Discussion The Changing Relationship Between Commodity Prices and Prices of Other Assets with Global Market Integration by Barbara Rossi Domenico Giannone Université libre de Bruxelles, ECARES and CEPR IMF-TMB Conference Policy Responses to Commodity Price Movements Istanbul, April 212 1 / 32
Once upon a time... American Economic Association meetings, 29 2 / 32
Once upon a time... American Economic Association meetings, 29 2 / 32
Once upon a time... Discussion of Can exchange rates forecast commodity prices? Hélène Rey London Business School, CEPR, NBER 3 / 32
Once upon a time... Intriguing result Exchange rates of small commodity exporters have forecasting power for global commodity index. True in and out of sample. Reverse is not true, i.e. commodity prices do not seem to forecast exchange rates very well (and not out of sample). 4 / 32
Once upon a time... Conclusions Very interesting set of results Calls for extensions Stock market index has also some forecast ability in and out of sample. Seems a slightly noisier predictor than the exchange rate. 5 / 32
The Data 5.5 (log) Global Commodity Price 5 4.5 4 95 Q1 Q1 5 Q1 1 Q1 35 (log) NZ Equity Price 3 25 2 15 1 95 Q1 Q1 5 Q1 1 Q1 6 / 32
The Sample 5.5 (log) Global Commodity Price 5 4.5 4 95 Q1 Q1 5 Q1 1 Q1 35 (log) NZ Equity Price 3 25 2 15 1 95 Q1 Q1 5 Q1 1 Q1 7 / 32
Forecasting Global Commodity Prices (CP) Forecasting Global Commodity Prices (CP) using asset prices of small commodity producers (e.g. New Zealand (NZ)) - Exchange Rate (EXR) - Equity Prices (EP) This paper (EP) E t CP t+1 = α + β EP NZ t + ρ CP t Barbara s previous work ( EXR): E t CP t+1 = α + β EXR NZ t The Naive Benchmark: Random Walk E t CP t+1 = + ρ CP t 8 / 32
Forecasting Global Commodity Prices (CP) Forecasting Global Commodity Prices (CP) using asset prices of small commodity producers (e.g. New Zealand (NZ)) - Exchange Rate (EXR) - Equity Prices (EP) This paper (EP) E t CP t+1 = α + β EP NZ t + ρ CP t Barbara s previous work ( EXR): E t CP t+1 = α + β EXR NZ t The Naive Benchmark: Random Walk E t CP t+1 = + ρ CP t 8 / 32
Forecasting Global Commodity Prices (CP) Forecasting Global Commodity Prices (CP) using asset prices of small commodity producers (e.g. New Zealand (NZ)) - Exchange Rate (EXR) - Equity Prices (EP) This paper (EP) E t CP t+1 = α + β EP NZ t + ρ CP t Barbara s previous work ( EXR): E t CP t+1 = α + β EXR NZ t The Naive Benchmark: Random Walk E t CP t+1 = + ρ CP t 8 / 32
Forecasting Global Commodity Prices (CP) Forecasting Global Commodity Prices (CP) using asset prices of small commodity producers (e.g. New Zealand (NZ)) - Exchange Rate (EXR) - Equity Prices (EP) This paper (EP) E t CP t+1 = α + β EP NZ t + ρ CP t Barbara s previous work ( EXR): E t CP t+1 = α + β EXR NZ t The Naive Benchmark: Random Walk E t CP t+1 = + ρ CP t 8 / 32
The Input 1 Commodity Price Equity Price 5 5 1 95 Q1 Q1 5 Q1 1 Q1 9 / 32
The Output 15 1 Model Forecast Commodity Price Random Walk 5 5 4 Q1 6 Q1 8 Q1 1 Q1 1 / 32
The Output 15 1 Model Forecast Commodity Price Random Walk 5 5 4 Q1 6 Q1 8 Q1 1 Q1 11 / 32
The Evaluation: Quadratic Loss 15 1 Model Forecast Commodity Price Random Walk 5 5 4 Q1 6 Q1 8 Q1 1 Q1 12 1 8 6 4 2 Squared Errors Model Random Walk 4 Q1 6 Q1 8 Q1 1 Q1 12 / 32
The Evaluation: Fluctuation Test 15 1 Model Forecast Commodity Price Random Walk 5 5 4 Q1 6 Q1 8 Q1 1 Q1 12 1 8 6 4 2 Squared Errors Model Random Walk Smooth Mod. Smooth RW 4 Q1 6 Q1 8 Q1 1 Q1 13 / 32
The empirical finding [...] the appearance of the out-of-sample predictive ability of the equity market predictor [...] dated around mid-2. Since the mid-2s marked a large increase in investment in commodity markets,...... our empirical evidence suggests a decrease in market segmentation at approximately the same time,...... and therefore the possibility that shocks in equity markets might have started to spill-over onto commodity markets in that period. 14 / 32
The empirical finding [...] the appearance of the out-of-sample predictive ability of the equity market predictor [...] dated around mid-2. Since the mid-2s marked a large increase in investment in commodity markets,...... our empirical evidence suggests a decrease in market segmentation at approximately the same time,...... and therefore the possibility that shocks in equity markets might have started to spill-over onto commodity markets in that period. 14 / 32
The empirical finding [...] the appearance of the out-of-sample predictive ability of the equity market predictor [...] dated around mid-2. Since the mid-2s marked a large increase in investment in commodity markets,...... our empirical evidence suggests a decrease in market segmentation at approximately the same time,...... and therefore the possibility that shocks in equity markets might have started to spill-over onto commodity markets in that period. 14 / 32
The empirical finding [...] the appearance of the out-of-sample predictive ability of the equity market predictor [...] dated around mid-2. Since the mid-2s marked a large increase in investment in commodity markets,...... our empirical evidence suggests a decrease in market segmentation at approximately the same time,...... and therefore the possibility that shocks in equity markets might have started to spill-over onto commodity markets in that period. 14 / 32
Other Predictors 5.5 (log) Global Commodity Price (log) Exch. Rate.2 5.4 4.5.6.8 4 Q1 1 Q1 1 Q1 1 Q1 35 (log) NZ Equity Price 5 (log) Global IP 3 4.9 4.8 25 4.7 2 4.6 15 4.5 4.4 1 Q1 1 Q1 4.3 Q1 1 Q1 15 / 32
Other Predictors 3 Equity Price 2.5 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 16 / 32
Other Predictors 3 Equity Price Exch. Rate 2.5 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 17 / 32
Other Predictors 3 Equity Price Exch. Rate Global IP 2.5 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 18 / 32
Other Predictors 3 2.5 Equity Price Exch. Rate Global IP Pool 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 19 / 32
Summing up The appearance of predictive ability is not specific to the the equity market predictor!! Factors other that the decrease in market segmentation and spillovers from equity markets might have been at work! 2 / 32
Forecasting Commodity Prices at the Time of the Great Recession 5.5 (log) Global Commodity Price (log) Exch. Rate.2 5.4 4.5.6.8 4 Q1 1 Q1 1 Q1 1 Q1 35 (log) NZ Equity Price 5 (log) Global IP 3 4.9 4.8 25 4.7 2 4.6 15 4.5 4.4 1 Q1 1 Q1 4.3 Q1 1 Q1 21 / 32
Forecasting Commodity Prices at the Time of the Great Recession 5.5 (log) Global Commodity Price (log) Exch. Rate.2 5.4 4.5.6.8 4 Q1 1 Q1 1 Q1 1 Q1 35 (log) NZ Equity Price 5 (log) Global IP 3 4.9 4.8 25 4.7 2 4.6 15 4.5 4.4 1 Q1 1 Q1 4.3 Q1 1 Q1 22 / 32
Commodity and Equity Prices at the Time of the Great Recession 15 Commodity Price Equity Price 1 5 5 1 15 2 25 3 95 Q1 Q1 5 Q1 1 Q1 23 / 32
Forecasting Commodity Prices at the Time of the Great Recession 3 2.5 Equity Price Exch. Rate Global IP Pool 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 24 / 32
Forecasting Commodity Prices at the Time of the Great Recession 3 2.5 Equity Price Exch. Rate Global IP Pool 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 25 / 32
Forecasting Commodity Prices at the Time of the Great Recession 3 2.5 Equity Price Exch. Rate Global IP Pool 2 1.5 1.5 4 Q1 6 Q1 8 Q1 1 Q1 26 / 32
Can we really forecast commodity prices?? 27 / 32
An Alternative Look at the Data: Modeling Structural Change Model: Time-Varying Vector Autoregressions (TV-VAR) Cogley and Sargent, 21, 25, Primiceri, 26 y t = A,t + A 1,t y t 1 + ε t, ε t N(, Σ t ) (1) Parameters evolve according to Coefficients : θ t = θ t 1 + ω t, ω t N(, Ω) (2) Variances : log σ t = log σ t 1 + ξ t, ξ t N(, Ξ) (3) Autocovariance : φ i,t = φ i,t 1 + ψ i,t, ψ i,t N(, Ψ i ) (4) ψ i,t, ξ t, ω t, ε t all mutually uncorrelated at all leads and lags. Reliable Forecasting Tool: D Agostino, Gambetti and Giannone, 29. Flexible but parsimonious model of structural changes 28 / 32
An Alternative Look at the Data: Modeling Structural Change Model: Time-Varying Vector Autoregressions (TV-VAR) Cogley and Sargent, 21, 25, Primiceri, 26 y t = A,t + A 1,t y t 1 + ε t, ε t N(, Σ t ) (1) Parameters evolve according to Coefficients : θ t = θ t 1 + ω t, ω t N(, Ω) (2) Variances : log σ t = log σ t 1 + ξ t, ξ t N(, Ξ) (3) Autocovariance : φ i,t = φ i,t 1 + ψ i,t, ψ i,t N(, Ψ i ) (4) ψ i,t, ξ t, ω t, ε t all mutually uncorrelated at all leads and lags. Reliable Forecasting Tool: D Agostino, Gambetti and Giannone, 29. Flexible but parsimonious model of structural changes 28 / 32
An Alternative Look at the Data: Modeling Structural Change Setting the Prior Parameters evolve according to y t = A,t + A 1,t y t 1 + ε t, ε t N(, Σ t ) (5) θ t = θ t 1 + ω t, ω t N(, Ω) (6) Estimate the model by ML over the training sample: 1992-21 ( θ) θ = θ V (θ ) = V ( θ) Ω = V (θ t θ t 1 ) = V ( θ) λ 2 Quite loose prior to favor time variation: λ 2 = 1 1 Estimation: Gibbs Sampling on entire sample 29 / 32
An Alternative Look at the Data: Time Varying Coefficients.5 Commodity to Commodity.15 Equity to Commodity.4.3.2.1 Q1 Q1 5 Q1 1 Q1 15.1.5.5.1 Q1 Q1 5 Q1 1 Q1 15 1 Commodity to Equity.3 Equity to Equity.2.5.1.1.5 Q1 Q1 5 Q1 1 Q1 15.2 Q1 Q1 5 Q1 1 Q1 15 If anything, Commodity Prices are becoming Less Predictable 3 / 32
An Alternative Look at the Data: Stochastic Volatility 2 Variance of Commodity Price Innovations 15 1 5 Q1 2 Q1 4 Q1 6 Q1 8 Q1 1 Q1 12 6 Variance of Equity Price Innovations 5 4 3 2 1 Q1 2 Q1 4 Q1 6 Q1 8 Q1 1 Q1 12 Strong comovement in volatility: might be exploited in risk assessment 31 / 32
Conclusions Fascinating, relevant and innovative research agenda Vey interesting and well executed paper. I strongly advice you to read it!! 32 / 32