Revisiting Feldstein–Horioka Puzzle: Evidence from SAARC Economies

Abstract

In the macroeconomic perspective, domestic saving and investment constitute two important wheels that keep the growth process moving on towards stability. But when domestic savings tend to fly away to foreign countries for excess returns, the warranted domestic investment remains unattained thereby which impair the growth trend unless the gap is bridged by the foreign investment inflows. However, excess inflows of foreign investments may deteriorate current account balances, if not appropriately absorbed. In this pretext, this article revisited the Feldstein–Horioka puzzle with the twin-deficit hypothesis for SAARC countries. The results lend to support the validity of the twin-deficit hypothesis along with a low degree of association between domestic savings and investment in the region thereby justifying the argument that international capital movements or financial integration have increased in the post-reform era. This observation has significant policy implications for the sustained growth of the SAARC nations.

Keywords

Saving investment twin-deficit Feldstine–Horioka puzzle SAARC

JEL Code: F21, F30, F36, O16

I. Introduction

In the macroeconomic perspective, capital accumulation has been occupying a central place in the growth process of economies across time and space. It, by pooling together the savings of surplus-spending units and mobilising the same through an efficient financial and credit system, provides resources for the production of capital goods (Meier & Baldwin, 1957). Thus, saving and investment form two critical cogs in the wheels of long-run economic growth which help to achieve price stability with near full-employment in economies across the globe (Mishra, 2011). Precisely, the saving–investment relationship is pivotal as any divergence between planned saving and preferred investment can lead to cyclical fluctuations in an economy (Gutiérrez & Solimano, 2007). At this juncture, it quite pertinent to ask: why do domestic saving and investment diverge in an economy? The most plausible answer drawn from the macroeconomic literature is the cross-border flight of capital (Dzhumashev & Cooray, 2016; Lucas, 1990). Assuming that the capital owners can easily invest anywhere in the world, they generally prefer countries where the marginal productivity of capital is higher, and such investment continues until the return per unit of investment is equalised across different countries. In line with this argument, an Indian capital owner will have no incentive for investing in India when the return per unit of his investment is highest, let us say in Japan. The implication is that an increase in domestic savings need not necessarily increase domestic investment. The bottom line is that in the absence of regulations in the international financial markets and the presence of perfect capital mobility, domestic savings of a country can fly to countries with the most productive investment opportunities thereby making domestic saving and investment rates uncorrelated. In contrast to this basic macroeconomic argument, Feldstein and Horioka (1980) in a seminal article, observed the existence of a high degree of correlation between saving and investment rates in a sample of 16 Organisation for Economic Co-operation and Development (OECD) countries. And, thereafter this contradictory observation has been widely discussed in the literature of international finance as Feldstein–Horioka puzzle, henceforth F–H puzzle (Apergis & Tsoumas, 2009; De Wet & Vab Eyden, 2005; Obstfeld & Rogoff, 2000).

Feldstein and Horioka (1980), based on their empirical observation concluded that about 85–95 per cent of domestic savings are domestically invested. This observation goes against the conventional argument that industrialised economies such as OECD have fewer restrictions on cross-border movements of capital. Precisely, it points to the international capital immobility and supports the existence of strong home-bias. Such home-bias argument is quite evident in the works of Adam Smith. In the Smithian version, the capital owners prefer to invest in their own country as much as possible because such investments are safe and secured in domestic lands. Thus, in the Smithian world, invisible hands guide investment decisions of capital owners in pursuit of security rather than profit, and thus, domestic savings and investment are highly correlated (Yasutomi & Horioka, 2010). Recently, Horioka et al. (2015) revisited this home-bias phenomenon and concluded about the presence of such bias (international capital immobility), especially in the case of short-term debt securities with varying degrees of strength across countries and regions. Specifically, the study observed that the home-bias is strong in advanced Asia, especially in Japan, and relatively weak in developing Asia. This finding has made it imperative to revisit the paradox for the economies of developing Asia. So, in continuation to the findings of Horioka et al. (2015), we have considered a subset of the South Asian countries, namely, Afghanistan, Bangladesh, Bhutan, India, Nepal, Pakistan and Sri Lanka, which constitute the SAARC region along with the Maldives, to revisit the F–H puzzle.

In this era of globalisation, the financial integration in developing countries has become a major challenge because the degree of international capital mobility can pose the problems of fiscal and current account deficits; henceforth twin-deficits, and can also affect the extent of the growth of their economies. Since the twin-deficit issue is related to the degree of international capital mobility, F–Z puzzle is related to the issue of twin-deficits. So, Obstfeld and Rogoff (2000) called F–Z puzzle as the mother of all puzzles. In line with the observations of Horioka et al. (2015) for the low degree of correlation between domestic savings and investment in developing Asia, there will be a high degree of international capital mobility in these economies. Such a high degree of capital mobility is likely to lead to fiscal and current account deficits in these countries thereby creating the issue of twin-deficits (Bagheri & Hazrati, 2012). Therefore, it is quite pertinent to revisit the F–H puzzle in the perspectives of saving–investment relationship and international capital mobility in developing Asian economies such as SAARC nations. In this context, this article addressed the following questions: Does the F–H puzzle valid in SAARC economies? If yes, what does it imply for international capital mobility or financial integration? Given that the F–H puzzle holds for SAARC economies, what are the macroeconomic factors that delineate the savings–investment relationship?

In this study, we have adopted two approaches to revisit the F–H puzzle—one approach stems from the empirical model of Feldstein and Horioka (1980), while another stems from the empirical model of Fidrmuc (2003). These two empirical models have been augmented by taking into consideration other relevant macroeconomic variables including the growth rate of income, rate of inflation (INF), exchange rate and global financial shocks. The estimation of these augmented empirical models in the panel fixed effects framework lends to support the validity of the twin-deficits hypothesis as well as F–H puzzle for the selected SAARC economies. This, on the one hand, implies the presence of a tendency for international capital mobility, and on the other hand, indicates fiscal deficit induced deteriorating current account balance (CAB) position in South Asian economies. Given the argument that countries with relatively higher capital mobility are more likely to be exposed to the adverse effects of international market fluctuations (Ketenci, 2016), the novelty of this study is the inclusion of the global financial recession shock (2007–2009) in explaining F–H puzzle in the post-liberalisation era. The remaining of the study is organised as follows: Section II reviews the relevant literature to justify the research problem and set the building block of the study; Section III presents observations from the trends in domestic savings and investment in the SAARC region; Section IV elaborates the data and methodology used in the study; Section V draws inferences from the empirical analyses; Section VI makes the concluding remark; and Section VII notes down the scope for further study while pointing out the limitations.

II. Literature Review

In the macroeconomics literature, the validity of the F–H hypothesis has been studied in specific country cases as well as cross-country cases. This hypothesis predicts that the degree of association between domestic savings and investment rates is high (Feldstein & Horioka, 1980; Feldstein, 1983). It, thus, implies the absence of perfect capital mobility across countries and exhibits home-bias (Caprio & Howard, 1984; Murphy, 1984; Obstfeld, 1986a, 1986b; Wong, 1990; Yasutomi & Horioka, 2010). Later, this hypothesis constituted the subject matter of many researchers. But the research outcomes are mixed in explaining this hypothesis. Several studies concluded the existence of a high degree of savings–investment association and thus, low level of international capital mobility thereby supporting F–H hypothesis in large and small economies (Akkoyunlu, 2020; Lam, 2012; Bagheri & Hazrati, 2012; Eyuboglu & Uzar, 2020; Hassan et al., 2014; Irandoust, 2019; Petreska & Mojsoska-Blazevski, 2013; Tasar, 2017; Yildirim & Orman, 2018). Nevertheless, this correlation was found to be weaker for developing economies (Adeniyi & Egwaikhide, 2013; Bangaké & Eggoh, 2010; Chan et al., 2011; Chang & Smith, 2014; Coakley et al., 1999; De Wet & Van Eyden, 2005; Dooley et al., 1987; Horioka et al., 2015; Kasuga, 2004; Mamingi, 1997; Mumtaz & Munir, 2016; Patra & Mohanty, 2020; Payne & Kumazawa, 2005; Raheem, 2017; Shahbaz et al., 2010; Sinha & Sinha, 2004; Vamvakidis & Waczairg, 1998; Wong, 1990; Younas, 2007).

Conversely, a bunch of studies reveal disagreement with this hypothesis of co-movement of domestic savings and investment. The argument put forward is that most of the countries in the late twentieth century have had accepted the policies of general deregulations and capital market integration, and thus, capital movements across the borders arguably occupy a significant place in the modern world economy. Precisely, investors reveal their apparent indifference in choosing investment avenues across the borders. In other words, the rates of domestic savings and investment cannot be highly correlated to validate the F–H hypothesis (Aristovnik & Djuric, 2010; Baharumshah et al., 2009; Baxter & Crucini, 1993; Fidrmuc, 2003; Frankel & MacArthur, 1988; Ghosh, 1990; Marinheiro, 2008; Obstfeld, 1986a, 1986b; Saeed & Khan, 2012; Tesar, 1991; Tobin, 1983; Westphal, 1983). Miller (1988) provides evidence that savings and investment in the US economy were co-integrated throughout the regime of the fixed exchange rate, but not in the regime of the flexible exchange rate. Otto and Wirjanto (1989) reveals that the savings–investment in the United States and Canadian economy are not co-integrated. Montiel (1994) elaborated about the vulnerability of the F–H hypothesis for indirect correlations between domestic savings and investment that did not reflect capital mobility. A few recent studies conclude that the financial openness has increased the capital mobility in the world and invalidated the F–H hypothesis (Georgopoulos & Hejazi, 2005; Isaksson, 2001; Younas, 2007; Younas & Chakraborty, 2009). Khundrakpam and Ranjan (2010) found a weaker relationship between saving and investment rates with the inclusion of the post-reform period which depicts the era of more liberalised capital flows. Ketenci (2016) found lower saving retention coefficient and thus, higher international capital mobility in the post-Asian crisis period. In another study, Phiri (2017) found increase in international capital mobility in the post-global financial crisis period thereby weakening the F–H high correlation between domestic savings and investment. Therefore, countries with higher international capital mobility are likely to be affected by the contagion effects of global financial crisis.

The empirical literature also relates the F–H puzzle with the twin-deficit hypothesis and puts forth two strands of argument. While one strand of studies supports the F–H hypothesis of the strong correlation between domestic savings and investment and low level of international capital mobility (Aristovnik, 2005; Bagheri & Hazrati, 2012; Lam, 2012; Wirasti & Widodo, 2017), another strand of studies found the evidence against this hypothesis (Aristovnik & Djuric, 2010; Baharumshah et al., 2009; Fidrmuc, 2003; Khan & Saeed, 2012; Marinheiro, 2008).

A bunch of empirical literature reveals the role of macroeconomic factors such as country size, CAB, performance of financial sector, growth rate of income, foreign direct investment (FDI) inflows, productivity shocks, INF, exchange rate regime, and technological and demographic variables in influencing the international capital mobility to evaluate the F–H puzzle (Akkoyunlu, 2020; Ang, 2007; Attanasio et al., 2000; Baxter & Crucini, 1993; Chu, 2012; Coakley et al., 1996; Corbin, 2001; Dash, 2019; Engle & Kletzer, 1987; Fouquau et al., 2008; Ho, 2003; Jansen, 1996; Jiranyakul & Brahmasrene, 2009; Kasuga, 2004; Murphy, 1984; Obstfeld, 1985; Patra & Mohanty, 2020; Pelgrin & Schich, 2008; Razak & Masih, 2017; Shahbaz et al., 2010; Summers, 1985; Taylor, 1994).

Owing to the observation of Wahid et al. (2010) regarding the invalidity of the F–H hypothesis for the South Asian region, Riaz (2011) regarding inconclusive outcome about the F–H dilemma and F–H dilemma Horioka et al. (2015) regarding low saving retention coefficient for developing Asian countries, given their dynamics of financial integration, we agree that the validity of F–H hypothesis remains a conundrum in the SAARC region. Shahbaz et al. (2010) conclude about low saving retention coefficient for Pakistan while controlling for inflation, exchange rate and openness. Rahman (2015) and Suruga and Rahman (2016), by examining the long-run relationship between domestic savings and investment in SAARC economies in the F–H framework, provide the evidence of a low degree of the positive association between domestic savings and investment as well as reasonable mobility of capital thereby making the F–H hypothesis relatively weaker. In another study, Mumtaz and Munir (2016) observed that the twin-deficits hypothesis does not hold in South Asian countries while F–H hypothesis holds only for Bangladesh and Sri Lanka, but not for India and Pakistan. In a recent study, Patra and Mohanty (2020) observed that domestic savings play a significant role in influencing the domestic investment in the South Asian region while controlling for openness and financial sector performance, and that the positively significant saving retention coefficient is dependent on regime shifts.

It is inferred from the review of past studies that

Only a few studies are focusing on the South Asian region for evaluating the F–H puzzle.

Only a few studies are there which incorporate other macroeconomic variables while evaluating the robustness of the F–H puzzle in the region.

To the best of our knowledge, no studies are there which control for the global contagion effects emanated from the USA financial recession in the SAARC region.

Given the argument that the financial openness has increased in the post-reform period (Mishra & Mishra, 2015), studies are lacking to focus on this period.

The validity of the F–H hypothesis has remained a controversial one even in the SAARC region.

Thus, it is essential to revisit the issue for SAARC economies. Therefore, this study attempts to fill these gaps keeping in view that the SAARC economies are taking their saving–investment decisions in a more liberalised and globalised world where international financial markets are more or less integrated.

III. Trends in Savings and Investment Rates in SAARC Economies

In this section, we have examined the trends in the share of domestic savings and investment in GDP in the selected SAARC region over the study period. Table 1 summarises the 5-yearly average¹ of the share of domestic savings and investment in GDP for SAARC economies. In addition, Figure 1 plots the trend in the share of domestic savings and investment in GDP for these economies over the sample period.

Table 1.

Savings and Investment (% of GDP) in SAARC, 2002–2019

Country/Year	2002–2006		2007–2011		2012–2016		2017–2019
Country/Year	Savings	Invest	Savings	Invest	Savings	Invest	Savings	Invest
Afghanistan	–36.53	18.78	-18.94	18.34	3.63	17.93	7.51	18.63
Bangladesh	20.94	25.20	20.34	26.45	22.50	28.75	24.31	31.11
Bhutan	31.68	58.03	32.47	51.09	30.41	56.86	24.70	47.56
India	30.17	30.74	33.34	34.41	31.46	30.35	29.27	28.16
Nepal	10.07	20.10	10.90	21.58	9.35	24.72	16.35	33.22
Pakistan	15.34	15.65	9.99	15.49	8.28	13.61	6.16	14.77
Sri Lanka	16.66	22.20	18.53	24.75	24.06	27.70	22.89	26.35
Mean	12.62	27.24	15.23	27.44	18.53	28.56	18.74	28.54
Mean Gap	–14.62		–12.21		–10.03		–9.80
S.D.	23.05	14.41	17.68	12.08	11.30	13.88	8.99	10.67
Correlation	0.51		0.72		0.75		0.67

Source: Authors’ calculation.

There are certain key observations on the trends in savings and investment in the SAARC region made from Table 1 and Figure 1. First, on average, both the shares of domestic savings and investment in GDP were low within and across the countries in the region. Second, the average savings-GDP and investment-GDP ratios depict an increasing trend over the periods across the countries. Third, there has been a falling trend in the savings–investment gap over the periods in the SAARC region. Fourth, the average fluctuations (measured by S.D.) in savings-GDP and investment-GDP ratios depict a continuous decline over the periods. Fifth, the correlation between the savings-GDP and investment-GDP ratios is throughout positive and reveals a rising trend over the periods. This is an indication of co-movement of the savings-GDP and investment-GDP ratios in the SAARC region. Sixth, the savings-GDP and investment-GDP ratios for Afghanistan, Bangladesh, Nepal and Sri Lanka show rising trends over the years. Seventh, the saving-GDP and investment-GDP ratios for Bhutan, India and Pakistan are showing declining trends over the years.

Despite the increasing trends of the share of domestic savings and investment in GDP, the share of domestic savings falls below the share of domestic investment for all SAARC economies over the study period. This is an indication of inadequate domestic savings to finance the domestic investment projects and therefore, relying on foreign capital may increase current account deficits. Based on these observations, a discernible relationship between the savings-GDP and investment-GDP ratios in the SAARC region is difficult to be concluded. This noticeable relationship between these two variables might be due to the effects of other macroeconomic factors including the growth rate of income, FDI inflows, INF, exchange rate dynamics, fiscal and current account imbalances. Therefore, in this study, we examined the savings–investment relationship in a panel framework in the perspective of twin-deficits and F–H puzzle.

IV. Data and Methodology

In this article, the domestic savings and investment relationship as hypothesised by Feldstein and Horioka is revisited for SAARC economies over the period from 2002 to 2019. For this purpose, we have formed a panel dataset consisting of seven SAARC economies, namely, Afghanistan, Bangladesh, Bhutan, India, Nepal, Pakistan and Sri Lanka and 19 annual periods. The reasons for considering the study period from 2002 to 2019 include the following:

During this period, the SAARC countries have opened up their economies to rests of the world in terms of trade and investments which is likely to affect the domestic saving–investment relationship,

During this period, SAARC economies have weathered the contagion effects of US financial recession having the obvious repercussions for the domestic saving–investment association.

For this period, data on required macroeconomic variables are available to form a panel dataset.

Following the empirical conventions for studying F–H puzzle (Akkoyunlu, 2020), we have taken Gross domestic savings as a percentage of GDP and Gross fixed capital formation as a percentage of GDP as measures of domestic savings and domestic investments, respectively. The annual data for these two variables for the sample period have been compiled from the world development indicators database of the World Bank. Since the degree of association between domestic savings and domestic investment has implications for capital mobility (Patra & Mohanty, 2020), we have included in the study foreign direct investment as measured by net inflows of foreign direct investment as a percentage of GDP, and its annual data have been compiled from the world development indicators database of the World Bank. Furthermore, the domestic saving–investment correlation and its consequential impacts on international capital mobility or financial integration in the era of globalisation can be affected by the long-term solvency constraints on current accounts of economies (Fouquau et al., 2008; Pelgrin & Schich, 2008), the extent of economic growth (Fouquau et al., 2008), and domestic and global production shocks (Chu, 2012). Therefore, we have included CAB as a percentage of GDP, the growth rate of GDP per capita, INF, the real exchange rate (RExR) of domestic currency against the USD and US recession dummy (URD) (2007–2009) in the study. The annual data on CAB have been compiled from the Asian Development Outlook (2007–2020) of the Asian Development Bank. The annual data on the growth rate of GDP per capita and the INF have been compiled from the world development indicators database of the World Bank. The annual data on the RExR of domestic currency against the USD have been calculated by multiplying the nominal exchange rate to the ratio of CPI of USA and CPI of the national economy. For this purpose, the required annual data have been compiled from the International Financial Statistics database of IMF.

The entire empirical exercise followed in this study can be segregated into two parts—in the first part, we have augmented the basic Feldstein and Horioka (1980) equation to revisit the F–H puzzle, and in the second part, we have augmented the basic Fidrmuc (2003) empirical equation to revisit the issue while taking into consideration the ‘twin deficit hypothesis’ in SAARC countries (for description of model variables, please refer to Appendices A and B).

The basic Feldstein and Horioka (1980) equation is as follows:

{(\frac{I}{Y})}_{i t} = ξ_{1} + ξ_{2} {(\frac{S}{Y})}_{i t} + ε_{i t}

(1)

Here, I is the gross domestic investment as measured by gross fixed capital formation; Y is the gross domestic product; and S is the gross domestic savings; In Equation (1), $ξ_{2}$ , the savings-retention coefficient, is expected to have a positive sign in line with the observations of Feldstein and Horioka (1980). If $ξ_{2}$ is equal to 1, then Equation (1) indicates imperfect international capital mobility or a high degree of correlation between domestic rates of savings and investment, and if $ξ_{2}$ is equal to 0, then Equation (1) indicates perfect international capital mobility or low degree of correlation between domestic rates of savings and investment. The empirical literature supports that the strong association between domestic savings and domestic investment may be due to several factors including macroeconomic growth, degree of trade openness, CAB, country size, domestic and global production shocks, age dependency ratio, etc., (Chu, 2012; Fouquau et al., 2008; Patra & Mohanty, 2020). In this study, we have augmented the basic Feldstein and Horioka (1980) equation by including annual growth in GDP per capita (G), FDI, INF, CAB and URD. The use of URD (2007–2009) would indicate the strength of the empirical relationship amid global financial shocks. Therefore, we have estimated the following panel equation:

{(\frac{I}{Y})}_{i t} = ξ_{1} + ξ_{2} {(\frac{S}{Y})}_{i t} + ξ_{3} G_{i t} + ξ_{4} F D I_{i t} + ξ_{5} I N F_{i t} + ξ_{6} C A B_{i t} + ξ_{7} U R D_{i t} + ε_{i t}

(2)

In our panel dataset, the cross-section dimension is smaller (7) than the time period (19). Thus, it is essential to check whether the panel dataset is cross-sectionally independent or not. Baltagi and Pesaran (2007) argued that cross-section dependence can arise due to spatial or spillover effects or could be due to unobserved common factors. For this purpose, we have used the Cross-Sectional Dependence (CD) test as proposed by Pesaran (2004) as no other test performs better than CD (Moscone & Tosetti, 2009). The CD test statistic is given by

C D = \sqrt{\frac{2 T}{N (N - 1)}} \sum_{i = 1}^{N - 1} \sum_{j = i + 1}^{N} {\overset{⌢}{ρ}}_{i j}

(3)

Here ${\overset{⌢}{ρ}}_{i j}$ is the average of the pair-wise correlation coefficients of ordinary least squares (OLS) residuals regressions under fixed/random effect model. The null hypothesis under the CD test is cross-sectional independence (it means errors in different cross-section units are not correlated). Baltagi and Pesaran (2007) mentioned that the first generation unit root tests are adequate in the presence of cross-sectional independence. Therefore, we have used the panel unit root methods as proposed by Im et al. (2003) and Levin et al. (2002). The Levin, Lin and Chu (LLC) test assume that the autoregressive coefficient is common across cross-sections whereas the Im, Pesaran and Shin (IPS) test assume that the said coefficient varies freely across cross-sections. In other words, LLC test assumes that there is a common unit root process, whereas IPS test assumes that there is an individual unit root process. In both these tests, the null hypothesis is that the variable contains a unit root (i.e., it is not stationary). It will be seen from Section V that there is no issue of residual CD in the panel model (2), and that all variables are integrated of order zero, that is, all variables are level stationary. Furthermore, the result of the Hausman test indicates that the fixed effects regression is the best choice under the present circumstances.

Furthermore, as a robustness check, we have also augmented the basic Fidrmuc (2003) empirical equation to test the validity of F–H hypothesis for SAARC economies. Derived from the national income identity, it is known that the current account deficit is the sum of the saving–investment gap and the fiscal deficit in an economy. Slightly modifying this identity, Fidrmuc formulated the following empirical equation:

C A D_{i t} = ψ_{1} + ψ_{2} F s D_{i t} + ψ_{3} I_{i t} + ε_{i t}

(4)

Here, CAD is the current account deficit, FsD is the fiscal deficit and I is the domestic investment. In Equation (4), the coefficient of FsD is assumed to be positive ( $ψ_{2} > 0$ ) to validate the ‘twin deficit’ hypothesis as high fiscal deficit increases the volume of current account deficit, and that of the domestic investment is assumed to be negative ( $ψ_{3} < 0$ ). If the absolute value of $ψ_{3}$ is significantly lower than one, then (4) validates the F–H hypothesis. The existing empirical literature suggests for augmenting the basic Fidrmuc (2003) equation for getting robust outcomes. Evidence supports the use of domestic production (Bagheri & Hazrati, 2012; Mohanty, 2019), rate of interest (Banday & Aneja, 2019; Kim & Roubini, 2008; Mohanty, 2019), trade openness (Mohanty, 2019), the exchange rate (Kim & Roubini, 2008; Mohanty, 2019; Shahbaz et al., 2010), INF (Banday & Aneja, 2019; Shahbaz et al., 2010), money supply (Banday & Aneja, 2019) and population growth (Bagheri & Hazrati, 2012). In this study, we have also augmented the basic Fidrmuc (2003) equation by including FDI, INF, RExR and URD. The inclusion of foreign direct investment would indicate the impact of inward capital mobility on the current account deficit. Therefore, we have estimated the following panel equation:

C A D_{i t} = ψ_{1} + ψ_{2} F s D_{i t} + ψ_{3} I_{i t} + ψ_{4} F D I_{i t} + ψ_{5} I N F_{i t} + ψ_{6} E x R_{i t} + ψ_{7} U R D i t + ε_{i t}

(5)

It will also be seen from Section IV that all the variables in the panel specification (5) are integrated of order zero, and their use does not lead to residual CD. Furthermore, the result of the Hausman test indicates that the fixed effects regression is the best choice under the present circumstances.

V. Results and Discussion

At the outset, we have checked the residual CD in both the panel specifications (2) and (5) using the Pesaran’s CD test, and the results are summarised in Table 2. In both these panel models, the CD test failed to reject the null hypothesis of no residual CD thereby justifying the use of the first generation panel unit root tests. The results of the first generation unit roots tests—LLC and IPC tests—are summarised in Table 3.

Table 2.

Pesaran’s Cross-Sectional Dependence Test—SAARC Panel

Panel Specification	Panel Data Model—SAARC	CD Test Stat.	p-value
(2)	Fixed Effect	–2.229	0.257
(2)	Random Effect	–1.658	0.972
(5)	Fixed Effect	–0.459	0.646
(5)	Random Effect	–0.211	0.832

H₀: No cross-sectional dependence.

Source: The authors.

It is observed from Table 3 that all the variables used in the panel specification (2) are integrated of order zero, that is, they are all level stationary. Therefore, we have to choose between fixed effects and random effects estimation for the panel specification (2). The results are presented in Table 4. The Hausman test outcomes indicate that the fixed effects model is the better choice under the given conditions. The values of R-sq. and adj. R-sq. also indicate better explanatory power of the fixed effects model. The saving-retention coefficient in the fixed effects model (0.2753) is positively significant at 0.01 levels. It means only 0.2753 per cent of domestic investment is financed through domestic savings when there is an increase in domestic savings by 1 per cent. It implies a very low degree of association between domestic savings and domestic investment while controlling for the growth rate of income, FDI flows, INF, CAB and global financial recession shocks in SAARC economies. The savings-retention coefficient being positive and low in value has several implications, which are as follows:

In SAARC economies, there are relatively high degrees of capital immobility at domestic levels.

In these economies, capital has relatively high degrees of mobility at international levels.

In selected South Asian countries F–H puzzle holds (invalidates the F–H hypothesis of a high degree of correlation between domestic savings and investment) indicating the absence of home-bias.

The selected South Asian economies have fair degrees of financial integration in the post-liberalisation era.

In SAARC economies, the substitution possibility of domestic savings by foreign savings is more.

If the selected SAARC economies prefer to depend on international finance, then their current account solvency would deteriorate causing BOPs imbalances.

Table 3.

Results of Panel Unit Root Tests: SAARC Panel

Variables	IPS W-stat. (Level)	LLC t-stat (Level)	Order of Integration
Gross Fixed Capital Formation (% of GDP)	–1.8535**	–1.7535**	I(0)
Gross Domestic Savings (% of GDP)	–1.4677***	–2.4142*
GDP Per Capita (% Annual)	–3.9152*	–5.7376*
FDI, Net Inflows (% of GDP)	–2.5605*	–2.8012*
Rate of Inflation	–1.4983***	–2.7839*
Current Account Balance (% of GDP)	–2.7288*	–4.0201*
Fiscal Balance of Central Government (% of GDP)	–3.6959*	–3.5747*
US Recession Dummy	–2.3344*	–0.9308
Real Exchange Rate	–1.6767**	–3.1746*

Source: The authors.

Note: *p < .01, **p < .05, ***p < .10

Table 4.

Estimates of Augmented Feldstein–Horioka Empirical Model

Coefficients	Fixed Effects Model	Random Effects Model
Gross Domestic Savings (% of GDP)	0.2753*	0.2901*
GDP Per Capita (% Annual)	0.0986	0.5995*
FDI, Net Inflows (% of GDP)	1.3244*	0.4576
Rate of Inflation	0.0464	–0.2691*
Current Account Balance (% of GDP)	–0.5684*	–0.9675*
US Recession Dummy	–0.8495	0.9330
Constant	19.5592*	18.7276*
No. of Panel Observations	125	125
F-Stat. (p-val.)	131.51 (0.000)*	59.507 (0.000)*
R-sq.	0.9337	0.7516
Adj. R-Sq.	0.9266	0.7389
Hausman Test (p-val.)	307.8262 (0.000)*

Source: The authors.

Dependent Variable: Gross Domestic Investments (% of GDP).

Note: *p < .01, **p < .05, ***p < .10

Table 5.

Estimates of Augmented Fidrmuc Empirical Model

Coefficients	Fixed Effects Model	Random Effects Model
Fiscal Balance of Central Govt. (% of GDP)	0.6077*	0.3518**
Gross Domestic Investment (% of GDP)	–0.5639*	–0.5384*
FDI, Net Inflows (% of GDP)	0.5285	0.3229
Rate of Inflation	–0.1858***	–0.2235**
Real Exchange Rate	–0.0143	–0.0155
US Recession Dummy	1.7937	1.7768
Constant	16.3087*	15.1688*
No. of Panel Observations	125	125
F-Stat. (p-val.)	28.5507 (0.000)*	31.0545 (0.000)*
R-sq.	0.7536	0.6122
Adj. R-Sq.	0.7272	0.5925
Hausman Test (p-val.)	64.2704 (0.000)*
Wald Test ( $ψ_{2} = ψ_{3} = 1$ )	Chi-Sq. stat. = 274.84 (0.000)*

Source: The authors.

Dependent Variable: Current Account Balance (% of GDP).

Note: *p < .01, **p < .05, ***p < .10

It is observed that the coefficient of FDI is 1.32 which is significantly positive at 0.01 levels. It means 1.32 per cent of domestic investment is financed through foreign savings when there is an increase in net FDI inflows by 1 per cent. It implies a high degree of association between domestic savings and foreign investment. This observation supports all the implications drawn from the low value of savings retention coefficient for SAARC economies. The coefficient of the CAB is significantly negative at 0.01 levels as expected thereby indicating adverse effects on domestic saving–investment relationship. Furthermore, the coefficients of other control variables such as growth rate of income, INF and URD have expected signs but not statistically significant. However, the robustness of the study has not been deteriorated because the fixed effects model has passed the necessary diagnostics.

For a cross-check of the validity of the F–H puzzle for SAARC economies, we further estimated the augmented twin-deficit equation of Fidrmuc (2003). It is observed from Table 3 that all the variables used in the panel specification (5) are integrated of order zero, that is, they are all level stationary. Therefore, we have to choose between fixed effects and random effects estimation for the panel specification (5). The results are presented in Table 5. The Hausman test outcomes indicate that the fixed effects model is the better choice under the given conditions. The values of R-sq. and adj. R-sq. also indicate better explanatory power of the fixed effects model. In the fixed effects model, the coefficient of fiscal balance is positively significant at 0.01 levels to validate the ‘twin deficit’ hypothesis because high fiscal deficit can increase the volume of the current account deficit. It means, in SAARC economies increase in fiscal deficit is likely to induce the domestic absorption thereby causing expansion of imports and current account deficits. The coefficient of the domestic investment is negatively significant at 0.01 levels, and its absolute value is significantly lower than one. This observation supports the validity of the F–H puzzle for SAARC economies while controlling for net FDI inflows, INF, RExR and the effects of the global financial recession of 2007–2009. It is observed from Table 3 that the investments in SAARC economies are primarily financed by international capital inflows. Now, assuming that the fiscal deficits are also financed by foreign capital (based on a fair degree of financial integration as implied from Table 4), the coefficients of FsB and GCF must be equal to unity. Therefore, it is necessary to test the null hypothesis that H₀: $ψ_{2} = ψ_{3} = 1$ , to check the validity of the F–H puzzle. We have tested this hypothesis through Wald test of coefficient restrictions and the result is noted at the bottom of Table 4. It is evidenced that the test rejects the null hypothesis at 0.01 levels of significance thereby supporting the validity of F–H puzzle for SAARC economies. Furthermore, the coefficient of the INF in the fixed effect model is negative and significant at 0.10 levels thereby indicating a deteriorating effect on the CAB in SAARC economies. The other control variables such as FDI inflows, RExR and URD are not statistically significant.

So, it is inferred from the results that the saving retention coefficient for South Asian economies is small thereby implying weaker capability of these countries to generate sufficient amount of domestic savings to finance the domestic investment projects which plausibly force them to rely on foreign capital. Hence, the results indicate a higher degree of international capital mobility. Furthermore, the low degree of association between domestic saving and investment in SAARC economies might be due to a smaller growth rate of income, and adverse effects of domestic inflation. In this connection, the negative effects of twin-deficits cannot be entirely ignored.

VI. Concluding Remarks

This article revisited the F–H puzzle for the selected SAARC economies over the period from 2002 to 2019. The empirical observations based on panel model estimations lend to support the validity of the twin-deficits hypothesis and also the F–H puzzle in the region. The validity of the twin-deficit hypothesis for the SAARC economies supports the Keynesian argument that an increase in fiscal deficit can deteriorate the CAB. So, the policy focus should be on fiscal and trade sector reforms to control twin-deficits in the region. Furthermore, we found a very low degree of association between domestic savings and investment thereby indicating the tendency for international capital outflows from the SAARC countries. It lends to conclude a fair degree of financial integration in these countries in the post-liberalisation era. Such an observation of low saving retention coefficient for the SAARC countries corroborates to the earlier similar findings for developing economies (Adeniyi & Egwaikhide, 2013; Ang, 2007; Bangaké & Eggoh, 2010; Coakley et al., 1999; De Wet & Van Eyden, 2005; Mumtaz & Munir, 2016; Ozmen, 2007; Payne & Kumazawa, 2005; Shahbaz et al., 2010; Younas, 2007). This observation has implications for interest rates, exchange rates, INF, CAB and macroeconomic growth. First, the real interest rates in SAARC countries are relatively less than that of Asian and World averages with apparent exceptions for Afghanistan and Bhutan (see Figure 2). So, the international financial market players can bid up the value of the currency of Afghanistan and Bhutan which in turn can cause trade deficits in these economies. Second, if international capital mobility is high, pursuing independent monetary policies by the SAARC nations becomes difficult as each country’s interest rate cannot be determined domestically. Third, since interest rate differential is one of the key factors influencing exchange rates, the relative effectiveness of fiscal and monetary policies becomes difficult. Fourth, since low saving retention coefficient implies increased international mobility of financial capital and the cross-border capital transfer requires transactions in international financial markets and goods markets as well, terms trade, exchange rates and commodity prices will be affected leading to risk of inflation. Fifth, the low saving retention coefficient implies significant changes in the behaviour of economic agents when they do not find suitable and productive investment avenue domestically. Thus, a tendency of outflow of domestic savings for a higher return on investments in foreign lands crops up, and SAARC countries have to rely on foreign investment inflows to sustain their long-run growth. Last, the tendency of higher capital mobility can exert negative effects of global financial shocks on the domestic investments as is evidenced from the negative coefficient of URD although not significant in this study. In the SAARC region, the low degrees of saving–investment relationship can be attributed to smaller growth rate of income, adverse effects of domestic inflation, lack of appropriate investment opportunities having higher return prospects, lack of matured capital markets and the presence of large scale rural economies having dominance of unorganised or informal sectors. Therefore, it is essential to enhance domestic savings and foreign direct investments as well to sustain long-term growth in South Asian economies. Thus, the policy circle needs to focus on designing fiscal and/or monetary policies that ensure economic and political stability in these countries which on the one hand, deepen financial markets, create productive investment avenues, attracts foreign capital, and on the other hand stabilises commodity prices, exchange rates and interest rates.

VII. Limitations and Scope for Further Research

This study has certain limitations—first, due to non-availability of data, the Maldives is excluded from the analysis; second, due to non-synchronous macroeconomic data structure, time periods prior to 2002 were excluded; third, the assumption of linear relationship is an obvious methodological limitation in the analysis; fourth, the study is only an indicative of the international capital mobility, but it has not been explicitly modelled; fifth, the factors such as country-specific structural and institutional dynamics, financial sector deepening and prevalence of domestic uncertainties and instabilities which affect saving–investment relationship are not included in the study; and last, the possibility that the presence of frictions in both goods and financial markets can inhibit international capital mobility and explain F–H hypothesis has not been taken care of. Therefore, scope is there to carry forward this piece of research work in a non-linear framework by explicitly modelling the international capital mobility by considering market frictions and other factors which drive saving–investment relationship.

Appendix A.

Theoretical Descriptions/Construct of Model (2)

Variables	Notation	Short Description	Data Source	Expected Sign in Regression
Gross Domestic Investment	GDI	It is the Gross Fixed Capita Formation taken as the percentage of GDP.	WDI, World Bank	Dependent variable
Gross Domestic Savings	GDS	It is the Gross Domestic Savings taken as the percentage of GDP. It is expected to have a positive impact on gross domestic investment in the SAARC economies.	WDI, World Bank	+
Growth Rate of Income	GRY	It is the annual growth rate of GDP per capita. It is assumed that the higher the annual growth rate of GDP per capita, stronger is the economic growth and Higher is the domestic investment because sustained productivity changes raise the volume of domestic savings and investment. It is expected to have a positive impact on gross domestic investment in the SAARC economies.	WDI, World Bank	+
Foreign Direct Investment	FDI	It is net inflows of foreign direct investment taken as the percentage of GDP. It is assumed to increase the availability of domestic savings, through its multiplier effects of employment creation and income generation, for economic growth by improving the capital efficiency. It is expected to have a positive impact on gross domestic investment in the SAARC economies.	WDI, World Bank	+
Rate of Inflation	INF	It is the rate of inflation as measured by GDP deflator. It is presumed that rising inflation can reduce the incentives for individuals to save more which in turn induces individuals to invest their money on the high return yielding projects. Thus, it is expected to have a positive impact on gross domestic investment in the SAARC economies.	WDI, World Bank	+
Current Account Balance	CAB	It is the current account balance taken as The percentage of GDP. Since imbalances in the BOPs reflect the gaps between domestic savings and domestic investment, higher absolute value of current account balance is expected to lead to the smaller F–H coefficient.	Asian Development Outlook, Asian Development Bank	–
US Recession Dummy	URD	It is included to observe the effect of global financial shocks on the saving–investment relationship. It is the proxy variable for capturing likely contagion effects of US financial recession (2007–2009) on SAARC economies. The lower the value of this indicator, better is the impact on the gross domestic investment in the SAARC economies.	Authors’ Definition (URD = 1, for Years 2007–2009 & URD = 0, for other years)	–

Source: The authors.

Appendix B.

Theoretical Descriptions/Construct of Model (5)

Variables	Notation	Short Description	Data Source	Expected Sign in Regression
Current Account Balance	CAB	It is the current account balance taken as the percentage of GDP.	Asian Development Outlook, Asian Development Bank	Dependent variable
Fiscal Balance	FsB	It is the fiscal balance of the Central government taken as the percentage of GDP. Increase in fiscal deficit raises the current account deficit due to imbalances in S–I relationship. So, FsB is expected to have a positive sign.	Asian Development Bank	+
Gross Domestic Investment	GDI	It is the Gross Fixed Capital Formation taken as the percentage of GDP. According to national income identity, a rise in investment, Ceteris paribus, deteriorates the current account balance. So, GDI is expected to have a negative sign.	WDI, World Bank	–
Foreign Direct Investment	FDI	It is net inflows of foreign direct investment taken as the percentage of GDP. If national investment increases due to increased inflows of FDI, savings remaining same, current account deficit increases. So, FDI is expected to have a positive sign.	WDI, World Bank	+
Rate of Inflation	INF	It is the rate of inflation as measured by GDP deflator. Increase in inflation due to increased money supply would reduce the rate of interest thereby causing an increase in the domestic investment which in turn would raise imports and reduce trade surplus. So, rate of inflation and current account balance are inversely related.	WDI, World Bank	–
Real Exchange Rate	RExR	It is the real exchange rate of a domestic Currency against the USD and calculated by multiplying nominal exchange rate to the ratio of CPI of USA and CPI of a national economy. An increase in exchange rate means depreciation of domestic currency which would enhance exports and reduce imports causing fall in trade deficits. So, real exchange rate and current account deficit are inversely related.	Authors’ Calculation from IMF data	+
US Recession Dummy	URD	It is included to observe the effect of global financial shocks on the current account balance. It is the proxy variable for capturing the contagion effects of US financial recession (2007–2009) on SAARC economies. During recession, exchange rate depreciates causing increase in exports and decrease in imports thereby reducing current account deficits.	Authors’ Definition (URD = 1, for Years 2007–2009 & URD = 0, for other years)	+

Source: The authors.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Note

ORCID iD

P. K. Mishra

References

Adeniyi

Egwaikhide

F. O.

(2013). Saving-investment nexus in developing countries: Does financial development matter? Journal of Economic Development , 38(2), 119–140.

Akkoyunlu

Ş.

(2020). Revisiting the Feldstein–Horioka puzzle for Turkey. Journal of Applied Economics , 23(1), 129–148.

Ang

J. B.

(2007). Are saving and investment cointegrated? The case of Malaysia (1965–2003). Applied Economics , 39(17), 2167–2174.

Apergis

Tsoumas

(2009). A survey of the Feldstein–Horioka puzzle: What has been done and where we stand. Research in Economics , 63(2), 64–76.

Aristovnik

(2005). Twin deficits hypothesis and Horioka–Feldstein puzzle in transition economies. EconPapers , 1–20. http://econpapers.repec.org/RePEc:wpa:wuwpif:0510020

Aristovnik

Djuric

(2010). Twin deficits and the Feldstein–Horioka puzzle: A comparison of the EU member states and candidate countries (MPRA Paper No. 24149). University Library of Munich.

Attanasio

O. P.

Picci

Scorcu

A. E.

(2000). Saving, growth, and investment: A macroeconomic analysis using a panel of countries. Review of Economics and Statistics , 82(2), 182–211.

Bagheri

Hazrati

F. D.

(2012). Twin deficits and Feldstein–Horioka puzzle in the case Iran. Journal of Social and Development Sciences , 3(5), 167–171.

Baharumshah

A. Z.

Ismail

Lau

(2009). Twin deficits hypothesis and capital mobility: The ASEAN-5 perspective. Jurnal Pengurusan , 29, 15–32.

10.

Baltagi

B. H.

Pesaran

M. H.

(2007). Heterogeneity and cross section dependence in panel data models: Theory and applications introduction. Journal of Applied Econometrics , 22(2), 229–232.

11.

Banday

U. J.

Aneja

(2019). Twin deficit hypothesis and reverse causality: A case study of China. Palgrave Communications , 5(1), 1–10.

12.

Bangaké

Eggoh

(2010). International capital mobility in African countries: Do the legal origins matter. Economics Bulletin , 30(1), 73–83.

13.

Baxter

Crucini

(1993). Explaining saving-investment correlations. American Economic Review , 83(3), 416–436.

14.

Caprio

Howard

D. H.

(1984). Domestic saving, current accounts, and international capital mobility (International Finance Discussion Paper No. 244). Board of Governors of the Federal Reserve System.

15.

Chan

Dang

Lai

Yan

(2011). Regional capital mobility in China: 1978–2006. Journal of International Money and Finance , 30(7), 1506–1515.

16.

Chang

Smith

(2014). Feldstein–Horioka puzzles. European Economic Review , 72, 98–112.

17.

Chu

K. H.

(2012). The Feldstein–Horioka puzzle and spurious ratio correlation. Journal of International Money and Finance , 31(2), 292–309.

18.

Coakley

Kulasi

Smith

(1996). Current account solvency and the Feldstein–Horioka puzzle. The Economic Journal , 106 (436), 620–627.

19.

Coakley

Hasan

Smith

(1999). Saving, investment, and capital mobility in LDCs. Review of International Economics , 7(4), 632–640.

20.

Corbin

(2001). Country specific effect in the Feldstein–Horioka paradox: A panel data analysis. Economics Letters , 72(3), 297–302.

21.

Dash

S. K.

(2019). Has the Feldstein–Horioka puzzle waned? Evidence from time series and dynamic panel data analysis. Economic Modelling , 83, 256–269.

22.

De Wet

A. H.

Van Eyden

(2005). Capital mobility in sub-Saharan Africa: A panel data approach. South African Journal of Economics , 73(1), 22–35.

23.

Dooley

Frankel

Mathieson

D. J.

(1987). International capital mobility: What do saving-investment correlation tell us? IMF Staff Papers , 34(3), 503–530.

24.

Dzhumashev

Cooray

(2016). The Feldstein–Horioka hypothesis revisited. B.E. Journal of Macroeconomics , 77(1), 1–30.

25.

Engle

Kletzer

(1987). Savings and investment in an open economy with no-traded goods [NBER working paper no. 2141]. National Bureau of Economic Research.

26.

Eyuboglu

Uzar

(2020). Is the Feldstein–Horioka puzzle valid in lucky seven countries? The Journal of International Trade & Economic Development , 29(4), 399–419.

27.

Feldstein

(1983). Domestic saving and international capital movements in the long-run and the short-run. European Economic Review , 21(1–2), 129–151.

28.

Feldstein

M. S.

Horioka

C. Y.

(1980). Domestic saving and international capital flows. Economics Journal , 90(358), 314–329.

29.

Fidrmuc

(2003). The Feldstein–Horioka puzzle and twin deficits in selected countries. Economics of Planning , 36(2), 135–152.

30.

Fouquau

Hurlin

Rabaud

(2008). The Feldstein–Horioka puzzle: A panel smooth transition regression approach. Economic Modelling , 25, 284–299.

31.

Frankel

J. A.

MacArthur

A. T.

(1988). Political vs. currency premia in international real interest rate differentials. European Economic Review , 32(5), 1083–1121.

32.

Georgopoulos

G. J.

Hejazi

(2005). Feldstein–Horioka meets a time trend. Economics Letters , 86(3), 353–357.

33.

Ghosh

A. R.

(1990). International capital mobility and optimal current account behavior: An empirical investigation [Discussion paper no. 50]. Department of Economics, Princeton University.

34.

Gutiérrez

Solimano

(2007). Savings , investment and growth in the global age: Analytical and policy issues [The AUP visiting scholar working paper series no. 43]. 1–25.

35.

Hassan

I. B.

Azali

Lee

(2014). Feldstein–Horioka puzzle and international capital mobility in high income countries: A pool mean group approach. Engineering Economics , 25(5), 480–486.

36.

T. W.

(2003). The saving-retention coefficient and country-size: The Feldstein Horioka puzzle reconsidered. Journal of Macroeconomics , 25, 387–396.

37.

Horioka

C. Y.

Terada-Hagiwara

Nomoto

(2015). Explaining foreign holdings of Asia’s debt securities: The Feldstein–Horioka paradox revisited (Discussion Paper No. 950). Japan: The Institute of Social and Economic Research, Osaka University.

38.

K. S.

Pesaran

M. H.

Shin

Y. C.

(2003). Testing for units roots in heterogeneous panels. Journal of Econometrics , 115(1), 53–74.

39.

Irandoust

(2019). Saving and investment causality: Implications for financial integration in transition countries of Eastern Europe. International Economics and Economic Policy , 16, 397–416.

40.

Isaksson

(2001). Financial liberalization, foreign aid, and capital mobility: Evidence from 90 developing countries. Journal of International Financial Markets, Institutions and Money , 11, 309–338.

41.

Jansen

W. J.

(1996). Estimating saving-investment correlations: evidence for OECD countries based on an error correction model. Journal of International Money and Finance , 15(5), 749–781.

42.

Jiranyakul

Brahmasrene

(2009). An exploratory inquiry of the Feldstein–Horioka puzzle in selected Southeast Asian countries. Journal of Transnational Management , 14(4), 259–276.

43.

Kasuga

(2004). Saving-investment correlation in developing countries. Economics Letters , 83(3), 371–376.

44.

Ketenci

(2016). The Feldstein–Horioka puzzle and structural breaks: Evidence from the largest countries of Asia. Margin The Journal of Applied Economic Research , 10(3), 337–354.

45.

Khan

M. A.

Saeed

(2012). Twin deficits and saving-investment nexus in Pakistan: Evidence from Feldstein–Horioka puzzle. Journal of Economic Cooperation and Development , 33(3), 1–36.

46.

Khundrakpam

J. K.

Ranjan

(2010). Saving-investment nexus and international capital mobility in India: Revisiting Feldstein–Horioka hypothesis. Indian Economic Review , 45(1), 49–66.

47.

Kim

Roubini

(2008). Twin deficit or twin divergence? Fiscal policy, current account, and real exchange rate in the US. Journal of international Economics , 74(2), 362–383.

48.

Lam

N. X.

(2012). Twin deficits Hypothesis and Feldstein–Horioka puzzle in Vietnam. International Research Journal of Finance and Economics , 101(9), 169–179.

49.

Levin

Lin

C. F.

Chu

C. S. J.

(2002). Unit root tests in panel data: Asymptotic and finite-sample properties. Journal of Econometrics , 108(1), 1–24.

50.

Lucas

R. E.

(1990). Why doesn’t capital flow from rich to poor countries? The American Economic Review , 80(2), 92–96.

51.

Mamingi

(1997). Savings-investment correlations and capital mobility: The experience of developing countries. Journal of Policy Modeling , 19(6), 605–626.

52.

Marinheiro

C. F.

(2008). Ricardian equivalence, twin deficits, and the Feldstein–Horioka puzzle in Egypt. Journal of Policy Modeling , 30(6), 1041–1056.

53.

Meier

G. M.

Baldwin

(1957). Economic Development: Theory, history, policy . John Wiley & Sons, New York.

54.

Miller

M. S.

(1988). International capital mobility: What do saving-investment correlations tell us? Comment on Dooley, Frankel, and Mathieson. IMF Staff Papers , 35(2), 391–396.

55.

Mishra

P. K.

(2011). Savings and investment nexus in the perspective of foreign capital: The Feldstein–Horioka paradox. International Journal of Finance and Policy Analysis , 3(2), 40–49.

56.

Mishra

P. K.

Mishra

S. K.

(2015). Integration of capital markets in SAARC region: An empirical study. Uttaranchal Business Review , 5(1), 81–90.

57.

Mohanty

R. K.

(2019). An empirical investigation of twin deficits hypothesis: Evidence from India. Journal of Quantitative Economics , 17, 579–601.

58.

Montiel

P. J.

(1994). Capital mobility in developing countries: Some measurement issues and empirical Estimates. World Bank Economic Review , 8(3), 311–353.

59.

Moscone

Tosetti

(2009). A review and comparison of tests of cross‐section independence in panels. Journal of Economic Surveys , 23, 528–561.

60.

Mumtaz

Munir

(2016). Dynamics of twin deficits in South Asian countries [MPRA paper no. 74592]. University Library of Munich.

61.

Murphy

R. G.

(1984). Capital mobility and relationship between saving and interest rates in OECD Countries. Journal of International Money and Finance , 3(3), 327–342.

62.

Obstfeld

Rogoff

(2000). The six major puzzles in international macroeconomics: Is there a common cause? [NBER working paper no. 7777]. National Bureau of Economic Research, Inc.

63.

Obstfeld

(1985). Capital mobility in the world economy: Theory and measurement [NBER working paper no. 1692]. National Bureau of Economic Research, Inc.

64.

Obstfeld

(1986a). Capital mobility in the world economy: Theory and measurement. Carnegie-Rochester Conference Series on Public Policy , 24, 55–104.

65.

Obstfeld

(1986b). How integrated are world capital markets: Some new tests [NBER working paper no. 2075]. National Bureau of Economic Research, Inc.

66.

Otto

Wirjanto

(1989). National saving and domestic Investment in the long run: Some time series evidence for the U.S. and Canada [Working paper no. 754]. Institute for Economic Research, Queen’s University.

67.

Ozmen

(2007). Financial development, external rate regimes and the Feldstein–Horioka puzzle: evidence from MENA region. Journal of Applied Economics , 39(9), 1133–1138.

68.

Patra

S. K.

Mohanty

R. K.

(2020). Does the Feldstein–Horioka puzzle exist among South Asian countries? A regime-switching approach. Journal of Public Affairs , 20(3), 1–14.

69.

Payne

J. E.

Kumazawa

(2005). Capital mobility, foreign aid, and openness: Further panel data evidence from sub-Saharan Africa. Journal of Economics and Finance , 29(1), 122–126.

70.

Pelgrin

Schich

(2008). International capital mobility: What do national saving investment dynamics tell us? Journal of International Money and Finance , 27, 331–344.

71.

Pesaran

M. H.

(2004). General diagnostic tests for cross section dependence in panels [Cambridge working papers in economics no. 0435]. University of Cambridge.

72.

Petreska

Mojsoska-Blazevski

(2013). The Feldstein–Horioka puzzle and transition economies. Economic Annals , 58(197), 23–45.

73.

Phiri

(2017). The Feldstein–Horioka puzzle and the global recession period: Evidence from South Africa using asymmetric cointegration analysis [MPRA paper no. 79096]. University Library of Munich.

74.

Raheem

(2017). More finance or better finance in Feldstein–Horioka puzzle: Evidence from SSA countries. Global Business Review , 18(1), 1–12.

75.

Rahman

(2015). Savings-investment Nexus: Experience of major SAARC economies based on the basic version of the Feldstein–Horioka specification. Journal of Economic and Sustainable Development , 6(16), 154–173.

76.

Razak

L. A.

Masih

(2017). Revisit Feldstein–Horioka puzzle: Evidence from Malaysia (1960–2015) [MPRA paper no. 79407]. University Library of Munich.

77.

Riaz

(2011). Financial integration in South Asia: An application of Feldstein–Horioka puzzle [Independent thesis advanced level master degree in economics]. School of Social Sciences, Sodertorn University.

78.

Saeed

Khan

M. A.

(2012). The Feldstein–Horioka puzzle and twin deficits in Pakistan. Academic Research International , 2(2), 525–532.

79.

Shahbaz

Ahmad

Wahid

A. N.

(2010). Savings–Investment correlation and capital outflow: The case of Pakistan. Transition Studies Review , 17(1), 80–97.

80.

Sinha

(2004). The mother of all puzzles would not go away. Economics Letters , 82(2), 259–267.

81.

Summers

L. H.

(1985). Tax policy and international competitiveness. In Frenkel

J. A.

(Ed.), International aspects of fiscal policy . Chicago Press.

82.

Suruga

Rahman

(2016). Feldstein–Horioka Paradox: The case of South Asian association for regional cooperation. European Journal of Business and Management , 8(2), 22–30.

83.

Tasar

(2017). Empirical analysis of savings and investments relation in Turkey: Cointegration test with structural breaks approach. Journal of Economics Finance and Accounting , 4(2), 106–111.

84.

Taylor

(1994). Domestic saving and international capital flows reconsidered [NBER working paper series 4892]. National Bureau of Economic Research, Inc.

85.

Tesar

(1991). Saving, investment, and international capital flows. Journal of International Economics , 31(1–2), 55–78.

86.

Tobin

(1983). Domestic saving and international capital movements in the long-run and the short-run by M. Feldstein: Comment. European Economic Review , 21(1/2), 153–156.

87.

Vamvakidis

Wacziarg

(1998). Developing countries and the Feldstein–Horioka Puzzle [IMF working paper no. WP/98/2]. IMF Publications, 1–24.

88.

Wahid

A. N. M.

Salahuddin

Noman

A. M.

(2010). Savings and investment in South Asia. Journal of Economic Studies , 37(6), 658–666.

89.

Westphal

(1983). Domestic saving and international capital movements in the long-run and the short-run by M. Feldstein: Comment. European Economic Review , 21(1/2), 157–159.

90.

Wirasti

Widodo

(2017). Twin deficit hypothesis and Feldstein–Horioka hypothesis: Case study of Indonesia [MPRA paper no. 77442]. University Library of Munich. https://mpra.ub.uni-muenchen.de/77442/

91.

Wong

D. Y.

(1990). What do savings-investment relationship tell us about capital mobility? Journal of International Money and Finance , 9(1), 60–74.

92.

Yasutomi

Horioka

C. Y.

(2010). Adam Smith’s answer to the Feldstein–Horioka paradox: The invisible hand revisited [discussion paper no. 778]. The Institute of Social and Economic Research, Osaka University.

93.

Yildirim

Orman

E. E.

(2018). The Feldstein–Horioka puzzle in the presence of structural breaks: Evidence from China. Journal of the Asia Pacific Economy , 23(3), 374–392. https://doi.org/10.1080/13547860.2017.1396640

94.

Younas

(2007). Capital mobility, foreign aid, and openness: A reappraisal. Economics Bulletin , 6(12), 1–8.

95.

Younas

Chakraborty

(2009). Globalization and the Feldstein–Horioka puzzle. Applied Economics , 83(16), 2089–2096.