Sequencing Negotiating Partners: Implications for the Two‐Level Game?

What determines the way in which states play the two‐level game when negotiating bilateral free trade agreements (FTAs)? A country’s need to simultaneously satisfy domestic constituencies while attempting to bargain internationally (known as the “two‐level game”) suggests a number of questions about the incentives that shape the way a particular country approaches these negotiations. Under what conditions will a government take a harder bargaining position with its own domestic constituencies and make concessions to its international bargaining partner? What are the conditions under which a government protects its domestic constituencies and attempts to claim more value from its international negotiating partner? To arrive at more complete answers to these questions, particular FTA negotiations can be best understood within the context of a country’s entire set of FTA negotiations. In this view, the effects of sequencing dynamics — that is, the order in which a specific country negotiates with others — must be taken seriously. This article explores two different cases of FTA partner sequencing and generates the hypothesis that sequencing dynamics have important implications for the conditions under which negotiators play the two‐level game in any particular negotiation.

Between 1999 and 2004, Japan and South Korea each negotiated two bilateral FTAs: Japan with Singapore and Mexico, and South Korea with Singapore and Chile. Bilateral FTAs had emerged as viable options for trade liberalization following the breakdown of several multilateral trade negotiations as well as in the aftermath of the Asian, and subsequently global, financial crisis of the late 1990s.¹ At the time of both Japan and South Korea’s first negotiations, each was under pressure from certain domestic factions to participate in this emergent bilateral trend. In both countries, opinion was growing that FTAs were becoming more of a global economic necessity. The perception existed that in the absence of action, they would be shut out. As the rest of the world continued to sign bilateral trade agreements, political and business leaders perceived that their industries and exports would become more disadvantaged in international markets.

But within both Japan and South Korea, some groups opposed FTAs. Both countries have culturally and historically important agricultural sectors that have traditionally been protected from the forces of free trade. In addition, the ambiguities associated with Article 24.8 of the General Agreement on Tariffs and Trade (GATT) (and subsequently the World Trade Organization [WTO]) concerning the requirement that preferential trade agreements should cover “substantially all trade” have opened up opportunities for FTA partners to negotiate the exclusion of potentially sensitive sectors from such agreements.² Both of these realities stood as major impediments to the negotiation of any FTA that might include the reduction of preexisting economic protection mechanisms.

The pool of potential partners that were initially considered by Japan and South Korea, however, posed varying levels of threat. Singapore, for example — with its limited agricultural sector — was a relatively benign FTA partner. The flow of agricultural products into Japan or South Korea from Singapore would be minimal. Mexico and Chile, on the other hand, were potential FTA partners whose agricultural production could threaten domestic agriculture in Japan and South Korea if protection was not guaranteed. Thus, for the governments of both Japan and South Korea, the Singapore FTA was an “easy case,” relative to the other possible partners, in terms of achieving the necessary support for ratification. Mexico and Chile, however, were potentially “difficult cases” because the domestic agricultural sectors in Japan and South Korea would potentially need to be placated during the ratification process.

Despite these similarities, Japan and South Korea concluded FTAs using opposing partner sequencing strategies. Japan, for example, concluded the easy FTA with Singapore first, followed by one with Mexico. South Korea, on the other hand, concluded the more difficult FTA with Chile first, followed by one with Singapore. In their two difficult cases, Japan (Mexico) and South Korea (Chile) played the two‐level negotiating game very differently. Japan, for example, held fast at the international negotiating table, insisting that it be allowed to continue to support its domestic pig and orange farmers with tariffs on pork and orange juice. When it came time to ratify the Japan–Mexico FTA, the Japanese Diet, the country’s legislature, passed the measure with relative ease. South Korea, however, made a number of concessions at the international negotiating table with Chile, particularly the elimination of tariffs on grapes, choosing to reach a deal first and then placate its domestic constituents. When it came time to ratify the deal in the South Korean National Assembly, the government faced almost a year of vocal opposition from the agricultural sector and had to set aside more than $80 billion in public and private side payments to farmers before ratification took place.

These differing sequencing dynamics suggest important but often‐ignored consequences for the way in which countries negotiate internationally and domestically. Building on the cases as well as the two‐level game and negotiation strategies literature, this article develops a model of the effects of particular sequencing decisions on bargaining. The conclusion of the relatively easy case first alleviated one source of domestic pressure for the Japanese — the pressure to get any agreement done and get into the “FTA game.” With an FTA already concluded, the government could approach the more difficult FTA negotiation knowing that a “no‐agreement” outcome would not lead to a further increase in domestic pressure from the business community, which favored rapid entry into the bilateral FTA market. This alternative to no agreement put the government in a stronger bargaining position for its negotiation with Mexico, allowing it to fight harder at the international bargaining table to claim value for potentially affected domestic constituencies. On the other hand, without an easy FTA under its belt, South Korea approached its difficult negotiation with the Chileans knowing that in the absence of an agreement, the relative domestic pressure to get into the FTA game could increase. With a poor no‐agreement alternative, South Korea conceded more at the international negotiating table, thereby requiring greater side payments in order to gain support for ratification among its own domestic interest groups.

Following the contentious conclusion of the Uruguay Round of trade negotiations in the GATT, countries in the late 1990s began to move toward bilateral free trade negotiations.³ In this context, Japan began to reevaluate its overall trade policy. Traditionally fearful that bilateral trade agreements could lead to the types of economic blocs that became problematic in the 1930s, Japan’s preference had been to conduct economic negotiations in a multilateral setting such as the WTO (Mansfield and Reinhardt 2003; Pempel and Urata 2006). But by 2000, Japan had begun to move toward a “multilayered” policy in which bilateral and regional trade agreements might complement the multilateral negotiations at the WTO (Japanese Ministry of Economy Trade and Industry 2000; Pekkanen, Solís, and Katada 2007). With the breakdown of talks at the Seattle WTO meetings in 1999, the process of negotiating multilateral trade agreements had become even more difficult, and the trend toward bilateral negotiations was quickly accelerating. As one of only three WTO member countries — along with South Korea and Mongolia — that had not already conducted bilateral or regional FTAs, Japan was in danger of falling behind (Shimizu 2000).

Inside Japan, particularly within the Japanese business federation Nippon Keidanren, the pressure to get on the free‐trade bandwagon was building (Keidanren 2000a, 2000b; Yoshimatsu 2005). In a keynote address to the annual Japan–U.S. business conference in July 2000, Keidanren chairman Takashi Imai pointed to the global trend of regional free trade networks and expressed a hope that Japan would begin to enter into FTAs “in a timely manner” (Takashi 2000). Imai expressed concern that Japanese exports were being shut out of markets that had already negotiated FTAs. A Keidanren report released in the weeks after Takashi’s speech argued: “Japan, however, has yet to undertake a single free trade agreement. As a result, Japanese companies are losing out on business opportunities in the international arena, and also finding themselves placed at a competitive disadvantage in doing business with countries that have already concluded FTAs elsewhere” (Keidanren 2000b).

Aside from concerns about preferential market and investment access, Keidanren also viewed the negotiation of FTAs as a means to secure the implementation of important structural adjustment reforms in the Japanese economy (Yoshimatsu 2005). The selective and successful negotiation of FTAs might be used as a way to diminish the political influence of domestic industries that opposed domestic reform, as those industries might be expected to shrink once exposed to the international competition introduced by FTAs.⁴ By the middle of 2000, Keidanren was arguing that domestic regulatory reforms and the negotiation of FTAs were important and mutually reinforcing parts of Japan’s economic future (Keidanren 2000a; Yoshimatsu 2005).

Despite the pressure in favor of FTAs, Japan could not jump into these agreements without worrying about its agricultural sector. Overall, agriculture accounts for less than 2 percent of Japan’s gross domestic product (GDP) while employing less than 4 percent of the country’s workers, a very small part of the Japanese economy (Pempel and Urata 2006: 87). But the political power of the sector far outweighs its economic production. The perception among Japanese politicians is that the agricultural sector generates the highest number of votes of any particular economic sector (Pempel and Urata 2006). This perception has created a tradition of agricultural protectionism in Japan led by the rural‐based Liberal Democratic Party (LDP) (Solís and Katada 2007a).

It was against this Japanese economic and political backdrop that Japan and Singapore approached each other in December of 1999 and agreed to conduct a joint study of the possible economic benefits of an FTA. Six months later, the two countries had come to a consensus that an FTA would be mutually beneficial to both countries, and they agreed to begin negotiations. From the beginning, Singapore was always perceived as an attractive partner; for Japan, it represented a relatively “easy case” (Terada 2005: 10). The original goals for the agreement were to focus on liberalization and cooperation in various service sectors considered to be high growth — such as finance, transportation, and information and communication technologies — while leaving the domestically sensitive agricultural sector in Japan relatively unaffected.

Singapore has virtually no agricultural sector and possesses limited and relatively insignificant potential to export farm goods to Japan. The Japan–Singapore Joint Study Group concluded that agricultural products comprised only 1.7 percent of Japan’s total imports from Singapore, as of 1999 (Japan–Singapore Joint Study Group 2000). In addition, the Japanese submitted a tariff reduction schedule that listed 486 agricultural items. Of these, 428 items had already been eliminated as part of Japan’s WTO commitment, and the other 58 products were ones in which the customs duties were already effectively zero. Thus, in advance, Japan was aware that a Japan–Singapore FTA would likely require only a minimum level of tariff reduction to satisfy the WTO’s conditions that entire sectors not be excluded from agreements (Aoki 2004). Thus, the Japanese leadership was confident that a potential FTA with Singapore would be relatively well received domestically.

Negotiations began in January 2001 and proceeded fairly smoothly. It quickly became clear that the parties would be able to reach an agreement within a year. While there were some agricultural products, namely cut flowers and goldfish, whose export into Japan were reported to have posed potential problems for Japan in their FTA negotiations (Pempel and Urata 2006; Ravenhill 2006), several of the Japanese negotiators involved have argued that there was virtually no opposition involving agricultural issues during the negotiation itself (Terada 2005). In addition, the potential domestic opposition in Japan from its agricultural sector was much lower for an FTA with Singapore than it would be with other partners with larger agricultural sectors, particularly given the fact that for many agricultural products, customs duties were already effectively zero prior to the agreement (Aoki 2004).

In the middle of October 2001, Japan and Singapore reached a basic accord on an FTA, and the two heads of state, Japanese Prime Minister Junichiro Koizumi and his Singaporean counterpart Goh Chok Tong, officially signed the deal on January 13, 2002. Singapore waived tariffs on about 99 percent of products (on a value basis) imported from Japan and eliminated levies on remaining products such as beer and alcohol for medicinal use. Japan waived tariffs on about 94 percent of products imported from Singapore including key sectors such as textiles and petrochemicals (Japan–Republic of Singapore 2002).

By the end of February 2002, the Japanese government presented the treaty to the Diet for official ratification. Treaty proponents needed to convince several LDP party members that their constituents’ agricultural interests remained protected, but apart from that, they faced no significant need to link issues or offer particular domestic sectors significant side payments. The House of Representatives approved the pact at the beginning of April 2002, and the upper chamber, the House of Councillors, lent its approval a month later. The FTA officially went into effect on November 30, 2002.

In 2001, just prior to the beginning of the Japan–Singapore FTA negotiation, Japan’s exports to Singapore were valued at $14.7 billion, amounting to 3.65 percent of Japan’s overall exports to the world (International Monetary Fund 2006). While the percentage of total exports is small, Singapore represented Japan’s seventh largest export market and its fifth largest market for investment by Japanese corporations (Kin 2000). In addition, Japanese expectations about the benefits of an FTA with Singapore were clear. A joint Japan–Singapore study group identified a number of areas of a possible benefit prior to the negotiation. These included sharing and enhancing communications infrastructure, joint initiatives in science and technology, and the expansion of trade into critical service sectors that were important to both the Japanese and Singaporean economies (Japan–Singapore Joint Study Group 2000). The conclusion of the FTA with Singapore represented an important step forward for a number of Japanese business interests that were loudly advocating the conclusion of an FTA.

Riding the momentum of the Singapore negotiations and pressing forward with its new multilayered trade agenda, the Japanese government next began negotiations over what it knew would be a domestically harder case, an FTA with Mexico. The initial discussions with Mexico had actually begun in 2000, at Mexico’s initiative. The Japanese were initially quite cold to the idea, but early on a number of Japanese domestic interests pushed for a Japan–Mexico FTA as part of a larger Japanese need to become part of the FTA game (Japan External Trade Organization 2000). The two countries began an official study of the potential for an FTA, but by early 2002 signs that that this negotiation would be much more contentious had already begun to appear.

In March, Mexican officials participating in the study officially asked Japan not to exclude any farm products in a potential FTA. In the case of Mexico, Japanese agricultural interests were much more vulnerable than they had been to Singapore: the farm trade with Mexico represented 20 percent of the total bilateral trade between the two countries (Solís and Katada 2007a). Japanese farmers quickly voiced their concern that an FTA with Mexico might lead to a flood of cheap Mexican agriculture into Japan and continued to invoke the LDP policy of negotiating agricultural concessions only at the WTO and not bilaterally (Japan–Mexico Joint Study Group 2002). Treading carefully, the Japanese responded to the Mexican request by replying that certain steps, such as the mutual acceptance of emergency import restrictions and a phase‐out period for some tariffs of between five and ten years, might meet the Mexican request, but the Japanese continued to insist that the “five fingers”— pork, beef, chicken, oranges, and orange juice — be excluded from tariff concessions (Solís and Katada 2007a). These products were particularly sensitive to Japan, particularly pork, which comprised 50 percent of Mexico’s primary (nonmanufactured) goods exports to Japan and 10 percent of total exports (Solís and Katada 2007a).

By November 2002, the two countries had moved beyond the study phase and into the actual negotiations. The Japanese Ministry of Agriculture, Forestry, and Fisheries (MAFF) negotiated with domestic agricultural interests prior to negotiating with Mexico in order to establish a minimally acceptable deal that could be ratified by the Japanese government (Solís and Katada 2007a). While the Japanese negotiators were able to come to some acceptable terms with their domestic constituencies, agricultural sectors remained a sticking point. Mexico continued to insist that Japan remove tariffs on all its agricultural products; Japan continued to insist that it be allowed to protect its key domestic agricultural sectors from the cheaper Mexican imports that might flood the domestic economy if tariffs were to be lowered.

The two sides continued to argue over this issue for the better part of a year. The definition of import quotas on pork became a particular source of frustration. The Mexican government presented a number of compromise proposals that involved the establishment of a pork quota that would have allowed the Mexican government to export 250,000 tons of tariff‐free pork per year. It was an amount over six times the forty thousand tons of pork that Mexico was exporting to Japan at the time (Asia Pulse 2003b). Japan countered by offering a low‐tariff quota of eighty thousand tons of pork. In addition, Japan offered to add another seventy agricultural products to the list of tariff‐free goods, including bananas and honey, so long as the two could agree to postpone the discussion of pork for at least a couple of years. The Mexicans rejected this compromise as well (Nikkei Weekly 2003).

After six months of back‐and‐forth offers and rejections, the two countries finally came to a substantive agreement in March 2004. Japan would create a low‐tariff quota of approximately eighty thousand tons for pork and one of 6,500 tons for orange juice. In return, Mexico offered to eliminate automobile tariffs in seven years and those on steel in ten years, setting up in the interim a tariff‐free import quota for Japanese automobile exports equivalent to 5 percent of Mexico’s domestic market in the initial year (Japan–United Mexican States 2004). The two sides officially signed the agreement in September 2004.

While the final agreement covered 99.6 percent of total bilateral trade between the two countries, Japanese agricultural interests were able to avoid conceding to a number of Mexico’s strongest demands. Aside from the previously mentioned arrangement on pork quotas, the Japanese government negotiated graduated quota reductions on beef, chicken, oranges, and orange juice, and delayed discussions about particular tariff rates on these products until two years after the FTA was entered into force (Solís and Katada 2007a).

When it came time to ratify the treaty in the Japanese Diet, the Japanese government made some general promises to help farmers that might be hurt by the FTA if it became necessary to do so. Yoshiyuki Kamei, the Minister of Agriculture, Forestry and Fisheries, announced that the government would “study the matter when problems actually emerge” (Japan Economic Newswire 2004). Largely because of the relatively small concessions made at the negotiating table with Mexico, however, no specific issues were linked to the agreement for domestic sectors, and there were no specific side payments that were negotiated with the Japanese agricultural sector. Just two months after the agreement was signed, the Japanese government ratified the agreement with little vocal domestic opposition.

In the late 1990s and early 2000s, South Korea faced many of the same pressures to join the FTA club as did Japan prior to its first successful FTA negotiation. The multilateral trade negotiations in the WTO had stalled, and many countries had already begun the process of negotiating bilateral FTAs. Additionally, the Asian financial crisis of 1997–1998 had exposed a number of structural weaknesses within the South Korean economy and placed significant pressure on the country’s exports to perform better in the global economy (Sohn 2001; Bong and Jung 2005; Koo 2006; Park and Koo 2007). In late 2001, a joint report issued by various business groups and the state‐run Korea Institute for International Economic Policy (KIEP) projected that South Korea’s exclusion from emerging bilateral FTA networks was costing the country approximately $34.4 billion in exports and a loss of growth in its GDP of 1.33 percentage points (Cheong 2001).⁵

The Korean leadership also perceived an interest in utilizing FTAs as a means to lock in various domestic economic reforms. President Kim Dae‐jung had dealt with the global calls for liberalization and transparency that followed the Asian financial crisis by enacting a series of reforms: recapitalizing banks, buying up bad loans through a public asset management company, pressuring chaebols (Korean business conglomerates) to rid their balance sheets of any debt, and encouraging investment in industries such as information technology (Koo 2006; Park and Koo 2007). In addition, Kim reorganized many Korean bureaucracies as a means to consolidate support for these reforms (Park and Koo 2007). The country’s new trade strategy, which included the completion of FTAs, was seen as an integral part of the overall reform process.

As these types of pressures mounted, South Korean leaders were continuously reminded that their country was one of three WTO member countries, including Japan and Mongolia, not to have negotiated any FTAs (e.g., Cheong 2001; Yoo 2002). The cumulative effect of these various pressures was a perception among South Korean leaders that an FTA was needed, at least to create the impression both internally and externally that they were on the bandwagon and would not be closed out of the FTA game. South Korean policy makers began to develop a genuine interest in FTAs as a complementary trade arrangement (Sohn 2001; Park and Koo 2007).

Similar to Japan, however, South Korea had a historically and culturally important agricultural sector that had traditionally been protected by the government from outside trade forces. Throughout the 1970s and 1980s, Korea’s agricultural policy banned most agricultural imports with the exception of wheat, corn, cotton, soybeans, hides, and rubber (Diao et al. 2002). Although several countries successfully challenged these quantitative import restrictions before the GATT in 1989, Korea continued to protect rice, which is culturally and historically the country’s most important agricultural commodity (Diao et al. 2002). In addition, despite the elimination of a larger number of quantitative restrictions, Korea limited agricultural imports through a number of other trade barriers such as tariffs, quotas, and other nontariff barriers (Diao et al. 2002).

Within this context, South Korea began to consider a number of possible FTA partners. But South Korea differed from Japan in that it was not able to conclude an “easier” FTA with Singapore prior to the conclusion of its FTA with Chile. In November 1998 at the Asia‐Pacific Economic Cooperation (APEC) summit meeting, South Korean President Kim Dae‐jung proposed to Chilean President Eduardo Frei that the two countries deliberate on the possibility of negotiating an FTA. Within a month, a joint South Korea–Chile task force was formed to begin to identify the bargaining zones on such issues as market access, trade rules, intellectual property, services, and specific legal procedures (Park and Koo 2007). The two presidents agreed in principle to the idea of a free trade pact in September 1999, and official negotiations on the details began shortly afterward in December.

Between December 1999 and December 2000, negotiators were able to agree upon approximately 80 percent of the text of the FTA (Chung 2003). At that point, however, tariff concessions for a range of agricultural products became a sticking point. Many of Korea’s agricultural products were protected by tariffs of 100 percent or higher (Chung 2003). One of Chile’s motivations in pursuing the FTA was clearly greater access to these markets, but Chilean export competitiveness in such products as grapes, apples, and pears threatened Korean farmers (Chung 2003). In 2001, Chile exported $1.08 billion worth of fresh fruit, its second largest global export behind copper.⁶ Of this, grapes accounted for more than 41 percent, apples for more than 23 percent, and pears slightly less than 6 percent (Central Bank of Chile 2009). Disagreements over tariff reductions and market access for these products threatened negotiations from the start, as Korean grape growers, for example, publicly demanded that the South Korean negotiators take the removal of grape tariffs off the negotiating table. (This opposition occurred despite the fact that the Northern and Southern hemispheres have opposite growing seasons. Thus, direct competition between South Korean and Chilean farmers would be somewhat mitigated by the fact that Chile’s agricultural products would arrive during the South Korean winter and spring.)

The discussions on the particulars of the FTA continued for almost two years, bogged down by the domestic interests of South Korean farmers. Early in the negotiations, the South Koreans had taken rice tariffscompletely off the negotiating table; however, there still seemed to be some room to reach a compromise on grapes, apples, and pears. But as the negotiations became prolonged, the pressure steadily mounted on the South Korean government to conclude the FTA. The president of the Korea Chamber of Commerce and Industry, Park Yong‐sung, urged the South Korean government to move the FTA talks forward, arguing, “We will be a laughing stock if Seoul moves to abandon an FTA with Chile, in the face of pressures from local farmers” (Yoo 2002).

With the pressure building, the South Korean government came back to the negotiating table in early 2002 and offered concessions on agricultural tariffs to the Chilean government. While the tariffs on sensitive agricultural products still remained a nonnegotiable point for South Korea, there was apparently some progress made on other agricultural tariffs (Chung 2003). In the end, however, the Chileans insisted on even greater access, as Osvaldo Rosales, Chile’s general director for International Economic Affairs, argued: “There is still room for improvement in some categories such as fruit. We cannot exclude our leading fruit export items such as apples, pears, and grapes from an agreement, just as Korea cannot exclude automobiles” (Korea Times 2002).

After another six months of negotiation over these issues, the South Korean government concluded an agreement with Chile in October 2003. Under the agreement, South Korea agreed to eliminate tariffs on two‐thirds of imported products. As for the more sensitive agricultural sectors, South Korea was able to keep tariffs on rice and apples off the table, but it did make significant concessions on grape tariffs, agreeing to a ten‐year phase‐out of the tariffs with the exception of the season in which Korean grapes are harvested (Chung 2003; Park and Koo 2007). Because of these concessions, Korean farmers reacted angrily to the agreement, arguing that they would fight the agreement during the ratification process. A statement from the Korean Advanced Farmers Federation said, “We cannot stop our anger from exploding. We will fight to the end to frustrate its ratification at the National Assembly” (Agence France Presse 2002).

Over the next year, they did exactly that. Massive demonstrations of Korean farmers thwarted at least three attempts to ratify the agreement in the South Korean Assembly. Assembly members from predominantly rural districts were successfully able to filibuster a number of potential votes on the ratification bill either by physically blocking the voting proceedings or by invoking any number of legal technicalities in an attempt to slow down the process. Farmers and their agricultural lobbies also threatened to vote against any member of the assembly voting for the bill in the upcoming elections (Lee 2004).

All the while, the pressure from business groups and various economic ministers to ratify the FTA continued to mount (Park and Koo 2007). Just as they had been before the negotiation of the FTA, the arguments were still of a very general nature, pointing out that Korea’s lack of any FTA was “abnormal for a country that depends on exports and trade for growth, and is suffering from a weak consumer domestic market” (Yonhap News Agency 2004). But as the ratification process dragged on, South Korean manufacturing exports continued to drop as other market countries began to trade with countries other than South Korea as part of their own FTAs (Park and Koo 2007). In July 2003, the heads of the four major business groups in South Korea sent letters to all 272 of the countries’ assembly members, asking for an end to the delays over the ratification of the Korea–Chile FTA. The letter stated that failure to ratify the FTA would put South Korea “in the orphanage of the international market” (Asia Pulse 2003a). In early 2004, a report by the Korea International Trade Association (KITA) pointed out that Korean inaction placed it at a significant disadvantage in regional export markets (KITA 2004). Finally, by February 2004, Korean Deputy Prime Minister Kim Jin‐pyo verbalized his own concern that a failure to ratify the Korea–Chile FTA would damage his country’s external credit (Asia Pulse 2004).

In the end, in the face of these competing pressures, the South Korean government agreed to significant side payments to South Korean farmers. The initial offer was more than $620 million in compensation, but after a number of unsuccessful ratification attempts in 2003, President Roh increased the offer significantly. A special law was passed that earmarked more than $80 billion in public and private for side payments to Korean farmers and fishermen over the ten years following the agreement (Seo 2003; Bong and Jung 2005; Park and Koo 2007). This amount included $138 million in farm aid during the first year as well as significant expenditures to improve farm infrastructure and enhance South Korea’s farmers’ competitiveness abroad. Finally, the offer included a reduction in the agricultural loan interest rate from 8 percent to 5 percent (Bong and Jung 2005: 149).

The differences between the Japanese and South Korean experiences suggest the following question: what impact does a preexisting FTA have on the domestic ratification process when a country seeks to negotiate a second, more difficult FTA? Reconsideration of the literature on bargaining — specifically the simultaneous bargaining that occurs domestically and internationally during trade negotiations (i.e., the two‐level game literature) — yields some insights into and suggests a possible interpretation of the effects of partner sequencing in the Japanese and South Korean cases, respectively. The theoretical implications of this analysis call for further testing. Better understanding of partner sequencing dynamics might lead to a more precise understanding of FTA partner selection processes and their ramifications.

Robert Putnam’s (1988) articulation of the two‐level game analogy has had important consequences for the way in which international negotiation is studied. Building on earlier work (Schelling 1960 and 1980; Walton and McKersie 1965), Putnam observed that an international negotiator must negotiate simultaneously at both the international level and the domestic level and that this has consequences for the range of possible acceptable agreements. Upon this foundation, other scholars have attempted to more fully specify the nature of this process. Some have explicitly attempted to formalize the model: among other ideas, specifying different levels of certainty or information at the domestic level (Iida 1993 and 1996; Mo 1994), different preferences for the international negotiator (Mo 1994), and different levels of veto power within varying domestic constituencies (Mo 1995).

Others have focused more specifically on the nature of the domestic politics game and its effects at the international level. Early work in this area has examined the strategies available to governments seeking to gain domestic support for an international agreement. These include the use of side payments (Mayer 1992; Friman 1993) and issue reframing (Friman 1993). More recent work attempts to specify the effects associated with institutional linkages across various sectors (Davis 2004) and the effects of making domestic commitments to a particular plan of action (Leventoglu and Tarar 2005).

Despite these many advances, two important criticisms can be leveled against the two‐level game literature. First, much of this work has modeled negotiations as single‐shot, single‐issue, two‐party bargaining situations in which the actors are assumed to have very specific preference structures. Its tendency has been to focus on either singular, specific international negotiations or singular, hypothetical international negotiations, treating them as if they were isolated deals occurring outside the context of any series of negotiated agreements. As a consequence, the predictions derived from these models are highly context dependent (Hug and Konig 2002). But the branch of game theory that builds on the work of Robert Axelrod (1984) and others, which focuses on the important consequences of repetition for strategic interaction, is suggestive of the idea that two‐level games with multiple iterations might have a different dynamic than a single‐shot game.

In addition to these criticisms, much of the work on two‐level games has overlooked the scholarship that studies the processes of negotiation rather than focusing on particular outcomes. Scholars of international relations who have attempted more recently to link some of this diverse bargaining literature with earlier work in international relations have found, for example, that market conditions, belief systems, domestic politics, and domestic institutions help shape the particular strategies that negotiators employ in international economic negotiations (Odell 2000, 2006).

In taking seriously the politics that underpin particular negotiation strategies, the literature offers two important insights for thinking about international economic negotiations. The first is the idea that negotiating strategies are often shaped as much by alternatives to a negotiated agreement — that is, the options or position which are available to each party if negotiation fails — as they are by the range of possible outcomes within the bargaining space itself.⁷ The underlying logic is that as a particular party to a negotiation finds itself in a position in which walking away from the negotiating table without a deal will leave it in a bad position (either relative to previous no‐deal positions or relative to the no‐deal position of the opposing party), the minimal deal that party will accept — its resistance or reservation point — will lower. This has two practical effects. First, the party’s lower resistance point affects the number and type of concessions it will make. By definition, if a party will accept less as its bottom line, it will concede more to other negotiators. Second, the potential zone of possible agreement (ZOPA) between the negotiating parties is widened, ceteris paribus (all other things being equal) (Odell 2000).

Second, the literature on the process of negotiation also suggests that it is important to consider the linkages that might exist between negotiations (Watkins and Passow 1996; Crump 2007). Among other possibilities, these might be single bargaining instances within the context of what may be either a longer bargaining relationship (Sheppard 1995) or a series of similar issue negotiations with different partners (Sebenius 1996; Noe and Wang 2004). In this context, the notion of sequencing becomes particularly important. Theoretically, particular sequencing strategies can affect payoffs (Noe and Wang 2004); thus, such strategies should reflect parties’ maximization of a calculation that includes the expected value of gaining the assent of another party, the expected cost of the recruitment of that party, the probability of the successful recruitment of that party, and the cost of a failed effort to recruit that party (Sebenius 1996).

Given the different partner sequences in the cases of Japan and South Korea’s first two FTA negotiations, a model that attempts to capture the effects of sequencing dynamics on bilateral FTA negotiations and outcomes is in order. Much of the existing literature on two‐level games generally fails to fully address the iterative nature of bargaining and bargaining relationships — the model I propose could help fill this gap. Figure One demonstrates a set of partner sequencing dynamics and how they underpin a hypothesis that might be derived from these cases.

In a case in which a country faces domestic pressure to conduct an FTA and in which there is a defined set of possible negotiating partners, the country has a number of possible partner sequencing combinations depending on the size of the set. Assuming that much of the negotiation with possible partners in the set takes place simultaneously, however, each possible agreement can be broadly defined according to the ease with which it can be ratified domestically — that is, the amount of opposition it generates among various domestic constituencies — relative to another possible agreement. Here, the focus is on how a country conducts bargaining at both levels of the two‐level game in those cases that are relatively more “difficult,” for example, those that are domestically contentious.

Thus, Figure One depicts the potential bargaining space between hypothetical countries: X and Z. It is assumed that a potential FTA with country Z would be domestically contentious, a “difficult” case relative to the other possible agreement partners for country X. In the context of the two cases discussed above, country X could represent either Japan or South Korea, and country Z would be either Mexico or Chile for each partner, respectively.

As seen in the inset, at stage one, where country X has yet to conclude a single agreement, various and competing domestic pressures act on that country. First, assuming that an agreement with country Z would be domestically difficult, domestic pressures against the conclusion of an FTA will develop, but there will also be competing domestic pressures to conclude an FTA in order to “get in the FTA game” as soon as possible. As time continues to pass without a concluded FTA, this source of domestic pressure builds on the government. This mix of pressures for and against the conclusion of FTA agreements was visible in both Japan and South Korea as they attempted to restructure their domestic economies and overall trade policies.

The cases demonstrate the real economic and political consequences for country X’s “getting into the game.” Economically, entering the preferential FTA “game” should provide a country the opportunity to protect its own export markets (Mansfield and Reinhardt 2003), and this was a particularly acute concern in pre‐FTA Japan and South Korea given the environment of stalled multilateralism and proliferating FTAs in which they found themselves. In addition, preferential FTAs provide particular competitive advantages to multinational firms, as the reduction in tariffs allows businesses to more easily vertically integrate their production lines, using FTA partners as an export platform (Manger 2005). Finally, bilateral FTAs generate specific economic rents in the form of lower‐cost production in the partner country, and the exclusionary nature of such agreements vis à vis other countries provide competitive advantages for the industries that seek to capture those benefits (Manger 2005). All of these reasons were figured into the calculus of the Japanese and South Korean business communities as they pressured their governments to negotiate FTAs.

Politically, entering the FTA game also offers ways to lock in domestic economic reforms. International trade can have powerful effects on domestic political coalitions (Rogowski 1987 and 1989) and the proreform sectors in both countries hoped that FTAs would serve to significantly weaken antireform sectors and reduce their political leverage as a consequence (Irwin 1993 and 1995; Mansfield and Milner 1999). In addition, entering the FTA game can help reinforce security arrangements (Gowa 1994; Mansfield and Milner 1999) and improve a country’s international status (Ibarra‐Yunez 2003), two other potential political benefits. Finally, successful conclusion of an FTA can also have a “capacity building” benefit as a country’s negotiators are able to gain skills that they might utilize in future FTA negotiations (Okamoto 2003; Aggarwal and Koo 2005; Solís and Katada 2007b).

At the international negotiating table, these various domestic pressures result in a relatively low resistance point. Because of the continually building domestic pressure to sign a deal in order to break into the game, country X’s negotiators face a relatively poor alternative to any negotiated agreement. A breakdown in negotiations between countries X and Z still leaves country X without any concluded FTAs, standing outside the FTA game. Thus, because the alternatives to making the deal are poor, we would expect country X to be more likely to accept concessions and less likely to make strong demands of country Z on behalf of its own domestic constituencies. In addition, once the international portion of the deal with country Z is concluded, country X is left to seek domestic ratification. Given its willingness to make concessions at the international bargaining table to alleviate one domestic pressure, it is likely that X will have to employ any number of strategies, including the use of large side payments and possibly issue linkage, in order to mollify the remaining source of domestic pressure — those interests against the FTA.

Assume, on the other hand, that prior to the conclusion of the FTA with country Z, country X is able to conclude an “easy” FTA with another country.⁸ The existence of this easy, concluded agreement essentially diminishes a key source of domestic pressure that had been acting on country X in its negotiations with country Z. This has the effect (ceteris paribus) of improving country X’s no‐agreement alternative. If the negotiations between country X and country Z break down at this point, country X is still left with a preexisting FTA, and the domestic pressure to “get in the game” has been somewhat alleviated relative to what it would have been in the absence of the easy FTA. From this position, country X is in a better position to improve the minimal deal that it will be willing to accept. It is likely to employ tougher value‐claiming tactics at the international negotiating table with country Z, as it attempts to protect those domestic constituencies that oppose the agreement. Having fought hard to protect these interests at the international negotiating table, the government of X should find that it has to use relatively fewer side‐payment strategies to secure domestic ratification.

The logic of this argument is in line with negotiation theory, but it can also be derived from the cases themselves. Because it had already inked an FTA with Singapore, Japan’s no‐deal alternative to the Mexico negotiations was stronger than it otherwise would have been, and Japan was able to negotiate from a relatively stronger position that it might have otherwise been able to do. Because of its prior FTA agreement and the resulting reduced pressure to join the club of countries negotiating FTAs, the Japanese were able to stand by their domestic industries and concentrate the bulk of their value claiming efforts at the international negotiating table.

Alternatively, the significant pressure facing South Korean negotiators to conclude an FTA meant that South Korea approached the international negotiating table with a worse no‐agreement alternative than they otherwise might have if they had started their sequence of FTA negotiations with a domestically “easier” case like Singapore — with whom the South Koreans completed an FTA later, in 2004. In 2000, when South Korea was negotiating an FTA with Chile, its exports to Singapore totaled $4.079 billion, equaling 2.72 percent of its total world exports (International Monetary Fund 2006). Although this trade relationship is relatively small, when the two countries did finally begin negotiations on an FTA in 2003, both sides expressed hope that mutual benefits would arise from any potential agreement. The Koreans hoped that an FTA with Singapore would lead to increased trade in specific service sectors such as financial services, increased investment opportunities for Korean businesses, and increased cooperation in scientific and technological research (Korea–Singapore Free Trade Agreement Joint Study Group 2003). In a general sense, many of these perceived economic benefits were the same as those demanded by the Korean business community prior to the Chile FTA negotiations. Thus, there is reason to believe that had Korea negotiated a Singapore FTA first, the business community might have been relatively more satisfied about the overall direction of Korea’s trade policy, and it might not have been as vocal throughout the Chile negotiations about the need to get an agreement signed and ratified.

At the time of the negotiations with Chile, the best alternative to any negotiated agreement would have resulted in the status quo of zero negotiated FTAs. Because of the growing regional and global trend toward bilateral FTAs, a breakdown of the negotiation — a result that would have kept South Korea on the outside looking in — was increasingly unacceptable to South Korean business associations. As this source of domestic pressure continued to increase, South Korean negotiators’ resistance point lowered, and various concessions at the international negotiating table were more acceptable. Because of these concessions, the final FTA faced a much more protracted domestic negotiation with South Korean farmers during the ratification process. This pressure needs to be considered as partial explanation for the significant side payments the government laid out in order to ensure ratification.

The exploration of the two cases, buttressed by the theoretical insights from the negotiation and two‐level game literatures, yields the following hypothesis about the consequences that partner sequencing may have for the interaction of domestic and international negotiation:

Thinking about the process of bilateral FTA negotiations in this particular way helps address both of the criticisms of the two‐level‐game literature. First, the fact that many countries have looked not just to single FTA negotiations but are attempting to develop an FTA strategy, in which they foresee multiple bilateral and minilateral FTA agreements, provides an opportunity to think about these negotiations not as singular events but as one of a set of possible negotiations. Here, the two‐level game becomes an iterated game rather than a single‐shot bargaining game. The model provides a “first cut” at thinking through how the consequences of any particular negotiation affect future negotiations.

Second, the process of negotiation is as important to the model as are negotiation outcomes. Here, negotiation strategies act as the mechanism through which sequencing choices affect political outcomes. Studying the sequential choices about trade partners allows scholars to begin to specify the effects that choosing one partner over another might have on a country’s bargaining position and strategy in particular negotiations. Thus, it could conceivably lead to better understandings about the conditions in which a country would choose a particular sequence of partners.

The two cases presented here have been used to derive this theory about the effects of partner sequencing choices. As such, they do not, in and of themselves, represent a proper empirical test of the theory. An adequate evaluation of the theory’s robustness must explore its applicability to other cases. Given the dangers of selection bias associated with choosing cases based on the dependent variable (King, Keohane, and Verba 1994), it will be initially useful to study cases that vary with respect to the question of whether they had completed a “domestically easy” FTA prior to negotiating the “domestically difficult” FTA.

In this respect, there are some pairs of FTA cases that present interesting opportunities to test the theory. For example, Thailand began its foray into the world of FTAs by negotiating and ratifying an FTA with Bahrain in 2002. On the surface, Bahrain represented a “domestically easy” case for Thailand, as in a strictly economic sense there was little to be gained in the deal. But Thai elites were under considerable pressure from various domestic constituencies to climb aboard the global FTA bandwagon. Many of the same pressures at work in the Japan and South Korea cases appear to have been relevant to the Thai case: concerns about their outsider status given the proliferation of preferential trade agreements, the need to lock in domestic economic reforms in the wake of the Asian financial crisis, and the need for the potential security benefits that could accrue (Hoadley 2007). After the negotiation with Bahrain, Thailand and China negotiated what would seem to be a much more domestically difficult FTA in 2003.⁹

The domestic negotiation in Thailand over the Thai–China FTA could be compared with the effects of the partner sequencing decisions made by Malaysia. Facing many of the same pressures as Thailand to join the growing FTA network as well as some of the same incentives to negotiate a “domestically easier” FTA first (Hoadley 2007), Malaysia’s first FTA partner was Japan; the agreement entered into force in July 2006. On the surface this would appear to have been a relatively more challenging FTA to ratify, and thus the effects of Malaysia’s partner sequencing dynamics are ripe for comparison with those of Thailand.

Once this type of exploration is undertaken, however, a complete exploration of the theory might also involve choosing cases that share a similar outcome — for example, two cases involving difficult domestic FTA ratification processes with heavy domestic side payments or, alternatively, two cases involving relatively easy domestic FTA ratification negotiations. This process of exploring cases with similar outcomes might be useful in specifying the necessary and sufficient conditions of the model (George and Bennett 2005). In this vein, it might be useful to compare two countries whose second FTA negotiations involved similarly difficult domestic ratification processes such as Peru (partner was United States) and Thailand (partner was China). Does the presence or absence of a preexisting FTA that was relatively easy to ratify domestically lead us to any conclusions about the necessary and/or sufficient conditions under which hypothesized partner sequencing effects hold?

Even prior to more rigorous testing, a number of concerns need to be addressed and possibly incorporated into the model. First, the observable implications of the model seem to present an important obstacle. Given the different effects that the sequencing dynamics appear to have had in these two cases, one might be led to ask why countries do not always pursue an “easy‐first” FTA as a strategy. If Japan’s conclusion of the Singapore FTA had the effect of lessening one source of domestic pressure for the following negotiation, allowing it to utilize stronger international negotiating tactics in protection of its domestic interests, then why would other countries not do the same?

If one assumes that the South Koreans were equally as rational as the Japanese, then theoretically they should have had access to the logic of this argument. It seems logical that many of the pressures associated with getting in the game would have been equally satisfied had it negotiated an FTA with Singapore first. As seen in the case of Japan, in terms of partner choice, Singapore would have offered South Korea an opportunity both to lock in various domestic structural reforms as well as an opportunity to gain valuable FTA negotiating experience with a seasoned partner. These benefits would have come at relatively low economic risk given the size of the Singaporean markets. So why would the South Koreans choose to negotiate the more difficult FTA with Chile first? Why would any country negotiate the more difficult FTA first?

One explanation might lie in the fact that the choice of negotiating partner is better modeled as a strategic interaction. Decisions on whether and when to form a negotiating partnership are almost never made by just one party to the negotiation; they are often the product of initial joint agreements to negotiate that are based on mutually constituted exploratory committees. Initially, this reality is likely to limit the pool of willing partners for any one country. For example, at the time that South Korea opened itself to bilateral FTA negotiations, it had received positive responses from only Turkey, Thailand, New Zealand, Israel, and Chile (Park and Koo 2007).

Second, once the possible pool of partners is identified, the notion that “it takes two to tango” implies that it is just as important to consider the positions of partners such as Singapore and Chile as that of South Korea when attempting to explain particular sequencing decisions. For Singapore, both the choice of FTA partners and its overall trade strategy were motivated by many of the same concerns as South Korea. Frustrated by the lack of progress in multilateral trade negotiations, the leaders of Singapore generally sought increased market access and protection in the form of dispute resolution mechanisms, in addition to the diplomatic benefits of alliance formation and profile enhancement (Lee 2006; Hoadley 2007). The case is equally similar in Chile, where their possible and realized FTAs were seen as carrying a combination of economic, security, and credibility benefits (Schiff 2002).

The relevant question for the more general model is whether the potential partners — Singapore, Mexico, and Chile in this article — considered the negotiations with Japan and South Korea as domestically easy or difficult in their own right. Here, the relative comparisons of best alternative to a negotiated agreement might be more insightful if done vis à vis the opposing party’s best alternative rather than vis à vis other possible agreements. Adding in a strategic component to the model would help to more clearly illuminate the cases that appear to fall outside of the observable implications of the model that this article has presented. At the very least, it would help to specify the model’s necessary and sufficient conditions.

Second, if the logic of the model holds for a country entering the FTA game, how do its partner sequencing dynamics affect its ability to bargain once more than one FTA has been concluded?¹⁰ Presuming that some domestic pressure is alleviated upon the completion of the first FTA and that this has impacts on the two‐level game while negotiating the second, what is the cumulative impact of having two FTAs concluded on the two‐level game played during the negotiation over a third? One might imagine a scenario of gradually diminishing pressures that render these dynamics increasingly less important with the conclusion of each new FTA. But the model and the cases also seem to indicate that the specific choice of partners may have important implications for the rate at which those pressures diminish. A series of swiftly negotiated, easy FTAs may cumulatively alleviate pressure from the pro‐FTA community leaving the government in a better position to bargain hard to protect its domestic interests at each successive negotiation. Further theoretical and empirical explorationof the iterative and cumulative effects of partner sequencing dynamics is clearly warranted.

Finally, what is the empirical importance of this model relative to alternative explanations for why a country may approach the international or domestic bargaining situations from positions of relative strength or weakness? Many of these explanations come out of the formal two‐level game literature discussed earlier. A large body of literature, for example, explores the various strategies governments might employ domestically to strengthen their international bargaining position. Tactics such as tying hands or engaging in costly signaling can generate domestic audience costs for negotiators that can serve to strengthen their relative positions at the international bargaining table (Fearon 1994 and 1997). Making various verbal commitments to domestic publics in promise of particular benefits from the international negotiation is also a type of audience‐cost strategy that can generate international bargaining leverage (Leventoglu and Tarar 2005).

In the cases presented above, the South Koreans did not really utilize any of these tactics in their negotiations with Chile. There were few, if any, ex ante promises to their domestic constituencies, and it was only during the domestic ratification process that various and escalating levels of side payments were promised to South Korean farmers. While it is conceivable that various domestic tactics could have placed South Korea in a stronger international bargaining position with Chile, even in the absence of a preexisting FTA with Singapore, it seems equally conceivable that these tactics would have been countered by the reality of South Korea’s no‐agreement position. An attempt by South Korea to claim value in their negotiations with Chile through ex ante verbal commitments to their domestic constituencies, for example, would likely have been viewed as a weak signal by the Chilean negotiators, given the fact that South Korean business groups were pressuring the government to take a first step toward getting into the game by completing and signing an FTA.

Thus, it would appear that despite the need for further testing and the inevitable tweaking of the model that will result, the core idea that partner sequencing choices have implications for future relative bargaining strength is important enough to merit further theoretical and empirical study. On first blush, it would seem that the implications of the general core of the argument might even extend beyond the negotiation of FTAs. For example, it is conceivable that governments under domestic pressure to negotiate multilateral environmental agreements would do well to “get in the game” with a relatively “easy” agreement from the perspective of its domestic audience. Doing so could confer enough benefits to reduce this source of domestic pressure, putting the government in a relatively strong bargaining position to negotiate harder agreements both internationally and domestically.

Individual negotiations are often part of a sequence of negotiations, and the particular sequence undertaken by any one country has implications for the way that it plays the two‐level game. This article has utilized two case studies to propose a theory about the effects that partner sequencing may have on the simultaneous international and domestic negotiations of FTAs. A country negotiating FTAs, for example, can improve its ability to support its domestic interests and claim more value at the international negotiating table by negotiating and concluding an easy agreement first. With the easy agreement in hand, the country has reduced the pressure to achieve any agreement and improved on its set of no‐deal alternatives as it moves to negotiate the more difficult agreement.

Because it had already inked an FTA with Singapore, Japan’s no‐deal alternative to the Mexico negotiations was stronger than it otherwise would have been. Some of the economic and political incentives for getting into the FTA game were satisfied by the first FTA with Singapore. Thus, Japan was able to negotiate from a relatively stronger position than it might have otherwise been able to do. South Korea generally faced the same set of economic and political incentives and resulting pressures to get into the FTA game, but its lack of a preexisting deal meant that it approached the Chile negotiations with a relatively worse no‐deal alternative. This had the effect of weakening its international bargaining position and requiring large concessions at the domestic bargaining table as a result.

Although further empirical testing is necessary and revisions to the initial model presented here are inevitable, the study of partner sequencing dynamics and their consequences for politics and bargaining seems overdue. This research program has the potential to contribute to a more contextualized understanding of negotiation — particularly international negotiation — where the iterated and linked nature of the bargaining environment has important consequences for the negotiated outcomes.