AdobeStock_297235820.jpeg
VoxEU Column

US monetary policy, international risk spillovers, and policy options

The question of how shifts in US monetary policy affect other nations is key to central bank policymaking around the world. This column combines a simple theoretical framework with macro data from 70+ countries and micro data from over a million non-financial firms in 43 of these countries to show that monetary policy spillovers from the US to the rest of the world operate through changes in risk premia, with emerging market economies most vulnerable. It also discusses the policy options available to countries to deal with the effects of these risk spillovers.

How do shifts in the US monetary policy affect other nations? This question is imperative to central bank policymaking around the world. Standard economic theory postulates that any external shock, such as changes in the US interest rates, can be smoothed out by the flexible exchange rates. Rey (2013) argues that exchange rate flexibility does not matter for the spillover effects of the US monetary policy in a world characterised by the global financial cycle. She argues that since US monetary policy affects leverage of global banks and credit growth in the international financial system, financial conditions are transmitted from the US to the rest of the world through gross capital flows, irrespective of other countries’ exchange rate regimes. Subsequent work by Bruno and Shin (2014, 2015) and Ziang et al. (2019) shows that extensive dollar borrowing in the world and the associated currency mismatches on global banks’ balance sheets are important drivers of the global financial cycle. Work by di Giovanni et al. (2018) shows a strong transmission of the global financial cycle to domestic credit creation in emerging market economies through lower borrowing costs that fuel a large expansion not only in foreign but also in local currency borrowing.

In recent work (Kalemli-Ozcan 2019), I show that monetary policy spillovers from the US to the rest of the world operate through changes in risk premia. The effects of the shifts in US monetary policy on countries’ financial and real outcomes materialise through what I call ‘risk spillovers’ in a world characterised by a high degree of trade and financial integration. 

 I ask three questions.

  • First, which countries are more vulnerable to spillovers created by the shifts in US monetary policy and why? You might not be surprised that the answer to this question is emerging market economies. However, the reason is not that most of these countries have volatile growth and high current account deficits, but rather that capital flows in and out of emerging markets are more risk-sensitive conditional on their fundamentals. 
  • The second question I ask is about the implications of monetary policy divergence on exchange rates and economic activity. I find that that exchange rates and interest rates are disconnected and the degree of this disconnect can be explained by fluctuations in risk premia. Emerging markets which are exposed more to risk shocks will suffer more from negative output effects. 
  • The third question is whether exchange rate regimes matter in mitigating the effect of spillovers. I find that they do since risk shocks can be absorbed better in emerging market economies with free floats who do not suffer from negative output effects as a result.

To arrive at these answers, I use a simple theoretical framework combined with empirical evidence based on macro data from 70+ countries and micro data from over a million non-financial firms in 43 of these countries.

I start by showing extensive dispersion in the policy rate differentials (vis-a-vis the US) for emerging market economies relative to advanced countries (see Figure 1). This dispersion captures differential exposure of these countries to investors’ risk sentiments (proxied by VIX) and their monetary policy responses to these sentiments. The positive relationship between policy rate differentials and the VIX implies that the relationship between monetary policy divergence and capital flows will also change over time depending on risk shocks.

Figure 1 Monetary policy divergence 

a)         Emerging market economies

b)         Advanced economies

Of course, an important common driver of both capital flows and monetary policy divergence are fundamentals. Conditioning on fundamentals, captured by growth and inflation differentials among countries, the conditional correlation between policy rate differentials and capital flows is negative in emerging market economies and zero in advanced economies. Only after I control for the role of global and country-specific risk, through adding the VIX and the EMBI indices, does the conditional correlation between policy rate differentials and capital flows become positive in both set of countries. This is the ‘search for yield’ result. These results highlight the importance of recognising the sensitivity of capital flows in and out of emerging market economies not only for global but also for country-specific risk shocks.

Based on this evidence, I add ‘risk’ to a basic two-country model of spillovers through country risk premium. This premium has two parts: global risk and country-specific risk. Under the global financial cycle, global risk aversion is a function of US monetary policy. In addition, I assume that there is a wedge between the domestic policy rate and the interest rates that govern saving/borrowing behaviour. The wedge is given by the country risk premium. I also assume that the same wedge enters into the uncovered interest parity (UIP) condition. Higher country risk depreciates the domestic currency, creating a disconnect between exchange rate adjustment and interest rate differentials, leading to UIP deviations. These two mechanisms can deliver a powerful force such that when the US tightens, global risk sentiments and risk aversion rise, leading to a rise in the country-risk premium in the small open economy, which in turn depreciates the exchange rate and implies higher interest rates in the small open economy. As a result, contractionary US policy has a contractionary effect on output of the small open economy. The result is symmetric: when the US loosens, output in the small open economy expands.

The key question is whether these two mechanisms have any bite in the data? I show that they do. I regress short rates in a given economy on policy rates. I use 3-month and 12-month deposit rates and short-term lending rates. In an economy with perfect transmission of monetary policy, we expect to find a coefficient close to 1. And in fact, this is the case for advanced economies. For emerging market economies, the estimated coefficients are between 0.5 and 0.8, hence there is a short rate disconnect. Once I account for the role of global and country-specific risk on short rates, adding the VIX and EMBI indices to the regression, the estimated coefficients on the policy rate for emerging market economies for all types of short-term interest rates rise to 0.9, similar to advanced countries. This evidence of a short rate disconnect implies that pass-through of monetary policy rates into domestic saving and borrowing rates is less than full in emerging market economies. And the degree of pass-through of monetary policy rates to short rates in emerging market economies is a function of global and country-specific risk. 

An important implication of incomplete monetary policy transmission is the direct effect of capital flows on domestic lending spreads. As I show, capital inflows decrease the lending spreads and capital outflows increase the lending spreads. The negative effect of capital flows on lending spreads is present both for emerging market economies and advanced economies. The difference between these two set of countries is the fact that monetary policy has limited effectiveness in dealing with these effects in emerging markets, whereas in advanced countries monetary policy can mitigate these effects. Another interpretation of these results is the response of domestic monetary policy to capital outflows by easing the policy. Since then short-term government borrowing rates will go down, leading to an increase in the spreads, delivering a negative relationship between capital flows and spreads. If the estimated coefficients are due to this type of endogenous response of monetary policy and the policy is fully effective, then the effect should still disappear upon controlling for monetary policy, which is not the case for emerging markets.

To further show this point, I use exogenous shocks to US monetary policy. Building on the US monetary policy transmission literature, I use high-frequency identification where US monetary policy shocks are identified from changes in fed funds futures in a 30-minute window around the policy announcement.  The local projections for 12-month government bond rate differentials show that these differentials increase in emerging market economies as a response to a 1 percentage point contractionary US monetary policy shock, whereas as a response to the same contractionary US shock, the bond rate differentials go down in advanced economies (see Figure 2). The effects are large: a 1 percentage point rise in US rates imply 2.3 percentage point rise in emerging market rate differentials and a 0.5 percentage point decline in advanced country rate differentials after three quarters. I control for domestic monetary policy responses and country fundamentals in these impulse responses. These results suggest that if changes in US monetary policy lead to changes in investors’ risk sentiments, interest rate differentials in emerging market economies pick up the associated risk premia. The country-specific risk increases in emerging markets, regardless of the policy response in these countries when US tightens. The results will be symmetric for an expansionary monetary policy shock in the US.

Figure 2 Responses of 12-month government bond rate differentials

a)         Emerging market economies

b)         Advanced economies

In terms of the evidence for the risk-adjusted UIP condition, I show that UIP deviations co-move with investors risk perceptions, proxied by the VIX index in both emerging markets and advanced economies. However, there are stark differences in the drivers of this correlation between these countries. To show this, I decompose the deviations into in an interest rate differential term and an exchange rate adjustment term. If there are no UIP deviations, then these two terms offset each other so that UIP holds. If a country has a positive interest rate differential with the US, its currency is expected to depreciate so there will be no deviations. If interest rate differentials are higher than the expected depreciation, then there are positive UIP deviations, meaning the country in question is expected to provide excess returns to global investors. The UIP deviations in emerging market economies comove almost one-to-one with interest rate differentials. This situation is in stark contrast to advanced economies, where the exchange rate adjustment term now moves one-to-one with UIP deviations (see Figure 3, taken from Kalemli-Ozcan and Varela 2019). These facts imply that, when there is a surprise contractionary shock to US monetary policy, the UIP deviations increase in emerging markets economies, capturing the higher risk premia in interest rate differentials, whereas in advanced economies they do not respond as exchange rates adjust (see Figure 4). 

Figure 3 Sources of UIP deviations

a)         Emerging market economies

b)         Advanced economies

Figure 4 Responses of UIP deviations

a)         Emerging market economies

b)         Advanced economies

  

Policy implications

These results have two key policy implications. The first implication is that flexible exchange rates can help to smooth out real effects of monetary policy spillovers by smoothing risk shocks. I show that the positive correlation between risk sentiments and monetary policy divergence practically becomes zero for free-floating emerging market economies and the strong negative effect of VIX on growth in managed floats disappears in free floats. This means that by managing the exchange rate, there is a potential danger of turning nominal exchange rate volatility into real output volatility. The key intuition behind this result is the fact that UIP deviations are correlated with risk sentiments, and they will be affected from domestic and foreign monetary policy. A country that tries to limit exchange rate volatility has to change the interest rate by a large amount to achieve this goal, compared to an environment where UIP holds.

Thus, when monetary policy spillovers work through changes in risk perceptions, then exchange rate flexibility might help a lot in combatting such risk spillovers. The reason why the case for flexible exchange rates is stronger in a world of international risk spillovers is different from the standard Mundell-Fleming channel. Under international risk spillovers, exchange rate adjustment can smooth out shocks to risk sentiments. Trying to limit exchange rate volatility can be counterproductive as this policy response requires a large change in domestic interest rates, turning a nominal exchange rate volatility into real output volatility in terms of lower GDP growth.

In fact, using monetary policy to fight a depreciating exchange rate or prevent an appreciation to involve in a currency war may ultimately lead to self-inflicted wounds. The standard Mundell-Fleming channel – that is, with an expansionary monetary policy US can depreciate its currency and gain in a currency war against emerging market economies – has been at the centre of the ‘currency wars’ debate in the aftermath of the 2008 crisis that depreciated the US dollar. Brazil’s finance minister, Guido Mantega, accused the US of waging a currency war on other countries in September 2010. In June 2019, US president Donald Trump similarly complained to ECB President Mario Draghi that the accommodative policies of the ECB had depreciated the euro, which made it unfairly easier for Europeans to compete against the US. As shown by Gopinath’s (2016) Jackson Hole paper, we should not give too much weight to the view that easing monetary policy can weaken a country’s currency enough to bring a lasting improvement in its trade balance. This is because the standard Mundell-Fleming channel of expenditure switching effects work mostly via imports, and not via exports, if trade is invoiced in a dominant currency (the dollar).   

In a world where there is a growing concern that the global economy may be both slowing and becoming more volatile as a result of trade wards, countries may be tempted to use their central banks to influence the value of their currencies and gain a trade advantage. My work suggests that any effort to use monetary policy to buffer the currency may be a losing battle. US policy moves shift the perceived risks of investing in other countries, making local monetary policy less effective with the changes needed to influence currency markets made all the larger as a result. 

Why, then, do countries manage their exchange rate? One of the key reasons for exchange rate management, or fear of floating, is the extent of foreign currency debt in an economy. This is of course not the only reason, but too much exchange rate volatility can certainly hurt an economy that has too much foreign currency debt. My findings show that countries that manage their exchange rates can still be vulnerable to monetary policy spillovers if these are risk spillovers. I show that risk spillovers lead to boom-bust cycles in firm leverage regardless of the extent of foreign currency debt in a given economy. Exchange rate appreciation, on the other hand, is associated with higher leverage only in countries with high levels of foreign currency debt.  In addition, by managing the exchange rate, countries might make the foreign currency debt problem worse as the amount of foreign currency debt in an economy changes with risk shocks. For a representative emerging market economy with managed float, I show that the extent of foreign currency debt might be endogenous to fluctuations in risk sentiments (see Figure 5). The strong positive correlation between them is clear. Recall that UIP deviations comove positively with the VIX index and country-specific risk. When there are risk-off shocks, the risk premium on local currency is higher, so the share of local currency debt is lower and the share of foreign currency debt is higher.

Figure 5 

What policy options are available to countries to deal with the effects of risk spillovers related to changes in US monetary policy? Countries can act on the transmission channel cyclically by limiting credit growth and leverage during booms and doing the reverse during downturns. This can be achieved by the use of macroprudential policies. In the case of emerging market economies, the policies that limit unhedged foreign currency-denominated liabilities not only in the financial sector but also in the non-financial corporate sector must be a priority. The rationale for these policies is to provide insulation from spillovers that arise from balance sheet effects of exchange rate fluctuations with large levels of unhedged foreign currency denominated debt. It is better to deal with this debt directly, with the help of macroprudential polices, rather than trying to use a relatively blunt instrument like monetary policy. 

However, dealing with excessive credit growth and foreign currency-denominated debt may not be enough. A significant component of international risk spillovers for emerging market economies is related to country-specific risk. A long-term objective that would act on the transmission channel structurally entails reducing this inherent country risk. In order to achieve higher GDP growth and mitigate volatility related to monetary policy spillovers, countries need to decrease the risk-sensitivity of capital flows through reducing inherent country risk by improving institutional quality. A reform agenda aimed at improving transparency, fighting corruption, and protecting institutional integrity with an emphasis on central bank independence will be beneficial in terms of attracting long-term, stable capital flows. These policies would reduce the sensitivity of capital flows to changes in the US policy, and hence the associated risk sentiments. 

My findings do not imply emerging markets will always be more vulnerable to monetary policy spillovers stemming from shifts in US policy. Monetary policy actions of any country have the potential to spill over to other countries as long as international investors’ risk perceptions change with changes in monetary policy. This is why international risk spillovers present a serious challenge for monetary policymaking across the world.

References 

Bruno, V and H S Shin (2014), “Cross-Border Banking and Global Liquidity”, The Review of Economic Studies 82(2): 535–564. 

Bruno, V and H S Shin (2015), “Capital Flows and the Risk-taking Channel of Monetary Policy”, Journal of Monetary Economics 71: 119 – 132.

di Giovanni, J, S Kalemli-Ozcan, M F Ulu, and Y S Baskaya (2018), “International Spillovers and Local Credit Cycles”, NBER Working Paper 23149.

Gopinath, G (2016), “The International Price System”, Jackson Hole Symposium Proceedings.

Jiang, Z, A Krishnamurthy, and H N Lustig (2019), “Dollar Safety and the Global Financial Cycle”, NBER Conference Paper f123919.

Kalemli-Ozcan, S (2019), “US Monetary Policy and International Risk Spillovers”, Jackson Hole Symposium Proceedings.

Kalemli-Ozcan, S and L Varela (2019), “Exchange Rate and Interest Rate Disconnect: The Role of Capital Flows and Risk Premia”.

Rey, H (2013), “Dilemma not Trilemma: The Global Financial Cycle and Monetary Policy Independence”, Jackson Hole Symposium Proceedings.

4,199 Reads