A Quasi–Maximum Likelihood Approach for Large, Approximate Dynamic Factor Models

By Catherine Doz, Domenico Giannone, L. Reichlin. Review of Economics and Statistics, 2012

516 citations recorded by the source corpus.

Abstract Is maximum likelihood suitable for factor models in large cross-sections of time series? We answer this question from both an asymptotic and an empirical perspective. We show that estimates of the common factors based on maximum likelihood are consistent for the size of the cross-section (n) and the sample size (T), going to infinity along any path, and that maximum likelihood is viable for n large. The estimator is robust to misspecification of cross-sectional and time series correlation of the idiosyncratic components. In practice, the estimator can be easily implemented using the Kalman smoother and the EM algorithm as in traditional factor analysis.

Subfields: Artificial Intelligence, Economics and Econometrics

View Paper: https://www.semanticscholar.org/paper/dca2dd345f4a082d006ba576478cccbc28854c91

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Economics and Econometrics

Citations in the corpus

Regional Policy Evaluation: Interactive Fixed Effects and Synthetic Controls (2016, unknown): “Moreover, we do not specify the dynamics of factors in the spirit of Doz, Giannone, and Reichlin (2012).”
Diagnosing the Financial System: Financial Conditions and Financial Stress (2012, methodology): “…financial conditions indexes in the literature, however, the NFCI uses a very flexible estimation procedure for these weights that builds off the work of Stock and Watson (2002), Doz, Giannone, and Reichlin (2006), and Aruoba, Diebold, and Scotti (2009) and is described in detail in the appendix.”
DYNAMIC FACTOR MODELS (2018, methodology): “This merged estimation procedure occurs in two steps, which are described in more detail in Giannone, Reichlin, and Small (2008) and Doz, Giannone, and Reichlin (2006).”
Measuring the response of gold prices to uncertainty: An analysis beyond the mean (2018, methodology): “We suppose that the number of fundamental shocks μt is similar to the that of the common factors r. Ultimately, we estimate the Eqs (7) and (8) by means of the Quasi Maximum Likelihood (QML) procedure developed by Doz et al. (2012).”
On estimation of the noise variance in high dimensional probabilistic principal component analysis (2013, unknown): “For related recent references on inference of large approximate (or dynamic) factor models, we refer to Bai (2003), Forni et al. (2000) and Doz et al. (2012).”
Now-Casting and the Real-time Data Flow (2013, unknown): “How large is large enough is an empirical question and Monte Carlo exercises of Doz, Giannone, and Reichlin (2011) and Doz, Giannone, and Reichlin (2012) show that substantial robustness is achieved already with a handful of variables.”
Identification and Bayesian Estimation of Dynamic Factor Models (2015, methodology): “The maximization can be implemented through the expectation–maximization (EM) method as in Doz, Giannone, and Reichlin (2012).”
Common Factors of Commodity Prices † (2018, methodology): “However, Doz et al. (2012) ECB Working Paper Series No 2112 / November 2017 7 have shown that under the approximate factor structure (i.e. pervasive factors and limited cross-sectional correlation among idiosyncratic components), maximum-likelihood estimates of the model are robust to misspeci…”
Generalized Shrinkage Methods for Forecasting Using Many Predictors (2012, methodology): “Doz, Giannone, and Reichlin 2011 ), improve upon the first-five principal components forecasts used here, the results here understate forecasting potential of improved DFM variants.”
World Asset Markets and the Global Financial Cycle (2015, methodology): “Consistency of the Maximum Likelihood estimator is proven under this type of misspecification in Doz et al. (2011). (9)A similar specification has been adopted by Kose et al.”
Surprise and Uncertainty Indexes: Real-Time Aggregation of Real-Activity Macro Surprises (2016, methodology): “To gauge the importance of the various indicators in constructing the surprise and uncertainty indexes, I consider two different standpoints in analyzing the weights: (i) I construct the cumulative weights as in equation (17) and (ii) I analyze, at each time t, the vector of t× 1 weights, w t , that”
Estimation and Inference of FAVAR Models (2014, methodology): “In factor analysis literature, many studies, such as Bai and Li (2012, 2016), Doz, Giannone, and Reichlin (2012), investigate the finite sample properties of the QMLE, that is, ̃, F̃ , ̃ee.”