Accessible Version
Monetary Policy Stance and Commodity Cycles in Emerging Economies, Accessible Data
Figure 1. Global factors of commodity prices and sovereign spreads
The figure displays two unobserved common factors extracted from two Dynamic Factor Models (DFMs) estimated over the period 2002Q1–2025Q4 using data on commodity export prices and sovereign spreads for a sample of 13 emerging economies. The horizontal axis (X-axis) spans from 2002Q1 to 2026Q1 in quarterly frequency, while the vertical axis (Y-axis) measures the estimated factors in standardized units (standard deviations from the mean), ranging from approximately −6 to +6. The blue solid line tracks the common factor extracted from country-specific commodity export price indices, while the red dashed line tracks the common factor extracted from sovereign spreads (proxied by the ICE BofA US Dollar Sovereign Index). Both factors follow AR(2) processes.
The figure reveals a strong negative co-movement between the commodity price factor and the sovereign spread factor throughout the sample period, with an overall correlation of −0.35. Several key episodes stand out. During the commodity supercycle of the early-to-mid 2000s, the commodity price factor rises substantially while the sovereign spread factor declines markedly, consistent with the compression of borrowing costs during periods of elevated commodity revenues. Around the Global Financial Crisis (2008–2009), the commodity price factor spikes sharply before collapsing, while the sovereign spread factor rises abruptly, reflecting the sudden widening of emerging market borrowing conditions. The recovery period of 2010–2011 shows a partial reversal, with commodity prices recovering and spreads declining. The broad commodity price decline of 2015–2016 is also visible, associated with a renewed increase in sovereign spreads. Finally, the COVID-19 pandemic of 2020 is captured by a sharp joint movement in both factors, followed by a commodity price surge in 2021–2022 and a gradual normalization thereafter. The figure provides empirical motivation for the inclusion of the country-risk channel in the model and is consistent with evidence in Bastourre et al. (2012) and Fernández et al. (2018).
Figure 2. Calibration of income and country risk effects
The figure is a scatter plot displaying the estimated income effect and country risk effect for the 13 emerging economies included in the calibration exercise. Each dot represents a country. The horizontal axis (X-axis) measures the income effect coefficient ($$b_5$$), which captures the elasticity of the output gap with respect to the commodity export price index gap, ranging from approximately 0.00 to 0.20. The vertical axis (Y-axis) measures the country risk semi-elasticity ($$\phi_q$$), which captures the semi-elasticity of the sovereign spread with respect to the commodity export price index gap, ranging from approximately −0.18 to 0.02. Numerical tick marks are displayed on the right-hand side secondary axis for the Y-axis and on the bottom for the X-axis, both with increments of 0.02.
The 13 countries in the sample are Peru, Ecuador, Panama, Bolivia, Trinidad and Tobago, Kazakhstan, Morocco, Brazil, Colombia, Chile, South Africa, Costa Rica and Indonesia. Two dashed lines — one horizontal and one vertical — mark the median values of each coefficient across the 13 countries: the median income effect ($$b_5$$) is 0.079 and the median country risk semi-elasticity ($$\phi_q$$), is −0.029. These median values are used as the baseline calibration in the model. The figure shows that many country-specific estimates of the country risk effect are positive, consistent with higher commodity prices generating expansionary demand effects. Similarly, most country-specific estimates of the country-risk effect are negative, consistent with higher commodity prices compressing sovereign spreads. The scatter plot provides direct empirical support for the two channels incorporated in the model and illustrates the cross-country heterogeneity in both effects.
Figure 3. Response to a one-standard deviation increase in commodity prices
The figure displays the impulse response functions (IRFs) of 12 model variables to a one-standard-deviation shock to the commodity export price index gap under the baseline calibration ($$\phi_q$$ = 0.029). The shock corresponds to a 12 percent deviation of commodity prices from their trend, which equals the median standard deviation of the commodity price index gap across the 13 countries in the calibration sample. The horizontal axis (X-axis) spans from 0Q1 to 6Q1 (approximately 6 years or 24 quarters) and the vertical axis (Y-axis) measures the response of each variable in its natural unit (percentage or percentage points per annum, as indicated in each panel title). The figure contains 12 panels arranged in a 4×3 grid, each displaying the response of one model variable. From left to right and top to bottom, the panels show: the commodity price index gap (%), the output gap (%), the real monetary conditions index (in % per annum), the real interest rate gap (%), the real exchange rate gap (%), the nominal interest rate (in % per annum), CPI inflation QoQ (in % per annum), nominal exchange rate depreciation QoQ (in % per annum), the natural real interest rate (in % per annum), the real interest rate (in % per annum), the neutral interest rate (in % per annum), and the monetary policy stance (in % per annum). A single black solid line tracks the response of each variable under the baseline calibration.
The results show that a positive commodity price shock generates an expansion of the output gap on impact, which gradually closes over the subsequent quarters. Inflation initially rises due to above-normal demand but subsequently falls below its long-run level as the real exchange rate appreciates — a channel reinforced by the compression of the country risk premium. The natural real interest rate declines by approximately 50 basis points in the short run, driven by the fall in the country risk premium through the long-run UIP condition. Since expected inflation remains relatively stable, the neutral nominal interest rate falls by a similar magnitude. The central bank responds by lowering the policy rate by approximately 20 basis points within four quarters of the shock, despite the positive output gap, reflecting the dominant downward pull of the declining neutral rate. Overall, the monetary policy stance turns more restrictive, as the policy rate cut is insufficient to offset the sharp decline in the neutral nominal interest rate.
Figure 4. Response to a one-standard deviation increase in commodity prices: baseline vs. counterfactual calibration
The figure compares the impulse response functions (IRFs) of 12 model variables to a one-standard-deviation shock to the commodity export price index gap under two calibrations: the baseline specification ($$\phi_q$$= 0.029, black solid line) and a counterfactual specification in which the country risk channel is shut down ($$\phi_q$$ = 0, red dashed line). The layout, axes, and panel structure are identical to Figure 3. A legend in the upper-left panel identifies the two lines. The shock magnitude, horizon, and variable selection are identical to Figure 3.
The comparison highlights the role of the country risk channel in shaping the model's dynamics. The response of the output gap is identical across both specifications, since both are subject to the same income effect from higher commodity prices and $$b_5$$ is unchanged. However, because higher commodity prices no longer compress the country risk premium in the counterfactual, the nominal exchange rate appreciates by approximately 0.7 percentage points less on impact. This weaker appreciation attenuates the disinflationary pressures associated with the real appreciation of the currency, causing inflation to rise by approximately 0.10 percentage points more on impact in the counterfactual. Crucially, the natural rate of interest no longer responds to the shock in the counterfactual, and the neutral nominal interest rate remains unchanged. As a result, the challenging trade-off faced by the central bank in the baseline — where the decline in the neutral nominal interest rate coincides with an increase in the output gap — is absent. The central bank instead raises the policy rate by approximately 30 basis points, consistent with the Taylor rule prescription. The monetary policy stance is less restrictive than under the baseline calibration, and the policy rate takes more than a year to return to its pre-shock level.
Figure 5. Response to a one-standard deviation increase in commodity prices: baseline vs. alternative calibrations
The figure compares the impulse response functions (IRFs) of 12 model variables to a one-standard-deviation shock to the commodity export price index gap under three alternative calibrations of the country risk semi-elasticity ($$\phi_q$$): the baseline value of 0.029 (black solid line), the estimate from Shousha (2016) of 0.014 (red dashed line), and the estimate from Fernández et al. (2018) of 0.008 (blue dotted line). The layout, axes, panel structure, shock magnitude, and horizon are identical to Figures 3 and 4. A legend in the upper-left panel identifies the three lines.
The figure illustrates how the calibration of the country risk channel affects the quantitative implications of the model. Relative to the baseline, inflation rises by approximately 0.05 and 0.08 percentage points more in the short run under the calibrations of Shousha (2016) and Fernández et al. (2018), respectively, as a weaker country risk channel implies a smaller disinflationary contribution from real exchange rate appreciation. The implications for the policy rate are also quantitatively important: under the Shousha (2016) and Fernández et al. (2018) calibrations, the Taylor rule prescribes policy rate increases of approximately 13 and 21 basis points, respectively, compared with a decline of 20 basis points under the baseline. The monetary policy stance becomes progressively less restrictive as the sensitivity of the risk premium to commodity prices declines in absolute value: the policy rate gap rises to approximately 30 and 27 basis points under the alternative calibrations, compared with approximately 42 basis points in the baseline. The natural real interest rate and the neutral nominal interest rate show smaller declines under the alternative calibrations, consistent with a weaker country risk channel. Overall, the figure underscores the quantitative importance of the calibration of the commodity-spread semi-elasticity for the conduct of monetary policy.