and the relatively low export proportion in the truck market, the KFTC determined that anti-competitive effects dominated efficiency effects in that market. As a result, the KFTC did not impose any conditions in the passenger car or bus market, but ordered the combined Hyundai/Kia not to raise domestic truck prices at higher rates than export prices for the next three years. Notably, the KFTC did not order a partial divestiture (the sale of Hyundai’s or Kia’s production facilities for trucks to third parties), probably because at the time it made its decision (March 1999), Korea had not still recovered from the severe recession of 1998 and Kia desperately needed a quick infusion of capital and thus could not wait for a buyer of its truck capacity.

B. SK Telecom-Shinsegi Telecom Merger (2000)[31]

There were five mobile telephone network operators in Korea in 1999. SK Telecom was the former state monopoly using cellular technology from April 1984 and Shinsegi began operation as the second cellular licensee in April 1996. In October 1997, three Personal Communication System (PCS) licensees began providing services, bringing full competition to the Korean mobile telephony market. Fierce competition among the five carriers for recruiting subscribers (both through massive investments in networks and through heavy promotions including handset subsidies) in the explosively growing market[32] resulted in heavy losses for the new entrants. By the end of 1999, Shinsegi had suffered a cumulative loss of 479 billion KRW (roughly 479 million US dollars, assuming an exchange rate of 1,000 KRW for one US dollar) and the three PCS carriers each had lost from 286 billion to 539 billion KRW.

SK Telecom, on the other hand, was profitable with accumulated profits of 1.39 trillion KRW. Three things set SK Telecom apart from its competitors:

(i) It had a high-paying business customer base.[33]

(ii) It had already made the bulk of its network investments (while the newcomers had to make huge investments building up networks).[34]

(iii) The spectrum it used (800MHz for cellular technology) was more cost efficient in call transmission so as to require only 1/3 to 1/2 base stations than the PCS spectrums bands (1.8 MHz) do for comparable call quality.[35]

In 1999, Korea was preparing for the assignment of 2MHz band spectrum for the secondgeneration (‘2G’ or ‘IMT-2000’) digital mobile telephony services, which would require even more massive investments than the first generation (‘1G’) analog services. Given the huge losses that the new entrants had suffered (for example, Shinsegi had lost more than half of its total capital of 800 billion KRW) and the resulting high debt burdens (Shinsegi’s debt ratio was 575 per cent at the end of 1999 and 9.8 per cent of its total revenue was used to pay interests),[36] it was widely believed that not all of them could raise new capital to finance investments to provide 2G services on a full scale.

Hence, the environment was ripe for industry consolidation. The question was what form it would take: would the new entrants merge among themselves or would SK Telecom, the number one carrier, take over some of the weaker competitors? It was SK Telecom which made the first move. In December 1999, SK Telecom announced that it would acquire 51.2 per cent of Shinsegi’s stocks from its two major shareholders (Posco and Kolon) for cash payments of 1.1 trillion KRW and 6.5 per cent of SK Telecom’s new shares.[37] This five-to-four merger would result in a post-merger HHI of 3,882 with an increase of 1,213, with the combined firm controlling 56.9 per cent of subscribers, as shown in Table 2 below.

The KFTC identified three anti-competitive concerns (the first two of which are, in our view, legitimate, but the third is not). First, because SK Telecom’s price was regulated by the Ministry of Information and Communications, it might not be able to raise prices directly. However, the loss of competition between SK Telecom and Shinsegi could still

Table 2: Market Shares and Concentration Ratios in the Korean Mobile Telephony Industry, 1999

Source: KFTC Decision 2000-76, ‘Concerning Violation by SK Telecom of Article on Combination of Enterprises’ (May 2000). The market shares are based on the numbers of subscribers. SK Telecom’s and Shinsegi’s market shares in terms of revenue are higher: 46.5 and 14.5 per cent respectively (in 1999). Hence, revenue-based calculations would result in even higher post-merger HHI (4,252) and increase in HHI (1,349) than reported in the table.

harm consumers by delaying price cutting, by reducing the quality of current services or by delaying the launch of new services. Second, the combined firm would have a monopsonistic power over the purchase of cellular handsets and SK Telecom might favour its own handset manufacturer over competitors.[38] Third, because SK Telecom had advantages over competitors in terms of financial situation, distribution network, research and development (R&D) capabilities and the cost efficiency of spectrum, the combined firm might be able to attract more subscribers at the expense of the other three competitors.

The KFTC accepted some of SK Telecom/Shinegi’s arguments on efficiencies (namely, cost savings from integrating the two cellular networks and sales forces; price reduction/volume increase for handset purchases (from the supra-competitive levels); and realising R&D synergies) but decided that anti-competitive effects dominated the pro-competitive effects. The KFTC imposed two remedies. First, the combined firm was ordered to limit its purchase of cellular handsets from its affiliate to 1.2 million units per year from 2000 to 2005. Second, it was further ordered to reduce its market share in terms of subscribers to below 50 per cent by June 2001.

The first remedy and the theory of anti-competitiveness that underlies it (monopsonistic power over cellular handsets) appear to be well grounded in economic theory and reasonable. However, the second remedy and the liability theory (a form of ‘efficiency offence’) are hard to justify. In particular, the limit on market shares violates the core principle of competition law: the combined firm was ordered to reduce its marketing campaign and efforts to recruit new subscribers. This remedy seems to be motivated by the fact that the Korean law has explicit clauses on the presumption of anti-competitive merger: if the combined firm’s market share exceeds 50 per cent of the total market and is above 1.33 times of the second largest firm’s share, that merger is presumed to be anti-competitive (MRFTA articles 4 and 7). Ordering the combined firm to reduce its market share below 50 per cent by June 2001 creates the temporary appearance that the presumption clause of the law was not violated. However, even from the perspective of evaluating the effectiveness of the remedy (assuming that it was a right remedy), it had no lasting effect: the combined firm was simply ordered to reduce its market share below 50 per cent by a certain point in time, but with no follow-up measures. By May 2002, the combined SK Telecom/Shinsegi’s market share had climbed back to 53.4 per cent.

C. Muhak-Daesun Merger (2003)[39]

Traditionally, each Korean province has had a local soju (Korean hard liquor) brand. There is also a national brand called Jinro. In 2002, Muhak, a leading soju producer in Gyeongnam province launched a hostile takeover of Daesun, a rival soju manufacturer in the neighbouring metropolitan city of Busan by acquiring 41 per cent of Daesun’s stocks. Table 3 below summarises market shares in Gyeongnam, Busan and Korea as a whole.

On the surface, it was clear that this merger would create anti-competitive effects in the two regions. Daesun sold its soju only in Busan and Gyeongnam and Muhak’s revenue from soju sales outside of Busan and Gyeongnam accounted for only 1.1 per cent of its total revenue. Hence, both in the Busan and Gyeongnam regions, a very dominant firm (with market share of about 85 per cent) would merge with one of the two significant competitors. In each region, the combined firm would face Jinro only, with about three to seven per cent of market share. One would normally expect such a loss of competition to generate substantial anti-competitive effects. However, matters were more complicated due to the unique history of the soju industry and government tax regulations. First, because of tax regulations, Jinro had to set its prices uniformly across Korea. As a result, contrary to the KFTC’s simplistic assertion that this three-to-two merger would increase the risk of collusion in Busan/Gyeongnam regions, localised coordinated effects between the combined firm and Jinro were not feasible. (In other words, Jinro could not raise its prices in Busan and Gyeongnam only in explicit/tacit coordination with the combined firm.) Rather, coordination between the Muhak/Daesun and Jinro would require Jinro to raise it prices

Table 3: Market Shares and Concentration Ratios in the Korean Soju Industry, 2001

Source: KFTC Decision 2003-27, ‘Concerning Violation by Muhak on Combination of Enterprises’ (January 2003).

in all provinces. However, in each province, it faced a different competitor (a total of seven in addition to Muhak/Daesun). Since sales in the Busan/Gyeongnam regions accounted for only 1.8 per cent of Jinro’s total revenue, unless the other soju producers also participated, it did not make sense for Jinro to raise prices on a national scale in coordination with the combined firm’s price increases in the Busan/Gyeongnam provinces. In sum, it was not clear if the usual coordinated effects (between Jinro and the merged firm in the Busan/Gyeongnam regions) would materialise as a result of Muhak-Daesun merger.

Second, unilateral effects (ie the anti-competitive effects arising from the loss of competition between Muhak and Daesun) were certainly a serious concern. The key question here was if Muhak/Daesun would be able to raise its profits by unilaterally increasing their prices, holding Jinro’s prices fixed. (Notice that the above analysis casts doubt on the possibility that Jinro would raise its price in coordination with Muhak/Daesun.) The KFTC simply stated that Muhak and Daesun had competed head to head for a long time and the loss of competition between the two could not be offset by Jinro, as there was a strong preference for local brands in the two regions. The KFTC was clearly right about the loss of head-to-head competition between Muhak and Daesun. However, its observation on Jinro’s potential to rein in the combined firm was conclusory and was not based on any real analysis. A closer look reveals that Jinro’s competitive threats to Muhak/Daesun could not be dismissed, because as the following data show, Jinro’s market share was as high as 40 per cent in the two regions in the recent past.

Furthermore, in the region of Jeonnam, when the local brand (Bohae) pursued a high-price policy from 1997, Jinro had more than tripled its market share (from seven to over 25 per cent), as shown in Table 5 below.

Table 4: Market Shares in Busan and Gyeongnam Regions, 1993–2001

Source: KFTC; shares based on volume of sales.

Table 5: Market Shares in Jeonnam Region, 1993–2002

Source: KFTC; shares based on volume of sales.

Hence, without a more rigorous analysis, one could not say that the merged firm would be able to engage in profitable unilateral price increases in the Busan/Gyeongnam regions. Such analyses took place in the appeals process in the form of a geographical market definition exercise. Daesun, whose management resisted the hostile takeover attempt by Muhak, submitted an economic analysis report to the Seoul High Court that defended the KFTC’s position that the relevant geographical markets were the Busan and Gyeongnam regions.[40] This study was based on a survey of soju consumers in the two regions and explicitly applied the critical loss analysis technique for market definition.[41]

According to Daesun’s economic report, the combined firm could increase its profits by raising their prices by 10, 15 and ‘more than 20 per cent’ (but not by five per cent). Muhak’s economic experts[42] criticised several aspects of Daesun’s economic analysis. The most important one was about the classification of the two firms’ costs (based on accounting data) into fixed and variable costs. Muhak’s experts argued that the time horizon for determining if a cost is variable should be one to two years under the hypothetical monopolist’s small but significant and non-transitory increase in price (SSNIP) test, which the critical loss analysis technique operationalises by using data on consumer switching (‘actual loss’) and accounting profit data (‘critical loss’). Under the one to two-year criterion, several costs that Daesun’s experts classified as variable should be re-classified as fixed costs.

In particular, while Daesun’s experts classified all labour costs as variable, Muhak’s experts argued that salaries for regular workers (as opposed to temporary workers who could be fired when their contracts expired) were better classified as fixed costs in Korea under the one to two-year time horizon of the SSNIP test, because under the Korean labour law, it is impossible to lay off workers in Korea unless the company faces the dangerous prospects of bankruptcy. After such re-classification, no price increase was profitable for the combined firm. Daesun’s experts countered that the time horizon under the critical loss analysis should be ‘a sufficiently lengthy period of time during which a hypothetical monopolist is expected to exercise its monopoly power’.[43] Such a claim is not based on the hypothetical monopolist test, which is operationalised to be over a one to two-year time horizon. However, the Seoul High Court accepted Daesun’s position (without providing any analysis) and upheld the KFTC’s ban of the merger.

Setting aside these differences in expert opinions about the merged firm’s ability to engage in profitable unilateral price increases in the Busan/Gyeongnam regions, the KFTC’s structural remedy (prohibition of the merger) is the right one if one accepts the KFTC’s view that the loss of head to head competition between Muhak and Daesun would result in significant anti-competitive harm in the two provinces.

D. DCC-CCC Merger (2006)[44]

Carbon black is mainly used as a reinforcing filler in rubber products (such as tyres) and for special purposes such as laser printer/photocopier toner.[45] In 2006, DC Chemicals (DCC), a Korean producer of carbon black, acquired 100 per cent of stocks of Columbian Chemicals Company (CCC), a global carbon black firm with operations in nine countries (Korea, United States, Spain, Hungary, Italy, United Kingdom, Canada, Brazil and Germany). DCC’s capacity in 2004 was about 0.2 million tons per year and CCC’s global capacity was 1.1 million tons. CCC’s Korean subsidiary (Columbian Chemicals Korea (CCK)) had a capacity of 0.1 million tons. Hence, in terms of size, the Korean portion (CCK) was about 1/10 of the total transaction.

The KFTC defined two separate relevant product markets, one each for the two types of carbon black (‘carbon black used to reinforce rubber products’ and ‘specialty type carbon black’). The KFTC also defined the relevant geographical market as Korea. There was little controversy on the matter of product market definition. Not only was there little demand substitutability between the two uses of carbon black, but there was poor supply substitution between the two types of carbon black, because producing the high valued-added specialty type required sophisticated technology and equipment.

Geographical market definition was somewhat more complicated. Carbon black is bulky and thus costly to transport (especially over the land). Hence, the volume of internationally traded carbon blacks for reinforcing rubber was not high (except between neighbouring nations with short sea transport). For example, Korea imported only about 1.3 thousand tons of carbon black for reinforcing rubber products, less than 0.4 per cent of Korea’s total usage in 2004. While imports (especially from China) were rapidly increasing at the time (for example, the import volume more than doubled to 2.9 thousand tons), they accounted still a very small portion (less than one per cent) of total usage in Korea in 2005. Hence, the KFTC’s geographical market definition of Korea for carbon black used to reinforce rubber products was reasonable.[46]