Monetary Policy Spillovers through Invoicing Currencies (Figures)

Figure 1: Average Responses to U.S. Monetary Policy Shocks 3 4 ● ● AUS MLT ● GRC 0 ● CHE ● CYP ● SVN ● ● DNKISL ● EUR ● EST ● LTU ● MLT ● JPN −4 ● ● ● GBR CAN GRC Interest Rate Response (%) Exchange Rate Response (%) ● NOR 2 ● ● 1 EST ISR● ● CAN ● ● CYP ● ● ● EUR● AUS GBR DNK 0 CHE ● NOR ● JPN ● LTU ● ● LVA KOR −1 LVA ● SVN ● ISR −8 ● ● KOR 0.00 0.25 0.50 ISL 0.75 Dollar Invoiced Imports / Consumption 0.25 0.50 0.75 Dollar Invoiced Imports / Consumption Notes: I estimate the following regression ∆ytj = αj + δt + γ j mF RB,t + jt . ∆ytj represents the change in country j’s dollar exchange rate or the change in country j’s one-year nominal interest rate on date t, αj is a country-level fixed effect and δt is a date fixed effect. γ j measures the average response of country j’s dependent to monetary policy shocks from the FRB, mF RB,t . In the left-hand panel, I plot the coefficients γ j from the regression of changes in nominal exchange rates against the country’s dollar invoiced imports, normalized by consumption. β = −6.476, R2 = 0.13. In the right-hand panel, I plot the coefficients γ j from the regression of changes in nominal interest rates against the country’s dollar invoiced imports normalized by consumption. β = 3.142, R2 = 0.23. The sample consists of monetary policy announcement days between Jan. 1999 through Mar. 2014, excluding the height of the financial crisis (Jul. 2008 - Jun. 2009). Figure 2: U.S. Monetary Policy Shock Comparison ● Three−month Eurodollar Futures Shocks 0.2 ● 0.1 ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ●●● ● ●● ●● ●●● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ●●● ● ●●● ● ●● ●● ●● ● 0.0 ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● −0.1 ● ● ● ● ● ● ● ● ● ● ● −0.2 ● ● −0.15 −0.10 −0.05 0.00 0.05 0.10 Nakamura and Steinsson (2015) Shocks Notes: The figure plots U.S. monetary policy shocks derived from changes in three-month Eurodollar futures on U.S. monetary policy announcement days against those from Nakamura and Steinsson (2015). The slope of the regression line is 1.000 and the intercept is -0.008. The correlation between the two measures of monetary policy shocks is 0.74. Table 1: Nominal Exchange Rate Response: Nakamura and Steinsson (2015) Shocks Dependent variable: Nominal Exchange Rates (1) M$n × ∆if,t (2) −6.981∗∗∗ (2.193) (3) (4) −6.368∗∗∗ −12.340∗∗∗ (3.424) (2.146) (1.589) 0.667 (0.411) 0.664 (0.413) −2.603 (10.859) 0.949 (0.910) −1.096 (10.037) −0.045 (0.037) 1,675 0.638 1,675 0.638 1,501 0.620 −2.144∗∗ (0.881) −8.846∗∗∗ (3.177) (Imports / Cons.) × ∆if,t (Dollar Import Share) × ∆if,t Distance ×∆if,t US Import Share ×∆if,t (NetDebt$,t−1 /Y ) × ∆if,t Observations Adjusted R2 (5) −6.669∗∗∗ 1,675 0.638 1,675 0.639 Notes: This table presents regression results from estimating equation (18). The dependent variable is changes in nominal exchange rates. M$n is the value country n’s dollar invoiced imports normalized by consumption. (Imports / Cons.) is country n’s total imports as share of its consumption and (Dollar Import Share) represents the share of country n’s imports that is invoiced in dollars. M$n = (Dollar Import Share) × (Imports / Cons.). Distance is the log of the populaiton weighted distance between country n and the United States. US Import Share is the value of each country’s imports from the U.S. normalized by consumption. (NetDebt$,t−1 /Y ) is the country’s dollar denominated net debt position. Standard errors in parentheses are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Table 2: Nominal Interest Rate Response: Nakamura and Steinsson (2015) Shocks Dependent variable: Nominal Interest Rates (1Y) (1) M$n × ∆if,t (2) 2.844∗∗∗ (0.824) (3) (4) (5) 2.916∗∗∗ 2.761∗∗∗ 1.915∗∗∗ (0.686) (0.846) (0.904) 0.241∗ (0.145) 0.376 (0.234) 1.151 (1.304) 0.473 (0.289) 2.604 (1.755) 0.007 (0.005) 1,366 0.157 1,366 0.158 1,280 0.175 1.002∗∗∗ (0.380) 1.969∗∗∗ (0.653) (Imports / Cons.) × ∆if,t (Dollar Import Share) × ∆if,t Distance ×∆if,t US Import Share ×∆if,t (N etDebt$,t−1 /Y ) × ∆if,t Observations Adjusted R2 1,366 0.156 1,366 0.153 Notes: This table presents regression results from estimating equation (18). The dependent variable is changes in one-year nominal interest rates. M$n is the value country n’s dollar invoiced imports normalized by consumption. (Imports / Cons.) is country n’s total imports as share of its consumption and (Dollar Import Share) represents the share of country n’s imports that is invoiced in dollars. M$n = (Dollar Import Share) × (Imports / Cons.). Distance is the log of the populaiton weighted distance between country n and the United States. US Import Share is the value of each country’s imports from the U.S. normalized by consumption. (NetDebt$,t−1 /Y ) is the country’s dollar denominated net debt position. Standard errors in parentheses are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Table 3: Nakamura and Steinsson (2015) Shocks, Controlling for Product Pricing Nominal Exchange Rate (1) (2) −5.251∗ (3.022) −5.101 (3.533) M$n × ∆if,t Diff. Share ×∆if,t Nominal Interest Rate (1Y) (3) 2.663∗∗∗ (0.800) 0.616 (0.388) −3.679 (11.414) Diff. Share ×M$n × ∆if,t 1,675 0.639 1,675 0.638 −0.110 (1.284) 3.626∗∗ (1.742) −3.775 (11.252) Observations Adjusted R2 (4) 1,366 0.157 1,366 0.159 Notes: Table 3 estimates equation (18) controlling for how products are priced according to Rauch (1999). The outcome variable in columns (1) and (2) is the nominal exchange rate. The outcome variable in columns (3) and (4) is the one year nominal interest rate. Diff. Share is the share of each country’s imports that are differentiated products (not traded on an organized exchange or reference priced). Standard errors in parentheses are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Table 4: Nakamura and Steinsson (2015) Shocks, Alternative Specifications (1) (2) Panel A: (3) (4) Nominal Exchange Rates M$n × ∆i$,t −6.652∗∗∗ (2.120) −4.978∗∗∗ (1.859) −9.741∗∗∗ (2.648) −5.850∗∗ (2.396) Observations Adjusted R2 1,201 0.601 1,236 0.644 1,675 0.638 1,675 0.620 Panel B: M$n × ∆i$,t Observations Adjusted R2 Nominal Interest Rates (1Y) ∗∗∗ 2.877 (0.816) 3.296∗∗∗ (0.916) 4.383∗∗∗ (1.356) 2.712∗∗∗ (0.561) 892 0.249 1,068 0.315 1,366 0.155 1,294 0.040 Notes: This table estimates equation (18) using various alternative specifications. The dependent variable in Panel A is the change in nominal exchange rate and the dependent variable in Panel B is the change in the one year nominal interest rate. Column (1) only uses observations monetary policy shocks prior to the height of the financial crisis, June 2007. Column (2) restricts the sample of countries to those that maintained a Fitch sovereign bond credit rating above BBB- throughout the entirety of the sample. Column (3) normalizes M$n by GDP rather than consumption. Column (4) measures changes outcome variables from one day prior to the monetary policy announcement day to three days after the announcement day. Standard errors are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Table 5: Foreign Monetary Policy Rate Responses 30 Day Response 60 Day Response (1) (2) M$n × ∆if,t −0.020 (0.067) −0.036 (0.064) Observations Adjusted R2 1,142 0.015 1,133 0.047 Notes: This table estimates equation (18) using responses in foreign central bank monetary policy rates to Federal Reserve monetary policy shocks as the dependent variable. I measure changes in foreign central bank policy rates in the 30 days and the 60 days after Federal Reserve monetary policy announcements. Standard errors are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Table 6: Pooled Regressions: Multiple Central Banks Nominal Exchange Rate (1) Mfn × ∆if,t (3) (4) −4.881∗∗ (2.164) −4.234 (4.080) −1.466 (4.709) 1.682∗∗∗ (0.571) 1.510∗∗ (0.696) 0.195 (0.946) 6,460 0.692 6,460 0.692 5,075 0.128 5,075 0.128 Mfn × ∆if,t × F RB Observations Adjusted R2 (2) Nominal Interest Rate (1Y) Notes: Table 6 presents regression results from estimating equation (18) using data from multiple central banks. ∆if,t is the measurement of the monetary policy shock from central bank f on date t. Mfn is the value country n’s imports invoiced in country f currency and normalized by consumption. Nominal exchange rates are denominated in the currency of the central bank that issues the monetary policy shock. Standard errors in parentheses are clustered by date. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Figure 3: Relative Change in Industrial Production 0.3 γX,−1 0.0 −0.3 −0.6 0 10 20 30 Horizon (months) Table 7: Relative Change in Industrial Prodcution τ (months) X$n × ∆iU S,t Date Fixed Effects Country Fixed Effects Observations Adj. R-sqr 6 −0.033 (0.103) Y Y 3,358 0.285 12 −0.121 (0.132) Y Y 3,231 0.554 18 −0.333∗∗ (0.141) Y Y 3,105 0.649 24 −0.439∗∗∗ (0.182) Y Y 2,974 0.681 30 −0.495∗∗∗ (0.167) Y Y 2,842 0.722 36 −0.411∗∗∗ (0.157) Y Y 2,719 0.746 Notes: X$n is the share of country n’s exports that invoiced in dollars. Standard errors in parentheses are clustered by month. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01 Figure 4: Relative Change in Industrial Production Controlling for Lags 0.5 γX,−1 0.0 −0.5 −1.0 −1.5 0 10 20 30 Horizon (months) Table 8: Relative Change in Industrial Production Controlling for Lags τ (months) X$n × ∆iU S,t Date Fixed Effects Country Fixed Effects Observations Adj. R-sqr 6 −0.103 (0.156) Y Y 1,680 0.267 12 −0.152 (0.251) Y Y 1,616 0.547 18 −0.498∗ (0.257) Y Y 1,553 0.648 24 −0.872∗∗ (0.370) Y Y 1,488 0.682 30 −0.865∗∗∗ (0.288) Y Y 1,422 0.726 36 −0.638∗∗ (0.276) Y Y 1,360 0.744 Notes: X$n is the share of country n’s exports that invoiced in dollars. Standard errors in parentheses are clustered by month. ∗ p<0.1; ∗∗ p<0.05; ∗∗∗ p<0.01