Behavioral and Empirical Perspectives on FDI: International Capital Allocation Across Asia Behavioral and Empirical Perspectives on FDI: International.

  • Published on
    15-Jan-2016

  • View
    214

  • Download
    0

Transcript

Growth, Trade, and Investment Horizons for Asia

Behavioral and Empirical Perspectives on FDI: International Capital Allocation Across Asia Douglas H. Brooks, Asian Development Bank InstituteDavid Roland-Holst, UC BerkeleyFan Zhai, Asian Development Bank

Presented at the ADB /ITD ConferenceShaping the Future: Prospects for Asias Long-term Development over the Next Two Decades11-12 December 2006Bangkok, ThailandRoland-Holst #ContentsIntroduction and OverviewMethodological IssuesMacro Determinants of FDIFDI and Market ExpansionFDI and ProductivityConclusions12 December 2006Roland-Holst #1. Introduction and OverviewInternational capital allocation has been a primary driver of modern growth dynamics, particularly for emerging economies, and this relationship has nowhere been more fortuitous than in Asia. Together with disciplined commitments to domestic and external economic reform, the regions economies have leveraged foreign savings to achieve growth and modernization beyond the imagining of prior generations. Despite the pervasive influence FDI has had on Asias growth experience, the precise benefits of foreign investment remain challenging to quantify and the process of international capital allocation very difficult to predict. 12 December 2006Roland-Holst #International capital allocation has been a primary driver of modern growth dynamics, particularly for emerging economies, and this relationship has nowhere been more fortuitous than in Asia. Together with disciplined commitments to domestic and external economic reform, the regions economies have leveraged foreign savings to achieve growth and modernization beyond the imagining of prior generations. Despite the pervasive influence FDI has had on Asias growth experience, the precise benefits of foreign investment remain challenging to quantify and the process of international capital allocation very difficult to predict. Given the nearly universal appeal of FDI as a growth catalyst, however, it would clearly be desirable for policy makers to better understand its fundamental determinants. As Asia transits from a loose federation of emerging economies to a more fully integrated and mature economic region, the need to understand multilateral investment dynamics will only increase.

3Given the nearly universal appeal of FDI as a growth catalyst, however, it would clearly be desirable for policy makers to better understand its fundamental determinants. As Asia transits from a loose federation of emerging economies to a more fully integrated and mature economic region, the need to understand multilateral investment dynamics will only increase.

12 December 2006Roland-Holst #Trends in Global Agregates12 December 2006

Sources: Exports and GDP - IMF WEO September 2006 database; FDI Outflows - UNCTAD FDI September 2006 database.Roland-Holst #Asian Public and Private Inflows(Asian inbound Aid and FDI, USD Billions)12 December 2006

Clearly, we have entered an Age of Complementarity.Roland-Holst #6From this perspective, aid more closely resembles activities on the capital account, including one dramatically emergent phenomenon, foreign direct investment (FDI). While FDI is a private sector activity, and thus is animated by very different primary objectives, it has been known to confer many benefits on developing economies that are consistent with aid objectives, including human resource development, technology diffusion, and, ultimately, poverty alleviation and more sustainable growth. In this sense, it has long been recognized that there may be essential complementarities between private and public foreign investment in developing countries, where the latter means aid. The extent to which the complementarities matter is not intellectual, but dependent upon their real and potential economic significance. To get an idea of this, look at the trends above. These depict, for Asian Developing countries, levels of inbound Aid and FDI for the period 1973-2003. The most arresting feature of this data is of course the meteoric rise in FDI, which has increased about one hundredfold over the last three decades. This trend must inspire reflection on the appropriate strategy to public overseas investment, or development assistance, going forward.

FDI Inflows in Selected Developing Asian Economies, 2001-0512 December 2006

Roland-Holst #Annual FDI Inflows (US$ million)12 December 2006

Roland-Holst #FDI Inflows (as % of Gross Fixed Capital Formation)12 December 2006

Roland-Holst #FDI Inflows Per Capita (US$)12 December 2006

Roland-Holst #2. Methodological IssuesFinancial flows generally, and foreign capital flows in particular, have been one of the most challenging areas of empirical trade research. These flows offer important growth leverage to regional economies, particularly developing ones.Despite general agreement about what kind of phenomena deserve primary study, researchers have failed to develop empirically robust specifications.12 December 2006Roland-Holst #MethodologyIn this study, we approach the problem from three perspectives. Macroeconomic determinants FDI is a component of macro adjustmentMarket Expansion FDI is driven by supply chainsProductivity FDI is a portfolio decision

12 December 2006Roland-Holst #3. Aggregate Determinants of Inbound FDIA parsimonious representation of the basic macro-driver model would be

whereZ denotes a monotone index of the level of inbound FDIP is an index for capital cost or a forward price of savingsR is an index of local relative to global real interest ratesG is an index of local real GDP growthThis can be conveniently represented in elasticity form as

12 December 2006

Roland-Holst #Econometric EstimatesUsing annual data for twelve Asian countries, we experimented with a variety of proxies for P and R. We were unsuccessful in identifying variables to represent P, but for R the most useful proxy was the ratio of average domestic interbank rates to LIBOR. As in much of the literature on this subject, our estimates did not indicate conclusive macroeconomic interactions between FDI, growth, or rates of return.Instead, we find that flows are most dependent on initial conditions (national fixed effect coefficients), with high degrees of statistical significance and a very high R-square for pooled data.

12 December 2006Roland-Holst #Generally, our results indicate that variations in FDI are most dependent on initial conditions (national fixed effect coefficients), with high degrees of statistical significance and a very high R-square for pooled data. From a multi-country modeling perspective, the fixed effects are simply national calibration parameters and the GDP growth effect can be incorporated into dynamic transition equations. In these results that PRC was used as the omitted condition or intercept, and the remaining fixed effects are negative because of Chinas status as the largest recipient of inbound FDI.There is little empirical evidence here to support model calibration. Like the price level variable, our real interest rate proxy contributes insignificantly to either FDI levels (FDI) or rates of change (logFDI). Growth of GDP is significant in predicting the level of FDI, but not its rate of change. There is an obvious magnitude bias from PRCs massive FDI flows. This, coupled with PRCs above average growth rate, explains the role of log(GrGDP) in explaining sample variation for FDI levels. When FDI is re-scaled in logarithmic terms (Figure 63), this scale bias is substantially reduced, and fixed effects are sufficient to pick up the outlier economy.In the case of rate of return proxies, the results are even more ambiguous. When compared to FDI levels, logR is insignificant for most of the sample, although Figure 4 indicates it might be significant in a sub-sample including PRC. When examining elasticity effects, the results are completely indeterminate. When explaining log(FDI), the implied elasticity with respect to R/RW is actually negative, contradicting basic Fisherian investment theory, and nearly significant. More work with these data, particularly devising new proxies and experimenting with different lag structures, may better elucidate the macro drivers of FDI. Meanwhile, however, we accept both the intuition behind these three determinants and the need to better understand the mechanisms through which FDI works to affect growth and income distribution. To this end, we carried out a series of experiments with range values of elasticities for two of the explanatory variables, logR and LogG. We discuss the details of the simulation design in section 5.

14Log(FDI) and Log(GDP growth)12 December 2006

Roland-Holst #Generally, our results indicate that variations in FDI are most dependent on initial conditions (national fixed effect coefficients), with high degrees of statistical significance and a very high R-square for pooled data. From a multi-country modelling perspective, the fixed effects are simply national calibration parameters and the GDP growth effect can be incorporated into dynamic transition equations. In these results that PRC was used as the omitted condition or intercept, and the remaining fixed effects are negative because of Chinas status as the largest recipient of inbound FDI.

15FDI and Log(R/RW)12 December 2006

Roland-Holst #Like the price level variable, our real interest rate proxy contributes insignificantly to either FDI levels (FDI) or rates of change (logFDI). Growth of GDP is significant in predicting the level of FDI, but not its rate of change. There is an obvious magnitude bias from PRCs massive FDI flows. This, coupled with PRCs above average growth rate, explains the role of log(GrGDP) in explaining sample variation for FDI levels. When FDI is re-scaled in logarithmic terms (Figure 63), this scale bias is substantially reduced, and fixed effects are sufficient to pick up the outlier economy.

16Log(FDI) and Log(R/RW)12 December 2006

Roland-Holst #17FDI Flow Regression 112 December 2006

Roland-Holst #In the case of rate of return proxies, the results are even more ambiguous. When compared to FDI levels, logR is insignificant for most of the sample, although Figure 4 indicates it might be significant in a sub-sample including PRC. When examining elasticity effects, the results are completely indeterminate. When explaining log(FDI), the implied elasticity with respect to R/RW is actually negative, contradicting basic Fisherian investment theory, and nearly significant. 18FDI Flow Regression 212 December 2006

Roland-Holst #FDI Elasticity Regression 112 December 2006

Roland-Holst #There is little empirical evidence here to support model calibration. Like the price level variable, our real interest rate proxy contributes insignificantly to either FDI levels (FDI) or rates of change (logFDI). Growth of GDP is significant in predicting the level of FDI, but not its rate of change. There is an obvious magnitude bias from PRCs massive FDI flows. This, coupled with PRCs above average growth rate, explains the role of log(GrGDP) in explaining sample variation for FDI levels. When FDI is re-scaled in logarithmic terms (Figure 63), this scale bias is substantially reduced, and fixed effects are sufficient to pick up the outlier economy.20FDI Elasticity Regression 212 December 2006

Roland-Holst #More work with these data, particularly devising new proxies and experimenting with different lag structures, may better elucidate the macro drivers of FDI. Meanwhile, however, we accept both the intuition behind these three determinants and the need to better understand the mechanisms through which FDI works to affect growth and income distribution. To this end, we carried out a series of experiments with range values of elasticities for two of the explanatory variables, logR and LogG. We discuss the details of the simulation design in section 5.21FDI From a Simulation PerspectiveIn the absence of definitive econometric evidence regarding FDI behavior, a simulation framework may be able to elucidate the primary interactions between initial conditions and outcomes using a variety of alternative behavioral specifications. We used a global CGE model to examine how the ultimate effects of trade policy would vary under different hypothetical patterns of FDI behavior. Given the importance of private capital flows to the modern process of globalization, it is hardly surprising that trans-boundary investment behavior can strongly influence the effects of trade liberalization. Indeed, it is apparent even in this preliminary analysis that shifting FDI patterns can make the difference between success and failure for countries joining regional FTAs and larger trade reform initiatives.The model we use is a multi-country, dynamic CGE calibrated to the GTAP VI database. The present version includes an option for endogenous determination of FDI flows, based on the same logic as the estimating equation of the previous section. 12 December 2006Roland-Holst #Simulation ExperimentsWe conducted four experiments based on a scenario of global trade liberalization (GBL), assuming all tariffs and export subsidies are removed over the period 2005-2010. This scenario has the predictable results for global efficiency gains and growth, and then forms a policy reference for four FDI scenarios based on the following elaboration of the last equation

where for country r, Z denotes total investment, Pw/P denotes the relative price of future consumption, TR/WRR is a the ratio of domestic and global rental rates, and is the growth rate of real GDP. This specification explains domestic aggregate investment shares as a product of three components: a forward discount rate, an inter-country relative rate of return, and an accelerator mechanism (including a lagged investment term)12 December 2006

Roland-Holst #While we were unable to estimate the elasticities in this relationship with available data, it is still possible to analyse the economic implications of this specification of FDI behavior with counterfactual elasticity values.23Scenarios12 December 2006

Roland-Holst #24Equivalent Variation Aggregate Income(percent change from Baseline in 2025)12 December 2006

Roland-Holst #Table 3.2 shows equivalent variation aggregate income (EV) for each Asian country under the globalization reference (GBL) and the four other scenarios, with results in each expressed as percentage changes from Baseline values in 2025. The most arresting feature of this table is the negative results (positive results are shaded), yet it is also important to notice that country results both exceed and fall short of the GBL scenario, depending on the country and scenario. Thus some kind of growth transfer process appears to arise from the capital movement, which is precisely what one might expe...

Recommended

View more >