Advertisement

Comparative Economic Studies

, Volume 60, Issue 3, pp 361–387 | Cite as

Economic Transition and Growth Dynamics in Asia: Harmony or Discord?

  • Pui Sun Tam
Article

Abstract

This paper studies whether the economic transitional and growth dynamic experiences of Asian economies are in harmony or discord so as to serve as a catalyst for, or a hindrance to, fostering regional economic cooperation and integration. Empirical evidence is derived from the application of a nonlinear time-varying factor convergence approach in conjunction with nonparametric distribution dynamics techniques. Results are in favor of growth convergence, but not level convergence, among the economies along their transition dynamic growth paths, and provide insights on issues that have significant bearing on the prospect of deepening regional economic cooperation and integration.

Keywords

Asia Convergence Growth Regional integration Transition 

JEL Classifications

C23 O47 O53 

Introduction

In the new world economic order after the Second World War, three waves of regionalism in the world trading system can be distinguished (Goto 2002; Pomfret 2005). The first wave of regionalism was unveiled in the 1950s when European economies undertook a series of economic cooperation and integration activities that led to the ultimate formation of the European Union (EU). The second wave began in North America in the 1980s and peaked with the establishment of the North American Free Trade Agreement. The third wave currently underway is led by Asian economies. Asia used to be the least regionalized region in the world, due to its high degree of dependence on international trade and investment relationships with the USA instead of intra-regionally (Schott 1991). More progress in Asian economic regionalism occurred in the 1990s when the Association of Southeast Asian Nations (ASEAN) initiated the ASEAN Free Trade Agreement (FTA). Triggered by the 1997–1998 Asian Financial Crisis, the 2000s saw a proliferation of regional trade agreements being proposed, under negotiation and concluded (Ahn and Cheong 2007; Kawai 2007).

Amidst the new tide of regionalism in Asia, ASEAN’s ten members (ASEAN-10), including Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand and Vietnam, together with six of its trading partners, namely Australia, China, India, Japan, New Zealand and South Korea, jointly known as ASEAN + 6 (Kawai and Wignaraja 2008), formally launched negotiations for the formation of the Regional Comprehensive Economic Partnership (RCEP) in 2012. When realized, RCEP will become the largest economic bloc in the world. The prospective RCEP members have become more earnest in accelerating their ongoing negotiations upon the US withdrawal from the Trans-Pacific Partnership (TPP). Table 1 shows that the ASEAN + 6 economies are dispersed in terms of geographical location and pre-RCEP per capita income level. RCEP aims to achieve a mutually beneficial economic partnership agreement to boost economic growth and equitable economic development in the region. A crucial question then arises is whether the economic transitional growth dynamic experiences of these Asian economies in the past decades are in harmony or discord, so as to serve as a catalyst for, or a hindrance to, the deepening of regional economic cooperation and integration. Signs of economic convergence among the diverse Asian economies would suggest promising prospect for the proposed regional economic arrangement. Economies that are similar in per capita income derive greater benefits from economic integration through larger volume of trade and investment among them (Jovanovic 1998; Robson 1998). They are also better able to accommodate the redistribution of income and employment among members resulting from trade flow adjustments (Schott 1991) and have less burden on the need for creating aid programs such as the Regional and Cohesion Funds of the EU (Sala-i-Martin 1996). Furthermore, with similar level of economic development, economies are more likely to engage in time-consistent policy cooperation (de Brouwer et al. 2006). Against this background, this paper investigates the economic transition and growth dynamics of Asian economies from a renewed perspective by considering the convergence (or divergence) patterns of three important macroeconomic variables, namely income, consumption and government expenditure, through the application of a nonlinear time-varying factor convergence approach in conjunction with nonparametric distribution dynamics techniques.
Table 1

Geographic distribution and income classification of economies in 2016.

Source: World Development Indicators

World Bank income classification

Region

Northeast Asia

Southeast Asia

South Asia

Australasia

High

Hong Kong (HKG)

Brunei (BRN)

 

Australia (AUS)

Japan (JPN)

Singapore (SGP)

 

New Zealand (NZL)

Korea (KOR)

   

Macao (MAC)

   

Taiwan (TWN)

   

Upper middle

China (CHN)

Malaysia (MYS)

  
 

Thailand (THA)

  

Lower middle

 

Cambodia (KHM)

India (IND)

 
 

Indonesia (IDN)

  
 

Laos (LAO)

  
 

Myanmar (MMR)

  
 

Philippines (PHL)

  
 

Vietnam (VNM)

  

Classification based on per capita GNI in US$ (Atlas methodology) in 2016

There has been a growing literature in recent years that investigates whether the income convergence hypothesis holds in Asia, partly because of the connection between the equality of income across economies and the success of economic integration among them as mentioned above (Park 2000a; Moon 2006). However, studies vary substantially in their coverage of economies under investigation due to overlapping economic groupings in the region and/or availability of data. For instance, while Lim and McAleer (2004) and Jayanthakumaran and Verma (2008) examine the ASEAN-5 economies (including Indonesia, Malaysia, the Philippines, Singapore and Thailand), Solarin et al. (2014) focus on ASEAN-9 (excluding Myanmar), and Park (2000a) and Tam (2017) investigate ASEAN-10. Heng and Siang (1999) and Moon (2006) study ASEAN-5 plus selected Northeast Asian economies (such as Japan, Hong Kong and South Korea). Felipe (2000), Evans and Kim (2005), Ito (2016) and Aulia (2017) consider a selection of ASEAN, Northeast Asian and South Asian economies (such as India, Pakistan and Sri Lanka), whereas Park (2000b) and Wang (2012) analyze ASEAN-10 together with some Northeast Asian and Australasian economies (Australia and New Zealand) countries. With the endeavor to evaluate the prospect of RCEP, this paper pioneers the study of income convergence among the RCEP economies with adaptations discussed in the data description section. Moreover, unlike the extant literature, this paper applies the log t convergence test of Phillips and Sul (2007) in conjunction with the conditional density estimation and visualization methods of Hyndman et al. (1996) that are particularly suitable for analyzing the transitional and growth dynamic experiences of Asian economies. Merits of these econometric methods are discussed in the methodology section.

Despite being an important and popular indicator of economic development, income is an imperfect measure of well-being or living standards (Sen 1987; Dasgupta 1990; Khan 1991). Income is only one of the means, whereas consumption is the end (Wan 2005), and household consumption is a plausible demand-side measurement of economic well-being (Cuffaro et al. 2006). Therefore, assessing consumption convergence is vital in the broader context of economic development for the maintenance of regional cohesion (Ševela 2004; Liobikienė and Mandravickaitė 2013) and henceforth promising economic integration in the long run. In fact, more attention has been paid to the testing of consumption convergence in the EU region in recent years, especially with the accession of transition economies as new members. Examples of these studies are Ševela (2004), Fiaschi and Lavezzi (2005), Liobikienė and Mandravickaitė (2011) and Dudek et al. (2013). Empirical evaluation of consumption convergence is also of particular relevance to RCEP, given the presence of a number of less developed economies as shown in Table 1, but has largely been overlooked in the Asian context.

On top of the above-mentioned innovations, this paper also contributes by probing into the issue of fiscal convergence among the RCEP economies. Investigating fiscal convergence is of practical interest for the EU, since member states have to fulfill certain fiscal convergence criteria (along with other criteria) for entry into the Eurozone. Some relevant works in this vein include Afxentiou and Serletis (1996), Esteve et al. (2000) and Kočenda et al. (2008), which use different fiscal measures in their tests. Although the creation of a monetary union is not in the scope of RCEP, it is commonly held that such a monetary architecture would be a long-term objective for the Asian economies (de Brouwer et al. 2006). With this in mind, this paper also assesses the readiness of these Asian economies for monetary unification. To do so, government expenditure convergence (or divergence) is examined as in Afxentiou and Serletis (1996), de Bandt and Mongelli (2000), Blot and Serranito (2006), Apergis et al. (2013) and Perović et al. (2016). This type of convergence is theoretically supported by Skidmore et al.'s (2004) model that is consistent with the macroeconomic growth literature and therefore with the transitional and growth dynamic experiences of the Asian economies.

The remainder of the paper is organized as follows. The next section sets out the econometric methodologies, namely the log t convergence test and the conditional density estimation and visualization. The subsequent section describes the data used in analysis and outlines the observational patterns. The penultimate section presents the econometric results that trace out the transitional dynamic growth experiences of Asian economies and discusses the factors that drive the results. The final section concludes the paper and draws implications derived from the empirical findings.

Econometric Methodology

This section sets out the methodologies of the nonlinear time-varying factor approach and the nonparametric distribution dynamics techniques employed in this paper. Given their respective merits, a growing number of empirical studies has employed either methodology in recent years. As will be made clear below, while the former method models the trajectories of transition growth paths of individual economies, the latter approach studies the trajectory of cross-economy distributions. They can therefore be regarded as complementary methods and can provide a multi-dimensional analysis of the transition growth dynamics of economies when used in combination, which is taken up in this paper.

Log t Convergence Test

A main advantage of the log t convergence test of Phillips and Sul (2007) is that it is designed for use under technological heterogeneity. It is based on a nonlinear time-varying factor model that embodies both common and idiosyncratic components. Under this model, the process by which convergence takes place over time is identified by the individual-specific time-varying factor loadings, thus providing measures of individual economic transition. Consequently, this formulation is flexible to accommodate a wide spectrum of transition growth paths for individuals and is therefore particularly suitable for the present analysis since it is highly probable that the heterogeneous RCEP economies experience convergence to varying extents. In contrast, under the technological homogeneity assumption that is extensively used in growth studies, the convergence test approaches based on the conventional augmented Solow regression model (Barro and Sala-i-Martin 1992; Mankiw et al. 1992) and standard unit root and cointegration testing (Bernard and Durlauf 1996; Carlino and Mills 1993; Evans and Karras 1996) become inappropriate methods according to Phillips and Sul (2009). Other merits of the log t convergence test include its robustness with respect to the stationarity properties of the variables under consideration and its power against cases of club convergence.

Consider a panel data set of the logarithm of per capita income, zi,t, with i = 1,…,N and t = 1,…,T, where i and t index the cross section and time dimension, respectively, while N and T are the number of economies and number of time observations in the sample, respectively. Suppose that zi,t can be decomposed as
$$z_{i,t} = a_{i,t} + e_{i,t} ,$$
(1)
where ai,t and ei,t are, respectively, the systematic component and transitory component. Assume that in this panel of economies, there exists in zi,t a common trend component such as world technology, ft, which follows either a non-stationary stochastic trend with drift or a trend stationary process. Then, Eq. (1) may be transformed into
$$z_{i,t} = \left( {\frac{{a_{i,t} + e_{i,t} }}{{f_{t} }}} \right)f_{t} = b_{i,t} f_{t} ,$$
(2)
where bi,t measures the relative deviation of economy i from ft, with ft determining the common steady-state growth path. The growth dynamic experience is heterogeneous across economies and can be described by the relative transition parameter, hi,t, constructed as
$$h_{i,t} = \frac{{z_{i,t} }}{{(1/N)\sum\nolimits_{i = 1}^{N} {z_{i,t} } }} = \frac{{b_{i,t} }}{{(1/N)\sum\nolimits_{i = 1}^{N} {b_{i,t} } }}.$$
(3)

This measures the transition element bi,t for economy i relative to the panel average at time t. The evolution of hi,t over time traces out the trajectory of each economy relative to the average, which also measures the relative divergence of the economy from the common steady-state growth path. Despite possible transient diverging relative patterns, long-run convergence among economies is possible if \(\mathop {\lim}\nolimits_{t \to \infty } (z_{i,t} /z_{j,t} ) = 1\) for all i \(\ne\) j, or equivalently \(\mathop {\lim}\nolimits_{t \to \infty } b_{i,t} = b\) for all i. In other words, convergence occurs when \(\mathop {\lim}\nolimits_{t \to \infty } h_{i,t} = 1\) for all i.

To test for the null hypothesis of convergence for all i against the alternative of non-convergence for some i, the following time series regression is estimated
$$\log \left( {\frac{{H_{1} }}{{H_{t} }}} \right) - 2\log \left( {\log t} \right) = a + \gamma \log t + u_{t} ,\quad t = T_{0} , \ldots ,T,$$
(4)
where \(H_{t} = \left( {1/N} \right)\sum\nolimits_{i = 1}^{N} {\left( {h_{i,t} - 1} \right)^{2} }\) and T0 = [ϰT] for some ϰ > 0, so that the first ϰ% of the time series data is discarded before carrying out regression. Considering the sample size in this study, ϰ is set to be 0.3 according to the simulation results in Phillips and Sul (2007). Under the null of growth convergence, γ ≥ 0, whereas γ < 0 under the alternative. The null hypothesis is tested based on a robust t-statistic calculated on the slope coefficient γ in Eq. (4), tk. This is called the log t test due to the log t regressor. Level convergence among all i can also be tested under this framework by using the null hypothesis of γ ≥ 2 versus the alternative of γ < 2.

Since the log t test has power against cases of club convergence, Phillips and Sul (2007) also propose a clustering algorithm to test for this phenomenon. The main steps of the algorithm are outlined as follows. First, economies are ordered according to the values of their last observations. Second, the core group is identified. This is done by selecting the first k highest economies to form the subgroup, and then, the log t regression is run to obtain the log t-statistic tk. The core group size k* is chosen by maximizing tk over k subject to min {tk} > − 1.65. Third, individuals are sieved for club membership. Each of the remaining economy is added separately to the chosen core group and the log t regression run. The individual is included in the convergence club if tk > c, where c is set to zero as suggested by Phillips and Sul (2007). The final step searches the remaining economies in the panel for other convergence clubs by repeating the above steps. If no such clusters can be found, it can be concluded that these economies diverge.

Conditional Density Estimation and Visualization

The distribution dynamics approach advanced by Quah (1993) studies the trajectory of cross-economy distribution, tracks shifts of distribution, as well as examines intra-distribution dynamics (Quah 1997). In empirical work on distribution dynamics, one seeks a law of motion that describes the evolution of the distribution across economies. The discretized version of the law of motion amounts to the construction of a transition probability matrix (Quah 1993), which is replaced by the estimation of a stochastic kernel in the continuous case (Quah 1996).

Let X and Y denote the relative transition parameter at time t (the initial year of the sample period) and time t + τ (the final year of the sample period), respectively. Let f(x,y) be their joint density, f(x) the marginal density of X and g(y|x) = f(x,y)/f(x) the conditional density of Y|(X = x). The kernel estimator of g(y|x) is
$$\hat{g}(y|x) = \frac{{\hat{f}(x,y)}}{{\hat{f}(x)}},$$
(5)
where
$$\hat{f}(x,y) = \frac{1}{{Nh_{x} h_{y} }}\sum\limits_{i = 1}^{N} {K\left( {\frac{{\left\| {x - X_{i} } \right\|_{x} }}{{h_{x} }}} \right)K\left( {\frac{{\left\| {y - Y_{i} } \right\|_{y} }}{{h_{y} }}} \right)}$$
(6)
is the kernel estimator of f(x,y) and
$$\hat{f}(x) = \frac{1}{{Nh_{x} }}\sum\limits_{i = 1}^{N} K\left( {\frac{{\left\| {x - X_{i} } \right\|_{x} }}{{h_{x} }}} \right)$$
(7)
is the kernel estimator of f(x). Notice that hx and hy are bandwidth parameters controlling the smoothness of fit, while \(\| {_{{}}^{{}} \cdot_{{}}^{{}} }\|_{x}\) and \(\| {_{{}}^{{}} \cdot_{{}}^{{}} } \|_{y}\) are distance metrics on the spaces of X and Y, respectively. The kernel function, K(u), is assumed to be real, integrable, nonnegative, and even on ℜ concentrated at the origin such that \(\int_{\Re } {K(u){\text{d}}u = 1}\), \(\int_{\Re } {uK} (u){\text{d}}u = 0\) and \(\int_{\Re } {u^{2} K(u){\text{d}}u < \infty }\). The conditional density estimator in Eq. (5) can then be rewritten as
$$\hat{g}(y|x) = \frac{1}{{h_{y} }}\sum\limits_{i = 1}^{N} w_{i} (x)K\left( {\frac{{\left\| {y - Y_{i} } \right\|_{y} }}{{h_{y} }}} \right),$$
(8)
where
$$w_{i} (x) = \frac{{K\left( {\left\| {x - X_{i} } \right\|_{x} /h_{x} } \right)}}{{\sum\nolimits_{j = 1}^{N} {K\left( {\left\| {x - X_{j} } \right\|_{x} /h_{x} } \right)} }},$$
(9)
which can be a biased estimator. The bias is adjusted using a modified version of Eq. (8) proposed by Hyndman et al. (1996)
$$\hat{g}^{*} (y|x) = \frac{1}{{h_{y} }}\sum\limits_{i = 1}^{N} w_{i} (x)K\left( {\frac{{\left\| {y - Y_{i}^{ * } (x)} \right\|_{y} }}{{h_{y} }}} \right),$$
(10)
where \(Y_{i}^{ * } (x) = \varepsilon_{i} + \hat{r}(x) - \hat{l}(x)\), with \(\hat{r}(x)\) being the estimator of the conditional mean function \(r(x) = E[Y|(X = x)]\), \(\varepsilon_{i} = y_{i} - \hat{r}(x)\) and \(\hat{l}(x)\) the mean of the estimated conditional density of \(\varepsilon |(X = x)\). The conditional density estimation is implemented with a Gaussian kernel and bandwidths chosen using the normal reference rules of Bashtannyk and Hyndman (2001).

Hyndman et al. (1996) also propose two graphical methods for visualizing the estimated conditional densities, the stacked conditional density (SCD) plot and the high density region (HDR) plot. Since they highlight the conditioning, they are much more informative than the traditional three-dimensional perspective plots and contour plots. The SCD plot displays a number of densities side by side in a perspective fashion. For a given income level at time t, the probability density of the income level at time t + τ is traced, thus clearly showing the evolution of changes in the shape of the income distribution at time t + τ over the range of the income levels at time t. An HDR is the smallest region of the sample space covering a specified probability. The HDR plot consists of vertical strips, each representing the projection on the xy plane of the conditional density of y on x. Within each strip, the 50, 95 and 99% HDRs, in increasingly lighter-shaded regions, are reported. The mode for each conditional density is shown as a bullet. Each strip is thus essentially a visual summary of the characteristics of a probability distribution.

Strong convergence toward equality, analogous to level convergence, is supported if the probability distribution mass is clustered at unity on the horizontal axis in the SCD plot. This manifests in the 50% HDRs being crossed by the horizontal line at unity in the HDR plot. These patterns suggest that any existing deviation across i at time t disappears at time t + τ. Weak convergence toward equality is evidenced when the horizontal line at unity crosses the 95 and 99% HDRs. Furthermore, low persistence and some intra-distribution mobility is present if the diagonal crosses only the 95 and 99% HDRs. These features corroborate the notion of growth convergence. By contrast, high persistence is indicated by the mass of distribution being concentrated along the diagonal line in the SCD plot, and the 50% HDRs being crossed by the main diagonal in the HDR plot. In this case, existing deviations across i at time t continue to exist at time t + τ.

Data Description

Annual data for the period of 1970–2014 are collected from the Penn World Table 9.0. Data on gross domestic product measured in US dollars are at constant 2011 national prices. This version of income is suitable for use in cross-country growth comparison (Feenstra et al. 2015) and is also found to be an appropriate measure of economic activities (Pinkovskiy and Sala-i-Martin 2016). Data on consumption expenditure and government expenditure measured in US dollars at constant 2011 national prices are derived from their local currency counterparts. These three aggregate variables are converted into the per capita basis using the population data and are denoted as GDP, CEXP and GEXP, respectively. They are expressed in logarithms in analysis in the next section. The data set covers a total of 19 economies. In addition to the 16 prospective RCEP members, Hong Kong, Macao and Taiwan are also included due to their intimate economic trade and investment relationships with the RCEP economies, notably with China under the Closer Economic Partnership Arrangement regarding Hong Kong and Macao, and the Economic Cooperation Framework Agreement in the case of Taiwan. The significant relationship between China and these three economies in the context of income convergence has been demonstrated in Lei and Tam (2010, 2013).

Table 2 displays, for each economy, GDP for selected years and its average annual growth rate over the 45 years. It can be observed that economies with GDP over US$10,000 in 1970, including Australia, Brunei, Japan and New Zealand, tended to grow much more slowly than the others with lower initial GDP. During the sample period, the richest, Brunei, in 1970 experienced negative growth, whereas the poorest, Myanmar and Vietnam, in 1970 grew at the rates of 4.56 and 4.11%, respectively. China, with a relatively low initial GDP, grew at a high annual rate of 5.58%. Hong Kong, Macao, Singapore and Taiwan, with income levels lower than that of Japan in 1970, had grown faster than Japan over time, and they all became richer than Japan in 2014. However, the experience of Cambodia and the Philippines is exceptional. Although they had relatively low initial GDP, they only grew at a slow speed of less than 2% over time. This growth rate is similar to those of the high-income economies of Australia and New Zealand. It is also noteworthy that economies with initial GDP under US$2,000 typically remained poor in 2014 with GDP at around US$10,000 or less. These economies include Cambodia, India, Indonesia, Laos, Myanmar and Vietnam.
Table 2

Per capita gross domestic product (GDP), 1970–2014.

Source: Penn World Table 9.0

Economy

Amount (US$)

Growth rate (%)

1970

1980

1990

2000

2014

1970–2014

AUS

22,520

26,411

30,582

38,758

47,544

1.71

BRN

92,785

150,293

85,693

83,116

75,420

− 0.47

CHN

1150

1489

2311

4181

12,524

5.58

HKG

7589

14,069

23,522

29,603

46,641

4.21

IDN

1772

2977

4448

5792

10,009

4.01

IND

1195

1274

1801

2536

5534

3.55

JPN

14,666

20,305

30,287

32,946

36,250

2.08

KHM

1434

716

906

1292

2933

1.64

KOR

2609

5279

11,836

21,538

34,540

6.05

LAO

1031

1221

1681

2433

5249

3.77

MAC

12,677

27,526

38,825

42,094

139,989

5.61

MMR

673

797

741

1292

4793

4.56

MYS

3541

6898

9277

14,332

21,737

4.21

NZL

19,025

21,023

23,488

27,894

33,966

1.33

PHL

3359

4509

4078

4298

6774

1.61

SGP

8297

16,964

28,564

43,761

64,624

4.78

THA

2392

3581

6383

8986

14,642

4.20

TWN

3712

7731

14,772

25,833

41,378

5.63

VNM

925

1074

1473

2597

5440

4.11

GDP at constant 2011 national prices

It can be deduced from Table 3 that there was a tendency for CEXP to grow at a low rate over time for economies with high initial CEXP and vice versa. On the one hand, Brunei was the biggest consumer in 1970, but its CEXP only grew at a negative rate on average over the sample period. The large consumers of Australia, Japan and New Zealand also grew at the slow rates of 1.88, 2.07 and 1.26%, respectively. On the other, CEXP of China and Myanmar, which were economies with the lowest initial CEXP, grew at annual rates of 4.46 and 3.32%, respectively. On the per capita basis, Hong Kong, Macao, Singapore and Taiwan consumed less than Japan in 1970. By 2014, they had all turned around to consume more than Japan due to their faster CEXP growth rate than that of Japan. Cambodia and the Philippines serve as exceptional cases. Their CEXP grew at a relatively slow speed over time in spite of their relatively low initial levels of CEXP in 1970. It can be noted that although Brunei was the biggest consumer in 1970, it was only an average consumer in 2014 among the sample of economies. Besides, economies with low initial CEXP (lower than US$1,000) in 1970, namely China, India, Indonesia, Laos, Myanmar and Vietnam, typically continued to be small consumers with CEXP less than US$6,000 in 2014.
Table 3

Per capita consumption expenditure (CEXP), 1970–2014.

Source: Penn World Table 9.0

Economy

Amount (US$)

Growth rate (%)

1970

1980

1990

2000

2014

1970–2014

AUS

11,114

13,638

15,876

20,089

25,245

1.88

BRN

17,359

7631

7764

7663

11,159

− 1.00

CHN

674

753

1088

1763

4598

4.46

HKG

4542

9125

15,376

19,751

30,877

4.45

IDN

835

1477

2267

3539

5526

4.39

IND

832

918

1151

1521

3154

3.07

JPN

8846

12,510

17,533

19,439

21,817

2.07

KHM

1584

792

972

1092

2277

0.83

KOR

2336

3749

7296

12,206

17,115

4.63

LAO

830

984

1336

2030

2736

2.75

MAC

4793

10,406

14,090

16,844

29,333

4.20

MMR

583

632

584

736

2456

3.32

MYS

1745

3353

4095

5486

10,884

4.25

NZL

11,403

11,637

13,173

15,196

20,073

1.29

PHL

2283

2728

2793

3234

4866

1.74

SGP

5441

9051

13,166

19,342

22,838

3.31

THA

1516

2149

3292

4500

7826

3.80

TWN

2383

4723

9528

17,053

22,394

5.22

VNM

840

976

1313

1771

3593

3.36

CEXP at constant 2011 national prices

Table 4 shows that Australia, Brunei, Japan and New Zealand, which had the highest initial GEXP in 1970, experienced slow growth in GEXP during the sample period. Their growth rates were, respectively, 1.91, 2.54, 2.73 and 1.33%. China, which started off with a relatively low level of GEXP in 1970, grew at the highest GEXP rate of 6.11% over time. For the other economies with initial GEXP less than US$200, namely, India, Indonesia, Laos, Myanmar and Vietnam, their average annual growth rates of GEXP were in the range of 3.32 and 4.33%. Singapore and Taiwan began with over US$1,000 GEXP in 1970. Their GEXP grew at the rates of 4.22 and 3.53%, respectively, over time. Thus, by 2014, their GEXP had almost caught up with that of New Zealand. The Philippines is observed to be an outlier in that it had relatively low initial GEXP, but its GEXP also grew at the slowest rate. Cambodia, India, Indonesia, Laos, Myanmar and Vietnam, with the lowest GEXP (less than $200) in 1970, also fell in the group with the lowest GEXP (less than US$1000) in 2014.
Table 4

Per capita government expenditure (GEXP), 1970–2014.

Source: Penn World Table 9.0

Economy

Amount (US$)

Growth rate (%)

1970

1980

1990

2000

2014

1970–2014

AUS

3577

4671

5718

6800

8224

1.91

BRN

5200

10,873

11,782

11,571

15,679

2.54

CHN

122

194

297

628

1652

6.11

HKG

894

1480

2388

3165

4155

3.55

IDN

136

326

438

406

879

4.33

IND

115

141

224

331

602

3.84

JPN

2276

3350

4606

6078

7430

2.73

KHM

39

19

35

56

163

3.32

KOR

835

1309

2109

2952

5084

4.19

LAO

173

204

253

339

935

3.91

MAC

870

1878

4577

5550

9916

5.69

MMR

139

150

139

200

796

4.06

MYS

377

853

968

1263

2817

4.68

NZL

3575

4424

4846

4900

6400

1.33

PHL

403

595

517

484

685

1.21

SGP

1007

1747

2619

5054

6215

4.22

THA

314

569

766

1255

2429

4.76

TWN

1319

1987

3836

5295

6068

3.53

VNM

52

61

90

136

336

4.33

GEXP at constant 2011 national prices

The data presented in Tables 2 through 4 suggest that the sample of economies as a whole reveals signs of converging behavior. Those starting off with higher initial levels of GDP, CEXP and GEXP tended to grow more slowly than those with lower initial levels. However, there still exist substantial disparities across the board of economies in their respective levels of GDP, CEXP and GEXP. Economies that had relatively high levels in the beginning of the sample period stayed put at the end of the period. The reverse also holds true. These observations are indicative of the presence of growth convergence but not level convergence across economies, which is explored econometrically in the next section.

Econometric Results

Whether the Asian economies are converging in terms of GDP, CEXP and GEXP is formally tested using the log t convergence test. The distribution dynamics of these variables across economies over time are also examined with the estimated conditional densities visualized using SCD and HDR plots. In this section, empirical results for GDP, CEXP and GEXP are first reported, followed by a discussion of the empirical findings.

Income

Figure 1 provides a time plot of the relative transition parameters for GDP of the economies. With respect to most economies, a narrowing of the parameter values toward unity over time, particularly from the late 1990s onward, albeit at a very slow speed, can be observed. However, the transition growth paths of the economies traced out by these parameters are heterogeneous in nature. While some economies, such as Brunei, have transited from a high initial state relative to the group average, some others, such as China, have transited from a low initial state. There are also some economies like New Zealand that have followed roughly monotonic transition paths to unity, while some other economies like Myanmar have taken on paths that are characterized by transient phase(s) of divergence. Exceptions of prolonged divergence in the transition process, such as the Philippines, are also in place. It can also be noted that the relative transition parameters across economies were widely dispersed in values in 1970. In spite of signs of them converging to unity over time, their disparities were still apparent in 2014. Economies with GDP above the group average in 1970 persisted to possess above average GDP in 2014. Among economies with below average initial GDP, only Korea, Malaysia and Taiwan were able to grow and achieve above average GDP by 2014. The GDP levels of China and Thailand in 2014 were also relatively close to the group average level. The rest of these low GDP economies in 1970 continued to attain disparately low GDP levels in 2014.
Fig. 1

Relative transition paths of GDP, 1970–2014

The log t convergence test results for GDP are reported in Table 5. The null of growth convergence for all economies is first tested. As indicated by the t-statistic, the null hypothesis cannot be rejected at any of the conventional significance levels. Although the notion of growth convergence is supported, little evidence is found for level convergence. The estimated coefficient of the log t regressor for all economies is significantly below 2, and the null hypothesis of level convergence for the economies is strongly rejected.
Table 5

Log t convergence test results for GDP, 1970–2014

Group

Coefficient

t-statistic

Converging economies

Growth convergence

All

0.110

2.049

All

High-level club

0.399

5.811

AUS, BRN, HKG, JPN, KOR, MAC, MYS, NZL, SGP, TWN

Low-level club

3.371

2.120

CHN, IDN, IND, KHM, LAO, MMR, PHL, THA, VNM

Growth club 1

0.283

3.626

AUS, BRN, CHN, HKG, IND, JPN, KOR, MAC, MMR, MYS, NZL, SGP, THA, TWN, VNM

Growth club 2

0.564

5.124

IDN, KHM, LAO, PHL

Level convergence

All

0.110

− 35.204***

High-level club

0.399

− 23.331***

Low-level club

0.371

− 9.323***

Growth club 1

0.283

− 21.971***

Growth club 2

0.564

− 13.030***

***Indicates significance at the 1% level

Figure 2 presents the SCD and HDR plots in the left and right panels, respectively. As observed from the SCD plot, the distribution mass tends to cluster around the two sides of the value of 1 in 2014 instead of concentrating along the main diagonal, which is an indication of substantial mobility of GDP across economies over time between 1970 and 2014. In fact, as illustrated in the HDR plot, modes of the poorest economies in 1970 are located above the main diagonal line, while modes of the richest economies in 1970 lie below it. The main diagonal crosses a number of the 95 and 99% HDRs, whereas the horizontal line at unity usually crosses the 95 and 99% HDRs. In other words, the poorest economies grew faster over time and became relatively richer to catch up with the richest economies, whereas the richest economies slowed down their growth speed and became relatively poorer. Such transition behavior of the poorest and richest economies is representative of intra-distribution mobility and constitutes evidence for growth convergence among them, which is consistent with the log t growth convergence test result found above.
Fig. 2

Conditional density region plots of GDP, 1970–2014. a Stacked conditional density plot, b High density region plot

Notwithstanding, the SCD plot shows that there is a high concentration of the probability mass at two relative GDP levels in 2014, one above and one below the group mean. In the HDR plot, the horizontal line at unity mainly crosses the 95 and 99% HDRs. For economies with initial GDP above average, many modes are located around the relative GDP level of 1.07 in 2014. Modes for some economies with initial GDP below average correspond approximately to the relative GDP level of 0.87 in 2014. These patterns corroborate the observations in Fig. 1 and the rejection of the null of level convergence for all economies with the log t test. Also of interest is that the diagonal line crosses the 50% HDRs of economies with initial relative GDP between 0.8 and 1.15, which means stability and persistence for these economies over time. In other words, they are susceptible to the so called middle-income trap.

From the above findings, the sample of economies can be divided into two clubs according to their 2014 relative transition parameter values. The first club comprises economies with above average GDP, called the high-level club, and the second club consists of economies with below average GDP, called the low-level club. The log t convergence test results on these two clubs are reported in Table 5. Results show that there is significant evidence for growth convergence but not level convergence within each of the clubs. In fact, routine application of the Phillips and Sul (2007) clustering algorithm classifies the economies into one large and one small growth convergence club, displayed in Table 5 as growth clubs 1 and 2, respectively. Growth club 1 encompasses the high-level club members and some fast-growing members drawn from the low-level club. Figure 3 shows the time paths of the cross-sectional means of the relative transition parameter values of the clubs using both ways of classification, by GDP level in the left panel, and by GDP growth in the right panel. In both cases, the relative transition paths indicate that the clubs first diverge and then converge. In fact, the growth clubs can be merged to form the single growth club for the entire sample of economies (Phillips and Sul 2009), albeit falling short of achieving level convergence.
Fig. 3

Relative transition paths of GDP across clubs, 1970–2014. a Level clubs, b Growth clubs

Consumption

The relative transition parameters for all economies with respect to CEXP are displayed in Fig. 4. As a whole, the parameter values across economies became less dispersed over time, especially turning into the 2000s, in spite of economies taking on heterogeneous transition growth paths during the sample period. While economies like Australia have moved toward unity from an initial level that is above average, economies such as China have approached unity from a position below average. For economies such as Taiwan, relative transition has involved an initial phase of divergence followed by later convergence. In contrast, the Philippines has experienced contracted period of divergence away from the rest of the economies. The economies exhibited disparate CEXP in both the beginning and end of the sample period, despite mild evidence of converging to the group mean over time. Similar to GDP, economies with above average CEXP in 1970 continued to have above average CEXP in 2014. Korea, Malaysia and Taiwan started off to have below average CEXP in 1970, but have caught up substantially with the big consumers over time. By 2014, they have all achieved above average CEXP. Other economies had below average CEXP throughout the entire sample period, although Thailand’s CEXP was not far from the group average toward 2014, and China has been catching up fast.
Fig. 4

Relative transition paths of CEXP, 1970–2014

Table 6 contains the log t convergence test results for CEXP. According to the reported t-statistic, the null of growth convergence for all economies cannot be rejected at all conventional significance levels. However, the t-statistic for testing level convergence is significantly negative, thus lending no support for the convergence of the economies to equality.
Table 6

Log t convergence test results for CEXP 1970–2014

Group

Coefficient

t-statistic

Converging economies

Growth convergence

All

− 0.033

− 0.406

All

High-level club

0.197

2.130

AUS, BRN, HKG, JPN, KOR, MAC, MYS, NZL, SGP, TWN

Low-level club

0.045

0.239

CHN, IDN, IND, KHM, LAO, MMR, PHL, THA, VNM

Growth club 1

0.237

1.827

AUS, CHN, HKG, JPN, KOR, MAC, MMR, MYS, NZL, SGP, THA, TWN

Growth club 2

0.297

1.910

BRN, IDN, IND, KHM, LAO, PHL, VNM

Level convergence

All

− 0.033

− 25.341***

High-level club

0.197

− 19.493***

Low-level club

0.045

− 10.325***

Growth club 1

0.237

− 13.576***

Growth club 2

0.297

− 10.932***

***Indicates significance at the 1% level

Results from the log t convergence test above are consistent with the patterns observed in the SCD and HDR plots. The SCD plot in the left panel of Fig. 5 shows that the probability mass, instead of clustering along the diagonal line, is concentrated around the two sides of the CEXP group average in 2014. With reference to the HDR plot in the right panel of Fig. 5, the smallest consumers in 1970 grew to become relatively much bigger spenders in 2014, with their modes lying well above the diagonal line. On the contrary, the biggest consumers in 1970 tended to slow down their consumption growth rates and became relatively smaller consumers in 2014, as illustrated by their modes being located below the diagonal. In addition, the diagonal line crosses some 95 and 99% HDRs, while the horizontal line at unity crosses mainly the 95 and 99% HDRs. The observed intra-distribution mobility provides evidence for growth convergence with respect to CEXP among the economies.
Fig. 5

Conditional density region plots of CEXP, 1970–2014. a Stacked conditional density plot, b High density region plot

However, level convergence in CEXP is not supported by the patterns exhibited in the plots. In the SCD plot, most of the probability mass is concentrated at either high or low relative CEXP in 2014. Referring to the HDR plot, the horizontal line at unity mainly crosses the 95 and 99% HDRs. The modes characterizing economies with the highest (lowest) initial relative CEXP in 1970 are found to be clustering around the relative CEXP level of 1.1 (0.88) in 2014. It is also noteworthy that the 50% HDRs of economies with initial CEXP between 0.85 and 0.98 are crossed by the diagonal line. Stability and persistence of their relative CEXP over time suggest that these economies lag behind the others in improving and upgrading their CEXP levels and patterns.

The economies are next divided into two clubs according to their CEXP in 2014, one with CEXP above group average, the high-level club, and another with CEXP below average, the low-level club. Notice that the clubs of economies by CEXP level coincide with those by GDP level above. Furthermore, employing the log t clustering algorithm, economies can be classified into two CEXP growth clubs. The first growth club is comprised of all the high-level club members except Brunei with negative growth rate, and low-level club members with high growth rates. The second CEXP growth club includes all members from the second GDP growth club plus Brunei and two additional low GDP and CEXP economies. Thus, catch-up in terms of income tends to be a necessary, but not sufficient, condition for catch-up in terms of consumption. While an economy may be able to achieve spectacular economic growth as measured by income, the ultimate share of income that households can commandeer to fulfill the basic needs and to raise their living standards depends on factors not captured by income alone. Results presented in Table 6 suggest CEXP growth convergence, rather than level convergence, within each club. Examining the transition paths of the cross-sectional means of the relative CEXP transition parameters of the clubs presented in Fig. 6, there is evidence for the two clubs in each classification to converge to each other toward the end of the sample period, despite protracted period of persistence and/or divergence during the transition.
Fig. 6

Relative transition paths of CEXP across clubs, 1970–2014. a Level clubs, b Growth clubs

Government Expenditure

As shown in Fig. 7, the values of the relative transition parameters for GEXP across economies were widely dispersed throughout the entire sample period. There were signs of a narrowing of the gaps among the economies during the transition, and more so in the 2000s, although at a very slow speed. GEXP of the economies have evolved along stylized relative transition paths as described above for the other two macroeconomic variables. Some economies like Australia and New Zealand have transited from positions above average toward unity, whereas some other economies like China have moved toward unity from a low initial state. Economies such as Myanmar and Taiwan have shown transitory divergence initially before turning around to exhibit converging behavior. Economies like the Philippines have experienced prolonged deviating behavior from the rest of the economies. For all economies with the exceptions of Malaysia and Thailand, their GEXP levels relative to the group average stayed put in the beginning and end of the sample period. Malaysia and Thailand attained GEXP levels above average in 2014, although they started off to have below average levels in 1970.
Fig. 7

Relative transition paths of GEXP, 1970–2014

The log t convergence test results for GEXP are set out in Table 7. The null of growth convergence for all economies cannot be rejected at any conventional significance level, according to the reported t-statistic. By contrast, the t-statistic for the null of level convergence is highly significant, which means that no empirical support can be established for convergence of the economies in their GEXP levels. This is consistent with the observations found in Fig. 7 above.
Table 7

Log t convergence test results for GEXP, 1970–2014

Group

Coefficient

t-statistic

Converging economies

Growth convergence

All

0.048

1.042

All

High-level club

0.399

5.051

AUS, BRN, HKG, JPN, KOR, MAC, MYS, NZL, SGP, THA, TWN

Low-level club

0.371

9.323

CHN, IDN, IND, KHM, LAO, MMR, PHL, VNM

Growth club 1

0.278

3.125

AUS, BRN, CHN, HKG, JPN, KHM, KOR, MAC, MYS, NZL, SGP, THA, TWN, VNM

Growth club 2

0.572

1.241

IDN, IND, LAO, MMR, PHL

Level convergence

All

0.048

− 42.741***

High-level club

0.399

− 23.331***

Low-level club

0.371

− 9.323***

Growth club 1

0.278

− 19.379***

Growth club 2

0.572

− 3.097***

***Indicates significance at the 1% level

From the SCD plot in the left panel of Fig. 8, the relative GEXP levels in 2014 are found to be more dispersed for economies with low to medium 1970 relative GEXP than for those with high 1970 relative GEXP. The probably mass is in general not highly concentrated along the diagonal line. As can be seen from the HDR plot in the right panel of Fig. 8, mobility implying growth convergence is found for economies with high and low initial relative GEXP. While modes for the former are located well below the diagonal line, modes for the latter are situated above. The diagonal line crosses a number of the 95 and 99% HDRs, while at the same time the horizontal line at unity usually only crosses the 95 and 99% HDRs. There is little, if any, support for level convergence among the economies, however. The horizontal line at unity fails to cross most of the 50% HDRs. The findings from the distribution dynamics plots corroborate those from the log t convergence analysis.
Fig. 8

Conditional density region plots of GEXP, 1970–2014. a Stacked conditional density plot, b High density region plot

Same as for the other two macroeconomic variables, the sample of economies is divided into clubs by two classification schemes, viz. the level of GEXP and the growth of GEXP. According to the first scheme, the clubs with GEXP above and below the group average can be distinguished. The high GEXP level club includes all high-level economies in terms of GDP and CEXP, as well as Thailand. For both clubs, growth convergence, but not level convergence, among the club members is supported, as shown in Table 7. The second scheme amounts to an application of the clustering algorithm of log t convergence to GEXP, and economies are divided into two growth clubs. The null of level convergence, on the contrary, is strongly rejected within each club. Notice that the first growth club includes all members from the high-level club as well as some members with high GEXP growth rates from the low-level club. The cross-sectional means of the relative GEXP transition parameters of the converging clubs are plotted in Fig. 9. It can be deduced that growth convergence of GEXP for all economies can be supported. This is because although the two clubs under different classification schemes showed prolonged period of persistence and/or divergence during transition, they eventually exhibited the tendency to converge to each other.
Fig. 9

Relative transition paths of GEXP across clubs, 1970–2014. a Level clubs, b Growth clubs

In passing, notice that GEXP spans over a wider range of levels across economies in comparison with GDP and CEXP. This means that it may be a lengthier process for the economies to achieve convergence in GEXP level than in GDP and CEXP levels, which is reasonable given that fiscal convergence is related to a deep form of economic integration in terms of monetary unification.

Discussion

This subsection discusses some general possibilities that have led to the growth convergent and level divergent behavior of the Asian economies as found above. More in-depth analysis of the various structural and policy factors driving the relative performance of the economies is beyond the scope of this paper and is left for future investigation.

In general, more evident signs of growth convergence among the economies with respect to the three macroeconomic variables can be found from the late 1990s and turning into the 2000s. This finding is synchronous with the developments in regional economic integration. RCEP is envisaged as an ASEAN-centered regional free trade agreement (FTA). Although ASEAN was established in 1967, it was not until 1993 that the ASEAN FTA came into force. The poorest economies only joined ASEAN afterward, with Vietnam in 1995, Laos and Myanmar in 1997, and Cambodia in 1999. Furthermore, the ASEAN+3 (ASEAN with China, Japan and South Korea collectively) cooperation process began only in 1997, while the ASEAN + 1 FTA initiatives (ASEAN with Australia and New Zealand, China, India, as well as Japan separately) became effective in stages as late as between 2005 and 2010.

Nevertheless, the prospective RCEP members are far from convergence toward equality. Although the poorest and least developed economies have achieved significant improvements in economic performance in recent years, substantial gaps still exist between them and the rest of the economies. Some potential factors underlying disparate levels of development across the board of economies are laid out in Table 8. First, according to the World Bank (1993), the success of the eight high-performing Asian economies (HPAEs) in achieving phenomenal economic growth hinges on the accumulation of physical and human capital. As can be seen, low GDP economies generally lag behind high GDP economies substantially in terms of both physical and human capital build-up measured by per capita capital stock and the human capital index, respectively. Second, Wu and Davis (1999) and Li and Xu (2007) show that economic freedom promotes economic growth and therefore convergence. The low GDP economies are typically associated with low degree of economic freedom according to the economic freedom index. Third, Sachs and Warner (1995) find that openness promotes economic convergence. In addition, Dholakia and Talukdar (2004), Kónya and Ohashi (2007) and Aizenman and Brooks (2008) suggest that globalization drives consumption convergence through social and cultural influences. As given in Table 8, poorer economies and smaller consumers are very often those that are less integrated with the rest of the world as indicated by the globalization index. It can also be noted that smaller consumers that are less globalized are those with around half of their total household consumption expenditure devoted to food and beverages consumption. Fourth, Mauro (1995), Mo (2001), Pellegrini and Gerlagh (2004) and D’Agostino et al. (2016) show that corruption has significant negative effects on growth, while Jajkowicz and Drobiszová (2015) demonstrate that the allocation of government expenditure into different functions hinges on the level of corruption. As shown in Table 8, economies with low GDP and GEXP levels are usually associated with high level of corruption as measured by the government integrity index. These economies are also usually those with low levels of government spending on education, health care and social protection (Estrada et al. 2014).
Table 8

Economic conditions of economies.

Sources: Penn World Table 9.0; 2017 Index of Economic Freedom, World Heritage Foundation; KOF Index of Globalization (Dreher 2006); Global Consumption Database, World Bank

Economy

Capital stock per capita (2014)

Human capital index (2014)

Economic freedom index (2017)

Globalization index (2013)

Share of food and beverages consumption in total, % (2010)

Government integrity index (2017)

AUS

162,790

3.80

81.02

81.93

74.81

BRN

283,096

2.74

69.77

67.38

40.90

CHN

49,356

2.47

57.40

60.73

36.22

41.63

HKG

218,900

3.20

89.82

80.34

IDN

51,839

2.36

61.94

57.75

48.55

44.73

IND

16,906

2.05

52.64

51.26

44.63

44.29

JPN

139,808

3.54

69.58

67.86

86.06

KHM

5940

1.82

59.52

50.32

46.20

12.85

KOR

132,989

3.59

74.26

65.42

67.34

LAO

15,293

1.87

54.01

30.38

52.34

32.60

MAC

233,194

2.73

70.67

41.35

37.10

MMR

7453

1.78

52.48

34.42

29.60

MYS

63,346

2.97

73.78

79.14

51.83

NZL

98,561

3.29

83.75

78.15

89.87

PHL

20,183

2.65

65.61

57.86

47.12

38.66

SGP

284,572

3.52

88.58

86.93

87.91

THA

53,495

2.66

66.22

70.45

39.63

40.70

TWN

142,288

3.20

76.51

70.50

VNM

14,894

2.62

52.44

49.91

53.24

24.65

Capital stock and government expenditure are in constant 2011 national prices measured in US$. Human capital is based on years of schooling and returns on education

Among the low-level economies, Cambodia, Laos, India, Indonesia and the Philippines often fall into the small growth convergence club. Cambodia and Laos had experienced a prolonged period of war, political instability and regime change. India had protracted period of protectionist policies with widespread state intervention and regulation. Indonesia was one of the HAPEs identified by the World Bank (1993), but was hardest hit by the 1997–1998 Asian Financial Crisis. Slow economic progress was brought about by political and economic instability, social unrest, corruption and terrorism. The economy of the Philippines stagnated under the rule of dictatorship, with economic mismanagement and political volatility. Myanmar and Vietnam belong to the small growth convergence club with respect to GEXP and CEXP only. Myanmar suffered from protracted period of stagnation, mismanagement and isolation, but economic development has benefited substantially from its membership into ASEAN. In recent years, Vietnam has undergone high rate of industrialization and has been receiving large amount of foreign direct investment flows. China and Thailand have always been associated with the high-level economies in the large growth convergence clubs. While Thailand was among the HPAEs according to the World Bank (1993), China’s economic growth has been spectacular upon its open door policy since four decades ago.

Conclusion

In the new wave of regionalism in Asia, ASEAN + 6 economies have recently embarked on negotiations for the formation of RCEP, the world’s largest economic bloc. This paper studies whether the economic transitional and growth dynamic experiences of these economies during the sample period have been conducive or detrimental to fostering regional economic cooperation and integration. To this end, converging (or diverging) patterns of not only income, but also consumption and government expenditure, are examined. Robust empirical results are obtained when employing a nonlinear time-varying factor convergence approach in conjunction with nonparametric distribution dynamics techniques.

Asian economies are found to evolve along heterogeneous transition dynamic growth paths in all the three macroeconomic variables under investigation. For each of these variables, there is evidence in favor of growth convergence among the economies. In other words, lower-level economies tend to catch up with higher-level ones. The slow pace of growth convergence among the prospective RCEP members during the sample period is expected to be speeded up, with the adoption of the ASEAN Economic Community Blueprint 2025 in 2015, and the acceleration of negotiations within RCEP upon the US withdrawal from TPP in 2017. Thus, there is optimism toward the successful pursuit of RCEP in fostering greater economic cooperation and integration in the region.

Notwithstanding, level convergence among the economies in terms of all three macroeconomic variables is not supported. Substantial disparities remain across the board of economies. Middle-income economies are susceptible to the middle-income trap. Income convergence is a necessary, but not sufficient, condition for consumption convergence. The majority of below average consumers lag behind the others in improving and upgrading their consumption levels and patterns. There is also a long way ahead for the economies to achieve fiscal convergence in level for monetary unification. The prospect of a deepening of economic cooperation and integration in the region thus hinges on the design of regional policies and coordinated measures to effectively increase the economic opportunities of the less developed economies so as to achieve more equitable share of benefits among the regional members.

Notes

Acknowledgements

We are grateful to an anonymous referee for suggesting additional conciliation of the methodologies employed, and a discussion on factors driving the convergence (or divergence) behavior of the economies. We are also indebted to Josef Brada, the editor, for his valuable comments. This research was supported by the Research Committee of the University of Macau under research grant no. MYRG2016-00170-FBA.

References

  1. Ahn, C.Y., and I. Cheong. 2007. A search for closer economic relations in East Asia. Japanese Economic Review 58: 173–190.CrossRefGoogle Scholar
  2. Afxentiou, P.C., and A. Serletis. 1996. Government expenditure in the European Union: Do they converge or follow Wagner’s law? International Economic Journal 10: 33–47.CrossRefGoogle Scholar
  3. Aizenman, J., and E. Brooks. 2008. Globalization and taste convergence: The case of wine and beer. Review of Development Economics 16: 217–233.Google Scholar
  4. Apergis, N., C. Christou, and C. Hassapis. 2013. Convergence in public expenditures across EU countries: Evidence from club convergence. Economics and Finance Research 1: 45–59.CrossRefGoogle Scholar
  5. Aulia, M.K. 2017. The convergence analysis of the economic growth of ASEAN+3 countries and its influencing factors. International Journal of Business and Management Review 5: 22–41.Google Scholar
  6. Barro, R.J., and X. Sala-i-Martin. 1992. Convergence. Journal of Political Economy 100: 223–251.CrossRefGoogle Scholar
  7. Bashtannyk, D.M., and R.J. Hyndman. 2001. Bandwidth selection for kernel conditional density estimation. Computational Statistics and Data Analysis 36: 279–298.CrossRefGoogle Scholar
  8. Bernard, A.B., and S.N. Durlauf. 1996. Interpreting tests of the convergence hypothesis. Journal of Econometrics 71: 161–174.CrossRefGoogle Scholar
  9. Blot, C., and F. Serranito. 2006. Convergence of fiscal policies in EMU: A unit-root tests analysis with structural break. Applied Economics Letters 13: 211–216.CrossRefGoogle Scholar
  10. Carlino, G.A., and L.O. Mills. 1993. Are US regional incomes converging? A time series analysis. Journal of Monetary Economics 32: 335–346.CrossRefGoogle Scholar
  11. Cuffaro, M., F. Cracolici, and P. Nijkamp. 2006. Economic Convergence Versus Socio-Economic Convergence in Space. New York: Mimeo.Google Scholar
  12. D’Agostino, G., J.P. Dunne, and L. Pieroni. 2016. Government spending, corruption and economic growth. World Development 84: 190–205.CrossRefGoogle Scholar
  13. Dasgupta, P. 1990. Well-being and the extent of its realisation in poor countries. Economic Journal 100: 1–32.CrossRefGoogle Scholar
  14. de Bandt, O., and F.P. Mongelli. 2000. Convergence of fiscal policies in the euro area. ECB Working Paper No. 20. European Central Bank.Google Scholar
  15. de Brouwer, G., A. Ramayandi, and D. Turvey. 2006. Macroeconomic linkages and regional monetary cooperation: Steps ahead. Asian Economic Policy Review 1: 284–301.CrossRefGoogle Scholar
  16. Dholakia, U.M., and D. Talukdar. 2004. How social influence affects consumption trends in emerging markets: An empirical investigation of the consumption convergence hypothesis. Psychology and Marketing 21: 775–797.CrossRefGoogle Scholar
  17. Dreher, A. 2006. Does globalization affect growth? Empirical evidence from a new index. Applied Economics 38: 1091–1110.CrossRefGoogle Scholar
  18. Dudek, H., G. Koszela, and M. Krawiec. 2013. Changes in households expenditures structures in the European Union–is there convergence? Roczniki Naukowe Ekonomii Rolnictwa i Rozwoju Obszarów Wiejskich 100: 43–50.Google Scholar
  19. Esteve, V., S. Sosvilla-Rivero, and C. Tamarit. 2000. Convergence in fiscal pressure across EU countries. Applied Economics Letters 7: 117–123.CrossRefGoogle Scholar
  20. Estrada, G., Lee, S.-H., and D. Park. 2014. Fiscal policy for inclusive growth: An overview. ADB Economics Working Paper Series No. 423. Asian Development Bank.Google Scholar
  21. Evans, P., and G. Karras. 1996. Convergence revisited. Journal of Monetary Economics 37: 249–265.CrossRefGoogle Scholar
  22. Evans, P., and J.U. Kim. 2005. Estimating convergence for Asian economies using dynamic random variable models. Economics Letters 86: 159–166.CrossRefGoogle Scholar
  23. Feenstra, R.C., R. Inklaar, and M.P. Timmer. 2015. The next generation of the Penn World Table. American Economic Review 105: 3150–3182.CrossRefGoogle Scholar
  24. Felipe, J. 2000. Convergence, catch-up and growth sustainability in Asia: Some pitfalls. Oxford Development Studies 28: 51–69.CrossRefGoogle Scholar
  25. Fiaschi, D., and A.M. Lavezzi. 2005. Growth and Convergence Across European Regions: An Empirical Investigation. Mimeo: Department of Economics, University of Pisa.Google Scholar
  26. Goto, J. 2002. Economic preconditions for monetary cooperation and surveillance in East Asia. In Report on the Study Group on Strengthening Financial Cooperation and Surveillance, ed. Institute for International Monetary Affairs. Tokyo.Google Scholar
  27. Heng, T.M., and T.C. Siang. 1999. A neoclassical analysis of the ASEAN and East Asian growth experience. ASEAN Economic Bulletin 16: 149–165.CrossRefGoogle Scholar
  28. Hyndman, R.J., D. Bashtannyk, and G. Grunwald. 1996. Estimating and visualizing conditional densities. Journal of Computational and Graphical Statistics 5: 315–336.Google Scholar
  29. Ito, T. 2016. Growth convergence and the middle income trap. Working Paper Series No. 349. Center on Japanese Economy and Business, Columbia University in the City of New York.Google Scholar
  30. Jajkowicz, O., and A. Drobiszová. 2015. The effect of corruption on government expenditure allocation in OECD countries. Acta Universitatis Agriculturae et Silviculturae Mendelianae Brunensis 63: 1251–1259.CrossRefGoogle Scholar
  31. Jayanthakumaran, K., and R. Verma. 2008. International trade and regional income convergence. ASEAN Economic Bulletin 25: 179–194.CrossRefGoogle Scholar
  32. Jovanovic, M. 1998. International Economic Integration: Critical Perspectives on the World Economy. London: Routledge.Google Scholar
  33. Kawai, M. 2007. Evolving economic architecture in East Asia. Kyoto Economic Review 76: 9–52.Google Scholar
  34. Kawai, M., and G. Wignaraja. 2008. EAFTA or CEPEA: Which way forward? ASEAN Economic Bulletin 25: 113–139.CrossRefGoogle Scholar
  35. Khan, H. 1991. Measurement and determinants of socioeconomic development: A critical conspectus. Social Indicators Research 24: 153–175.CrossRefGoogle Scholar
  36. Kočenda, E., A.M. Kutan, and T.M. Yigit. 2008. Fiscal convergence in the European Union. North American Journal of Economics and Finance 19: 319–330.CrossRefGoogle Scholar
  37. Kónya, I., and H. Ohashi. 2007. International consumption patterns among high-income countries: Evidence from the OECD data. Review of International Economics 15: 744–757.CrossRefGoogle Scholar
  38. Lei, C.K., and P.S. Tam. 2010. A panel data approach to the income convergence among Mainland China, Hong Kong and Macao. Journal of the Asia-Pacific Economy 15: 420–435.CrossRefGoogle Scholar
  39. Lei, C.K., and P.S. Tam. 2013. Stochastic convergence of the Greater China economies: A panel unit root approach. In China’s Development and Harmonisation: Toward a Balance with Nature, Society and International Community, ed. B. Wu, S. Yao, and J. Chen. London: Routledge.Google Scholar
  40. Li, H., and Z. Xu. 2007. Economic convergence in seven Asian economies. Review of Development Economics 11: 531–549.CrossRefGoogle Scholar
  41. Lim, L.K., and M. McAleer. 2004. Convergence and catching up in ASEAN: A comparative analysis. Applied Economics 36: 137–153.CrossRefGoogle Scholar
  42. Liobikienė, G., and J. Mandravickaitė. 2011. Achievements of Lithuanian sustainable development during the integration process into the European Union. Technological and Economic Development of Economy 17: 62–73.CrossRefGoogle Scholar
  43. Liobikienė, G., and J. Mandravickaitė. 2013. Convergence of new members of the EU: Changes in household consumption expenditure structure regarding environmental impact during the prosperous period. Environment, Development and Sustainability 15: 407–427.CrossRefGoogle Scholar
  44. Mankiw, G., D. Romer, and D. Weil. 1992. A contribution to the empirics of economic growth. Quarterly Journal of Economics 107: 407–437.CrossRefGoogle Scholar
  45. Mauro, P. 1995. Corruption and growth. The Quarterly Journal of Economics 110: 681–712.CrossRefGoogle Scholar
  46. Mo, P.H. 2001. Corruption and economic growth. Journal of Comparative Economics 29: 66–79.CrossRefGoogle Scholar
  47. Moon, W. 2006. Income convergence across nations and regions in East Asia. Journal of International and Area Studies 13: 1–16.Google Scholar
  48. Park, D. 2000a. Intra-Southeast Asian income convergence. ASEAN Economic Bulletin 17: 285–292.CrossRefGoogle Scholar
  49. Park, D. 2000b. Is the Asia-Australasia region a convergence club? Asian Economic Journal 14: 415–419.CrossRefGoogle Scholar
  50. Pellegrini, L., and R. Gerlagh. 2004. Corruption’s effect on growth and its transmission channels. Kyklos 57: 429–456.CrossRefGoogle Scholar
  51. Perović, L.M., S. Golem, and M.M. Kosor. 2016. Convergence in government spending components in EU15: A spatial econometric perspective. Amfiteatru Economic 18: 240–254.Google Scholar
  52. Phillips, P.C.B., and D. Sul. 2007. Transition modeling and econometric convergence tests. Econometrica 75: 1771–1855.CrossRefGoogle Scholar
  53. Phillips, P.C.B., and D. Sul. 2009. Economic transition and growth. Journal of Applied Econometrics 24: 1153–1185.CrossRefGoogle Scholar
  54. Pinkovskiy, M. and X. Sala-i-Martin. 2016. Newer need not be better: Evaluating the Penn World Tables and the World Development Indicators using nighttime lights. NBER Working Paper No. 22216. National Bureau of Economic Research.Google Scholar
  55. Pomfret, R. 2005. Sequencing trade and monetary integration: Issues and applications to Asia. Journal of Asian Economics 16: 105–124.CrossRefGoogle Scholar
  56. Quah, D.T. 1993. Empirical cross-section dynamics in economic growth. European Economic Review 37: 426–434.CrossRefGoogle Scholar
  57. Quah, D.T. 1996. Convergence empirics across economies with (some) capital mobility. Journal of Economic Growth 1: 95–124.CrossRefGoogle Scholar
  58. Quah, D.T. 1997. Empirics for growth and distribution: Stratification, polarization, and convergence clubs. Journal of Economic Growth 2: 27–59.CrossRefGoogle Scholar
  59. Robson, P. 1998. The Economics of International Integration. London: Routledge.Google Scholar
  60. Sachs, J.D., and A. Warner. 1995. Economic reform and the process of global integration. Brookings Papers on Economic Activity 1995: 1–118.CrossRefGoogle Scholar
  61. Sala-i-Martin, X.X. 1996. Regional cohesion: Evidence and theories of regional growth and convergence. European Economic Review 40: 1325–1352.CrossRefGoogle Scholar
  62. Schott, J.J. 1991. Trading blocs and the world trading system. World Economy 14: 1–17.CrossRefGoogle Scholar
  63. Sen, A. 1987. The standard of living: Lecture I, concepts and critiques. In The Standard of Living, ed. G. Hawthorn. Cambridge: Cambridge University Press.CrossRefGoogle Scholar
  64. Ševela, M. 2004. Convergence of household expenditures of the EU-member and acceding countries in the years 1995–2002. Agriculture Economics—Czech 50: 301–307.Google Scholar
  65. Skidmore, M., H. Toya, and D. Merriman. 2004. Convergence in government spending: Theory and cross-country evidence. Kyklos 57: 587–620.CrossRefGoogle Scholar
  66. Solarin, S.A., E.M. Ahmend, and J. Dahalan. 2014. Income convergence dynamics in ASEAN and SAARC blocs. New Zealand Economic Papers 48: 285–300.CrossRefGoogle Scholar
  67. Tam, P.S. 2017. On intra- and extra-ASEAN income convergence. Mimeo, University of Macau.Google Scholar
  68. Wan, G.H. 2005. Convergence in food consumption in Rural China: Evidence from household survey data. China Economic Review 16: 90–102.CrossRefGoogle Scholar
  69. Wang, M.S. 2012. Income convergence within ASEAN, ASEAN+3: A panel unit root approach. Applied Economics Letters 19: 417–423.CrossRefGoogle Scholar
  70. World Bank. 1993. The East Asian Miracle: Economic Growth and Public Policy. Washington, D.C.: World Bank.Google Scholar
  71. Wu, W., and O.A. Davis. 1999. The two freedoms, economic growth and development: An empirical study. Public Choice 100: 39–64.CrossRefGoogle Scholar

Copyright information

© Association for Comparative Economic Studies 2018

Authors and Affiliations

  1. 1.Department of Finance and Business Economics, Faculty of Business AdministrationUniversity of MacauMacauChina

Personalised recommendations