By Annamaria de Crescenzio, Etienne Lepers, Zoe Fannon
Emerging markets have increasingly been using macroprudential policy to mitigate risks from high currency exposures in both banks’ and firms’ balance sheets. One such tool, widely used in Latin America and to some degree in Central and Eastern Europe, is foreign exchange (FX)-differentiated reserve requirements. This column presents findings from a recent OECD study of the benefits and side effects of currency-differentiated reserve requirements. Unlike most of the existing literature, which uses binary policy change variables, we are able to capture the magnitude of policy changes, leveraging on a new granular dataset of reserve requirement rates.
Reserve requirements, which require banks to hold a portion of their deposits or liabilities as reserves at the central bank, have been part of the monetary policy toolkit for a long time. They belong to the category of “currency-based measures” (CBMs), i.e. measures that apply less favourable treatment on the basis of the currency of an operation rather than on the residency, such as limits on the net FX positions or differentiated liquidity ratios, which have proliferated in the post-global financial crisis period (De Crescenzio et al 2015, Ahnert et al 2020). Indeed, if the use of reserve requirements for monetary policy purposes has fallen out of fashion in many countries, they have recently made a comeback in academic debates and policymaking as useful macroprudential tools. This is notably the case when central banks apply different, usually higher, reserve requirements to deposits made in foreign currency (FX), in order to target the specific risks to financial stability associated with currency exposures and dollarization.
While the use of such FX reserve requirements is widespread, notably in Latin America and to a lesser degree in Central and Eastern Europe, the empirical evidence on their effectiveness is limited. Existing empirical work on the use of reserve requirements have not focussed specifically on currency-differentiation, or only broadly considered the direction of a given policy change, but not its magnitude. It is difficult to interpret the estimates from these studies, as the models hold as equivalent a 1% and a 10% increase in the gap between the FX and local currency (LC) reserve requirements, and can thus be of limited guidance for policymakers.
Using a unique OECD dataset combining previous data efforts (Federico et al., 2014) with a detailed survey of OECD members and partner countries’ authorities (OECD 2019), we analyse the benefits and side effects of FX reserve requirements, accounting for the intensity of policy changes (De Crescenzio et al 2021). Our dataset records the magnitude of FX and LC reserve requirements for 58 countries over a period from 1999Q1 to 2015Q3.
Our analysis shows that adjusting the gap between FX and LC reserve requirements (the FX-LC RR gap) can be an effective policy tool for addressing risks from high currency exposures in bank balance sheets.
Central banks’ use of FX reserve requirements as a macroprudential policy tool to mitigate risks to financial stability
FX reserve requirements (FX RR) are a widely used a tool to mitigate risks arising when banks take in foreign currency (OECD 2019). They require the bank to hold a fixed percentage of its FX liabilities as – generally cash – reserves at the central bank. Countries usually require higher reserve ratios for FX liabilities than LC ones to discourage further issuance of FX liabilities. Figure 1 displays the difference between the FX rate and the LC rate, hereafter “the gap” for key selected countries since 1999.
Figure 1: Currency-differentiated reserve requirements gap for selected countries
Since many domestic banks’ assets are primarily in the form of loans to domestic businesses and individuals with income in LC, they can suffer from currency mismatches between their assets and liabilities if they fund themselves in – usually cheaper – foreign currency. In the event of a decline in the local currency, these banks may struggle to cover their FX liabilities with their LC income, posing risks to the banking system. The hope of the central bank is that increasing the FX-LC RR gap will reduce the attractiveness of conducting business in FX. If banks are forced to hold a larger portion of FX liabilities in reserve than for LC liabilities, they may not offer as high a return to FX depositors.
The effectiveness of FX reserve requirements as a macroprudential tool
As discussed above, central banks can reduce currency risks by changing the FX-LC RR gap. Now, we test that theory with our data. Leveraging on our reserve requirement dataset, we are able to capture the magnitude of the policy change and isolate the impact of the currency differentiation net of volume effects. Our main policy variable of interest is the quarterly change in the FX-LC RR gap, measured in percentage points. Assuming the impact of the policy change on our outcomes of interest happens with a lag, we study the implications of an FX-RR change in the preceding four quarters.
First, we test a number of outcome variables that may be targeted by changes to the FX-LC RR gap. These include: total capital inflows to domestic banks; the share of bank liabilities that are in FX; the share of bank assets that are in FX; and the net FX position of the banking sector.
Second, we investigate whether there are identifiable side effects, e.g. on exchange rate deviations from trends and on capital flows to other sectors than the banking sector.
Finally, we investigate where FX-RR changes can be circumvented domestically or internationally. Previous studies suggested that changes in FX macroprudential tools might shift the currency risk to other sectors within the country, hence circumventing the controls (Ahnert et al 2020), or deflect that risk to other countries. We check this by evaluating additional outcome variables, including non-financial corporation debt growth and inflows to non-banks, as well as alternative specifications modelling potential capital flow deflection effects to other countries internationally.
FX reserve requirements are an effective tool that central banks can use to achieve currency risk reduction
Our empirical analysis shows that central banks can alter FX exposure indicators by changing the FX-LC RR gap. Our analysis is also the first to indicate the magnitude of the change to the gap required to bring about the desired effect on FX exposure indicators. According to our main statistical model, increasing the FX-LC RR gap by 1 percentage point has the following statistically significant effects over the subsequent four quarters:
- A 0.1 percentage point reduction in the share of domestic bank liabilities that are in FX;
- A 1.62 percentage point reduction in the net FX position of domestic banks.
We do not find statistically significant effects on the share of FX-denominated domestic banks’ assets, or on total cross-border capital inflows to banks. This is broadly consistent with the direct story outlined earlier: the policy change has made FX-denominated funding more expensive, triggering a shift to local currency liabilities and a reduction in banks’ net FX position and exposure to currency risk.
On the other hand, we demonstrate a broader negative impact on cross border capital flows beyond the targeted FX flows to banks. In particular, increases in the FX-LC portfolio debt flows are associated with a significant reduction in portfolio debt flows.
Finally, we do not find evidence of domestic or international circumvention. We did not observe a statistically significant increase in debt growth of domestic non-financial corporations.
Overall, our research points to the usefulness of FX reserve requirements as a macroprudential policy tool for reducing risks associated with FX exposures, while being mindful of the broader impact on cross-border capital flows. These findings support the decision under the recently revised OECD Capital Movements Code to assess the use of these tools by Code’s adherents on a case-by-case basis, taking into account country-specific context and different reserve requirement design to ensure that their use is well targeted.
References
De Crescenzio, A., E. Lepers and Z. Fannon (2021), Assessing the effectiveness of currency-differentiated tools : The case of reserve requirements, (2021/01; OECD Working Papers on International Investment)
Ahnert, T., Forbes, K., Friedrich, C., and Reinhardt, D. (2020), Macroprudential FX regulations: Shifting the snowbanks of FX vulnerability?, Journal of Financial Economics
OECD (2019), Reserve Requirements: Current Use, Motivations and Practical Considerations, Technical Note by the OECD Secretariat,
De Crescenzio, A., Golin, M., & Ott, A.-C. (2015). Currency-based measures targeting banks- Balancing national regulation of risk and financial openness (2015/03; OECD Working Papers on International Investment)
Federico, P., C. Vegh and G. Vuletin (2014), Reserve Requirement Policy over the Business Cycle, National Bureau of Economic Research, Cambridge, MA,
On the level 