Markov-switching Dynamic Factor Model Research Articles

Approximating Botswana's financial cycle: Expanding the macroprudential toolkit

AbstractThis study estimates the financial cycle for Botswana using the Christiano–Fitzgerald band‐pass filter, unobserved components approach and a Markov switching dynamic factor model. Using real credit and equity prices, together with several macroeconomic variables from 2001Q1 to 2021Q4, we find that the domestic financial cycle generally captures movements along the business cycle, with peaks reflecting the cumulative build‐up of risks during the boom that ultimately burst, coinciding with periods of financial distress. These findings shed light on the implications of financial fluctuations on domestic financial stability, hence, inform effective calibration of macroprudential tools, while also providing traction for implementation of the recently approved macroprudential policy framework for Botswana.

Macroprudential and Monetary Policy Interactions and Coordination in South Africa: Evidence from Business and Financial Cycle Synchronisation

The article reports on the interactions and possibility of coordination between macroprudential and monetary policies in South Africa, based on business and financial cycles synchronisation. To this end, relying on financial and economic time series indicators spanning the period 2000M01–2018M12, a two-step Markov switching dynamic factor model was adopted for the measurement of composite indices, while both the dynamic conditional correlations and asymmetric generalised dynamic conditional correlations models were adopted for synchronisation analysis, together with the Metcalfe scale of coordination. The empirical evidence obtained is such that, under conditions of financial and real economic stress in South Africa, when there is crisis management rather than crisis prevention, macroprudential policy and monetary policy decisions are complementary. Therefore, there will be limited/no need for coordination between the two policy decisions. However, under normal times when there is crisis prevention rather than management, macroprudential policy and monetary policy decisions are noncomplementary; hence, the greatest degree of coordination is warranted, even though it might not be easy. Therefore, we conclude that it is possible to the coordinate the conduct of macroprudential, and monetary policies based on the synchronicity of business and financial cycles.

The more, the better? Forecasting gains from high-frequency housing prices in a Markov-switching dynamic factor model

AbstractThis paper extracts housing boom-bust cycle signals from metropolitan statistical area (MSA)-level housing prices using a Markov-switching dynamic factor model. To mitigate the estimation bias, it utilizes high-frequency housing prices that follow the methodology of the monthly Case–Shiller house price indices. The housing bust phases specified from weekly and daily housing prices precede those based on monthly prices by approximately 2 years. MSAs with top signal-to-noise ratios offer greater marginal contributions to improvements in forecasting housing cycles than MSAs with bottom ratios for all frequencies. The results highlight the importance of indicator quality and provide evidence against “The more, the better” since incorporating more MSA-level housing prices into housing factors does not guarantee more satisfactory housing cycle forecasts.

Predicting ordinary and severe recessions with a three-state Markov-switching dynamic factor model: An application to the German business cycle

We estimate a Markow-switching dynamic factor model with three states based on six leading business cycle indicators for Germany, preselected from a broader set using the elastic net soft-thresholding rule. The three states represent expansions, normal recessions and severe recessions. We show that a two-state model is not sensitive enough to detect relatively mild recessions reliably when the Great Recession of 2008/2009 is included in the sample. Adding a third state helps to distinguish normal and severe recessions clearly, so that the model identifies all business cycle turning points in our sample reliably. In a real-time exercise, the model detects recessions in a timely manner. Combining the estimated factor and the recession probabilities with a simple GDP forecasting model yields an accurate nowcast for the steepest decline in GDP in 2009Q1, and a correct prediction of the timing of the Great Recession and its recovery one quarter in advance.

Improved recession dating using stock market volatility

We offer an improved dating of U.S. business cycle turning points both retrospectively and in real time. This improvement is made possible by augmenting existing Markov-switching dynamic factor models with additional information on the stock return volatility. The model improves the prediction of the state of the economy using fully revised data significantly. Real-time identification can be made noticeably earlier than the NBER announcements, beating both the peak and trough announcements for recent recessions by several months.

Improved Recession Forecasts Considering Stock Market Volatility

We offer improved dating of U.S. business cycle turning points both retrospectively and in real time. This improvement is made possible by augmenting existing Markov-switching dynamic factor models with additional information on stock return volatility. The model significantly improves prediction of the state of the economy using fully revised data. Real-time identification can be made noticeably earlier than NBER announcements, beating both peak and trough announcements for recent recessions by several months.

The Asymmetric Effects of Monetary Policy on Stock Market

This paper shows that the effects of expansionary monetary policy on the U.S. stock market are asymmetric across different monetary policy phases and different stock market regimes. A Markov-switching dynamic factor model dates the periods of stock market regimes, and generates a new composite measure for overall stock market movements. A time-varying parameter analysis finds that an expansionary monetary policy such as an increase in monetary aggregates or a decrease in the Federal funds rate has positive impacts on stock returns only during the periods in which they are used as monetary policy targets by the Federal Reserve.

Markov-switching dynamic factor models in real time

We extend the Markov-switching dynamic factor model to account for some of the specificities of the day-to-day monitoring of economic developments from macroeconomic indicators, such as mixed sampling frequencies and ragged-edge data. First, we evaluate the theoretical gains of using data that are available promptly for computing probabilities of recession in real time. Second, we show how to estimate the model that deals with unbalanced panels of data and mixed frequencies, and examine the benefits of this extension through several Monte Carlo simulations. Finally, we assess its empirical reliability for the computation of real-time inferences of the US business cycle, and compare it with the alternative method of forecasting the probabilities of recession from balanced panels.

Large Shocks and the Business Cycle: The Effect of Outlier Adjustments

This study examines the impact of outlier-adjusted data on business cycle inferences using coincident indicators of the composite index (CI) in Japan. To estimate the CI and business cycles, this study proposes a Markov switching dynamic factor model incorporating Student’s t-distribution in both the idiosyncratic noise and the factor equation. Furthermore, the model includes a stochastic volatility process to identify whether a large shock is associated with a business cycle. From the empirical analysis, both the factor and the idiosyncratic component have fat-tail error distributions, and the estimated CI and recession probabilities are close to those published by the Economic and Social Research Institute. Compared with the estimated CI using the adjusted data set, the outlier adjustment reduces the depth of the recession. Moreover, the results of the shock decomposition show that the financial crisis in mid-2008 was caused by increase of clustering shocks and large unexpected shocks. In contrast, the Great East Japan Earthquake in 2011 was derived from idiosyncratic noise and did not cause a recession. When analyzing whether to use a sample that includes outliers associated with the business cycle, it is not desirable to use the outlier-adjusted data set.

Estimating Dynamic Connectivity States in fMRI Using Regime-Switching Factor Models.

We consider the challenges in estimating the state-related changes in brain connectivity networks with a large number of nodes. Existing studies use the sliding-window analysis or time-varying coefficient models, which are unable to capture both smooth and abrupt changes simultaneously, and rely on ad-hoc approaches to the high-dimensional estimation. To overcome these limitations, we propose a Markov-switching dynamic factor model, which allows the dynamic connectivity states in functional magnetic resonance imaging (fMRI) data to be driven by lower-dimensional latent factors. We specify a regime-switching vector autoregressive (SVAR) factor process to quantity the time-varying directed connectivity. The model enables a reliable, data-adaptive estimation of change-points of connectivity regimes and the massive dependencies associated with each regime. We develop a three-step estimation procedure: 1) extracting the factors using principal component analysis, 2) identifying connectivity regimes in a low-dimensional subspace based on the factor-based SVAR model, and 3) constructing high-dimensional state connectivity metrics based on the subspace estimates. Simulation results show that our estimator outperforms -means clustering of time-windowed coefficients, providing more accurate estimate of time-evolving connectivity. It achieves percentage of reduction in mean squared error by 60% when the network dimension is comparable to the sample size. When applied to the resting-state fMRI data, our method successfully identifies modular organization in the resting-statenetworksin consistencywith other studies. It further reveals changes in brain states with variations across subjects and distinct large-scale directed connectivity patterns across states.

Dynamical Interaction between Financial and Business Cycles

We adopt the Dynamical Influence model from computer science and transform it to study the interaction between business and financial cycles. For this purpose, we merge it with Markov-Switching Dynamic Factor Model (MS-DFM) which is frequently used in economic cycle analysis. The model suggested in this paper, the Dynamical Influence Markov-Switching Dynamic Factor Model (DI-MS-FM), allows to reveal the pattern of interaction between business and financial cycles in addition to their individual characteristics. More specifically, this model allows to describe quantitatively the existing regimes of interaction in a given economy and to identify their timing, as well as to evaluate the effect of the government policy on the duration of each of the regimes. We are also able to determine the direction of causality between the two cycles for each of the regimes. The model estimated on the US data demonstrates reasonable results, identifying the periods of higher interaction between the cycles in the beginning of 1980s and during the Great Recession, while in-between the cycles evolve almost independently. The output of the model can be useful for policymakers since it provides a timely estimate of the current interaction regime, which allows to adjust the timing and the composition of the policy mix.

Evidencia de un ciclo económico común en la Integración Económica Centroamericana: estimación y caracterización mediante un modelo dinámico factorial con cambios de regímenes de Markov

Evidencia de un ciclo económico común en la Integración Económica Centroamericana: estimación y caracterización mediante un modelo dinámico factorial con cambios de regímenes de Markov

Predicting Ordinary and Severe Recessions with a Three-State Markov-Switching Dynamic Factor Model. An Application to the German Business Cycle

We estimate a Markow-switching dynamic factor model with three states based on six leading business cycle indicators for Germany preselected from a broader set using the Elastic Net soft-thresholding rule. The three states represent expansions, normal recessions and severe recessions. We show that a two-state model is not sensitive enough to reliably detect relatively mild recessions when the Great Recession of 2008/2009 is included in the sample. Adding a third state helps to clearly distinguish normal and severe recessions, so that the model identifies reliably all business cycle turning points in our sample. In a real-time exercise the model detects recessions timely. Combining the estimated factor and the recession probabilities with a simple GDP forecasting model yields an accurate nowcast for the steepest decline in GDP in 2009Q1 and a correct prediction of the timing of the Great Recession and its recovery one quarter in advance.

Monitoring the world business cycle

We propose a Markov-switching dynamic factor model to construct an index of global business cycle conditions, to perform short-term forecasts of world GDP quarterly growth in real time and to compute real-time business cycle probabilities. To overcome the real-time forecasting challenges, the model accounts for mixed frequencies, for asynchronous data publication and for leading indicators. Our pseudo real-time results show that this approach provides reliable and timely inferences of the world quarterly growth and of the world state of the business cycle on a monthly basis.

Monitoring the World Business Cycle

We propose a Markov-switching dynamic factor model to construct an index of global business cycle conditions, to perform short-term forecasts of world GDP quarterly growth in real time and to compute real-time business cycle probabilities. To overcome the real-time forecasting challenges, the model accounts for mixed frequencies, for asynchronous data publication and for leading indicators. Our pseudo real-time results show that this approach provides reliable and timely inferences of the world quarterly growth and of the world state of the business cycle on a monthly basis.

Monitoring the World Business Cycle

We propose a Markov-switching dynamic factor model to construct an index of global business cycle conditions, to perform short-term forecasts of world GDP quarterly growth in real time and to compute real-time business cycle probabilities. To overcome the real-time forecasting challenges, the model accounts for mixed frequencies, for asynchronous data publication and for leading indicators. Our pseudo real-time results show that this approach provides reliable and timely inferences of the world quarterly growth and of the world state of the business cycle on a monthly basis.

Extracting Nonlinear Signals from Several Economic Indicators

Summary We develop a twofold analysis of how the information provided by several economic indicators can be used in Markov switching dynamic factor models to identify the business cycle turning points. First, we compare the performance of a fully nonlinear multivariate specification (one-step approach) with the ‘shortcut’ of using a linear factor model to obtain a coincident indicator, which is then used to compute the Markov switching probabilities (two-step approach). Second, we examine the role of increasing the number of indicators. Our results suggest that one step is generally preferred to two steps, especially in the vicinity of turning points, although its gains diminish as the quality of the indicators increases. Additionally, we also obtain decreasing returns of adding more indicators with similar signal-to-noise ratios. Using the four constituent series of the Stock–Watson coincident index, we illustrate these results for US data. Copyright © 2014 John Wiley & Sons, Ltd.

Green shoots and double dips in the euro area: A real time measure

In order to perform real-time business cycle inferences and forecasts of GDP growth rates in the euro area, we use an extension of the Markov-switching dynamic factor models that accounts for the features of the day-to-day monitoring of economic developments, such as ragged edges, mixed frequencies and data revisions. We provide examples that show the nonlinear nature of the relationships between data revisions, point forecasts and forecast uncertainty. Based on our empirical results, we think that the real-time probabilities of recession inferred from the model are an appropriate statistic for capturing what the press call green shoots, and for monitoring double-dip recessions.

Markov-Switching Dynamic Factor Models in Real Time

We extend the Markov-switching dynamic factor model to account for some of the specifi cities of the day-to-day monitoring of economic developments from macroeconomic indicators, such as ragged edges and mixed frequencies. We examine the theoretical benefi ts of this extension and corroborate the results through several Monte Carlo simulations. Finally, we assess its empirical reliability to compute real-time inferences of the US business cycle.

Extracting Non-Linear Signals from Several Economic Indicators

We develop a twofold analysis of how the information provided by several economic indicators can be used in Markov-switching dynamic factor models to identify the business cycle turning points. First, we compare the performance of a fully non-linear multivariate specification (one-step approach) with the “shortcut” of using a linear factor model to obtain a coincident indicator which is then used to compute the Markov-switching probabilities (two-step approach). Second, we examine the role of increasing the number of indicators. Our results suggest that one step is generally preferred to two steps, although its marginal gains diminish as the quality of the indicators increases and as more indicators are used to identify the non-linear signal. Using the four constituent series of the Stock-Watson coincident index, we illustrate these results for US data.