Controls on capital inflows have seen a renaissance since the beginning of the global financial crisis in 2008 (Williamson, Jeanne, and Subramanian 2012) and many countries, including Thailand, South Korea, Peru, Indonesia, Brazil and Iceland, have imposed controls.
The IMF’s evolving position on capital controls
The IMF once advocated the removal of all controls on outflows and inflows in the heydays of the Washington Consensus in the 1990s. The Asian Crisis of 1997-1998, however, initiated a slow process of conversion that culminated with the IMF’s recent decision to explicitly and openly support the imposition of controls on capital inflows (euphemistically naming them ‘capital flow management policies’). The basic premise of this new IMF stance on capital controls is that controls should be imposed when countries are facing a capital inflow surge and after all other policy alternatives have been exhausted1. Given the IMF’s new stance, we ask in Jinjarak et al. (2012) whether capital controls on inflows are effective.
If they are, what are their exact effects? We attempt to answer these questions using the Brazilian experience in imposing new, price-based, controls during the global financial crisis of 2008-2011. There are five possible impacts of capital controls on inflows:
- Reduce the volume of inflows.
- Change their composition.
- Affect the real exchange rate.
- Enable a more independent pursuit of monetary policy.
- Increase/decrease financial stability2.
Two recent survey papers on capital controls find the evidence inconclusive, with some observed effects on the composition of flows, but very little effect on volumes of flows. In these papers, there is even less agreement on the impact of controls on the exchange rate and interest rates3. Brazilian capital controls Brazil liberalised its capital flow regimes gradually. Starting from the early 1990s, liberalisation culminated in an almost completely open capital account by the mid-2000s4. After a fairly brief period of no taxes on foreign capital transactions, taxes were reintroduced in March 2008 at the rate of 1.5% on fixed-income investments5. Investments related to equities remained exempt from taxes until a lot later6. The tax was reduced to zero in October 2008, the peak of the global financial crisis, when the exchange rate came under depreciation pressures. A 2% tax on fixed-income and equity inflows was reintroduced in October 2009. The tax was then increased to 6% in two stages in October 2010; but then reduced back down to 2% in January 20117.
In our research, we use micro-level data on capital flows from US and European mutual funds investing internationally, and a new methodology to estimate the counter-factual with no imposition of controls8. For every episode of capital control, we study 12 weeks (which is approximate to a quarter of the year) before and after the control date.
What happened to capital inflows in Brazil?
We graph the actual evolution of capital inflows and the counterfactual that assumes no change in policy. We summarise these results chronologically for each change in Brazil’s capital account policies in a set of five figures that correspond to the five policy changes.
For the first act – March 2008, taxing fixed income only – we observe a decline in flows in the run-up to the placing of controls, but net funds start flowing in again about two weeks before the episode. This budding inflow may be the impetus for the placing of controls, see Figure 1. The placing of controls did not appear to have a large influence, causing only a small and temporary slowdown in the inflow episode that resulted from the controls. While we observe a continuation of the inflow for the counterfactual scenario, Brazil experienced a similar dramatic rise, but with about a month’s delay. This delay, however, is also present in inflows to other Latin American destinations.
Figure 1. 12 March 2008: Capital flows when taxing fixed income investment
In Figure 2, we report on the second act – October 2008, removing the fixed-income tax – before which inflows were decreasing rapidly throughout the pre-crisis period, starting in July. We observe evidence of a slowdown in the capital outflows as a result of this removal of controls in October. The counterfactual Brazil, without the relaxation of controls, would have experienced a continuing capital flight, that is, a reduction of net capital inflows.
Figure 2. 23 October 2008: Capital flows when cutting fixed income tax
As shown in Figures 3 and 4, in both the third and fourth acts in 2009 and 2010 respectively the post-control inflow boom episode seem to be large and unique. It seems the controls did not manage to stem the volume of these inflows.
Figure 3. 20 October 2009: Capital flows when taxing stock and bond investment at 2%
Figure 4. 4 October 2010: Capital flows when increasing taxes 2% to 4%
For the fifth act – January 2011, reducing taxes on equities – we observe a short-run surge in equity investment that is unique to the Brazil funds, as shown in Figure 5. The relaxation of controls did appear to have a short-term but statistically and economically meaningful impact on capital flows. In the longer term (three months in our framework) there does not seem to be any significant impact. Capital controls as a signal After controlling for the counterfactual – that is, Brazil with no capital account policy change – for each event in which Brazil modified its capital controls during the first three years of the global financial crisis, we find no evidence that tightening of controls was effective in reducing the magnitudes of capital inflows into the country. We do observe some modest success in preventing further declines in inflows when the capital controls are relaxed.
Figure 5. 3 January 2011: Capital flows when reducing taxes from 6% to 4%
These results are consistent with survey responses described in Forbes et al. (2012). In these surveys, it appears that investment managers’ reactions to limited capital account policy changes are muted and remarkably heterogeneous. Given this heterogeneity, it may not be a surprise that we find so little impact when the capital account is tightened.
Why did we find an asymmetric impact? The Forbes et al. (2012) interviews suggest that many money managers were more interested in the signal content of the policy change rather than in the direct impact of the actual changes on their tax liability. Brazil, throughout this period, was controlled by the left-of-centre Workers’ Party, headed by Lula9. Our hypothesis is that the price-based mild capital controls’ only perceptible effects are to be found in the content of the signal they broadcast regarding the government’s larger intentions and sensibilities. In Lula’s case, the government was widely perceived as ambivalent to markets, and especially to the international capital markets. Thus, an imposition of controls was not perceived as ‘news’ and thus had no impact. A willingness to remove controls, however, as happened in October 2008 in the middle of the post-Lehman panic and then again in January 2011, were both apparently perceived as noteworthy indications that the government was not as hostile to the markets as many expected it to be (cf. Chang 2010).
In a meeting discussing the IMF’s guidelines for supporting the use of capital controls, the Brazilian finance minister Guido Mantega declared: "We oppose any guidelines, frameworks or 'codes of conduct' that attempt to constrain, directly or indirectly, policy responses of countries facing surges in volatile capital inflows" (Reddy 2011). The Brazilian government, as well as other representatives from emerging markets, found the IMF’s limited support of capital controls as too constraining, and argued for a broader mandate to use them.
We do not find that these controls have much impact, despite the rationale for their intended use and the vocal support they have garnered from many corners. The reasons for instituting these policies, of course, may be political and electoral in nature. Indeed, it may be that policymakers fully understand the inability of these controls to make any substantial impact, but nevertheless resort to adopting them.
Abadie, A., A. Diamond, and J. Hainmueller (2010), “Synthetic control methods for comparative case studies: Estimating the effect of California's tobacco control program”, Journal of the American Statistical Association, 105(490), 493-505.
Chang, Roberto (2010), “Elections, Capital Flows, and Politico Economic Equilibria”, American Economic Review, 100(4), 1759-77.
Forbes, Kristin, Marcel Fratzscher, Thomas Kostka and Roland Straub (2012), Bubble Thy Neighbor: Portfolio Effects and Externalities from Capital Controls”, NBER Working Paper, 18052.
Glick, Guo and Michael Hutchison (2006), “Currency Crises, Capital Account Liberalization, and Selection Bias”, Review of Economics and Statistics, 88(4), 698-714.
Jinjarak, Yothin, Ilan Noy and Huanhuan Zheng (2012), “Capital Controls in Brazil: Stemming a Tide with a Signal?”, working paper, School of Economics and Finance, Victoria, University of Wellington.
Magud N E, Carmen M Reinhart and Kenneth S Rogoff (2011), “Capital controls: Myth and reality - a portfolio balance approach”, NBER, 16805.
OECD (2011), “OECD Economic Surveys: Brazil”, Paris.
Ostry, Jonathan D, Atish R Ghosh, Marcos Chamon and Mahvash S Qureshi (2011), “Capital Controls: When and Why?”, IMF Economic Review, 59(3).
Reddy, Sudeep (2011), “Emerging Nations Reject Capital Plan: IMF Delays Program to Help Countries Manage Short-Term Inflows; as Geithner Blames China, Others Blame the Fed”, The Wall Street Journal, newspaper, 18 April.
Williamson, John, Olivier Jeanne and Arvind Subramanian (2012), “International rules for capital controls”, VoxEu.org, 11 June.
1 Figure 1 in Ostry et al. (2011) provides a parsimonious summary of these caveats and preconditions.
2 The evidence on financial stability in general, and in particular about the impact of controls on the likelihood of financial crises is quite mixed (cf. Glick et al. 2006). For reasons that are better explained in our full report (Jinjarak et al. 2012), we chose to focus on the volume of capital inflows.
3 Magud, et al. (2011) and Ostry et al. (2010).
4 Baba and Kokenyne (2011) provide an evaluation of Brazil’s capital account regime in the run-up to just before the global financial crisis.
5 This tax, known as the IOF (imposto sobre operações financeiras), has been used during the 1990s as well.
6 In May 2008, the tax was extended to cover ‘simultaneous operations’ to prevent circumvention of the inflow tax (circumvention which was apparently widespread).
7 Tax was also expanded to cover margin calls on derivative positions and foreign borrowing with maturities below one year.
8 We use a methodological innovation recently formalized in Abadie, Diamond and Hainmueller (2010). The synthetic counterfactual’s construction in their methodology is based on the statistical relationship to a control group. The algorithm does not presume to impose any ad hoc assumptions about the likely control group, but rather derives this control group as a weighted average of all non-treated country observations (with weights estimated from pre-treatment, pre-GFC data). The procedure allows us to construct a no-policy-change counterfactual and thus measure in detail the impact of the controls themselves.
9 Luiz Inácio Lula da Silva. Replaced on 1 January 2011 by Dilma Rousseff from the same left-of-centre political party.