Endogenous Financial and Trade Openness

Transcription

1 Review of Develoment Economics, 3(2), 75 89, 2009 DOI:0./j x Endogenous Financial and Trade Oenness Joshua Aizenman and Ilan Noy* Abstract The authors study the endogenous determination of financial and trade oenness. They construct a theoretical framework leading to two-way feedbacks between financial and trade oenness and identify these feedbacks emirically. They find that one standard deviation increase in commercial oenness is associated with a 9.5% increase in de facto financial oenness (% of GDP). Similarly, an increase in de facto financial oenness has owerful effects on future trade oenness. De jure restrictions on caital mobility have only a weak imact on de facto financial oenness, while de jure restrictions on the current account have a large adverse effect on commercial oenness. The authors investigate the relative magnitudes of these directions of causality using Geweke s (982) decomosition methodology. They conclude that in an era of raidly growing trade integration, countries cannot choose financial oenness indeendently of their degree of oenness to trade. Dealing with greater exosure to turbulence by imosing restrictions on financial flows is likely to be ineffectual.. Introduction and Overview Salient features of the international economy during the last 20 years are the growing financial and commercial integrations of develoing countries and the recurring financial instabilities and crises. These develoments have led to contentious debates regarding the desirability of financial oenness. Prominent economists have concluded that the gains from financial integration are illusive and caution develoing countries against rushing towards financial oenness (Rodrik, 999). Yet, other studies have rovided tentative suort for the resence of significant gains (Henry, 2003). While some observers conclude that greater exosure to financial turbulence should be dealt with by curbing financial flows, others caution against caital controls noting their ineffectiveness. These issues are of crucial imortance to, for examle, China and India, as both have increased their trade oenness while sustaining financial reression. This debate resumes that countries can choose their desirable level of financial oenness indeendently of their degree of oenness to trade. Past studies tyically focus on the formal acts associated with de jure financial oening, such as changing regulations. Yet, as has been noted by Prasad et al. (2003), de facto financial integration is by itself of considerable interest. The actual level of financial oenness is the outcome of the interaction between market forces and the enforcement of existing regulations. We utilize the concets of de facto trade and financial integration to investigate this resumtion, and we study the two-way feedbacks between de facto financial and trade oenness and the residual role of de jure oenness. After outlining a model, we show that de facto financial oenness deends ositively on lagged trade oenness. * Aizenman: University of California, Santa Cruz, CA, 95064, USA. jaizen@ucsc.edu. Noy: University of Hawaii, Manoa, HI 96822, USA. noy@hawaii.edu. We thank Michael Hutchison, Hiro Ito, Nancy Marion, and Shang-Jin Wei for data, two anonymous referees, and the many useful comments we received from seminar articiants at the Hebrew University, UCSC, UCLA, the University of Haifa, the University of Illinois (Urbana-Chamaign), the University of Hawaii, Tel Aviv University, the University of Washington, and LACEA Any errors are ours. This aer summarizes the results in Aizenman and Noy (2004)., 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 0248, USA

2 76 Joshua Aizenman and Ilan Noy In section 2 we outline a model in which financial oenness is determined endogenously. The model integrates ublic finance and olitical-economy considerations, focusing on a develoing country characterized by limited tax caacity and olitical instability. Fiscal outlays are financed by means of two taxes: a direct income tax and an imlicit tax induced by caital controls. Both taxes are costly: the income tax is associated with deadweight losses due to the collection and enforcement costs. The tax caacity is imacted by the resources devoted to enforcement and administration. Similarly, enforcing caital controls entails costly olicing to revent illicit caital flight. There are two tyes of olicymakers. The resonsible one faces the canonic ublic finance roblem: determining tax olicies in order to maximize the reresentative consumer s welfare, subject to the need to fund a fiscal revenue stream. The second is of the extractive tye, interested in obtaining resources for a narrow interest grou. In this set-u, financial oenness is endogenously determined by the authority s choice of financial reression. This olicy is tantamount to a tax on domestic saving, generating the incentive to engage in illicit caital flight in order to avoid the tax. Caital flight is intermediated via trade misinvoicing. Enforcing financial reression requires direct exenditure on monitoring and olicing trade invoices, thereby greater trade oenness increases enforcement costs. We show that financial reression characterizes countries that are below a certain threshold of fiscal efficiency high enough cost of tax collection would induce the imlementation of financial reression as a means of taxation. Countries tyified by low trade oenness and high saving are more likely to use financial reression. In these circumstances, higher tax collection costs, higher fiscal exenditure, and lower commercial oenness would increase the otimal financial reression. We also find that lower robability of staying in ower reduces the otimal investment in tax caacity, and increases the likelihood that financial reression would be art of the menu of taxes. If the olicymaker is maximizing the reresentative consumer s welfare, a higher robability of staying in ower increases the investment in future tax caacity, reducing the otimal financial reression. If the olicymaker is of the extractive tye, reflecting the interests of narrow ressure grous, the tax would be set at the eak of the corresonding tax Laffer curve. The attitude of this olicymaker towards caital controls is mixed caital controls would be imosed on the reresentative consumer, while the fiscal surlus would be ut in off-shore accounts, as insurance against losing ower. In these circumstances, lower robability of staying in ower increases caital flight and reduces financial reression. We examine emirically some of the hyotheses suggested by our model. We estimate the level of de facto financial oenness as a function of lagged trade oenness and other variables. We find that de facto financial oenness deends ositively on lagged trade oenness. These results suggest de facto sequencing, where greater de facto trade oenness is associated with larger future de facto financial oenness. The reverse association may also hold. Hence, we exect to find two-way ositive linkages between financial and commercial oenness and confirm these redictions emirically. We investigate the relative magnitudes of these directions of causality using the decomosition test develoed in Geweke (982). We find that almost all of the linear feedback between trade and financial oenness can be accounted for by G-causality from financial oenness to trade oenness (53%) and from trade to financial oenness (34%). The residual is due to simultaneous correlation between the two annual measures. The evidence in the aer is in line with a recent contribution by Antràs and Caballero (2007). They introduce financial frictions to the 2 2 standard international trade model. Secifically, they assume that entrereneurs in the relatively comlex

3 ENDOGENOUS OPENNESS 77 sector are endowed with essential skills, but due to informational frictions they can borrow only a fraction of their caital endowment. They find that trade and caital flows are comlements, and that financially underdeveloed economies that liberalize trade may exerience caital inflows. Our results are also consistent with the notion that a significant share of the volume of financial flows to and from develoing countries is due to diversification of olitical risk, as in Dooley (988). This interretation may rovide an additional rationale for our finding concerning the negative marginal association of democracy and financial oenness. This finding also suggests that the home bias in the allocation of financial assets identified by the financial literature (in the OECD countries) may be less ronounced in develoing countries because of olitical risk considerations. 2. The Model This section extends Aizenman and Guidotti (994), describing the links between trade and financial oenness in a develoing country characterized by limited tax caacity and olitical uncertainty. We consider a country where fiscal outlays are financed by a direct income tax and an imlicit tax induced by caital controls. Both taxes are costly: the income tax is associated with deadweight losses due to collection and enforcement costs. The tax caacity is imacted by the resources devoted to enforcement and administration. Similarly, enforcing caital controls entails costly olicing to revent illicit caital flight. We assume a small, two goods, two-eriod α β α β economy. The consumer s utility is V = u( X Y )+ u( X2 Y2 ) ( + ρ ), u >; u 2 0; a + b =, where X is the domestic good and Y is the foreign, imorted good. There are two tyes of olicymakers. The resonsible one, denoted by w, chooses ublic finance olicies in order to maximize the reresentative consumer s welfare, subject to the need to fund the given fiscal revenue stream. The second tye of olicymaker, denoted by n, is of the extractive tye, interested in obtaining resources for narrow interest grous. Consumers are endowed each eriod with X units of the domestic good. The authorities tax the income from the endowment X at a rate t.the consumer saves in eriod one D, allocating it between domestic and foreign bonds, D and D*, resectively. The international real interest rate is r*. The authorities imose caital control in the form of a tax on the foreign bond. Let f denote the tax rate, imlying that the domestic interest rate, r, is determined by + r = ( + r*)( - f). Consequently, the tax determines the remium between the foreign and domestic real interest f = (r* - r)/( + r*). The remium f is also a measure of the intensity of financial controls. The existence of the remium imlies that consumers would have the incentive to engage in illicit caital flight, in order to avoid the tax. This caital flight is intermediated via the trade account by trade misinvoicing, hence its otential magnitude would be determined by the volume of imorts (Y ) and exorts (X X ). Preventing illicit caital flows induced by a remium f requires sending τ X on enforcement. Assume that enforcement and the resultant remium are linked by the following reduced form: 2 φ = φ[ τxy ]; f >0; f <0. This formulation assumes that larger trade oenness requires an equiroortionate increase in enforcement in order to suort the given remium. It also recognizes the diminishing marginal efficacy of enforcement. The tax on foreign bonds and the enforcement of financial reression imlies that the consumer is indifferent to the domestic and the foreign bond. 3 Enforcement of the income tax is associated with collection cost l i (i =, 2) er one dollar of gross taxes alied for tax rates below t, where t reflects the tax caacity.the net tax collected by a tax t is ( ) txfor t < t and ( ) txfor ti ti. The second λ i i i i λ i i

4 78 Joshua Aizenman and Ilan Noy eriod collection costs and tax caacity are determined by the first eriod investment rate in fiscal caabilities, denoted by y, where l 2 = l 2(y), l 2 0, l 3 0; t2 = t2( ψ ); t 0; t 0. The net revenue from the income tax, lus the revenue from the domestic bond sold in eriod one, finances the fiscal exenditure on exogenous ublic sending (G), lus the cost of the enforcement of caital controls and investment in fiscal caability [( τ + ψ) X ], lus the reayment of old debt. In these circumstances, financial reression imoses taxes on domestic savings at the remium rate, f. This tax has two comonents: first, it taxes foreign bonds, D*, directly, at the remium rate; second, it reduces the cost of financing the domestic debt, D, at the remium rate. The sum of both imlies that financial reression taxes domestic savings, D* + D. Unlike the rivate sector, the effective real interest rate facing the fiscal authorities equals the foreign one. We consider two ossible scenarios:. The first eriod olicymaker is of tye w, execting to survive for the second eriod with robability q. Policy maker w chooses the tax rates t, the enforcement t, and investment in tax caacity y that would maximize the exected utility of the reresentative agent, subject to the fiscal and the rivate budget constraints, and the uncertainty regarding the future olicymaker. With robability - q, the second eriod olicymaker would be of tye n. 2. The first eriod olicymaker is tye n, reflecting the interests of a narrow ressure grou. Policy maker n chooses t, t, y that would maximize the exected utility of the narrow interest grou, subject to the fiscal and the rivate budget constraints, and the uncertainty regarding the future olicymaker. With robability -, the second eriod olicymaker will be of tye w. The following results characterize the attern of otimal taxes and the resulting financial reression [see Aizenman and Noy (2004) for the derivation]:. Financial reression is art of the otimal ublic finance of olicymaker w if λd φ ( 0) τ = 0 > β. Hence, regimes characterized by costly collection of taxes, low trade oenness, and high saving are more likely to use financial reression. If λd φ ( 0) τ = 0 > β, the otimal financial reression increases with the collection costs associated with income taxes and dros with the economy s trade oenness. This follows from the observation that greater trade oenness increases the effective cost of enforcing financial reression, reducing thereby the usefulness of financial reression as an imlicit tax. 2. Lower robability of staying in ower (i.e. low and q) reduces the otimal investment in tax caacity, increases l, and increases the likelihood that financial reression would be art of the menu of taxes. If the first eriod olicymaker is of tye w, a higher robability of staying in ower for the second eriod would increase the first eriod investment in future tax caacity, reducing the otimal financial reression. If the first eriod olicymaker is of tye n, reflecting the interests of narrow ressure grous, the tax would be set at the eak of the corresonding tax Laffer curve, t = t. The attitude of tye n olicymaker towards caital controls is mixed caital controls would be imosed on the reresentative consumer, while first eriod fiscal surlus would be ut in off-shore accounts, as insurance against losing ower. Lower robability of staying in ower by the regime reresenting the narrow ressure grou, n, increases caital flight and reduces financial reression. 4 The model can be embodied in an overlaing generation structure, where the resent investment in tax caabilities determines the collection cost and the tax caacity next eriod. We can also extend the model to allow concave collection costs.

5 ENDOGENOUS OPENNESS 79 The robability of staying in ower can be modeled in the context of a more elaborated economy. These extensions would not change the main results regarding the association between trade and financial oenness. Following the aroach of Cukierman et al., (992), one exects less olarized societies and better functioning democracies to be characterized by more efficient tax collection systems (hence by lower l). Alying this conjecture, a more efficient tax system would be associated also with a lower tax rate, t, thereby reducing the attractiveness of caital flight. It can be verified that with low enough financial reression, the net effect of imroving the tax system is to lower the incidence of caital flight, thereby reducing de facto financial integration. Of course, de facto financial oenness is imacted by other considerations not addressed by the ublic finance model described above, such as differentials in discount rates and investment oortunities across countries, etc. One should view the above model as suggestive of ossible links between macro- and olitical-economy factors and de facto oenness, motivating the emirical research. Our discussion so far has focused on the ossibility that greater trade oenness will lead to higher financial oenness. It is reasonable to exect that the linkages between trade and financial oenness oerate in both directions, and that higher financial oenness would lead to greater trade oenness. 5 A likely channel is vertical foreign direct investment (FDI). FDI allows multinationals to fragment roduction otimally, benefiting from the cost advantage associated with locating labor intensive roduction stages in labor abundant countries. A by-roduct of this fragmentation is the growth of two-way trade: higher imorts of rimary and intermediate roducts, followed by higher exorts of the ugraded roducts. The ositive association between trade and financial oenness may also be the outcome of olitical-economy factors. They roose an interest grou theory of financial develoment whereby incumbents oose financial develoment because it breeds cometition. In these circumstances, the incumbents oosition will be weaker when an economy allows both cross-border trade and caital flows. They redict that a country s domestic financial develoment should be ositively correlated with trade oenness, and they identify the time varying nature of this association. Another interesting aroach linking trade and financial oenness is Portes and Rey (2003), who show that international trade in both goods and assets are exlained by similar gravity regressions. We therefore exect to observe two-way linkages between trade and financial oenness. In the next section, among other things, we will look at these causal links emirically. We first confirm the imortance of lagged commercial oenness in Granger-causing contemoraneous financial oenness, after controlling for macro- and olitical-economy variables. Next, we show that lagged financial oenness lays an imortant role in accounting for commercial oenness. We close the emirical evaluation of this question by decomosing the relative quantitative imortance of the two channels. 3. The Emirical Model We measure de facto financial oenness using the sum of total caital inflows and outflows (in absolute values) measured as a ercent of gross domestic roduct. Caital flows are the sum of FDI, ortfolio flows, and other investments. 6 In terms of financial instability (measured as standard deviation), develoing countries seem to exerience much higher volatility relative to their degree of oenness. This difference is most

6 80 Joshua Aizenman and Ilan Noy striking between develoed countries and the East Asian emerging markets that aear to have similar levels of average oenness but with much higher volatility. For our commercial oenness index, we average the sum of exorts and imorts as a ercentage of GDP over the revious 4 years (t - to t - 4). By averaging, we smooth out any fluctuations due to temorary changes in the terms of trade. Bivariate analysis clearly shows a artial correlation between the two annual measures of oenness (stronger for the 990s), but only a weak and unstable correlation between the gross and net financial flows. As the revious theoretical discussion suggests, one of the determinants of de facto financial oenness should be the legal imediments to financial flows. Accordingly, we include in our multivariate analysis two measures for restrictions on the caital account. The first, taken from Chinn and Ito (2006), is constructed using the data rovided in the IMF s Annual Reort on Exchange Arrangements and Exchange Restrictions. The second measure is constructed from Bekaert et al. (2005) by summing three binary indicators of the oenness of equity markets. For the olitical-economy determinants of financial oenness, we concentrate our emirical investigation on three olitical institutional measures: a democracy index, a measure of olitical cometition, and a Herfindahl index for government fractionalization. Cukierman et al. (992) argued that functioning democracies will tend to have more efficient tax collection systems. In our theoretical work, we concluded that the degree of tax collection costs will determine the degree of financial reression. To investigate this hyothesis we examine whether the caacity of the olitical system to revent friction is a relevant measure. We exect less olarized societies and those in which conflicts are solved eacefully within the olitical system to have more efficient tax collection mechanisms in lace. We emloy a variable that measures the degree of democratic rule, and a variable that measures the degree of olitical cometition within a olity. These indices are highly correlated even though they were constructed on different criteria and sources. Another olitical-economy variable we use to examine the robustness of our results comes from a olitical dataset constructed at the World Bank (Keefer, 2002). Because we hyothesized that more olarized social and institutional arrangements will affect the efficiency of tax collections, we use an index that measures the fractionalization within government. Following the work of Wei (2000) and Dreher and Siemers (2003), we examine whether corrution matters for the degree of financial oenness. To that end, we use a measure of corrution taken from the International Country Risk Guide. Besides democracy and cometition variables, the other olitical variables do not seem significantly correlated. We also include a host of macroeconomic control variables. In all regressions we use the inflation rate, er caita gross domestic roduct, the government s budget surlus (as share of GDP), and a world interest rate. In order to examine whether the occurrence of financial crises matters, as they might systematically change the relationshi between financial oenness and our control variables, we include crises measures in several regressions. We attemted to include all 205 countries and territories for which data are available. Most of the data on financial flows are tyically available only from the 980s and only for a much smaller set of countries. Our dataset is therefore an annual anel of 83 countries for the years We further investigate the robustness of our results by examining various subsamles. Notably, we hyothesize that results for OECD countries might be different from those for develoing countries. We thus reeat our regressions for develoed economies which we define as those economies that were members of the OECD in 990. We exclude island economies (as these are often used as off-shore banking

7 centers and their level of de facto financial oenness is often much higher). The Data Aendix in Aizenman and Noy (2004) rovides a detailed summary of the variables, sources, and samles. Based on our theoretical work, we estimate the statistical significance of various sources of financial reression by ositing a linear structure for the determination of the level of financial oenness whereby: FO = α + β X + β CO + β P + ε, with ε = ρε + μ. () it it 2 it 3 it it it it it ENDOGENOUS OPENNESS 8 The deendent variable (FO it), financial oenness for country i at time t, is assumed to be deendent on an intercet (or, alternatively, searate country or regional intercets), a vector X it of macroeconomic control variables, an average of lagged commercial oenness (COit ), a vector of olitical institutional variables (P it), and an error term. A Durbin-Watson statistic for all iterations of the model strongly indicates that the error terms are autocorrelated. The error term is thus assumed to have an AR() structure with m iid. 7 We estimate the model using the Prais-Winsten algorithm. Table includes results for our benchmark regressions. For the first stage regression, the R 2 is between 0.20 and 0.67 deending on the exact secification and samle used. 8 For the second stage, the model converges very quickly (within two iterations), and most of the coefficients for the benchmark control variables are robust to the inclusion Table. Benchmark Model Results () (2) (3) (4) (5) (6) Per caita GDP 0.64** *** 2.4*** 2.02***.4*** (2.4) (.09) (4.09) (4.28) (3.67) (3.) Budget surlus -0.26* 0.44*** -0.40** -0.28* -0.42** -0.26* (% of GDP) (-.70) (4.60) (-2.07) (-.62) (-2.6) (-.8) Inflation (CPI) (-0.6) (-.46) (-0.38) (-0.27) (-0.47) (-0.28) US Treasury bill rate (-0.88) (-0.4) (-0.53) (-0.70) (-0.38) (-0.32) Trade oenness 0.*** 0.09*** 0.07*** 0.08*** 0.08*** 0.09*** (average for (9.08) (7.99) (4.52) (5.5) (5.5) (7.9) t,...,t 4) Democracy/autocracy -0.44*** ** -0.26* ( 2.7) ( 0.37) (-2.48) (-.60) Corrution -2.0** ** -.86* (-2.23) (-0.25) (-2.24) (-.59) The 990s 4.89*** 3.04*** 4.65** 4.04** 3.52* 3.83** (2.99) (3.7) (2.0) (2.08) (.62) (2.7) r 0.88*** 0.86*** 0.88*** 0.88*** 0.88*** 0.88*** Observations Samle ALL OECD DEV DEV DEV DEV Notes: t-statistics for all variables are given in arentheses. The left-hand side variable is the sum of financial inflows and outflows (as % of GDP). Estimation using the Prais-Winsten algorithm assuming an AR() rocess for the error terms. r is the correlation coefficient for the AR() rocess: e it = re it- + m it.all denotes the whole samle; OECD includes only OECD countries; DEV denotes the develoing countries samle. * Significant at 0%; ** significant at 5%; *** significant at %.

8 82 Joshua Aizenman and Ilan Noy and exclusion of other variables. In column () of Table, which includes the full samle, we already observe many of the results that remain throughout. The coefficient for er caita GDP is always significantly ositive i.e. an increase in GDP er caita increases financial oenness (excet for a regression containing only OECD countries in which the coefficient is insignificant). We find that an increase in er caita GDP of PPP$000 will facilitate a 0.4 to 2.28 ercentage oints increase in the volume of caital flows (as ercent of GDP). The ratio of budget surlus to GDP is tyically significant and always negative for develoing countries. A bigger budget deficit will increase de facto financial oenness. Again, this result does not hold for our OECD subsamle; for this case, reorted in Table column (2), the budget surlus coefficient is ositive and significant. 9 The inflation rate and the world interest rate are always insignificantly different from zero. But, as with the revious results, the coefficients for inflation and the world interest rate seem to be different for the OECD subsamle; although these are still insignificant for standard significance levels, the effect of inflation on financial oenness is larger (and negative) for the OECD countries and the effect of the US T-Bill rate is smaller. Both these results corresond with our intuition. We also include a binary variable for the 990s and, as exected, the coefficient for this variable is always ositive and significant. This increase in caital flows is found to be between.3 and 4.9 ercent of GDP. Additionally, we find that the trade oenness coefficient is always ositive and highly significant. As this variable describes the average oenness over the revious four years, we find that a history of more commercial oenness will increase financial oenness significantly. This result is robust to all the iterations we resent in Tables and 2. Before discussing our emirical analysis of the olitical-economy determinants of international financial flows, we note that including the corrution variable in our regressions also yields negative and significant coefficients in almost all the iterations of the model. Similar results from different data are analyzed in detail in Wei (2000) and Dreher and Siemers (2003). We examine the effect of the nature of the olitical regime on financial oenness (this index is between 0 (full democracy) and 0 (full autocracy)). For the full samle (Table column ) and the develoing countries subsamles (Table column 3), the coefficient for this variable is negative, significant, and aarently large. Any one-oint increase in this index (out of the 20 oints difference between full autocracy and democracy) reduces financial oenness (international financial flows) by almost onehalf of a ercentage oint of GDP. Because the results for the OECD subsamle are consistently different, and our theoretical modeling is focused on develoing countries, we give most attention to the develoing countries subsamle (these include all non-oecd countries that are not islands/financial centers). Columns 5 and 6 in Table reeat our secification for the develoing countries samle but exclude the regime variable in column 5 and both the regime and corrution measures in column 6. In both cases, we find that all the other results reorted above remain robust to these omissions. We calculate the quantitative significance of our findings for the benchmark model. For the samle of develoing countries, we find that a one standard deviation increase in the commercial oenness is associated with a 9.5 ercentage oints increase in de facto financial oenness (% of GDP), a one standard deviation increase in the democratization index reduces financial oenness by 3.5 ercentage oints, and a one standard deviation increase in corrution is associated with a reduction of financial oenness by 3. ercentage oints. Similarly, the corresonding associations for the whole samle are 2.3, 3., and 2.9. Furthermore, a develoing country will have: higher

9 ENDOGENOUS OPENNESS 83 financial oenness (measured as 3 additional ercentage oints of GDP) were it to have the median level of trade oenness of an OECD country; would be 2.2% less oen were it as democratic as the tyical OECD country; and 4% more oen to financial flows were it less corrut, as the tyical develoed country is. In Table 2 columns () (4) we further investigate the olitical-economy nature of financial oenness by relacing the democracy/autocracy (regime) variable with two others: a measure of olitical cometition, and an index of government fractionalization. For the olitical cometition variable, we find that increased institutionalized cometition within the olity decreases financial oenness. Thus, a more oenly cometitive, free, and inclusive olitical system will lead to lower levels of financial oenness, after controlling for incomes, macroeconomic olicy (inflation and budget surluses), interest rates, and commercial oenness. For the government fractionalization index (reorted in Table 2 columns 3 and 4), we find that the more a government is fractionalized (the ruling coalition includes more olitical arties), the higher is the financial oenness. Quantitatively, the estimated coefficient of.4.9 does not seem to suggest a very large effect on the level of international financial flows. In columns (5) (6) of table 2 we re-estimate our benchmark secification (Table column 3) but also include the de jure measures of financial oenness (oenness of the caital account). We use two measures. The first, taken from Chinn and Ito (2006), is constructed using the IMF data rovided in its Annual Reort on Exchange Arrangements and Exchange Restrictions (column 5). The second measure is constructed by adding the three binary indicators of equity markets oenness available in Bekaert et al. (2005). Interestingly, the coefficient for the Chinn-Ito measure of restrictions is not significant in this secification, nor in the other secifications we ran. The Bekaert et al. (2005) index is ositive and significant at the 0% level. These results merit further inquiry because of the questions they ose; and the imortance of the de jure environment on the de facto amount of caital flows has not been examined thoroughly in the literature. Nevertheless, our main results with resect to commercial oenness and the olitical regime remain significant also when the de jure measures are included; though the corrution coefficient is no longer always significant, ossibly reflecting a correlation between corrution and the decision by the authorities to use financial reression. In column 7 of Table 2, we utilize a different measure of trade oenness. The traditional measure of the sum of exorts and imorts (% of GDP) does not cature accurately trade oenness in countries that romote exorts but discourage imorts. Thus, we comute an alternative index of oenness based only on imorts. Results remain very similar for all variables, both statistically and in magnitude, but for a larger coefficient for the trade oenness measure. 0 In addition to the secifications discussed above, we tested a number of alternative secifications of our emirical model in order to verify the robustness of our results. Because of sace considerations we do not include the full secifications in our tables, but all these results are available uon request. First, we hyothesized that financial crises (either banking or currency crises) might significantly affect the level of financial oenness in general and more secifically the use of financial reression. Interestingly, in all iterations of the model we attemted, none of the crises coefficients comes out significant. Second, besides including the average of ast commercial oenness, we also included in our secification the contemoraneous TRADE/ GDP variable and obtained the following. In all cases, the lagged commercial oenness variable remains ositive and highly significant. For the develoing countries samle, as well as the whole samle, the lagged average is ositive and highly significant with a now larger coefficient (0.20 and 0.2, resectively), while the contemoraneous variable is

10 84 Joshua Aizenman and Ilan Noy Table 2. Robustness: Political Variables () (2) (3) (4) (5) (6) (7) Per caita GDP 2.26*** 2.5*** 0.64*** 0.33**.7***.03***.75*** (3.94) (4.7) (5.70) (2.03) (2.73) (9.42) (4.50) Budget surlus(% of GDP) -0.44** -0.30* * (2.9) (.7) (0.60) (.70) (.43) (0.37) (.7) Inflation (CPI) (0.45) (0.32) (0.02) (0.29) (0.32) (0.45) (0.07) US Treasury bill rate *** (0.5) (0.59) (2.6) (.4) (.03) (.54) (0.74) Trade oenness a (average for t,...,t 4) 0.08*** 0.08*** 0.05*** 0.06*** 0.07*** 0.04*** 0.23*** (4.65) (5.47) (5.04) (2.88) (3.96) (3.03) (4.59) Political cometition -0.67* (.75) (.9) Government fractionalization -.87** -.40 (-2.22) (-.24) Democracy/autocracy -0.39* -0.3*** 0.34*** (.74) (3.37) (2.07) De jure restrictions on the caital account b * (0.57) (.70) Corrution -2.07* 0.73*** ** 0.8 (-.74) (2.93) (.2) (2.38) (0.82) The 990s 4.0* 3.70*.27***.89*** * (.79) (.88) (2.56) (2.78) (.48) (.07) (.85) r 0.88*** 0.88*** 0.73*** 0.83*** 0.89*** 0.73*** 0.89*** Observations Samle DEV DEV DEV DEV DEV DEV DEV Notes: t-statistics for all variables are given in arentheses. The left-hand side variable is the financial oenness variable. Estimation using the Prais-Winsten algorithm assuming an AR(). r is the correlation coefficient for the AR() rocess: eit = reit- + mit. DEV denotes the develoing countries samle. a The measure for trade oenness in column 7 includes only imorts and not the standard sum of imorts and exorts. b The de jure caital account oenness index in column (5) is the measure given in Chinn and Ito (2006), while the one in column (6) is constructed by the authors from the data given in Bekaert et al. (2005). * Significant at 0%; ** significant at 5%; *** significant at %.

11 ENDOGENOUS OPENNESS 85 Table 3. Geweke (982) Decomosition of Causality Decomosition of feedback Percent of overall linear feedback From financial oenness to commercial oenness (F FO,CO) 0.27*** 53 From commercial oenness to financial oenness (F CO,FO) 0.7*** 34 Simultaneous causality between financial and commercial oenness (F FO CO) 0.06*** 3 Notes: *** reresents rejection of H 0: no causality at the % significance level based on a c 2 test as in Geweke (982). The feedback is defined in equations (2) (4). The ercent of the total linear feedback between the two time series is defined in equation (3). negative and significant. For the OECD samle, the lagged average is still ositive and highly significant, while the contemoraneous variable is now ositive but insignificant. The sum of the two coefficients is always ositive and highly significant for the three different samles. 2 As the olitical and institutional variables we use do not vary sufficiently over time we do not resent results for the model estimated with country effects. Tyically, the goodness of fit is higher but the indeendent olitical institutional variables lose most of their statistical significance (as would be exected). We include regional effects (binary variables for Latin America and East Asia) in our large and develoing countries samles. Time effects do not rovide any additional exlanatory ower besides a significant finding for the 990s (reorted above). 4. Causality In the revious section we have established that ast trade oenness Granger-causes financial oenness (see Granger, 969). 3 The Granger framework assumes no instantaneous feedback between the two series and is oriented to insect for uni-directional causality (Geweke, 982). Yet, because causality may also run from ast financial oenness to resent trade oenness we also estimate the oosite secification: COit = γ + δxit + δ2foit + δ3foit + δ4 Pit + ηit. We use the same assumtions, methodology, definition of variables, and samles as before. Results for several secifications are reorted in the working aer. Our focus is the determination of financial oenness and we therefore concentrate our attention on the financial oenness index. In all the secifications reorted in Table 6 of Aizenman and Noy (2004) it is aarent that financial oenness also Granger-causes trade oenness. These results hold whether we examine a one-year lag of the financial oenness measure, or a four-year average of ast financial oenness for the various subsamles reviously described. We also imlement a different causality test that is more robust to alternative assumtions about the error terms that are more likely to hold for time series the test is described in Holmes and Hutton (992). Results indicate a more robust evidence of rima facie causality from finance to trade than from trade to finance. We also find an asymmetric effect of de jure restrictions on de facto oenness: while there is only weak evidence that de jure restrictions on caital mobility imact de facto financial oenness, de jure restrictions on the current account have large adverse effects on commercial oenness. These findings suggest that it s much easier to evade restrictions on caital account convertibility than restrictions on commercial trade.

12 86 Joshua Aizenman and Ilan Noy In Granger (969), the ossibility of simultaneous causality between the two time series is ignored. In rincile, dividing the time series into shorter eriods should enable us to identify accurately the exact chronology of effects this is often imossible in ractice. As we emloy annual data, and because financial flows resond quickly to exogenous shocks, it is reasonable to exect that our data will also contain what aears to be instantaneous causality between trade and financial oenness. Geweke (982) suggests a methodology to distinguish between (temoral) causality (from x to y and from y to x) and simultaneous causality (between the two). We briefly describe the methodology and rovide results. 4 First we estimate the following five equations using a anel fixed-effects leastsquares estimation for our develoing countries samle. FOit = αi + β sfoit s + β2 scoit s + εit 0 2 FOit = αi + β 2 sfoit s β scoit s + εit 3 FOit = i α β sfoit s + εit. (2) 4 COit αi β 4 scoit s β2 4 4 = + + sfoit s + εit 5 COit = αi + β 5 5 scoit s + εit Next, following Geweke s (982) notation, we define F CO,FO as the linear feedback (i.e. G-causality) from trade oenness to financial oenness, F FO,CO as the G-causality from financial oenness to trade oenness, and F FO CO as the instantaneous linear feedback between the two series. 5 F FO,CO, defined as the total measure of linear deendence between the two series, is therefore given by: F = F + F + F FO, CO FO CO CO FO FO CO. (3) Given these definitions, Geweke (982) concluded the following: 5 4 FFO CO = log[ var ( εit ) var( εit )] 3 2 FCO FO = log[ var ( εit ) var( εit )]. 2 FFO = log[ var ( ε ) var ( ε )] CO it it Geweke (982) showed that the null hyothesis (H 0: F = 0) can be statistically examined using the c 2 distribution. In estimating (3), we started with three lags ( = 3) and reduced ste-wise the number of lags using the Akaike Information criterion. In all cases, it turned out that a single lag contained all the information required to estimate the model; we set = throughout. Table 3 rovides our results. Most of the linear feedback between trade and financial oenness (87%) can be accounted for by Granger-causality from financial oenness to trade oenness (53%) and from trade to financial oenness (34%). Simultaneous correlation between the two only accounts for 3% of the total linear feedback between the two series. When we reeated this algorithm using the same methodology but including in the regressions the control variables reviously described we obtained qualitatively and quantitatively very similar results for the feedback measures. We also examined the (4)

13 robustness using a theoretically equivalent but slightly different set of equations. We relaced the first equation in system (2) with it i s it s 2 s it s 0 FO CO = log[ var ( εit 4 ) var ( εit )]. CO = α + β FO + β CO + ε and F it ENDOGENOUS OPENNESS 87 The results we obtained were very similar. Similar results are also obtained when, following Johnson and Soenen (2004), we estimate the equations using seemingly unrelated regressions methodology. Because we reorted evidence of more financial and trade oenness in the 990s, we examined the difference in the measured feedback effects between the two subsamles. While the measured magnitude of the feedback effects is significantly smaller, we found no evidence that the relative imortance of the two feedback channels (from trade to finance and from finance to trade) has changed.we alied similar analysis to investigate the intertemoral linkages between disaggregated measures of trade and FDI. We found that the strongest feedback between the trade subaccounts is between FDI and manufacturing trade see Aizenman and Noy (2006). (5) 5. Concluding Remarks Our analysis indicates that the de facto financial oenness of develoing countries is a comlex endogenous variable, systematically imacted by economic and oliticaleconomy factors. For a samle of develoing countries, we find that a one standard deviation increase in the commercial oenness index is associated with a 9.5% increase in de facto financial oenness (international financial flows as ercent of GDP). One should be careful in attaching normative imlications to these findings. Yet, the results reorted in this aer may rovide some guidelines to olicymakers. For examle, in an era of raidly growing trade integration, countries cannot choose financial oenness indeendently of their degree of oenness to trade. A country that undergoes raid commercial integration will find it imractical or even imossible to enforce rigid de facto financial reression. Hence, the question for China is not if, but rather when and how, to imlement de jure financial integration. A remaining emirical challenge is to disaggregate de facto financial oenness into its various comonents. Because each tye of flow can be taxed differently with varying degrees of efficiency in tax collection, and faces different degrees of exroriation risk, one can exect the determinants of oenness for each tye of flow to be quite different. References Aizenman, J. and P. Guidotti, Caital Controls, Collection Costs, and Domestic Public Debt, Journal of International Money and Finance (994):4 54. Aizenman, J. and I. Noy, Endogenous Financial and Trade Oenness, NBER Working Paer 0496 (2004)., FDI and Trade Two Way Linkages? Quarterly Review of Economics and Finance 46 (2006): Alesina, A. and G. Tabellini, External Debt, Caital Flight, and Political Risk, Journal of International Economics 27 (989): Antràs, P. and R.J. Caballero, Trade and Caital Flows: A Financial Frictions Persective, mimeo, MIT (2007). Bekaert, G., C.R. Harvey and C. Lundblad, Does Financial Liberalization Sur Growth? Journal of Financial Economics 77 (2005):3 55.

14 88 Joshua Aizenman and Ilan Noy Boyce, L. and J.K. Ndikumana, Public Debts and Private Assets: Exlaining Caital Flight from Sub-Saharan African Countries, World Develoment (2003): Calderón, C. and L. Liu, The Direction of Causality between Financial Develoment and Economic Growth, Journal of Develoment Economics 72 (2003): Chinn, M. and H. Ito, What Matters for Financial Develoment? Caital Controls, Institutions, and Interactions, Journal of Develoment Economics 8 (2006): Chong, A. and C. Calderón, Causality and Feedback Between Institutional Measures and Economic Growth, Economics and Politics 2 (2000):69 8. Claessens, S. and D. Naudé, Recent Estimates of Caital Flight, World Bank Policy Research Working Paer 86 (993). Cukierman, A., S. Edwards, and G. Tabellini, Seigniorage and Political Instability, American Economic Review 82 (992): Dooley, M., Caital Flight: A Resonse to Differences in Financial Risks, IMF Staff Paers 35(3) (988). Dooley, M. and K. Kletzer, Caital Flight, External Debt and Domestic Policies, Federal Reserve Bank of San Francisco Review 3 (994): Dreher, A. and L. Siemers, The Intriguing Nexus Between Corrution and Caital Account Restrictions, mimeo, University of Goettingen (2003). Geweke, J., Measurement of Linear Deendence and Feedback Between Multile Time Series, Journal of the American Statistical Association 77 (982):304 3., Measures of Conditional Linear Deendence and Feedback Between Time Series, Journal of the American Statistical Association 79 (984): Giovannini, A. and M. de Melo, Government Revenue from Financial Reression, American Economic Review 83 (993): Granger, C., Investigating Causal Relations by Econometric Models and Cross-sectral methods, Econometrica 37 (969): , Some Recent Develoments in a Concet of Causality, Journal of Econometrics 39 (988): Helman, E. and A. Razin, A Theory of International Trade Under Uncertainty, New York: Academic Press (978). Henry, P.B., Caital Account Liberalization, the Cost of Caital, and Economic Growth, American Economic Review 93 (2003):9 6. Holmes, J. and P. Hutton, A New Test of Money-Income Causality, Journal of Money, Credit and Banking 24 (992): Hutchison, M. and I. Noy, How Bad Are Twins? Outut Costs of Currency and Banking Crises, Journal of Money, Credit and Banking 37 (2005): International Monetary Fund, Annual Reort on Exchange Arrangements and Exchange Restrictions,, Washington, DC (various years). Johnson, B. and L. Soenen, The US Stock Market and the International Value of the US Dollar, Journal of Economics and Business 56 (2004): Keefer, P., DPI2000: Database of Political Institutions: Changes and Variable Definitions, mimeo, World Bank (2002). Pierce, D., Comment: Measurement of Linear Deendence and Feedback Between Multile Time Series, Journal of the American Statistical Association 77 (982):35 6. Political Risk Services, International Country Risk Guide, Syracuse, NY (various years). Portes, R. and H. Rey, The Determinants of Cross Border Equity Flows, MS, Princeton University (2003). Prasad, E., K. Rogoff, S.-J. Wei, and A. Kose, Effects of Financial Globalization on Develoing Countries: Some Emirical Evidence, mimeo, IMF (2003). Rodrik, D., Who Needs Caital-Account Convertibility? in P. Kenen (ed.), Should the IMF Pursue Caital Account Convertibility?, Essays in International Finance 207, Princeton: Princeton University Press ((999). Talvi, E. and C. Végh, Tax Base Variability and Procyclical Fiscal Policy, NBER working aer 7499 (2000).

15 Tornell, A. and A. Velasco, Tragedy of the Commons and Economic Growth: Why Does Caital Flow from Poor to Rich Countries? Journal of Political Economy 00 (992): Wei, S.-J., How Taxing is Corrution on International Investors? The Review of Economics and Statistics 82 (2000):. Wei, S.-J. and Y. Wu, The Life-And-Death Imlications of Globalization, mimeo, IMF (2002). Zellner, Arnold, Comment: Measurement of Linear Deendence and Feedback Between Multile Time Series, Journal of the American Statistical Association 77 (982): Notes ENDOGENOUS OPENNESS 89. See Giovannini and de Melo (993) for documenting and measuring financial reression as an imlicit tax on savings. See Dooley (988), Tornell and Velasco (992), and Dooley and Kletzer (994) on caital flight as a means of risk diversification. See Claessens and Naudé (993) and Boyce and Ndikumana (2003) for discussions on trade mis-invoicing and caital flight. 2. The otential caital flight is assumed to be roortional to imorts. Similar results would hold if one assumed that otential caital flight be roortional to (imorts + exorts). 3. To simlify notation, we assume that the enforcement revents all tax evasion. The analysis can be extended to account for random intercetion of caital flight, where a higher t increases the robability of intercetion (see an earlier version of our aer, Aizenman and Noy (2004)). 4. Russia in the early 990s may rovide a good case study. Alesina and Tabellini (989) discuss a model where olitical instability and olarization determine the incidence of caital flight. 5. See Helman and Razin (978) for an integrated theory of trade in goods and securities. 6. Wei and Wu (2002) reviously used this financial oenness variable and made it available to us. See [Aizenman and Noy (2004)] for a detailed descrition of our data for financial oenness. 7. E(m t) = 0; E(m t2 ) = s u2 ; and Cov(m t, m s) = 0 for t s. 8. The higher R 2 values are generally for the models that include more olitical institutional variables and for the develoing and OECD subsamles. 9. The disarity between the imacts of budget surlus in develoing and OECD countries may be exlained by the differential cyclical atterns of fiscal olicy. In contrast to the OECD countries, fiscal olicy tends to be rocyclical in develoing countries. Financial crises tend to lead to recessions in develoing countries, inducing abrut fiscal adjustment and reducing fiscal deficits. These observations may lead to the ositive association between smaller budget deficits and lower de facto financial oenness, as in Talvi and Végh (2000). 0. The increase in the magnitude of the coefficient is either because the feedback effects between imort flows and financial flows are stronger, or because there is a high correlation between exorts and imorts. Our new measure is therefore much smaller.. We utilized a number of variants of these binary crisis indicators (currency crisis and banking crisis) but never rejected the null (no effect). The crisis measures are described in Hutchison and Noy (2005). 2. One ossible interretation is that major recessions in develoing countries (otentially triggered by caital flight) are associated with a dro in commercial oenness, as would be the case if the dro in imorts dominated any increase in exorts. Likewise, caital flight may increase financial oenness. It is difficult to rovide a better rationale without desegregating financial oenness into its subaccounts. 3. Holmes and Hutton (992) use the term rima-facie causality. They distinguish between this and true causality which is imossible to infer without strong structural assumtions. 4. Readers may also consult Geweke (984) and Granger (988). The only alications we are aware of which aly this methodology to macroeconomic data series are Chong and Calderón (2000) and Calderón and Liu (2003). Several aers in finance have emloyed the same methodology. 5. Geweke (982) referred the term linear feedback. Pierce (982), in a comment on Geweke s work, argued that a more aroriate term to describe the measures defined in our equations (9) (0) in Aizenman and Noy (2004) would be G-causality. Zellner (982), in another comment, argues that the word causality should not be used if it is only based on statistical observed relationshis rather than together with economic theory.