ISSN 1178-2293 (Online)
University of Otago Economics Discussion Papers No. 1001
February 2010
Permanent and Transitory Shocks among Pacific Island Economies - Prospects for a Pacific Islands Currency Union*
Willie Lahari
Address for correspondence: Willie Lahari Department of Economics University of Otago PO Box 56 Dunedin 9054 NEW ZEALAND Email: [email protected] Telephone: +64 3 479 8131 Fax: +64 3 479 8174
*I am grateful to Alfred A. Haug for suggesting this study and comments provided. I am also grateful to Arlene Garces-Ozanne and seminar participants of the Economics Department at the University of Otago, for helpful comments and suggestions. Funding for this research was provided through a PhD scholarship award from the New Zealand Agency for International Development.
Abstract
This paper re-kindles the debate on the feasibility of a Pacific Islands currency union in view of the recent expansion and consolidation of regional strategies and agreements such as the ‘Pacific Plan’ and the Pacific Agreement on Closer Economic Relations Plus. These initiatives, including past efforts, have given limited consideration to the subject for a Pacific Islands currency union. This study exploits the optimal currency area theoretical framework and employs the Gonzalo and Ng (2001) decomposition method. This is the first time this method is used in the analysis relating to currency or monetary unions. Newly-constructed quarterly time series data are also applied. This paper investigates the dynamic effects of permanent and transitory shocks on key macroeconomic variables among Pacific Island countries (PICs). Evidence shows that the proposed union of six PICs (Fiji, PNG, Samoa, Solomon Islands, Vanuatu and Tonga) do not meet most of the preconditions for a union. However, further investigation shows evidence for the Melanesian countries (Fiji, PNG, Solomon Islands and Vanuatu) to possibly form a monetary union, preferably with the Australian dollar as the anchor currency. Nonetheless, further costs in terms of the alignment of policies by Melanesian countries are required.
JEL Classification: C3, C5, E3
Key words: Currency union, Gonzalo and Ng (2001) decomposition, Pacific Island countries
1. INTRODUCTION
In recent years, Pacific Island countries (PICs) and development partners in the region have begun developing new and consolidating past strategies aimed towards closer regional economic integration and growth in the region (Forum Economic Ministers Meeting (FEMM), 2000; Asian Development Bank, 2005). However, the issue for a possible Pacific Islands currency union has been given less attention in recent regional strategies and agreements such as the ‘Pacific Plan’ of 2005, and the newly negotiated Pacific Agreement on Closer Economic Relations (PACER) Plus.1 Although many PICs felt that the issue of a currency union was a challenge to national sovereignty, others, such as Fiji, thought the idea should be considered in the future. Australia was of the view that it was premature to emphasize a formal Pacific union given that PICs faced a period where further integration was required prior to any prospects for a Pacific union. While the strategies towards greater regional integration would help address regional problems that include sluggish economic growth, rising unemployment and negative external shocks (e.g., volatility in export commodity prices, and cyclones), such efforts can be seen as a necessary platform on which to re-kindle the debate on the prospects for a Pacific Islands currency union.
There are 22 Pacific Island countries and territories (PICTs) of which only six Pacific Island countries (PICs), namely, Fiji, Papua New Guinea (PNG), Samoa, Solomon Islands, Tonga and Vanuatu are fully independent sovereign states.2 This study will focus on these six PICs. These countries represent about 86% of the combined total population of all the PICTs and about 36% of the total combined gross domestic product (GDP), including 95% of the total land area of the 22 PICTs combined. The choice of these six PICs is relevant to this study given that these PICs have their own currencies (legal tender), exchange rate regimes and central banks. Hence, these six PICs determine their own domestic monetary (and fiscal) policies. Moreover, these six PICs share many common economic and physical characteristics, such as smallness in terms of GDP and population, narrow productive sectors, limited export diversification, geographic isolation, vulnerability to terms of trade fluctuations and frequent natural disasters that are important considerations for a possible currency union (Browne and Orsmond, 2006).
1 Other regional agreements include the Pacific Islands Trade Agreement (PICTA), the Pacific Agreement on Closer Economic Relations (PACER) signed in 2002, and the Melanesian Spearhead Group (MSG) Preferential Trade Agreement of 1993. 2 The 22 PICTs are grouped into three sub-regions of Melanesia (west Pacific), Polynesia (southeast Pacific) and Micronesia (north Pacific), on the basis of their ethnic, linguistic and cultural differences. The Melanesian sub-region consists of Fiji, New Caledonia, PNG, Solomon Islands and Vanuatu. The Polynesian sub-region consists of American Samoa, Cook Islands, French Polynesia (Tahiti), Niue, Pitcairn Islands, Samoa, Tokelau, Tonga, Tuvalu and Wallis and Futuna. The rest consist of Micronesian nations and territories 1
During different phases of economic development since the 1980s, economic (GDP) growth in the Pacific region has been subdued mainly due to negative external and internal shocks.3 Whether the effects of these shocks are asymmetric or not, and how PICs respond to these shocks will have implications for the feasibility of a currency union. Although inter-regional trade flows among the PICs, apart from Australia and New Zealand, has been low, trade has generally increased in the last two decades especially among the Melanesian countries.4 This has been attributed to increased regional efforts such as Pacific Island Countries Trade Agreement (PICTA) that have contributed to strengthening cooperation and efforts to increase trade and regional growth among PICs, Australia and New Zealand. In particular, the Melanesian countries have benefited from inter-country trade both as members of PICTA and their own Melanesian Spearhead Group (MSG) agreement. The continued success or failure of these current efforts will also impact on future considerations for a currency union.
This analysis draws from the Optimal Currency Area (OCA) theory flowing from the seminal contributions of Mundell (1961), McKinnon (1963) and Kenen (1969). A key property of the OCA refers to the asymmetry of shocks. As Mundell (1961) and McKinnon (1963) argued, countries interested in a currency or monetary union will have to be part of a zone where a single currency circulates or, if there are several exchange rates, their exchange rates are fixed to each other permanently. In such a zone, a single monetary authority implements a common monetary policy at the union level. Hence, governments will have limited control over the use of monetary and exchange rate policies to respond to country-specific shocks. A common union-wide response would be effective if the shocks were symmetric in nature across union member countries. However, if this was not the case, and some countries faced asymmetric shocks and were worse-off than the other countries within the same zone, then it would not be beneficial for these countries to forego domestic policy autonomy and form a union. Generally, a currency union refers to a zone consisting of several countries or regions where a single exchange rate regime prevails, a single currency circulates, and where a single monetary authority implements a common monetary policy. A monetary union refers to a group of countries that agree to permanently fix their exchange rates under centralisation of monetary authority. A monetary union may not necessarily be fully centralised or have formal integration but there are commitments for monetary policy coordination among members through arrangements such as currency boards. For the general purpose of this discussion, the term ‘currency union’ will be
3 Examples include the global oil crisis in the 1980s, the decline of major agricultural commodity prices in the mid-1980s, the Asian financial crisis in 1997 and natural disasters (e.g., cyclones). Internal factors relate to civil unrests such as the coups in Fiji and the lack of good governance among many public institutions within the PICs. 4 For instance, from 1990 and 2005, the Solomon Islands imports from PNG increased from 2.8% to 4.1%, Vanuatu imports from Fiji increased from 0% to 6%, and Tonga’s exports to Fiji doubled. 2 used interchangeably to refer to either a currency or monetary union, or both, unless otherwise specified.
The study will employ the method of Gonzalo and Ng (2001) in analysing the dynamic effects of permanent (P) and transitory (T) shocks, denoted as P-T shocks, in a system of key macroeconomic policy variables. Initial investigations in assessing the degree of convergence (cointegration) of the variables will also be undertaken. Specifically, this analysis will examine the extent of (a)symmetry in the behaviour of P-T shocks, and associated costs and benefits as preconditions for PICs forming a currency union. This is undertaken in terms of evaluating the degree of similar (dissimilar) behaviour of P-T shocks pertaining to the following conditions: (a) dominance of either a P or T shock from variance decomposition; (b) direction of the effects of P- T shocks; (c) magnitude of P-T shocks upon initial impact; (d) persistence of P-T shocks, and (e) correlation of P-T shocks. This study will draw insights from recent empirical studies on consumption and wealth that have applied the Gonzalo and Ng (2001) methodology such as Lettau and Ludvigson (2004), Chen (2006), Kishor (2007), de Veirman and Dunstan (2008), and Giancarlo and Konstantinou (2009), but applied instead to the analysis of currency unions. The application of Gonzalo and Ng’s (2001) decomposition to the study of currency or monetary unions is novel. To my knowledge, there are no studies on currency unions that have applied Gonzalo and Ng’s (2001) methodology.
2. RELATED STUDIES FOR THE PACIFIC
There is clearly a lack of comprehensive analysis on currency unions in the Pacific region. Earlier studies centred on appropriate currency regimes, such as Xu (1999) and Duncan and Xu (2000) who argued that PNG should adopt the Australian dollar given the lack of central bank independence during the 1990s and the likelihood of monetary indiscipline that may arise due to internal or external interference. They argued that the economy would perform better with the Australian dollar. Similarly de Brouwer (2000) supported the idea that PICs should adopt the Australian dollar, whilst Duncan (2002) argued more in favour of the Solomon Islands adopting the Australian dollar. However, the arguments of Duncan (2002) and de Brouwer (2000) were policy arguments and lacked advanced empirical analysis. On the other hand, Hughes (2003) advised against any form of currency union in the Pacific prior to real reforms being undertaken. Similar views were expressed by Fichera (2006) and Creane et al., (2006). However, at about the same time that Hughes’ (2003) work was being debated, the Australian Parliamentary Senate Committee (2003) proposed a Pacific economic and political community as a panacea to the economic problems and political instability in the Pacific region. The discussions did not go far,
3 with little debate due to a general lack of interest among political leaders in the region and negative assessments for a currency union (Chand, 2003).5 Furthermore, studies such as those of Jayaraman (2001, 2003, 2004), Bunyaratavej and Jayaraman (2005), Bowman (2005) and Browne and Orsmond (2006) argued that it was premature for PICs to form a currency union given the dissimilarities in the behaviour of macroeconomic policies and key measures such as inflation and exchange rates. A number of these studies, such as Jayaraman (2001, 2003), have been criticised in connection with the application of nominal and effective exchange rates given the varying structure of the weights and issues of endogeneity of currency unions and trade (Duncan, 2005). Bowman’s (2005) test for the relationships of currencies alone was not considered an appropriate test for an OCA given that all the six PICs’ currencies were fixed to individual baskets, except for PNG whose currency is freely (managed) floated, and thus the tests do not assess the extent to which these PICs and Australia share industry-specific shocks (Duncan, 2005). Other related studies focused mainly on regional co-operation and integration such as those by the Asian Development Bank (2005), and Australian Agency for International Development (2008). From the above studies, it is obvious that there is lack of consensus on the issue of currency union for PICs given the mixed results and conflicting arguments.
3. DATA AND EMPIRICAL METHODOLOGY
This section discusses the data aspects of the analysis and the model employed. The discussion of the data will include the newly-constructed series for the PICs and the choice of variables. The Gonzalo and Ng (2001) decomposition method will be discussed in the subsequent sub-section.
3.1 Data
Time series data for key macroeconomic variables from existing data sources and newly- constructed time series data at a quarterly frequency are employed in this analysis. Given the time series requirements for this study and the general lack of timely and reliable quarterly data, newly-constructed series for relevant variables were derived.6 The time series are mainly from 1980:1 to 2006:4. However, a number of variables had relatively shorter time spans such as the
5 Many PICs were not interested in the idea mainly due to issues of loss of sovereignty (relating to surrendering monetary authority and monetary policy). Also, the PICs’ non-involvement in the initial process of the Australian Parliamentary Senate Committee’s recommendation, implied that PICs did not having ownership of the idea. PICs viewed that the idea was more attune to the interests of Australia and to some extent, New Zealand (see, e.g., Chand, 2003; Duncan, 2005). 6 See Lahari et al. (2008) and the Secretariat of the Pacific Community (2003, 2005, 2007) for discussions about the data issues and problems in the Pacific. 4
GDP series for New Zealand that started from 1987:2 to 2006:4.7 The new data includes the quarterly real GDP series for the PICs constructed by Lahari et al. (2008). The other newly- estimated series include the fiscal (government budget) balance to GDP ratio (FGDP), and the public debt to GDP ratio (DebtGDP). The methodology used by Lahari et al. (2008) is based on the Chow-Lin (1971) framework, with modifications proposed by Fernandez (1981), Litterman (1983), and Abeysinghe and Rajaguru (2004). Following the same approach as in Lahari et al. (2008), the estimation of quarterly series for FGDP and DebtGDP have been undertaken. Data descriptions are provided in Table A2 in the Appendix. The other variables employed include consumer prices (CPI), interest rates (IRATE) and the money supply (M2) whose data descriptions are provided in Table A1 in the Appendix. For our purposes, newly-constructed trade weighted nominal exchange rate indices (TWIs) for each of the PICs were used instead of the exchange rates.8 The TWI for each PIC was constructed following the method used by the Reserve Bank of New Zealand (RBNZ).9 The TWIs are compiled for an independent country currency basis and in the counterfactual monetary union with Australia as anchor currency 10,11 denoted as TWI-AUD, and New Zealand as anchor denoted as TWI-NZD respectively. The latter case involves a hypothetical case to evaluate the choice among PICs for adopting the currency of either one of their closest major trading partners, Australia or New Zealand.12 The details of the TWIs and descriptions are presented in Table A1 in Appendix.
The choice of variables was essentially based on the criteria of the Maastricht Treaty of 1991 (European Union, 2006), applied in the context of the PICs.13 However, the focus in this study is on the underlying policy goals of the Maastricht Treaty, as measured by the variables, and not the
7 The time series data for New Zealand, mainly GDP and fiscal (government budget) balance, started from 1987:2 due to the massive financial reforms in the prior years. Hence, where these variables are analysed within a group of countries including New Zealand, the time period will be based on the series for New Zealand. 8 The use of the TWI based on a common weighting structure of currencies represent a better measure of the degree of economic competitiveness of a country relative to its major trading partners. 9 See the RBNZ website, http://www.rbnz.govt.nz/statistics/exandint/twi/index.html 10 Studies such as Scrimgeour (2001, 2002) have compiled similar TWIs in assessing New Zealand’s experience in a currency union with its major trading partners. 11 Given the lack of consensus on the impact of a currency union on trade, and the controversies around Rose’s (2000) findings, weights are held fixed. Similar arguments have been raised by Scriemgor (2001, 2002) and Grimes, Holmes and Bowden (2000). However, in this analysis, the weights are based on the average value of the total merchandise trade and nominal GDP from 1980 to 2006 to avert any bias from the selection of single- year weights (see, e.g., Sawides, 1996). 12 The nominal TWI is employed throughout this study given our interest on the impact of P-T shocks on the economy. In such a scenario, it is the nominal exchange rate (TWI) that acts as the immediate buffering variable. See also arguments by Grimes, Holmes and Bowden (2000). 13 The Maastricht Treaty criteria, often referred to as the convergence criteria, were based on five key economic pre-conditions for entry into the European Monetary Union. These includes: (i) a country’s budget deficit to GDP ratio should not exceed 3% per annum; (ii) a country’s total public sector public debt to GDP ratio should not exceed 60% per annum; (iii) a country’s exchange rate must be contained within the required fluctuation margins of 2.25% of the EU’s exchange rate mechanism (ERM); (iv) a country’s rate of inflation must not exceed 1.5% above the average rate in the three lowest inflation EU countries, and (v) a country’s long-term interest rates should not exceed 2% above the average of those in the three lowest rate EU countries. 5
specific targets set by the treaty.14 This is relevant in respect to setting a benchmark for comparison for PICs which intend to form a currency union.15 In this context, the measures of FGDP and DebtGDP are seen as key indicators for fiscal policy. The other variables such as the CPI, exchange rate (TWI) and IRATE, are regarded as key measures for monetary policy, although the CPI is also a key measure in assessing fiscal policy. GDP is an output measure of economic performance influenced by both fiscal and monetary policies. The extent to which a group of countries’ policies behave and react to shocks has implications on forming a currency union.
Cluster of Potential Currency Union Groups
The six PICs (Fiji, PNG, Samoa, Solomon Islands, Tonga and Vanuatu) together with Australia and New Zealand are clustered into potential union blocks on the basis of their prevailing trade agreements, culture, language and historical connections. Hence, the following potential groups are proposed: Group 1: Pacific Only (Fiji, PNG, Samoa, Solomon Islands, Tonga and Vanuatu); Group 1A: Pacific with Australia; Group 1B: Pacific with NZ, Group 2: Melanesia (Fiji, PNG, Solomon Islands and Vanuatu); Group 2A: Melanesia (Group3) with Australia; Group 2B: Melanesia with New Zealand. The challenge was to first examine the possibility of countries in Group 1 (the Pacific group) forming a currency union. Evaluating initial prospects for uniting the PICs in a union would help distinguish possible prospects for PICs (least developed/developing economies) first, prior to further assessment with Australia and/or New Zealand (advanced/developed economies).16 Should this hold, then a further examination to form a union with Australia and/or New Zealand is undertaken. If not, then out of the Group 1 countries, a group of Melanesian countries (Group 2) is investigated separately.17 If forming a union is
14 The pre-conditions (or targets), other than the set of measures (variables) are specific for EU countries. The targets are ambitious and may not be reachable by many PICs given the extent of heterogeneity between the economies of the EU (mainly advanced developed economies) compared with PICs (least developed/developing economies). For example, the annual inflation rates for Fiji was 4.2% (2003) and 2.3% (2005) and for the Solomon Islands, 10.3% (2003) and 6.8% (2005). These rates exceed the ‘less than 1.5% of the average’ of the three lowest EU countries. 15 A number of studies (e.g., MacDonald and Taylor, 1991; Hafer and Kutan, 1994; and Haug, 2001) have analysed macroeconomic policy convergences employing similar variables to the Maastricht Treaty. 16 While inter-county trade flows among PICs may be relatively low, except with Australia/NZ, trade has increased among PICs. Thus, arguments against prospects for a union on the basis of low trade alone are debateable because forming a currency union can lead to substantial increases in trade (Rose, 2000), although this is controversial. More importantly, the issue of trade is only one aspect of OCA theory. Among other key OCA properties, the asymmetries of shocks are main concerns for this study. 17 The Melanesian countries of the South Pacific are a sub-regional group established as the Melanesian Spearhead Group (MSG) in 1993. Recently, the MSG Secretariat was established in 2008 to implement the regional developmental agenda of the Melanesian countries. 6
feasible for Group 2 then further investigation for prospects for Group 2 countries forming a currency union with Australia (Group 2A) and/or New Zealand (Group 2B) will be undertaken.18
3.2 VECM MODEL
The Vector Error Correction Model (VECM) incorporates cointegration restrictions on the vector autoregression process in a similar methodological framework to that of Gonzalo and Ng (2001). The Gonzalo and Ng (2001) method provides a systematic framework in identifying and decomposing the permanent and transitory (P-T) shocks among key variables of interest. Thus,
the VECM representation of Zt is presented in the reduced form as:
ΔZt = γα′Zt−1 + Γ(L)ΔZt−1 + et (1)
where ΔZt represents first differences of a 1 vector of I(1) process of Zt; is the number of cointegrated vectors in the system with , and 1, where is the number of variables; is the matrix of adjustment coefficients or error-correction parameters; consist of the cointegrating vectors, and Γ(L) is the lag polynomial.
Permanent and Transitory Decomposition
Following from Gonzalo and Ng (2001), we can generally represent the final-form of the P-T shocks (orthogonalised) from equation (1) as:
~ ~ ΔZ t = D(L)ηt (2)
~ ~ where D(L) is a lag polynomial, and ηt is a 1 vector of the transformed orthogonalised P-T ~ shocks. The ηt s are serially and mutually uncorrelated, and have unit variance. According to
~ p Gonzalo and Ng, the 1 vector of shocks, ηt , are the common stochastic trends or