Myanmar's trade and its potential.
Myanmar's incipient transition to a more stable, open political system and economy has moved it into the spotlight of international attention. However, years of insulation from international reporting standards and obligations have left the public domain with a scant record of economic data on Myanmar, including international trade statistics. Partly as a result of data limitations, the country's trade profile and an assessment of its potential to emulate the success of other countries in the region has received little attention in the literature.
This paper seeks to fill this gap by describing Myanmar's trade profile based on partner countries' records in the United Nations COMTRADE database. It then fits an augmented gravity model to a panel data set of 2000-10 yearly exports by six member countries of the Association of Southeast Asian Nations (ASEAN) to key destinations in Asia and the rest of the world. The estimates are used to predict, out of sample, Myanmar's exports to these destinations in a scenario that assumes the country having the same market access as do other member countries of ASEAN, absent any special sanctions or restrictions to trade with Myanmar. Actual and predicted exports are then plotted and summarised for comparison.
The paper is structured for section II to provide an overview of Myanmar's trade pattern and for section III to discuss the gravity regressions and predictions for exports by Myanmar. Section IV concludes.
II. Trade Profile
Myanmar records of bilateral merchandise trade in the COMTRADE database are missing for some years and incomplete for others. To provide a trade profile for Myanmar despite data limitations, it is derived here from trading partners' records of cross-border transactions with Myanmar at the three digit level of the Standard International Trade Classification (SITC).
In terms of current U.S. dollars, Myanmar's goods exports rose from roughly US$500 million in 1990 to US$2 billion in 2000 and to more than US$6 billion in 2010. Imports in 2010 were nearly US$9 billion, up from US$2.5 billion in 2000 (Figure 1). A sizeable deficit on the 2010 merchandise balance resembles deficits of a similar magnitude during most of the 1990s. In real terms, that is deflated by a world trade price index, Myanmar's exports and imports both expanded by a factor of 4.4 between 1990 and 2000, and more than doubled during the 2000s. (1)
Myanmar's output and per capita income expanded considerably faster than its trade with the rest of the world. Its GDP expanded from roughly US$9 billion (6 trillion kyat) in the late 1990s and early 2000s to more than US$50 billion (21 trillion kyat) in 2010 (Figure 2). Annual real GDP growth in local currency units averaged more than 12 per cent between 2000 and 2010. (2) In terms of constant U.S. dollars, thus reflecting the market exchange rate of the kyat to the U.S. dollar, per capita GDP expanded nearly fourfold, from US$178 in 2000 to US$675 in 2010. During the same period, Myanmar's population expanded by about one-fifth, to nearly 60 million in 2010 (Figure 3).
Thailand, India and the People's Republic of China (PRC) account for more than three quarters of Myanmar's (cumulated) exports between 2006 and 2010 (Table 1). (3) Nearly half the country's exports are destined to Thailand alone. PRC, Thailand and Singapore together account for nearly three-quarters of Myanmar's imports. More than one-third of imports are sourced from PRC alone.
Myanmar's export basket is heavy in fuels (natural gas), food and other primary commodities (including precious stones and gems), which together constituted nearly 90 per cent of total exports between 2006 and 2010 (Table 2). By contrast, more than 70 per cent of Myanmar's imports are manufactured goods.
The composition of exports varies by trading partners. Exports to Thailand, Myanmar's largest market by far with a share of 48.4 per cent, are almost entirely fuels (natural gas), whereas exports to India and PRC are mostly food and agricultural commodities. By contrast, Myanmar exports predominantly low-skill manufactured goods to industrialized countries such as Japan, Germany and the United Kingdom.
Manufactured goods constitute more than 90 per cent of Myanmar's imports from the PRC, the Republic of Korea and Japan. Somewhat lower, in the range of 40 per cent to 60 per cent, is the share of manufactured goods imports from Thailand, Singapore, Malaysia and Indonesia, which are also key suppliers of food and petrol (fuels).
III. Myanmar's Export Potential
Earlier applications of the gravity framework to evaluating the export potential of countries (4)--mostly with a focus on Central and Eastern Europe --include Baldwin (1994); Gros and Gonciarz (1996); Briilhart and Kelly (1998); Egger (2002); International Trade Centre (ITC; 2003); and Papazoglou, Pentecost and Marques (2006). In this section, Myanmar's export potential is predicted by a gravity model fitted to the exports of six ASEAN member countries other than Myanmar: Cambodia, Indonesia, Malaysia, the Philippines, Thailand and Vietnam. Among the ASEAN members, not included are Laos, due to data constraints, as well as Brunei Darussalam and Singapore, whose economic and geographic characteristics are largely incompatible with a gravity empirical setting/ The 2000-10 panel data include yearly bilateral exports between the selected ASEAN countries and their top thirty-five trading partners. Included are all the world's leading trading nations, both industrialized and emerging (see Annex 1 for a list of countries).
Exports by the group of six ASEAN member countries provide a relevant counterfactual scenario for Myanmar's trade potential. For it implicitly assumes that trading partners will grant Myanmar the same access to their markets as were enjoyed by the other ASEAN countries during 2000-10 on average. Such a scenario thus envisages the effects of Myanmar trade being freed of all sanctions and restrictions above those facing ASEAN countries as a group.
The empirical strategy is to estimate a gravity equation, which applied to panel data is most simply stated as: (6)
[x.sub.ijt] = exp([ß.sub.0] + a.sub.ijt])[g.sup.ß.sub.ijt][u.sub.ijt], (1)
[u.sub.ijt] = [µ.sub.ij] + [v.sub.ijt] (2)
where [µ.sub.ij] the unobserved bilateral effect and [v.sub.ijt], is the remaining error.
Assuming fixed effects, the specification in (1) and (2) can also be stated as:
E [[t.sub.ijt] | [g.sub.ijt]] = exp([ß.sub.0] + [?.sup.n.sub.d][?.sub.d][d.sub.ijt])[g.sup.[ß.sub.ijt] (3)
where heterogeneity across n country pairs is absorbed by a corresponding number of dichotomous variables [d.sub.ijt] ? (1, 0) taking co-efficients [?.sub.d].
In the analysis to follow, the regressand [x.sub.iji] is bilateral exports (in values deflated to 2000 U.S. dollars) and the vector of regressors [g.sub.ijt] includes the fundamental gravity variables (GDP, GDP per capita, distance) as well as a number of dichotomous variables capturing whether or not country pairs trading with each other have in common a border, a colonial history, a language spoken, or have in effect a free trade agreement.
The core regression in this paper applies ordinary least squares, allowing for year- and dyad-specific fixed effects by pooling and clustering observations. What is termed here a pseudo-fixed effects approach (PSEUDOFE) circumvents the problem of fitting a fixed-effects model proper in the presence of time-invariant dichotomous variables--such as distance--that are of course perfectly collinear with the country-pair individual effects.
The inclusion of dyadic dummies is of particular importance in the context of a gravity model used to estimate trade potential. Absent such dummies, earlier studies inaccurately interpreted large residuals from gravity regression as an indication of high trade potential when in fact it was model misspecification that lead to large systemic residuals (Egger 2002). Moreover, the inclusion of dyadic dummies avoids the need for gravity models to include a multilateral trade resistance term in order to avoid model misspecification and biased estimates of bilateral resistance terms, such as geodesic distance (Anderson and van Wincoop 2003).
The robustness of estimates is checked against alternative model specifications. (7) A generalized least squares random effects estimator (REGLS) assumes that disturbances across panels ij are not identically distributed and that errors [u.sub.ijt] are serially correlated. Essentially, this is achieved through the application of the Huber-White sandwich estimator (Stock and Watson 2008). A feasible generalized least squares estimator (FGLS) is used to allow for heteroscedasticity and AR(1) autoregressive correlation specific to each panel, rather than assuming it as a common feature across country pairs. Finally, an unconditional fixed-effects Tobit estimator (TOBIT) is deployed to address the fact that the observations in the panel dataset are truncated due to missing observations in relation to dyads with zero trade flows in any given year, which may cause bias in the other approaches. (8)
Turning to the results, the elasticity coefficients of variables included in the core specification (PSEUDOFE) have the expected sign and are statistically significant (Table 3, column 1). For example, a country with a GDP 10 per cent larger than average is estimated to be exporting 15 per cent more than average, all else the same. Exports will expand by nearly 9 per cent on average if the importing country's GDP increases by 10 per cent. Moreover, exports will respond positively to: higher GDP per capita at home as well as abroad; sharing borders with trading partners; some sort of colonial relationship past or present; or a preferential trade agreement in effect. By contrast, the negative sign of the estimated distance coefficient means that such trade costs significantly deter trade.
Against expectations and the typical gravity findings, the common language coefficient has a negative sign. This is a selection effect, due to the language-ethnography dummy reflecting the Philippines' exports to countries such as the United States, Canada, Australia, United Kingdom, Pakistan, or Egypt, which are lower than average ASEAN exports to these countries. (9)
An adjusted [R.sup.2] of 0.8 indicates an especially tight fit of the model to the data, which is important especially in an empirical context where trade potential is derived from estimated regression residuals. Furthermore, the results are robust to changes in the model specification; the second and fourth columns yield coefficients that are consistent with the core model in the first column in terms of magnitude, sign and statistical significance. The coefficients on the free trade agreement dummy in the GLS random effects (2) and Tobit (4) models are an exception, as significance is beyond the conventionally accepted thresholds. However, the magnitudes of point estimates are confirmed across columns, which attest to the fundamental robustness of the core estimates in the first column. (10)
The elasticities estimated by the gravity model for the six benchmark exporters and thirty-five importers form the basis for predicting Myanmar's export volumes. Such prediction is said to be out-of-sample, as observations on Myanmar were excluded from the benchmark regression and only now are being combined with the elasticities previously estimated. Specifically, Myanmar's potential or notional export values are the values fitted to its GDP between 2000 and 2010 jointly with all the other gravity regressors listed in the first column of Table 3. Fitted or "gravity" values of Myanmar's exports can thus be plotted against actual bilateral exports to key markets from 2000 to 2010, in lieu of a counterfactual envisaging Myanmar as having had the same relative ability to export and facing trade restrictions no higher than did the six benchmark ASEAN countries during the period of reference.
Figure 4 shows that Myanmar's actual exports exceeded gravity exports until 2007. Low gravity exports reflect Myanmar's lesser income relative to the six reference countries, and actual exports were driven by high exports to a few single countries (notably natural gas to Thailand). After 2007, Myanmar's real output growth accelerated, raising its export potential in terms of the gravity predictions. Since Myanmar's exports were largely unresponsive to accelerated GDP growth, expanding only moderately, potential trade outgrew actual exports, by a factor of more than four by 2010.
The weakest exporter among the sample of seven ASEAN countries considered in this analysis, Myanmar is seen ranking lowest in terms of the ratio of actual to gravity exports (Table 4). During 2006-10, this ratio averaged only 0.38 across Myanmar's trading partners. Weighing outliers, such as exports to Thailand, the geometric mean of Myanmar's actual-to-potential exports ratio turns out even lower, at 0.15. That is, Myanmar exploited only 15 per cent of its (gravity) export potential, on average, in the five years to 2010.
By contrast, Table 4 shows Vietnam to have outstripped gravity exports by an average factor of 3, or 2.4 when outliers are weighed through geometric rather than arithmetic mean computation. (11) Roughly in line with predictions are exports by Malaysia, Thailand, Cambodia and Indonesia, when controlling for outliers such as Cambodia's disproportional exports volume to the United States. Besides Myanmar, only the Philippines' underwhelming export performance stands out against the group of successful ASEAN exporters. (12)
Table 5--jointly with the time series plots of Myanmar's actual vs. gravity exports in Annex 2--shows that the country's exports during 2006-10 fell substantially short of potential for all destinations but Thailand, Vietnam and Pakistan. Myanmar's actual/gravity ratio in relation to Thailand exceeds 3, and for the other two markets is about 1.5. While trade with all three countries contributes positively to closing Myanmar's gap to potential, only Thailand does so significantly, accounting for nearly 31 per cent of the total difference between potential and actual exports.
Among the developing countries, PRC is the export destination with the largest unexploited potential for Myanmar; 2006-10 exports to PRC were about one-third their potential level and constituted nearly 20 per cent of Myanmar's total gap to potential. Other destinations in developing Asia with a large unused potential are Malaysia (with a ratio of 0.50 and a share of 2.7 per cent) and India (0.79 and 4.3 per cent). Due to exceptionally high natural gas exports to Thailand, developing countries jointly account for less than 10 per cent of Myanmar's export gap to potential. More than 90 per cent of it is on account of the industrialized countries, particularly Japan (share of 46 per cent), Europe (22 per cent), and the United States (17 per cent). (13) Myanmar's gradual integration with the world economy and normalized access to the Japanese, European and American markets may thus be expected to fill its exports gap at least in part.
The analysis in the paper points to Myanmar's vast unexploited trade potential, the bulk of which is determined by weak trade with industrialized countries. Myanmar's gradual integration with the world economy and normalized, unsanctioned access to the European and American markets may thus be expected to fill this gap at least in part.
Clearly, the notion of gravity potential only weakly relates to a country's true export potential, which depends on a range of variables beyond the narrow scope of the gravity framework. On the domestic front, Myanmar's trade potential will be defined by its capacity to mobilize its resources through reforms aimed at maintaining a stable macroeconomic environment, investing heavily in infrastructure and human capital, reforming its financial sector and the foreign exchange regime. On the international front, of key importance to Myanmar as a trading nation will be the course taken by the future adjustment of regional production and trade networks to a changing global economic environment. While the regional reallocation of production processes may well open up new opportunities for latecomers like Myanmar, the challenges from tighter competition within the region against the backdrop of a possible re-shoring of production towards the industrialized countries may well pose hurdles to Myanmar's prospects to diversify its economy away from exclusive reliance on primary commodities and low-skill manufactures exports.
Notwithstanding these challenges and assuming that necessary economic and political reforms will move ahead, Myanmar's transition towards normalized trade with the rest of the world will benefit from its favourable predisposition to international trade; the country is a natural trade and transportation hub for the entire region, nestled between the PRC and India, with well-established trade channels also to Japan, Thailand, Malaysia and Singapore, and with access to the Indian Ocean shipping lanes. A boost to exports will likely derive also from the expansion of a tourism industry that can build on the country's rich natural and cultural heritage.
ANNEX 1 List of Countries Included in the Gravity Regressions EXPORTERS (as well as IMPORTERS): ISO Country IDN Indonesia KHM Cambodia MYS Malaysia PHL Philippines THA Thailand VNM Vietnam IMPORTERS: ISO Country ARE United Arab Emirates ARG Argentina AUS Australia BEL Belgium BGD Bangladesh BRA Brazil CAN Canada CHN People's Rep. of China CZE Czech Republic DEU Germany EGY Egypt ESP Spain FRA France GBR United Kingdom IND India ITA Italy JPN Japan KAZ Kazakhstan KOR Republic of Korea LKA Sri Lanka MEX Mexico NLD Netherlands NZL New Zealand PAK Pakistan RUS Russian Federation SAU Saudi Arabia TUR Turkey USA United States ZAF South Africa Note: The regressions include exports from 2000 to 2010.
Myanmar's Actual and Gravity Exports
ADB. Key Indicators for Asia and the Pacific. Manila: Asian Development Bank, 2012 <http://www.adb.org/ publications/key-indicators-asia-and-pacific-2012>.
Anderson, James E. and Eric van Wincoop. "Gravity with Gravitas: A Solution to the Border Puzzle". American Economic Review 93, no. 1 (2003): 170-92.
Baldwin, Richard E. Towards an Integrated Europe. London: Centre for Economic Policy Research, 1994.
Bergeijk, Peter A.G. van and Steven Brakman, eds., The Gravity Model in International Trade. Cambridge: Cambridge University Press, 2010.
Briilhart, Marius and Mary Kelly. "Ireland's Trading Potential with Central and Eastern European Countries: A Gravity Study". Economics Technical Papers 9815, Trinity College Dublin, Department of Economics, 1998.
Egger, Peter. "An Econometric View on the Estimation of Gravity Models and the Calculation of Trade Potentials". The World Economy 25, no. 2 (2002): 297-312.
Gros, Daniel and Andrzej Gonciarz. "A Note on the Trade Potential of Central and Eastern Europe". European Journal of Political Economy 12, no. 4 (1996): 709-21.
IMF. World Economic Outlook Database. Washington, D.C.: International Monetary Fund, April 2012 <https://www. imf.org/external/pubs/ft/weo/2012/01/weodata/index.aspx>.
ITC. TradeSim (second version)--A Gravity Model for the Calculation of Trade Potentials for Developing Countries and Economies in Transition-Explanatory notes. Geneva: International Trade Centre (United Nations Conference on Trade and Development/World Trade Organization), 2003.
Papazoglou, Christos, Eric J. Pentecost, and Helena Marques. "A Gravity Model Forecast of the Potential Trade Effects of EU Enlargement: Lessons from 2004 and Path-dependency in Integration". The World Economy 29, no. 8 (2006): 1077-89.
Stock, James H. and Mark W. Watson. "Heteroskedasticity-Robust Standard Errors for Fixed Effects Panel Data Regression". Econometrica 76, no. 1 (2008): 155-74.
I benefited from assistance with the data by Cindy Petalcorin and Eileen Capilit, and from comments by Martin Bodenstein, Enrique Galan, Eugenia Go, Arief Ramayandi, and to the participants of the ERD Seminar Series in Manila, 24 October 2012. Any remaining errors are my own.
(1.) Exports and imports values in this study are deflated by an index derived from World Trade Monitor data maintained by the Netherlands Bureau for Economic Analysis. The kyat/US$ exchange rate was highly volatile over the period of observation and the various sources tracing Myanmar's exchange rate do not fully overlap (see note 2 about data sources used). Figures 1-3 show data in both current and constant terms.
(2.) GDP data were drawn from ADB's Key Indicators for Asia and the Pacific (2012) and the IMF World Economic Outlook (WEO) Database (April 2012). To overcome data limitations, constant GDP series are constructed from real GDP growth rates. US$-denominated series are derived from GDP in local currency units (kyats) and the implicit market exchange rate, extrapolated from IMF WEO data. Constant US$ GDP series are deflated by the U.S. GDP deflator. Per capita GDP series are derived by combining GDP and population series.
(3.) Tables 1 and 2 list countries with at least 1 per cent share of imports or exports out of Myanmar's total imports and exports. The commodity aggregates include the following SITC categories: Food (0+1+22+4); Fuels (3); Agricultural (non-food) commodities (2-22+27+28+667+971); Manufactured goods (5+6-667-68+7+8).
(4.) A country's export potential depends on a broad array of factors related to its availability of resources, competitiveness and stage of development. As such, export potential mainly relates to supply factors, rather than international demand, which falls more directly within the realm of gravity regressions. Although the term "export potential" has been widely adopted in the gravity literature and also in this paper, it should be emphasized that it merely denotes estimates of export opportunities as explained by the gravity explanatory variables.
(5.) In particular, Singapore's exceptionally high ratio of trade to GDP runs counter to premising trade intensity on economic size, as does the gravity model. Similarly, the small size of Brunei Darussalam's economy and the narrow features of its trade are largely incompatible with the thrust of the gravity model.
(6.) The interested reader is referred to Bergeijk and Brakman (2010) for a basic exposition of the gravity model and its applications.
(7.) It should be noted that correlation coefficient in the sample between GDP and GDP per capita is 0.487, which points to a moderate degree of multicollinearity fully that falls within the accepted norm in the context of gravity estimation.
(8.) Truncation to a degree is unavoidable in this empirical setting, as zero trade flows are incompatible with logarithmic specification, since undefined. However, the likelihood of a serious bias appears sufficiently remote here, as the incidence of zero trade flows in the squared data set represents less than 1 per cent of total observations.
(9.) Additional robustness tests--omitted here for brevity but available from the author on request--show that the core coefficients estimated are robust to the exclusion of the language dummy.
(10.) It should be noted that this is true also for the estimated coefficient of the constant term, the accuracy of which is of considerable importance when trade flows are predicted. Here, the -38.34 point estimate of the constant term is associated with a standard error of 2.53, which indicates an exceptionally high degree of accuracy hence reliability of predictions.
(11.) Clearly, this does not imply that Vietnam's exports are too high and it should reduce its export volume. Rather, the results indicate that, compared to the reference countries, Vietnam' economy is strongly export oriented. Over the longer term, the country will have scope to reorient some of its economy toward internal demand and lower its reliance on external demand.
(12.) The inclusion of services data in the analysis would likely lead to some adjustment in the findings, particularly for countries with substantial services exports, such as the Philippines.
(13.) The industrialized countries in the data set comprise the United States, Canada, Japan, Republic of Korea, Germany, United Kingdom, France, Italy, Spain, Belgium, Czech Republic and Australia. It should be noted that jointly these countries account for more than 90 per cent of the differential between potential and actual trade, also because Thailand closes nearly 31 per cent of that gap and thereby significantly lowers the weight of emerging countries as a group.
Benno Ferrarini is Senior Economist at the Economics and Research Department of the Asian Development Bank, 6 ADB Avenue, Mandaluyong 1550, Metro Manila, Philippines; email: email@example.com. This paper is written not in an official capacity and solely represents the views of the author, not necessarily those of the Asian Development Bank, its Management, or any of the countries it represents.
TABLE 1 Myanmar's Exports Composition (2006-10) Importer Total US$m STX Food Fuels Agric. Manuf. THA 13,615 48.4 3.3 91.3 4.5 0.9 IND A,122 16.8 62.8 0.0 36.1 1.1 CHN 2,891 10.3 25.0 3.6 67.5 4.0 JPN 1,583 5.6 32.7 0.0 7.0 60.3 MYS 812 2.9 48.1 0.1 43.1 8.8 KOR 532 1.9 10.9 26.8 5.1 57.2 DEU 515 1.8 2.5 0.0 6.9 90.6 SGP 421 1.5 37.8 0.7 46.3 15.2 GBR 304 1.1 30.2 0.7 2.2 66.9 ROW 2,763 9.8 41.4 0.1 26.2 32.4 World 28,157 100.0 23.1 45.1 20.3 11.5 Note: The figures represent percentage shares unless indicated otherwise. Total US$m are total exports, in US$ million. STX is the share out of total exports. "Manuf." are manufactures and "Agrie." are non-food agricultural commodities. "ROW" (Rest of the World) are all countries with STX smaller than 1 per cent. Source: Author's estimates. TABLE 2 Myanmar's Imports Composition (2006-10) Exporter Total US$m STM Food Fuels Agric. Manuf. CHN 10,622 35.7 3.1 5.1 1.3 90.5 THA 6,659 22.4 23.4 16.9 1.5 58.2 SGP 4,677 15.7 11.6 40.3 2.4 45.7 KOR 1,542 5.2 0.2 1.6 5.0 93.1 MYS 1,268 4.3 39.4 15.3 3.3 42.0 IDN 1,110 3.7 58.6 0.3 0.3 40.8 IND 1,005 3.4 13.2 2.2 1.9 82.7 JPN 931 3.1 0.5 0.3 1.4 97.8 ROW 1,977 6.6 16.6 5.2 2.7 75.5 World 29,792 100.0 13.6 13.1 1.9 71.5 Note: The figures represent percentage shares unless indicated otherwise. Total US$m are total imports, in US$ million. STM is the share out of total imports. "Manuf." are manufactures and "Agrie." are non-food agricultural commodities. "ROW" (Rest of the World) are all countries with STM smaller than 1 per cent. Source: Author's estimates. TABLE 3 Gravity Regressions Regressant: (1) (2) (3) (4) Exports (log) PSEUDOFE REGLS FGLS TOBIT GDP Exporter (log) 1.486 *** 1.465 *** 1.488 *** 1.466 *** (0.0961) (0.0959) (0.0323) (0.0851) GDPpc Exporter (log) 0.358 *** 0.573 *** 0.379 *** 0.564 *** (0.0977) (0.134) (0.0232) (0.123) GDP Importer (log) 0.887 *** 1.002 *** 0.882 *** 0.997 *** (0.0637) (0.0777) (0.0208) (0.0662) GDPpc Importer (log) 0.471 *** 0.483 *** 0.490 *** 0.483 *** (0.0667) (0.0955) (0.0176) (0.0731) Distance (log) -1.140 *** -1.353 *** -1.226 *** -1.344 *** (0.155) (0.203) (0.0564) (0.189) Common Border (0,1) 0.830 *** 0.704 ** 0.739 *** 0.709 (0.284) (0.318) (0.0994) (0.455) Common Language (0,1) -0.792 *** -0.832 *** -0.861 *** -0.830 *** (0.195) (0.177) (0.0740) (0.322) Colonial Link (0,1) 0.830 *** 0.711 *** 0.893 *** 0.716 ** (0.240) (0.263) (0.0807) (0.321) Free Trade 0.484 ** 0.290 0.538 *** 0.298 Agreement (0,1) (0.206) (0.223) (0.0670) (0.246) Constant -38.34 *** -40.31 *** -37.54 *** -40.22 *** (2.525) (2.705) (0.942) (2.318) R-squared (adjusted) 0.808 0.807 n/a n/a Observations 2,233 2,233 2,233 2,233 Country pairs 203 203 203 203 Robust standard errors in parentheses: *** p<0.01, ** p<0.05, * p<0.1 Note: The likelihood ratio test of ou = 0 in relation to TOBIT truncated at zero is [chi square] = 5772 ***. GDP and exports are deflated and expressed as constant 2000 US$. Country/pair and/or time dummy coefficients are included in the regressions but not reported. GDP, GDPpc and exports are deflated, with base year 2000. Source: Author's estimates. TABLE 4 Actual and Gravity Exports (2006-10) Ratio (geom. Exporter Actual Gravity (mean) mean) VNM 168,454 105,424 3.08 2.38 MYS 488,231 584,285 1.53 1.16 THA 439,008 743,924 1.34 1.04 KHM 12,203 4,661 3.07 1.03 IDN 386,119 689,035 1.11 0.85 PHL 151,902 370,239 0.77 0.50 MMR 18,533 38,929 0.38 0.15 Note: Actual and gravity exports are in constant year-2000 US$ million. Source: Author's estimates. TABLE 5 MMR Actual/Gravity Ratio (2006-10) Exports to Actual Gravity Ratio Share THA 9,348 3,101 3.01 30.63 VNM 195 118 1.65 0.38 PAK 176 117 1.51 0.29 IND 3,253 4,136 0.79 -4.33 ESP 164 293 0.56 -0.63 ZAF 24 45 0.55 -0.10 MYS 557 1,107 0.50 -2.70 ARG 33 69 0.48 -0.18 IDN 97 220 0.44 -0.61 TUR 54 149 0.36 -0.47 CHN 1,965 5,995 0.33 -19.76 DEU 359 1,124 0.32 -3.75 LKA 17 62 0.27 -0.22 ARE 150 564 0.27 -2.03 KOR 365 1,532 0.24 -5.72 MEX 21 99 0.21 -0.38 PHL 21 124 0.17 -0.51 NLD 45 277 0.16 -1.14 EGY 6 39 0.15 -0.16 CZE 5 37 0.15 -0.15 RUS 21 143 0.14 -0.60 ITA 78 589 0.13 -2.51 BEL 22 167 0.13 -0.71 SAU 23 173 0.13 -0.74 FRA 82 735 0.11 -3.20 JPN 1,090 10,388 0.10 -45.59 AUS 59 580 0.10 -2.55 BGD 73 748 0.10 -3.31 GBR 211 2,218 0.10 -9.84 NZL 4 46 0.09 -0.21 BRA 4 101 0.04 -0.48 CAN 11 308 0.04 -1.46 KAZ 0 24 0.01 -0.12 KHM 0 20 0.01 -0.10 USA 0 3,478 0.00 -17.05 Note: Actual and gravity exports are in US$ million. "Ratio" is the ratio of actual over gravity exports. "Share" is partners' share of the actual-gravity differential. Source: Author's estimates.
|Printer friendly Cite/link Email Feedback|
|Publication:||Journal of Southeast Asian Economies|
|Date:||Aug 1, 2014|
|Previous Article:||Five growth strategies for Myanmar: re-engagement with the global economy.|
|Next Article:||Foreign exchange market reform in Myanmar: achievements and challenges.|