Research

Job Market Paper

Modeling Common Bubbles: A Mixed Causal-Non-Causal Dynamic Factor Model [pdf] 

Abstract: This paper introduces a novel dynamic factor model designed to capture common locally explosive episodes, also known as common bubbles, within large-dimensional, potentially non-stationary time series. The model leverages a lower-dimensional set of factors exhibiting locally explosive behavior to identify common extreme events. Modeling these explosive behaviors allows to predict systemic risk and test for the emergence of common bubbles.  The dynamics of the explosive factors are modeled using mixed causal non-causal models, a class of heavy-tailed autoregressive models that allow processes to depend on their future values through a lead polynomial. The paper establishes the asymptotic properties of the model and provides sufficient conditions for consistency of the estimated factors and parameters. A Monte Carlo simulation confirms the good finite sample properties of the estimator, while an empirical analysis highlights its practical effectiveness. Specifically, the model accurately identifies the common explosive component in monthly stock prices of NASDAQ-listed energy companies during the financial crisis in 2008, significantly outperforming alternative forecasting methods. This new approach offers a powerful tool for detecting common bubbles and predicting their evolution, providing relevant insights for policymakers, investors, and risk managers.

Working Papers

A Novel Test for the Presence of Local Explosive Dynamics. [pdf] [submitted]

(with Siem Jan Koopman, Francisco Blasques and Sean Telg.)

Abstract: In economics and finance, speculative bubbles take the form of locally explosive dynamics that eventually collapse. We propose a test for the presence of speculative bubbles in the context of mixed causal-noncausal autoregressive processes. The test exploits the fact that bubbles are anticipative, that is, they are generated by an extreme shock in the forward-looking dynamics. In particular, the test uses both path level deviations and growth rates to assess the presence of a bubble of given duration and size, at any moment of time. We show that the distribution of the test statistic can be either analytically determined or numerically approximated, depending on the error distribution. Size and power properties of the test are analyzed in controlled Monte Carlo experiments. An empirical application is presented for a monthly oil price index. It demonstrates the ability of the test to detect bubbles and to provide valuable insights in terms of risk assessments in the spirit of Value-at-Risk.  

Observation-Driven filters for Time-Series with  Stochastic Trends and Mixed Causal Non-Causal Dynamics. [pdf] [submitted]

(with Siem Jan Koopman and Francisco Blasques)

Abstract: This paper proposes a novel time-series model with a non-stationary stochastic trend, locally explosive mixed causal non-causal dynamics and fat-tailed innovations. The model allows for a description of financial time-series that is consistent with financial theory, for a decomposition of the time-series in trend and bubble components, and for meaningful real-time forecasts during bubble episodes. We provide sufficient conditions for strong consistency and asymptotic normality of the maximum likelihood estimator. The model-based filter for extracting the trend and bubbles is shown to be invertible and the extracted components converge to the true trend and bubble paths. A Monte Carlo simulation study confirms the good finite sample properties. Finally, we consider an empirical study of Nickel monthly price series and global mean sea level data. We document the forecasting accuracy against competitive alternative methods and conclude that our model-based forecasts outperform all these alternatives.

Work in Progress

New Early Warning Indicators of Country Risk using Machine Learning. (with several coauthors)

This paper  proposes a new mixture-of-experts (MoE) machine learning predictive algorithm capable of producing a reliable and interpretable early warning indicator for country risk across 218 countries with heterogeneous datasets. The model is designed to handle an imbalanced panel of continuous, discrete, and categorical inputs, featuring missing observations, different observational frequencies, and different update moments for the most recent data. The MoE decomposes risk consistently across Fiscal, Financial, and External risk components. 

Weighted Maximum Likelihood for Mixed Causal Non-Causal Models.  (with Siem Jan Koopman and Francisco Blasques)

This paper proposes a novel weighted maximum likelihood approach designed to improve the forecast precision of misspecified mixed causal non-causal autoregressive (MAR) models. In applications, speculative bubbles appear with different growth rates over time. We show how, when estimating a MAR model, weighting differently the information coming from our data allows us to target an estimator that improves the forecast with respect to the standard MLE estimator. 

Maximum Likelihood Estimation of Location Filters for alpha-stable autoregressive processes.

This paper establishes the properties of maximum likelihood estimation of observation driven filters, with a focus on location filters, in the context of autoregressive processes with alpha-stable errors. This paper allows to extend the decomposition into stochastic trend and bubble component of non-stationary explosive processes in the realm of alpha-stable distributions. The theory established in this paper, however, determines the properties also for more general filters in the context of processes without finite moments.

Estimating the date of a bubble collapse.

This paper introduces a novel estimator for the date of collapse of bubbles. The bubble dynamics, explained by mixed causal non-causal processes, are induced by the anticipation of a future shock. The proposed methods relies on conditional probabilities of future errors to evaluate the date that is more likely to originate the explosive episode. This paper proposes also confidence bounds for the potential date of collapse of an explosive event.