Regional consolidation, cross-regional tie-ups, or multilateralisation of east Asian FTAs? Relative impacts when preference utilisation is incomplete

Jayant Menon 10 April 2013



With the Doha Round of the WTO stalled indefinitely, bilateral free trade agreements have proliferated. While countries in Asia were relative latecomers to preferential liberalisation, they have been catching up rapidly. Indeed, the proliferation of bilateral free trade agreements has been greatest in Asia over the past decade. As of January 2013, there are 109 bilateral free trade agreements involving an Asian country already in effect, 132 that have been signed, 75 being negotiated, and 50 more proposals. The outcome of the proliferation of often overlapping bilateral free trade agreements has been described as the spaghetti-bowl effect, or in Asia, the noodle-bowl effect.

How do we remedy the situation? Concluding Doha would help, but the single undertaking appears increasingly unlikely, and may not even be enough. Therefore, a number of proposals that have been put forward (Baldwin 2006, 2008; Menon 2007, 2009), but they can be broadly grouped under two headings: consolidation and multilateralisation. The consolidation approach proposes the creation of a region-wide bilateral free trade agreement, in order to neutralise intra-regional bilateral free trade agreements. The multilateralisation approach suggests that preferences be offered to non-members on a non-discriminatory basis, thereby eliminating any margin of preference. While these two proposals are aimed at addressing the noodle bowl, the impasse at the multilateral level has led to another development designed to broaden reciprocal access to markets outside the region. Recently, there has been growth in cross-regional tie-ups of bilateral free trade agreements, linking blocs in Asia with other blocs or countries within them. For instance, proposals to create an ASEAN-EU bilateral free trade agreement, an ASEAN-US bilateral free trade agreement, and other similar linkages, are gathering momentum.

In this column, we seek to assess the relative merits of these two approaches, as well as the recent trend for cross-regional tie-ups, by addressing a number of limitations in previous studies. In particular, we try and take into account more realistic utilisation rates of preferences in estimating welfare impacts. Most previous studies, and all previous studies on East Asia, have assumed that utilisation is complete, or 100%.1 This is a serious limitation since the evidence suggests that utilisation rates of Asian bilateral free trade agreements are very low, usually ranging between 10% and 20%, and rarely above 25% (see Menon 2013). Securing reciprocity is an important motivation for pursuing bilateral free trade agreements over unilateral actions. We isolate the impact of reciprocity, in comparing the consolidation and cross-regional tie-ups against multilateralisation, and identify the conditions under which any additional benefits can be secured.

Model and method

The analysis is conducted using the ‘MONASH multi-country’ model, which is a dynamic computable general equilibrium model of the global economy (see Mai 2004). The version of the MONASH multi-country model used for this study has 57 commodities and industries, and 14 economies and regions. These consist of six ASEAN economies (Indonesia, Malaysia, Philippines, Singapore, Thailand, and Vietnam); the ‘+6 economies’ (China, Japan, South Korea, Australia, New Zealand, and India); the US; and a residual rest of the world (ROW) economy. We focus on the most commonly discussed proposal in this region – a bilateral free trade agreements involving the ASEAN+6 countries, the so-called Regional Comprehensive Economic Partnership.

Each of the simulations described below is a comparison with a business-as-usual (but moving background) scenario, the baseline. The baseline shows the growth of economic indicators without any trade liberalisation taking place. The deviations of economic indicators from the baseline are used as measures of the effects of the various trade-liberalisation scenarios.

We conduct six basic simulations:

(i) Simulation 1: ‘Preferential Liberalisation with Full Utilisation of Preferences’;
(ii) Simulation 2: ‘Preferential Liberalisation with Incomplete (25%) Utilisation of Preferences’
(iii) Simulation 3: Multilateralisation of Preferences’
(iv) Simulation 4: Cross-regional Tie-ups, with Full Utilisation of Preferences
(v) Simulation 5: Cross-regional Tie-ups, with Incomplete (25%) Utilisation of Preferences;
(vi) Simulation 6: Global Liberalisation.

The difference between simulation 1 and simulation 2 is straightforward, and relates purely to different rates of uptake of preferences. With simulation 3, the preferences are offered to non-members on a non-discriminatory basis. Reciprocity is introduced in simulation 4 and simulation 5, where non-members reduce tariffs on imports from member countries. The difference between simulation 4 and simulation 5 is the rate of utilisation of preferences. Since reciprocity is secured through the regional FTA linking up with other FTAs, either through a series of bilateral or plurilateral tie-ups, preference utilisation is still an issue. So far, tariffs on trade between non-members remain unchanged. This is where the final simulation, simulation 6, comes in. Under simulation 6, tariffs are also removed on trade between non-members and, therefore, we have global liberalisation where trade between all countries is tariff free. As with the WTO, where members simultaneously reduce tariffs on trade with each other, reciprocity is secured through the ‘most favoured nation’ principle, and therefore incomplete utilisation is not an issue.


The results of our simulations are presented in Table 1. We focus on real-GNP estimates in our discussion because it provides an indication of income available for current and future consumption by members of the economy (Menon (2013) contains more detailed results, including those for individual countries and the ASEAN+3 grouping). The first feature worth noting for all six simulations is how small the numbers are. These magnitudes are not uncommon in computable general equilibrium analyses of trade liberalisation. Furthermore, the size of the numbers reflects the fact that we deal with only goods and not services; focus on removing tariffs only; and do not allow for endogenous productivity improvement caused by the reforms. Therefore, these results should be taken to represent lower-bound estimates. What we focus on are the differences across the six scenarios.

Table 1. Percentage Deviation from Baseline—ASEAN+6, 2020 (%)

The Regional Comprehensive Economic Partnership raises real GNP by US$103 billion (0.42%) with full utilisation and US$32.4 billion (0.13%) with incomplete utilisation. With incomplete utilisation, real GNP is slightly more than one quarter of the full utilisation outcome for the positive results, and slightly less than one quarter for the negative results. This nonlinearity could be attributable to the fact that incomplete utilisation also reduces the extent of trade diversion, and therefore the reduction in welfare. The welfare of all non-members is reduced under both complete and incomplete utilisation of preferences. This is despite global welfare being enhanced in both cases, albeit by very small amounts.

Comparing preferential liberalisation with multilateralisation of preferences, we find that the latter is superior in all cases, and especially when incomplete utilisation is taken into account. When preferences are multilateralised, real GNP increases by US$130.1 billion (0.54%). In general, when members extend their preferential reductions to non-members on a non-discriminatory basis, welfare is enhanced because of three primary effects:

  • The extent of the liberalisation is greater.
  • The broader liberalisation undoes the welfare-reducing trade diversion resulting from the preferential liberalisation.
  • The productivity of scarce resources within each member country is allocated more efficiently across its industries.

Scenarios simulation 4 and simulation 5 are akin to cross-regional tie-ups of bilateral free trade agreements, whereby the ‘ROW’ grouping reciprocates by reducing its tariffs on exports from ASEAN+6. The benefits to members when reciprocity is introduced are greater than simulation 3 only when there is full utilisation of preferences simulation 4. If utilisation is incomplete simulation 5, then members benefit more from multilateralisation of preferences (simulation 3). However, the GNP of ROW falls by 0.15% under S4 (with much smaller declines under simulation 5). These are the largest reductions for non-members under any of the scenarios. The additional gains to members in this scenario, with full utilisation of preferences, appear to occur at the expense of non-members. This raises the potential for possible retaliatory actions by non-members, reducing the benefits to the world as a whole. If the maximum gains to members accrue at the expense of potential retaliatory actions, then the possibility of trade deflection raises the likelihood of low utilisation rates. Since tariffs between large trading blocs such as NAFTA and the EU, and other significant groupings such as South America and Africa, remain unchanged, there are significant opportunities, and benefits, from trying to deflect trade in order to obtain duty-free access

In the final scenario (simulation 6), we consider global liberalisation, which is similar to that of a conclusion of the Doha Round as originally intended. If this were possible, the GNP of all member countries would be increased. For Regional Comprehensive Economic Partnership members, there is little difference in the welfare effects of global liberalisation versus multilateralisation of preferences. This finding has important implications for policy. It suggests that it is very much within the control of member countries to initiate actions that will produce almost the best welfare outcomes from trade liberalisation. There is really no need to wait for an unlikely single undertaking Doha deal for members to reap the benefits from it. It also appears that there is little to be gained from reciprocity, given time delay, negotiating costs, and uncertainty in the magnitude of benefits if utilisation is incomplete.


Bilateral free trade agreements in Asia have been proliferating. Previous studies on the impacts of bilateral free trade agreements in east Asia have assumed full utilisation of preferences. The evidence suggests that this assumption is seriously in error, with estimated uptake particularly low in east Asia. It is not uncommon to find utilisation rates as low as 10%–20%, and rarely above 25%. In this paper, we assume a more realistic utilisation rate of 25%. We find that this significantly diminishes the benefits from preferential liberalisation, but in a nonlinear way. A utilisation rate of 25% reduces benefits by slightly less than 75%, due to reduced trade diversion and terms of trade effects.

Reciprocity is an important motivation for pursuing new bilateral free trade agreements or expanding existing ones over unilateral actions. Proponents of bilateral free trade agreements argue that unilateral actions reduce the bargaining capacity of countries looking to gain greater access to traditional and new markets. We isolate the impact of reciprocity and consider whether the additional benefits that flow from it are likely to be realised. While reciprocity has the potential to impart substantial benefits, this again depends on the extent of utilisation when it is pursued through preferential agreements. At 25% utilisation, multilateralisation of preferences (without reciprocity) still delivers greater benefits to members. Furthermore, the potential for trade deflection combined with possible retaliatory actions may further reduce benefits to members, and to the world as a whole. Multilateralisation of preferences is not subject to either trade deflection or retaliation. Therefore, in the absence of Doha, multilateralisation of preferences is the practical route that is most likely to deliver the greatest benefits to members. The fact that most of the ASEAN+6 countries have already concluded, or are in the process of concluding, bilateral free trade agreements with their main trading partners suggest that the time is ripe to consider multilateralisation- there is little point in holding out for a small residual of countries, especially when full utilisation cannot be assured, but trade diversion and deflection are real risks.


Ando, M (2009), “Impacts of FTAs in East Asia: CGE Simulation Analysis”, RIETI Discussion Paper Series 09-E-037, Tokyo, RIETI, available at

Baldwin, R (2006), “Multilateralising Regionalism: Spaghetti Bowls as Building Blocs on the Path to Global Free Trade”, The World Economy, 29(11), 1451–1518.

Baldwin, R (2008), “Managing the Noodle Bowl: The Fragility Of East Asian Regionalism”, The Singapore Economic Review. 53(3). 449–478.

Mai, Y (2004), “The Monash Multi-Country Model”, CoPS Working Paper No. G-150, Melbourne, Centre of Policy Studies, Monash University.

Menon, J (2007), “Bilateral Trade Agreements”, Asian–Pacific Economic Literature. 21(2).
29–47, available at

Menon, J (2009), “Dealing with the Proliferation of Bilateral Free Trade Agreements”, The
World Economy, 32(10), 1381–1407, available at

Menon, J (2013), “Preferential and Non-Preferential Approaches to Trade Liberalization in East Asia: What Differences do Utilization Rates and Reciprocity Make?”, ADB Working Paper Series on Regional Economic Integration, 109, Manila, Asian Development Bank.

Petri, P, M Plummer, and F Zhai (2011), “The Trans-Pacific Partnership and Asia–Pacific Integration: A Quantitative Assessment”, East–West Center Working Paper, 119, Hawaii.

1 See Ando (2009) for a summary of these studies, and Petri et al. (2011) that assess the Trans-Pacific Partnership (TPP) with incomplete utilisation.



Topics:  International trade

Tags:  bilateral free trade agreements

Lead Economist in the Economic Research and Regional Cooperation Department, Asian Development Bank


CEPR Policy Research