This paper explores the impact of green sentiment in US media on financial markets. Using textual... more This paper explores the impact of green sentiment in US media on financial markets. Using textual analysis with a dictionary-based approach, we retrieve several scores of attention, tonality and uncertainty in the coverage of environmental news of four major US newspapers. We consider various weighting schemes to account for the visibility and relevance of the text sources and several sets of newspapers to measure the possible impact of their editorial line. Our results establish that greater attention to environmental news in US media reduced the excess returns of carbon-intensive stocks and increased their volatility over the last decade, especially when the coverage was negative or uncertain. The opposite result holds for the most virtuous green assets. Restricting the corpus of texts to conservative newspapers mitigates the impact of the coverage. Overall, our findings illustrate how rising environmental concerns lead investors to shift their asset allocation.
This paper introduces a Markov-switching model in which transition probabilities depend on higher... more This paper introduces a Markov-switching model in which transition probabilities depend on higher frequency indicators and their lags through polynomial weight-ing schemes. The MSV-MIDAS model is estimated via maximum likelihood (ML) methods. The estimation relies on a slightly modified version of Hamilton's recursive filter. We use Monte Carlo simulations to assess the robustness of the estimation procedure and related test statistics. The results show that ML provides accurate estimates, but they suggest some caution in interpreting the tests of the parameters involved in the transition probabilities. We apply this new model to the detection and forecasting of business cycle turning points in the United States. We properly detect recessions by exploiting the link between GDP growth and higher frequency variables from financial and energy markets. The spread term is a particularly useful indicator to predict recessions in the United States. The empirical evidence also supports the use of functional polynomial weights in the MIDAS specification of the transition probabilities.
This paper merges two specifications recently developed in the forecasting literature: the MS-MID... more This paper merges two specifications recently developed in the forecasting literature: the MS-MIDAS model (Guérin and Marcellino, 2013) and the factor-MIDAS model (Marcellino and Schumacher, 2010). The MS-factor MIDAS model that we introduce incorporates the information provided by a large dataset consisting of mixed frequency variables and captures regime-switching behaviors. Monte Carlo simulations show that this specification tracks the dynamics of the process and predicts the regime switches successfully, both in-sample and out-of-sample. We apply this model to US data from 1959 to 2010 and properly detect recessions by exploiting the link between GDP growth and higher frequency financial variables.
OECD Journal: Journal of Business Cycle Measurement and Analysis, 2014
In recent years, central banks and international organisations have been making ever greater use ... more In recent years, central banks and international organisations have been making ever greater use of factor models to forecast macroeconomic variables. We examine the performance of these models in forecasting French GDP growth over short horizons. The factors are extracted from a large data set of around one hundred variables including survey balances and real, financial, and international variables. An out-of-sample pseudo real-time evaluation over the past decade shows that factor models provide a gain in accuracy relative to the usual benchmarks. However, the forecasts remain inaccurate before the start of the quarter. We also show that the inclusion of international and financial variables can improve forecasts at the longest horizons.
This paper explores the forecasting abilities of Markov-Switching models. Although MS models gene... more This paper explores the forecasting abilities of Markov-Switching models. Although MS models generally display a superior in-sample fit relative to linear models, the gain in prediction remains small. We confirm this result using simulated data for a wide range of specifications by applying several tests of forecast accuracy and encompassing robust to nested models. In order to explain this poor performance, we use a forecasting error decomposition. We identify four components and derive their analytical expressions in different MS specifications. The relative contribution of each source is assessed through Monte Carlo simulations. We find that the main source of error is due to the misclassification of future regimes.
In recent years, factor models have received increasing attention from both econometricians and p... more In recent years, factor models have received increasing attention from both econometricians and practitioners in the forecasting of macroeconomic variables. In this context, Bai and Ng (2008) find an improvement in selecting indicators according to the forecast variable prior to factor estimation (targeted predictors). In particular, they propose using the LARS-EN algorithm to remove irrelevant predictors. In this paper, we adapt the Bai and Ng procedure to a setup in which data releases are delayed and staggered. In the pre-selection step, we replace actual data with estimates obtained on the basis of past information, where the structure of the available information replicates the one a forecaster would face in real time. We estimate on the reduced dataset the dynamic factor model of Giannone, Reichlin and Small (2008) and Doz, Giannone and Reichlin (2011), which is particularly suitable for the very short-term forecast of GDP. A pseudo real-time evaluation on French data shows the potential of our approach.
The French wholesale market is set to expand in the next few years under European pressure and na... more The French wholesale market is set to expand in the next few years under European pressure and national decisions. In this paper, we assess the forecasting ability of several classes of time series models for electricity wholesale spot prices at a day-ahead horizon in France. Electricity spot prices display a strong seasonal pattern, particularly in France given the high share of electric heating in housing during winter time. To deal with this pattern, we implement a double temporal segmentation of the data. For each trading period and season, we use a large number of specifications based on market fundamentals: linear regressions, Markov-switching models, threshold models with a smooth transition. An extensive evaluation on French data shows that modeling each season independently leads to better results. Among non-linear models, MS models designed to capture the sudden and fast-reverting spikes in the price dynamics yield more accurate forecasts. Finally, pooling forecasts gives more reliable results.
This paper investigates the relationship between electricity demand and temperature in the Europe... more This paper investigates the relationship between electricity demand and temperature in the European Union. We address this issue by means of a panel threshold regression model on 15 European countries over the last two decades. Our results confirm the non linearity of the link between electricity consumption and temperature found in more limited geographical areas in previous studies. By distinguishing between North and South countries, we also find that this non linear pattern is more pronounced in the warm countries. At last, rolling regressions highlight the significant impact of climate change on electricity use in Europe.
This paper explores the existence of a bounce-back effect in inventory investment using the Europ... more This paper explores the existence of a bounce-back effect in inventory investment using the European Commission opinion survey on stocks of finished products in manufacturing and retail trade sectors for France, Germany and a European aggregate, from 1985q1 to 2011q4. Our empirical findings support the existence of a high recovery episode for inventory investment, during the quarters immediately following the recessions: it occurs later and lasts longer in manufacturing than in retail trade sector. Since a third phase of rapid recovery has not been found in final sales data so far, the rebound in inventories could in turn explain the GDP growth bounce-back pointed out in previous empirical studies. This calls for a careful modeling of the inventory investment behavior in any sensible theoretical explanation of aggregate business cycles.
We collect data from 29 separate papers estimating the equilibrium level and possible undervaluat... more We collect data from 29 separate papers estimating the equilibrium level and possible undervaluation of the Chinese currency, the renminbi. These papers yield a total of 97 individual observations on misalignment, which we analyse with the help of meta-analysis. We find that the vast majority of observations point to renminbi undervaluation in recent years and that the undervaluation is more pronounced when the US dollar exchange rate is used instead of the real effective exchange rate. We find several characteristics of papers and authors that clearly seem to influence the reported misalignments. For example, when the author is affiliated with an investment bank, the reported misalignment is smaller. Using time-series techniques also results in lower estimates of undervaluation. On the other hand, refereed journals seemingly are inclined to publish papers that report larger misalignments. Results caution against trusting too much in any one study concerning renminbi undervaluation.
Les modèles à facteurs sont de plus en plus utilisés pour la prévision de court terme du PIB par ... more Les modèles à facteurs sont de plus en plus utilisés pour la prévision de court terme du PIB par les banques centrales et les grands organismes internationaux. Ils semblent en revanche un peu moins utilisés en France. Cet article propose une application de ces techniques à la prévision du taux de croissance trimestriel du PIB français à très court terme. Nous utilisons une base constituée d’une centaine de variables parmi lesquelles des variables d’enquêtes, des indicateurs réels, des variables monétaires et financières et des indicateurs sur l’environnement international. Une évaluation hors échantillon montre que la qualité des prévisions issues des modèles à facteurs est satisfaisante, même si les prévisions restent fragiles lorsque l’horizon de prévision est éloigné.
Cet article développe des étalonnages du taux de croissance du PIB français destinés à produire d... more Cet article développe des étalonnages du taux de croissance du PIB français destinés à produire des prévisions de très court terme de l’activité. Ils sont construits exclusivement à partir de données d’enquête de l’Insee, dans l’industrie mais également dans les services et le bâtiment. Nous examinons deux stratégies de réduction de l’information, l’une fondée sur l’algorithme de sélection automatique GETS par blocs de Hendry et Krolzig (2005), l’autre sur la méthode de combinaison popularisée par Stock et Watson (2004). Ces deux méthodes sont évaluées hors échantillon au travers de régressions récursives et roulantes. Nous montrons la supériorité des étalonnages construits avec GETS et l’intérêt de considérer d’autres enquêtes que celle dans l’industrie dans les stratégies de modélisation et de prévision.
Cet article étudie l’interaction entre agents chartistes et fondamentalistes sur le marché des ch... more Cet article étudie l’interaction entre agents chartistes et fondamentalistes sur le marché des changes. A cette fin, nous étendons le modèle à changements de régimes markoviens de Vigfusson afin de pouvoir en tester les implications. La stratégie de tests obtenue est ensuite appliquée aux taux de change canadien, japonais et allemand. Nos résultats valident les conclusions de Vigfusson. Les périodes calmes sont associées à l’action des agents chartistes, tandis que les périodes de volatilité accrue des changes correspondent à l’activité rééquilibrante des agents fondamentalistes.
This paper explores the impact of green sentiment in US media on financial markets. Using textual... more This paper explores the impact of green sentiment in US media on financial markets. Using textual analysis with a dictionary-based approach, we retrieve several scores of attention, tonality and uncertainty in the coverage of environmental news of four major US newspapers. We consider various weighting schemes to account for the visibility and relevance of the text sources and several sets of newspapers to measure the possible impact of their editorial line. Our results establish that greater attention to environmental news in US media reduced the excess returns of carbon-intensive stocks and increased their volatility over the last decade, especially when the coverage was negative or uncertain. The opposite result holds for the most virtuous green assets. Restricting the corpus of texts to conservative newspapers mitigates the impact of the coverage. Overall, our findings illustrate how rising environmental concerns lead investors to shift their asset allocation.
This paper introduces a Markov-switching model in which transition probabilities depend on higher... more This paper introduces a Markov-switching model in which transition probabilities depend on higher frequency indicators and their lags through polynomial weight-ing schemes. The MSV-MIDAS model is estimated via maximum likelihood (ML) methods. The estimation relies on a slightly modified version of Hamilton's recursive filter. We use Monte Carlo simulations to assess the robustness of the estimation procedure and related test statistics. The results show that ML provides accurate estimates, but they suggest some caution in interpreting the tests of the parameters involved in the transition probabilities. We apply this new model to the detection and forecasting of business cycle turning points in the United States. We properly detect recessions by exploiting the link between GDP growth and higher frequency variables from financial and energy markets. The spread term is a particularly useful indicator to predict recessions in the United States. The empirical evidence also supports the use of functional polynomial weights in the MIDAS specification of the transition probabilities.
This paper merges two specifications recently developed in the forecasting literature: the MS-MID... more This paper merges two specifications recently developed in the forecasting literature: the MS-MIDAS model (Guérin and Marcellino, 2013) and the factor-MIDAS model (Marcellino and Schumacher, 2010). The MS-factor MIDAS model that we introduce incorporates the information provided by a large dataset consisting of mixed frequency variables and captures regime-switching behaviors. Monte Carlo simulations show that this specification tracks the dynamics of the process and predicts the regime switches successfully, both in-sample and out-of-sample. We apply this model to US data from 1959 to 2010 and properly detect recessions by exploiting the link between GDP growth and higher frequency financial variables.
OECD Journal: Journal of Business Cycle Measurement and Analysis, 2014
In recent years, central banks and international organisations have been making ever greater use ... more In recent years, central banks and international organisations have been making ever greater use of factor models to forecast macroeconomic variables. We examine the performance of these models in forecasting French GDP growth over short horizons. The factors are extracted from a large data set of around one hundred variables including survey balances and real, financial, and international variables. An out-of-sample pseudo real-time evaluation over the past decade shows that factor models provide a gain in accuracy relative to the usual benchmarks. However, the forecasts remain inaccurate before the start of the quarter. We also show that the inclusion of international and financial variables can improve forecasts at the longest horizons.
This paper explores the forecasting abilities of Markov-Switching models. Although MS models gene... more This paper explores the forecasting abilities of Markov-Switching models. Although MS models generally display a superior in-sample fit relative to linear models, the gain in prediction remains small. We confirm this result using simulated data for a wide range of specifications by applying several tests of forecast accuracy and encompassing robust to nested models. In order to explain this poor performance, we use a forecasting error decomposition. We identify four components and derive their analytical expressions in different MS specifications. The relative contribution of each source is assessed through Monte Carlo simulations. We find that the main source of error is due to the misclassification of future regimes.
In recent years, factor models have received increasing attention from both econometricians and p... more In recent years, factor models have received increasing attention from both econometricians and practitioners in the forecasting of macroeconomic variables. In this context, Bai and Ng (2008) find an improvement in selecting indicators according to the forecast variable prior to factor estimation (targeted predictors). In particular, they propose using the LARS-EN algorithm to remove irrelevant predictors. In this paper, we adapt the Bai and Ng procedure to a setup in which data releases are delayed and staggered. In the pre-selection step, we replace actual data with estimates obtained on the basis of past information, where the structure of the available information replicates the one a forecaster would face in real time. We estimate on the reduced dataset the dynamic factor model of Giannone, Reichlin and Small (2008) and Doz, Giannone and Reichlin (2011), which is particularly suitable for the very short-term forecast of GDP. A pseudo real-time evaluation on French data shows the potential of our approach.
The French wholesale market is set to expand in the next few years under European pressure and na... more The French wholesale market is set to expand in the next few years under European pressure and national decisions. In this paper, we assess the forecasting ability of several classes of time series models for electricity wholesale spot prices at a day-ahead horizon in France. Electricity spot prices display a strong seasonal pattern, particularly in France given the high share of electric heating in housing during winter time. To deal with this pattern, we implement a double temporal segmentation of the data. For each trading period and season, we use a large number of specifications based on market fundamentals: linear regressions, Markov-switching models, threshold models with a smooth transition. An extensive evaluation on French data shows that modeling each season independently leads to better results. Among non-linear models, MS models designed to capture the sudden and fast-reverting spikes in the price dynamics yield more accurate forecasts. Finally, pooling forecasts gives more reliable results.
This paper investigates the relationship between electricity demand and temperature in the Europe... more This paper investigates the relationship between electricity demand and temperature in the European Union. We address this issue by means of a panel threshold regression model on 15 European countries over the last two decades. Our results confirm the non linearity of the link between electricity consumption and temperature found in more limited geographical areas in previous studies. By distinguishing between North and South countries, we also find that this non linear pattern is more pronounced in the warm countries. At last, rolling regressions highlight the significant impact of climate change on electricity use in Europe.
This paper explores the existence of a bounce-back effect in inventory investment using the Europ... more This paper explores the existence of a bounce-back effect in inventory investment using the European Commission opinion survey on stocks of finished products in manufacturing and retail trade sectors for France, Germany and a European aggregate, from 1985q1 to 2011q4. Our empirical findings support the existence of a high recovery episode for inventory investment, during the quarters immediately following the recessions: it occurs later and lasts longer in manufacturing than in retail trade sector. Since a third phase of rapid recovery has not been found in final sales data so far, the rebound in inventories could in turn explain the GDP growth bounce-back pointed out in previous empirical studies. This calls for a careful modeling of the inventory investment behavior in any sensible theoretical explanation of aggregate business cycles.
We collect data from 29 separate papers estimating the equilibrium level and possible undervaluat... more We collect data from 29 separate papers estimating the equilibrium level and possible undervaluation of the Chinese currency, the renminbi. These papers yield a total of 97 individual observations on misalignment, which we analyse with the help of meta-analysis. We find that the vast majority of observations point to renminbi undervaluation in recent years and that the undervaluation is more pronounced when the US dollar exchange rate is used instead of the real effective exchange rate. We find several characteristics of papers and authors that clearly seem to influence the reported misalignments. For example, when the author is affiliated with an investment bank, the reported misalignment is smaller. Using time-series techniques also results in lower estimates of undervaluation. On the other hand, refereed journals seemingly are inclined to publish papers that report larger misalignments. Results caution against trusting too much in any one study concerning renminbi undervaluation.
Les modèles à facteurs sont de plus en plus utilisés pour la prévision de court terme du PIB par ... more Les modèles à facteurs sont de plus en plus utilisés pour la prévision de court terme du PIB par les banques centrales et les grands organismes internationaux. Ils semblent en revanche un peu moins utilisés en France. Cet article propose une application de ces techniques à la prévision du taux de croissance trimestriel du PIB français à très court terme. Nous utilisons une base constituée d’une centaine de variables parmi lesquelles des variables d’enquêtes, des indicateurs réels, des variables monétaires et financières et des indicateurs sur l’environnement international. Une évaluation hors échantillon montre que la qualité des prévisions issues des modèles à facteurs est satisfaisante, même si les prévisions restent fragiles lorsque l’horizon de prévision est éloigné.
Cet article développe des étalonnages du taux de croissance du PIB français destinés à produire d... more Cet article développe des étalonnages du taux de croissance du PIB français destinés à produire des prévisions de très court terme de l’activité. Ils sont construits exclusivement à partir de données d’enquête de l’Insee, dans l’industrie mais également dans les services et le bâtiment. Nous examinons deux stratégies de réduction de l’information, l’une fondée sur l’algorithme de sélection automatique GETS par blocs de Hendry et Krolzig (2005), l’autre sur la méthode de combinaison popularisée par Stock et Watson (2004). Ces deux méthodes sont évaluées hors échantillon au travers de régressions récursives et roulantes. Nous montrons la supériorité des étalonnages construits avec GETS et l’intérêt de considérer d’autres enquêtes que celle dans l’industrie dans les stratégies de modélisation et de prévision.
Cet article étudie l’interaction entre agents chartistes et fondamentalistes sur le marché des ch... more Cet article étudie l’interaction entre agents chartistes et fondamentalistes sur le marché des changes. A cette fin, nous étendons le modèle à changements de régimes markoviens de Vigfusson afin de pouvoir en tester les implications. La stratégie de tests obtenue est ensuite appliquée aux taux de change canadien, japonais et allemand. Nos résultats valident les conclusions de Vigfusson. Les périodes calmes sont associées à l’action des agents chartistes, tandis que les périodes de volatilité accrue des changes correspondent à l’activité rééquilibrante des agents fondamentalistes.
Uploads