The saving-investment puzzle revisited: Is a linear relationship unfounded?

The least developed countries (LDCs) such as Bangladesh are often hamstrung by a lack of resources - domestic and foreign - needed for investment. Achievement of a high rate of economic growth is the only way to improve the standard of living for the masses, aside the question of distributive justice. Here the role of investment is central to the aim. Domestic resource mobilisation is a possibility if the nation can save enough after meeting its consumption needs.
Given that the LDCs are characterised by low levels of income and thus low saving, this option may sound like a tall order. The alternative is to attract foreign funds to supplement the domestic gap. The LDCs face a formidable task of ensuring a smooth flow of such funds to bolster investment, required to promote and sustain economic growth. Adding to the prevailing complications, are the problems pertaining to ongoing macroeconomic and political uncertainties. These are hurting the ability of the nation to deliver an adequate amount of domestic savings.
Arguably, the low income, macroeconomic uncertainties and high population density in Bangladesh and many similar other developing economies, are among the principal obstacles on the way to mobilising domestic savings and channelling them into investment. Developing nations need foreign funds so they are likely to benefit from smooth capital flows, although opportunities may be limited.

The perception of an existence of relationship between saving investment in the context of the problem, especially for the LDCs, was first documented in the work of Feldstein and Horioka (FH, hereon), named as the FH Puzzle after the initials of the authors. In a seminal paper in 1980 the authors first shed light on the high correlation between domestic savings and investment. Using data from 16-countries of the Organization for Economic Cooperation and Development (OECD), they provided estimate of correlation between the series. Although viewed as an empirical artifact, the paper documented that saving constitutes some 85-95% of domestic investment. The difference between the two is met from the international capital market, determined by the returns. FH interpret this high correlation as an indicator of capital immobility.
This interpretation, while broadly supported by empirical evidence, poses an uncomfortable question which led to the birth of the FH Puzzle. Despite the unease among the academics, a bourgeoning literature ensued, aimed at knowing more about the puzzle. As part of the discourse, the question was raised about whether or not the estimated saving-investment correlations truly contain enough information to predict international capital mobility - an issue that sill continues to breathe fresh life into it.
The existing literature on the savings-investment nexus assumes that any long-run equilibrium between saving and investment does not undergo big change in its course - i.e., any interaction (relationship) between them, staying essentially linear in nature. While this may be true of large developed economies, it appears less tenable for developing countries where conditions such as policy reforms; changes in disposable income; and the ease of transaction due to advancement in the financial sector are not only routine, rather they are an ongoing continuous process. Such events do bring about much faster and possibly more substantial changes. In light of the above backdrop, it appears that a careful re-examination of the FH literature is warranted. The task of revisiting the nexus between the two thus appears timely, if not long overdue.
In a recent paper presented in the United States, aimed at revisiting the topic of the FH Puzzle in the context of four South Asian economies- Bangladesh, Pakistan, Sri Lanka and India - authors of this write-up -- Muhammad Shafiullah of the PRI and Faridul Islam of Morgan State University report that a long run equilibrium between savings and investment exists in Bangladesh. However, in a departure from the past tradition, the authors invoked a novel non-linear econometric (statistical) procedure to examine and analyse the relationship for the above-listed four South Asian economies. This approach did generate some interesting findings, however.
With respect to Bangladesh, the authors found that saving has a nonlinear causal effect on investment, an empirical fact the policymakers might keep in mind and consider, as needed.
A quick look at the graph below depicts a revealing picture of the trend of investment-saving scenario in Bangladesh. Since independence in 1971, the saving and investment rates as a percentage of the gross domestic product (GDP) of Bangladesh have shown on an upward trend. The country's saving and investment rates faced some tumultuous early days - including a dip into the negative territory for both series - but then virtually stabilized and even recorded an upward trajectory from the mid-1980s onwards. However, as can be seen, neither investment nor saving appears linear in nature. It might be oversimplification to assume a linear relationship between the two non-linear macroeconomic indicators, a priori. The emergence of new, albeit advanced econometric software, has to a large extent, have made empirical research much cheaper and easier.  
Based on the findings, it transpires that the postulated interactions between saving and investment are non-linear, at least in the context of the sample examined. Among others, this implies that, perhaps Bangladesh should consider more dynamic and nuanced approaches to policymaking in setting its long-term strategy in relation to saving and investment. This might help find ways that better suit the underlying realities of Bangladesh. An appropriate dynamic saving-investment policy and improvements in the financial sector appear necessary for Bangladesh to boost private investment, which has been stagnating in recent years. We think that this is needed if the cherished goal is to lift a larger fraction of the masses out of abject poverty that bedevil the policy makers! To get there, the nation would have to achieve and sustain an economic growth rate of 8.0% or higher for some time.
 [This article was jointly written by Arbab Muhammad Shah and
Dr. Muhammad Shafiullah and
Dr. Faridul Islam for The Financial Express-Policy Research Institute Economic Analysis Unit (FE-PRI EAU). Mr. Shah is a summer intern at PRI. He can be reached at [email protected]. Shafiullah is a Senior Economist at Policy Research Institute of Bangladesh (PRI) and can be reached at [email protected]. Dr. Islam is Chair, Department of Economics, Morgan State University, USA.
He can be reached at [email protected].]