You are looking at 1 - 10 of 10 items for :

  • Type: Journal Issue x
  • Neural Networks and Related Topics x
Clear All Modify Search
Chris Redl and Sandile Hlatshwayo
We produce a social unrest risk index for 125 countries covering a period of 1996 to 2020. The risk of social unrest is based on the probability of unrest in the following year derived from a machine learning model drawing on over 340 indicators covering a wide range of macro-financial, socioeconomic, development and political variables. The prediction model correctly forecasts unrest in the following year approximately two-thirds of the time. Shapley values indicate that the key drivers of the predictions include high levels of unrest, food price inflation and mobile phone penetration, which accord with previous findings in the literature.
Mr. Jorge A Chan-Lau
We introduce unFEAR, Unsupervised Feature Extraction Clustering, to identify economic crisis regimes. Given labeled crisis and non-crisis episodes and the corresponding features values, unFEAR uses unsupervised representation learning and a novel mode contrastive autoencoder to group episodes into time-invariant non-overlapping clusters, each of which could be identified with a different regime. The likelihood that a country may experience an econmic crisis could be set equal to its cluster crisis frequency. Moreover, unFEAR could serve as a first step towards developing cluster-specific crisis prediction models tailored to each crisis regime.
Marijn A. Bolhuis and Brett Rayner
We leverage insights from machine learning to optimize the tradeoff between bias and variance when estimating economic models using pooled datasets. Specifically, we develop a simple algorithm that estimates the similarity of economic structures across countries and selects the optimal pool of countries to maximize out-of-sample prediction accuracy of a model. We apply the new alogrithm by nowcasting output growth with a panel of 102 countries and are able to significantly improve forecast accuracy relative to alternative pools. The algortihm improves nowcast performance for advanced economies, as well as emerging market and developing economies, suggesting that machine learning techniques using pooled data could be an important macro tool for many countries.
Marijn A. Bolhuis and Brett Rayner
We develop a framework to nowcast (and forecast) economic variables with machine learning techniques. We explain how machine learning methods can address common shortcomings of traditional OLS-based models and use several machine learning models to predict real output growth with lower forecast errors than traditional models. By combining multiple machine learning models into ensembles, we lower forecast errors even further. We also identify measures of variable importance to help improve the transparency of machine learning-based forecasts. Applying the framework to Turkey reduces forecast errors by at least 30 percent relative to traditional models. The framework also better predicts economic volatility, suggesting that machine learning techniques could be an important part of the macro forecasting toolkit of many countries.
Jin-Kyu Jung, Manasa Patnam, and Anna Ter-Martirosyan
Forecasting macroeconomic variables is key to developing a view on a country's economic outlook. Most traditional forecasting models rely on fitting data to a pre-specified relationship between input and output variables, thereby assuming a specific functional and stochastic process underlying that process. We pursue a new approach to forecasting by employing a number of machine learning algorithms, a method that is data driven, and imposing limited restrictions on the nature of the true relationship between input and output variables. We apply the Elastic Net, SuperLearner, and Recurring Neural Network algorithms on macro data of seven, broadly representative, advanced and emerging economies and find that these algorithms can outperform traditional statistical models, thereby offering a relevant addition to the field of economic forecasting.
Ms. Yevgeniya Korniyenko, Manasa Patnam, Rita Maria del Rio-Chanon, and Mason A. Porter
This paper studies the interconnectedness of the global financial system and its susceptibility to shocks. A novel multilayer network framework is applied to link debt and equity exposures across countries. Use of this approach—that examines simultaneously multiple channels of transmission and their important higher order effects—shows that ignoring the heterogeneity of financial exposures, and simply aggregating all claims, as often done in other studies, can underestimate the extent and effects of financial contagion.The structure of the global financial network has changed since the global financial crisis, impacted by European bank’s deleveraging and higher corporate debt issuance. Still, we find that the structure of the system and contagion remain similar in that network is highly susceptible to shocks from central countries and those with large financial systems (e.g., the USA and the UK). While, individual European countries (excluding the UK) have relatively low impact on shock propagation, the network is highly susceptible to the shocks from the entire euro area. Another important development is the rising role of the Asian countries and the noticeable increase in network susceptibility to shocks from China and Hong Kong SAR economies.
Mr. Alexei P Kireyev and Andrei Leonidov
This paper proposes a method for assessing international spillovers from nominal demand shocks. It quantifies the impact of a shock in one country on all other countries. The paper concludes that the network effects in shock spillovers can be substantial, comparable, and often exceed the initial shock. Individual countries may amplify, absorb, or block spillovers. Most developed countries pass-through shocks, whereas low-income countries and oil exporters tend to block shock spillovers. The method is used to study demand shocks originating from a large and medium country, China and Ukraine respectively.
Mr. Nicolas Arregui, Mr. Mohamed Norat, Antonio Pancorbo, Ms. Jodi G Scarlata, Eija Holttinen, Fabiana Melo, Jay Surti, Christopher Wilson, Rodolfo Wehrhahn, and Mamoru Yanase
This paper reviews tools used to identify and measure interconnectedness and raises the awareness of policymakers as to potential cross-sectional implications of prudential tools aimed at controlling interconnectedness. The paper examines two sets of tools—developed at the IMF and externally—to identify the implications of interconnectedness in systemic risk and how these tools have been applied in IMF surveillance. The paper then proposes a preliminary framework to analyze some key internationally-agreed-upon and national prudential tools and finds that while many prudential tools are effective in reducing interconnectedness, the interaction among these tools is far less clear cut.
Mr. Alexander Massara and Mr. Luca Errico
The paper focuses on systemically important jurisdictions in the global trade network, complementing recent IMF work on systemically important financial sectors. Using the IMF's Direction of Trade Statistics (DOTS) database and network analysis, the paper develops a framework for ranking jurisdictions based on trade size and trade interconnectedness indicators using data for 2000 and 2010. The results show a near perfect overlap between the top 25 systemically important trade and financial jurisdictions, suggesting that these ought to be the focus of risk-based surveillance on cross-border spillovers and contagion. In addition, a number of extensions to the approach are developed that can provide a better understanding of trade dynamics at the bilateral, regional, and global levels.
Paolo Manasse
This paper assesses the roles of shocks, rules, and institutions as possible sources of procyclicality in fiscal policy. By employing parametric and nonparametric techniques, I reach the following four main conclusions. First, policymakers' reactions to the business cycle is different depending on the state of the economy-fiscal policy is "acyclical" during economic bad times, while it is largely procyclical during good times. Second, fiscal rules and fiscal responsibility laws tend to reduce the deficit bias on average, and seem to enhance, rather than to weaken, countercyclical policy. However, the evidence also suggests that fiscal frameworks do not exert independent effects when the quality of institutions is accounted for. Third, strong institutions are associated to a lower deficit bias, but their effect on procyclicality is different in good and bad times, and it is subject to decreasing returns. Fourth, unlike developed countries, fiscal policy in developing countries is procyclical even during (moderate) recessions; in "good times," however, fiscal policy is actually more procyclical in developed economies.