A study on the nonlinear dynamics of ASEAN financial integration

1. Introduction

Stock market integration is important to policymakers, investors and academics. For policymakers, an integrated regional stock market implies a broader investor base that leads to more efficient capital allocation within a region. The range of financial products is also expanded, fostering the ability of domestic capital markets to compete on a global scale (Umutlu et al., 2010) [1]. Stock market integration gives investors the opportunity to efficiently allocate capital to the most productive places in a region (Narayan et al., 2011). For investors outside a region, the higher the level of stock market integration is, the lower the benefit of portfolio diversification across countries, since separate markets appear to move together (Click and Plummer, 2005). The level of stock market integration across countries has received considerable attention from international finance scholars. Recent advances in time series analysis have led to the development of various co-integration techniques that enable the investigation of long-run co-movement and equilibrium among stock markets.

This study examines the level of stock market integration in Association of Southeast Asian Nation (ASEAN) (Cambodia, Indonesia, Malaysia, Philippines, Singapore, Thailand and Vietnam) during the 2009–2022 period. Some studies have found regional Asian stock market integration to be stronger after the turmoil of the 2008 financial crisis (Yoshida, 2010; Chien et al., 2015). However, others show that the regional integration of Asian stock markets was actually held back by the sub-prime financial crisis (Caporale et al., 2019). The focus of this study is on the postcrisis period and concentrates on the profound, long-lasting impacts of the 2008 global financial crisis on economies and equity markets at both the global and regional levels. We use different frequencies of data for comparison. For this purpose, advanced co-integration techniques are applied to different frequencies of data (daily, weekly and monthly) for comparison and completeness. Based on financial theories, an integrated regional stock market is more efficient than segmented national markets (Narayan et al., 2018).

The integration of ASEAN stock markets is interesting for several reasons. First, ASEAN is the seventh largest and one of the fastest-growing economic blocs in the world [2]. Established on August 8, 1967 with five founding members, ASEAN now has ten members: Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand and Vietnam. The region's 661.8 million inhabitants were approximately 8.5% of the world's population in 2020 and offer tremendous economic potential. In 2020, ASEAN's total gross domestic product (GDP) reached US$3.0 tn, 3.5% of the world's GDP; the bloc has achieved an annual growth rate of 5.0% for the last two decades [3] and is expected to sustain that growth level for the next five years

Second, the bloc's stock markets have experienced significant growth in recent decades. Singapore enjoys a solid position among regional stock markets and has long been the first choice of ASEAN companies when going public because it creates the best opportunity for reaching international investors. However, this situation is changing rapidly as economies like Indonesia, Thailand and the Philippines seek to compete with Singapore. With a market capitalization of listed domestic companies of US$652.6 bn last year, Singapore is among the largest Asia–Pacific stock markets. However, its growth is not as robust as markets in other ASEAN countries [4].

Third, capital account liberalization and financial market deregulation are consistent policies of the bloc's governments in order to promote cross-border financial transactions. In April 2011, the ASEAN exchange network, a collaboration of seven exchanges from Malaysia, Vietnam (two exchanges), Indonesia, the Philippines, Thailand and Singapore, was formed to increase regional cross-border trading and facilitate easier funding and investment in the region.

For the purpose of the present study, an empirical analysis is conducted using Enders and Siklos's (2001) co-integration and threshold adjustment procedure. This advanced co-integration technique is superior to other co-integration techniques because it permits asymmetry in the adjustment toward equilibrium. Prior to conducting co-integration tests, we examine the stationarity of the data. For robustness and completeness, we undertake empirical analyses using different data frequencies, subject to data availability, that is, daily, weekly and monthly. Overall, the study finds evidence of strong co-integration that in turn indicates strong co-movement among the seven ASEAN stock markets between 2009 and 2022, regardless of data frequency. This accords with Yoshida's (2010) finding that regional stock integration in Asia appeared to be reinforced in the wake of the sub-prime financial crisis.

The rest of this paper is organized as follows. Section 2 reviews the literature on stock market integration, with a focus on the ASEAN context, and section 3 describes the data and methodologies used in the study. Section 4 presents the empirical results, while section 5 concludes and discusses policy implications.

2. Related literature

A significant body of research has investigated stock market integration, particularly with the use of co-integration techniques (e.g. Johnson and Soenen, 2002; Click and Plummer, 2005; Kim et al., 2005; Horvath and Petrovski, 2013; Graham et al., 2012; Tiwari et al., 2013; Chien et al., 2015; Mollah and Mobarek, 2016; Batten et al., 2019). The majority of these studies have focused on examining the degree of correlation among individual stock markets.

However, less research has examined stock market integration in Asia, including ASEAN countries, than in other regions; taken together, conclusions regarding the integration of Asian stock markets are mixed (Click and Plummer, 2005). This might be partially attributable to different methodologies, even when using co-integration techniques, different investigation periods and/or different country samples. Focusing on the post-Asian financial crisis period, Click and Plummer (2005) found stock markets in five ASEAN countries to be co-integrated. However, only one co-integrating vector was found among those five markets, leaving four common trends among the five variables.

Graham et al. (2012) found co-movements of stock market returns between the USA and 22 emerging markets at low frequencies (i.e. long-term fluctuations). The degree of co-movement varied by country. Their findings suggest the time-varying degree of stock market integration; one important implication of their findings is that investing selectively in emerging markets may provide significant diversification benefits that depend on the specific investment horizon.

Chien et al. (2015) found at most one co-integrating vector among the stock markets of China and the ASEAN-5 (Indonesia, Malaysia, the Philippines, Singapore and Thailand) between 1994 and 2002. This finding suggests that regional financial integration among China and the ASEAN-5 countries gradually increased but remained limited. In the same vein, Wu (2020) shows that the high level of stock market integration in East and Southeast Asia is driven by common global factors but is not as strong as it appears.

Although there are some empirical studies on stock market integration in Asian countries and subregions, such as the ASEAN-5 (e.g. Click and Plummer, 2005), the ASEAN-5 with China (e.g. Chien et al., 2015; Wu, 2020) and other external markets (e.g. Mensah and Premaratne, 2018; Caporale et al., 2022), none has examined the relationships of ASEAN stock markets while including ASEAN countries outside the ASEAN-5 block, such as Cambodia and Vietnam, partly due to the unavailability of the stock market data for these countries. Furthermore, none of the above studies considered nonlinear adjustment mechanisms, which constitute a critical development in recent empirical time series literature (Enders and Siklos, 2001). This is because a large number of studies have documented an asymmetric adjustment over the course of the business cycle for key macroeconomic variables like GDP, unemployment, industrial production, stock price, exchange rates and commodity prices (e.g. Enders and Siklos, 2001; Chen et al., 2005; Yau and Nieh, 2009; Ghosh and Kanjilal, 2016). Considering the properties of a theoretical model, the specific spill-overs of real economic activities in ASEAN countries should display similar dynamics in their stock prices (Chien et al., 2015). The authors explore this possibility by examining stock market integration among ASEAN-5 member states in the aftermath of the global sub-prime financial crisis.

It is expected that increasing economic integration through trade linkages and investment flows will be reflected in the interdependence of stock markets (Narayan and Smyth, 2004; Caporale et al., 2019). Based on the existing literature, we argue that the economic integration of ASEAN countries is determined by multiple direct and indirect channels that can become manifest in a stronger co-movement pattern among the members' equity exchanges. For example, regional economic integration via trade or financial development can shape a common sentiment regarding investment opportunities in the integrated markets. Additionally, when expanding their operations to regional markets through direct investments or mergers and acquisitions, the performance of firms listed on multiple exchanges could steer the evolution of regional stock indices in the same direction. Therefore, a simple and intuitive test for market integration is to check for long-run equilibrium relationships among stock exchanges.

However, there is little reason to believe that the impact of positive and negative changes in sentiments or the investment environment more broadly among member states would affect every member to the same degree. One possible reason is the well-established leverage effect in financial markets, whereby negative changes in stock prices induce higher subsequent volatility than positive changes, due to a reduction in equity value and investors' confidence and increasing uncertainty about future economic prospects (Ait-Sahalia et al., 2013; Black, 1976). Additionally, while there is a perception that world capital markets have become more integrated over time, country-specific investigations suggest that this is not always the case and that there exists a large degree of heterogeneity among countries regarding financial integration (Bekaert and Harvey, 1995; Bekaert et al., 2009), especially during crises.

Overall, although co-integration analyses have been widely employed to investigate market integration, the existing literature does not adequately consider asymmetric adjustment dynamics. To fill this gap, we employ Enders and Siklos's (2001) tests for threshold co-integration and Shin et al.’s (2014) non-linear autoregressive distributed lag (NARDL) model for seven ASEAN nations' stock markets and investigate two related hypotheses: (1) co-integration exists among ASEAN markets and (2) adjustments toward long-run co-integrating relationship are asymmetric.

3. Data and methodologies

The stock index data of the seven ASEAN countries considered in this study were obtained from Thomson Reuters's financial data vendor at daily, weekly and monthly frequencies. All variables are log-transformed. The sample period runs from January 1, 2009 to May 4, 2022, excluding holidays, and contains 3,480 observations for each market. The choice of countries and sample period were subject to the availability of stock market data suitable for empirical analyses in the aftermath of the global financial crisis. Table A1 in Appendix describes the stock market indexes used in the sample and their data sources, while Table A2 provides statistical descriptions of the variables by level, log level and first difference of log. In terms of stock market returns, we note that all markets posted positive average performances and that markets in less developed countries (Cambodia, Indonesia, Philippines, Thailand, and Vietnam) appeared to outperform even the most developed countries (Singapore and Malaysia) during the sample period.

To explain the integration of stock markets in seven ASEAN countries, the first step is to test for a unit root type of non-stationarity; we begin by testing for the presence of a unit root in stock prices. We employ two groups of unit root tests for comparison and completeness. One is based on Fourier approximations (augmented Dickey–Fuller [“ADF”] and Kwiatkowski–Phillips–Schmidt–Shin [“KPSS”] tests) and considers the no-linearity issue, while the other group—ADF, Phillips–Perron (“PP”), KPSS and Zivot and Andrews (1992)—does not.

In analyzing stock market integration in ASEAN countries, we account for nonlinear adjustment mechanisms, an important recent development in the time series econometric paradigm. Specifically, we employ Enders and Siklos's (2001) tests for threshold co-integration [5].

Let Yi,t be the logarithmic transformation of the stock market index of the ith ASEAN country (j=1,…,7) and Xt={Xj,t}(j=1,…,6; j≠i) be the vector of the other six logarithmic ASEAN stock market indices, all of which are scalar I(1) variables. The co-integrating relationship in the first stage can be expressed as follows:

Although Enders and Siklos's tests allow us to investigate whether adjustment of the error-correction term εt is asymmetric (subject to shocks to the error correction term itself), these tests do not allow us to pinpoint the source of the shocks. To complement them, we seek to decompose the shocks into positive and negative changes in stock indices—rather than in εt—and thus gauge the degree to which each member country reacts to common shocks from the other ASEAN markets. As such, we augment the above analysis by adopting Shin et al.’s (2014) nonlinear autoregressive NARDL (p, q), which is specified as

4. Empirical results

5. Concluding remarks

This study finds strong evidence of co-integration relationships existing in all possible considerations of stock market indexes for seven ASEAN countries between 2009 and 2022. The co-integration exists regardless of data frequencies. Overall, our empirical evidence shows a strong regional financial integration among ASEAN countries. In the longer-term future, integration between the seven stock markets will result in a larger and more diversified market and pool of investors. Drawing on our empirical results, we provide some suggestions below.

The evidence of stock market integration implies that ASEAN policymakers can adopt a coordinated approach to boost sequential liberalization and integration processes and improve their regulatory standards to comply with international standards. In addition, integration in stock markets might help foster the development of less developed capital markets in the region, such as Vietnam and Cambodia. However, despite gains in market efficiency, integration also entails potential risks. To address this issue, financial integration should be executed in a systematic and complementary manner.

The ASEAN community still has much work to do. In addition to expanding its relations with external partners such as the six Indo-Pacific states with which ASEAN has existing free trade agreements (China, Japan, South Korea, India, Australia and New Zealand), the ASEAN bloc needs to consistently promote co-operation throughout the region. Each ASEAN country can offer leverage for other members. The less developed group of ASEAN states, such as Vietnam, Cambodia, Laos and Myanmar, also need to strengthen their macroeconomic policy reform efforts to improve the competitiveness of their economies and help preserve and enhance peace, stability and prosperity in the region.

In this study, we were particularly interested in examining the long-run relationships among ASEAN stock markets. Future studies could explore short-run causal linkages between ASEAN stock markets to see if there is a solid theoretical basis for that causality to exist. Furthermore, it would be helpful to have a thorough analysis of ASEAN stock market integration across different sectors of the economy. As the worst of the COVID-19 pandemic is expected to be over in the very near future, research could be conducted on ASEAN financial integration during the recovery period to address the pandemic's long-term impacts.