# Linear System Challenges of Dynamic Factor Models

^{1}

^{2}

^{3}

^{4}

^{*}

## Abstract

**:**

## 1. Introduction

- The theoretical underpinning for allowing modelling to focus on AR rather than ARMA modelling, and the simplifications this makes possible for the identification of models.
- Issues arising in practice from numerical problems in attempting to build models; some of these errors can be mitigated through the use of carefully chosen forms for the models.
- The difficulties arising from modelling time series of different but multiply related periodicities, e.g., monthly and quarterly, and existing tools for resolving, at least partly, the difficulties. These build on the ideas developed for single-periodicity modelling.

## 2. Dynamic Factor Models

**Assumption**

**1.**

#### 2.1. The Goal of Modelling

#### The Issue of Errors and Continuity

- If there is a true underlying process with an associated canonical spectral factor, and a canonical spectral factor obtained using approximate values of lagged covariance coefficients, there will be an error between the two.
- We should use a description of the computed canonical spectral factor that is appropriate to the application, in the sense that the error mechanisms in the relevant computations will not cause significant errors in the quantities computed as part of the application.

**Assumption**

**2.**

#### 2.2. Controlling the Number of Parameters to Be Estimated

**Assumption**

**3.**

- The dimension n of a minimal state-space realization of the canonical spectral factor (or its McMillan degree)
- The number of columns q in the minimal state-space factor, or the dimension of an innovations sequence (this being a white noise sequence, known in this context as the dynamic factor sequence)
- The rank r of ${\gamma}^{N}\left(0\right):=\mathbb{E}\left[{\chi}_{t}^{N}{\left({\chi}_{t}^{N}\right)}^{\top}\right]$, which is the dimension of the space spanned by ${\chi}_{1t},{\chi}_{2t},\dots ,{\chi}_{Nt}$.

**Assumption**

**4.**

**Theorem**

**1.**

#### 2.3. Strong and Weak Cross-Dependence

**Assumption**

**5.**

**Assumption**

**6.**

## 3. Obtaining the Common Component Process from the Measurement Process

#### Two Alternative Approaches to Additive Decomposition of the Spectral Matrix

## 4. Tall Stable Miniphase Spectral Factors and Singular Rational Spectra

**Theorem**

**2**

- Suppose that for polynomial matrices ${\overline{A}}^{N}\left(z\right)\in {\mathbb{R}}^{N\times N}\left[z\right],{\overline{B}}^{N}\left(z\right)\in {\mathbb{R}}^{N\times q}\left[z\right]$ with ${\overline{A}}^{N}$ and ${\overline{B}}^{N}$ left coprime8 there holds ${K}^{N}\left(z\right)={\left({\overline{A}}^{N}\right)}^{-1}{\overline{B}}^{N}$. Then zero freeness implies ${\overline{B}}^{N}\in {\mathbb{R}}^{N\times q}$ has full rank for all z. Accordingly, there exists a polynomial unimodular (constant determinant) common premultiplier ${U}^{N}\left(z\right)$ of ${\overline{A}}^{N},{\overline{B}}^{N}$ such that a new polynomial fraction description of ${K}^{N}\left(z\right)$ exists with$${K}^{N}\left(z\right)={\left({U}^{N}\left(z\right){\overline{A}}^{N}\left(z\right)\right)}^{-1}\left({U}^{N}\left(z\right){\overline{B}}^{N}\left(z\right)\right)={\left({A}^{N}\left(z\right)\right)}^{-1}{B}^{N}$$
- The system (3) is left invertible with unknown initial state, that is, there exists an integer $L\le n$ such that from the sequence ${\chi}_{k}^{N},{\chi}_{k+1}^{N},\dots ,{\chi}_{k+L-1}^{M}$ the state ${x}_{k}$ and the sequence ${w}_{k+1},{w}_{k+2},\dots ,{w}_{k+L-1}$ can be determined.

**Theorem**

**3**

- 1.
- The spectrum${\mathrm{\Phi}}_{\chi}^{N}$can be generated by a singular autoregressive process$$[I+{A}_{1}^{N}z+{A}_{2}^{N}{z}^{2}+\cdots +{A}_{m}^{N}{z}^{m}]{\chi}_{t}^{N}={B}^{N}{w}_{t}$$
- 2.
- Suppose the rational spectral matrix${\mathrm{\Phi}}_{\chi}^{N}\left({e}^{j\omega}\right)$is written in the form$${\overline{H}}^{N}{(I-z\overline{F})}^{-1}{\overline{Q}}^{N}+{\left({\overline{Q}}^{N}\right)}^{\top}{(I-{z}^{-1}{\overline{F}}^{\top})}^{-1}{\left({\overline{H}}^{N}\right)}^{\top}-{\overline{H}}^{N}{\overline{Q}}^{N},\phantom{\rule{1.em}{0ex}}z={e}^{j\omega}$$
- (a)
- there exists${\overline{G}}^{N}$such that the stable miniphase (canonical) spectral factor can be written in the form${\overline{H}}^{N}{(I-z\overline{F})}^{-1}{\overline{G}}^{N}$
- (b)
- ${\overline{G}}^{N}$is computable in a finite number of rational calculations from$\overline{F},{\overline{H}}^{N}$and${\overline{Q}}^{N}$, and
- (c)
- the triple$\{\overline{F},{\overline{H}}^{N},{\overline{G}}^{N}\}$is minimal.

- 3.
- In the event that${\overline{G}}^{N}$defined above is independent of N, say$G={\overline{G}}^{N}$, then${\overline{Q}}^{N}=P{\left({\overline{H}}^{N}\right)}^{\top}$for some positive definite P which is independent of N, with P the unique solution of$P-\overline{F}P{\overline{F}}^{\top}=G{G}^{\top}$.

**Theorem**

**4.**

## 5. Obtaining Parameters in the AR and State-Variable Models

## 6. Mixed Frequency Systems

#### 6.1. Blocked Linear Systems

#### 6.2. Systems with Multiple Periodicities in Their Outputs

**Theorem**

**5.**

- Depending on the value of J and the dimensions of ${u}_{t},{y}_{t}$, the dimension of ${U}_{t}$ may end up larger than that of ${Y}_{t}$, i.e., the transfer function of the blocked system may be fat, as noted in Anderson et al. (2016b).
- Depending on the value of J and the dimensions of ${u}_{t},{x}_{t}$, the dimension of ${U}_{t}$ may end up greater than that of ${x}_{t}$ (which is the state dimension of the blocked system).

## 7. Mixed Frequency System Identification

#### 7.1. Mixed Frequency System Identificaiton with an AR Underlying System

#### 7.2. A Simple AR(1) Example

#### 7.3. Mixed Frequency System Identification Exploiting Blocked System Structure

#### 7.4. VARMA and VMA System Identification

**Theorem**

**6.**

## 8. Conclusions

- The zero-free property has been applied to macroeconomic analysis as a means to solve the fundamentalness problem. Further work is ongoing on this and, we hope, the zero-free property will increasingly become familiar to macroeconomists, see the literature cited below Theorem 4.
- An important observation on the zero-free property in statements like Theorem 4 is that its genericity is usually obtained with respect to a parameterization in which each parameter is free to vary independently of all the others, see the comment below Theorem 2. However, in applications to macroeconomic models, structural restrictions may prevent such free variability and can therefore interfere with the genericity of zero-freeness. For an attempt to solve this issue, see Forni et al. (2020).
- Another observation on Theorem 4 is that an AR representation may exist also for non-generic parameters, for which zerolessness does not hold but the zeros of $B\left(z\right)$ lie outside the unit circle. However, in that case the AR would not be of finite length.
- In some cases, we have outlined algorithms which have not yet been tried on real data. It would be worthwhile to investigate them using real data.
- A full treatment (algorithms through to testing on real data) including the decomposition into common components and idiosyncratic processes is needed for systems with multiple frequencies.

## Author Contributions

## Funding

## Institutional Review Board Statement

## Informed Consent Statement

## Data Availability Statement

## Conflicts of Interest

## Abbreviations

AIC | Akaike Information Criterion |

AR | Autoregression |

ARMA | Autoregression Moving Average |

MA | Moving Average |

PCA | Principal Components Analysis |

SVD | Singular Value Decomposition |

VAR | Vector Autoregression |

VARMA | Vector Autoregression Moving Average |

VMA | Vector Moving Average |

## Appendix A. Proof of Theorem 3, Part 3

## Notes

1 | This paper underpins a lecture by the first author presented at the 5th Vienna Workshop on High-dimensional Time Series in Macroeconomics and Finance, and celebrating the 80th birthday of Manfred Deistler. |

2 | State space models have larger equivalences classes and therefore more flexibility when choosing “nice” representatives. Special representatives, in echelon form, can yield state space models with the same parameters as ARMA models. |

3 | This is an easy consequence of the celebrated solution of the partial realization problem for covariance sequences, Byrnes and Lindquist (1997). |

4 | In a practical situation, given that sensors normally are not noise free, the introduction of more sensors is likely to aid the de-noising process. Our remarks here pertain to the process of common components. |

5 | Such a spanning set of entries can be obtained by identifying the rows defining an $r\times r$ nonsingular principal submatrix of $\mathbb{E}\left[{\chi}_{t}^{N}{\left({\chi}_{t}^{N}\right)}^{\top}\right]$ |

6 | As an alternative, PCA can also be used to obtain a minimal static factor. |

7 | For a scalar rational transfer function written as a ratio of coprime polynomials $n\left(z\right)/d\left(z\right)$, the zero free property is equivalent to $n\left(z\right)$ being a nonzero constant. The idea generalizes to coprime matrix fraction representations of a rational transfer function matrix |

8 | Coprimeness means there exists no polynomial ${C}^{N}\left(z\right)\in {\mathbb{R}}^{N\times N}\left[z\right]$ with nonconstant determinant and matrices ${\tilde{A}}^{N}\in {\mathbb{R}}^{N\times N}\left[z\right],{\tilde{B}}^{N}\in {\mathbb{R}}^{N\times q}\left[z\right]$ such that ${\overline{A}}^{N}={C}^{N}{\tilde{A}}^{N}$ and ${\overline{B}}^{N}={C}^{N}{\tilde{B}}^{N}$ |

## References

- Amengual, Dante, and Mark W. Watson. 2007. Consistent Estimation of the Number of Dynamic Factors in a Large N and T Panel. Journal of Business & Economic Statistics 25: 91–96. [Google Scholar]
- Anderson, Brian D. O. 1969. The inverse problem of stationary covariance generation. Journal of Statistical Physics 1: 133–47. [Google Scholar] [CrossRef]
- Anderson, Brian D. O. 1985. Continuity of the spectral factorization operation. Matematica Aplicada e Computacional 4: 139–56. [Google Scholar]
- Anderson, Brian D. O., and Manfred Deistler. 2008. Properties of zero-free transfer function matrices. SICE Journal of Control, Measurement, and System Integration 1: 284–92. [Google Scholar] [CrossRef] [Green Version]
- Anderson, Brian D. O., and Manfred Deistler. 2009. Properties of zero-free spectral matrices. IEEE Transactions on Automatic Control 54: 2365–75. [Google Scholar] [CrossRef]
- Anderson, Brian D. O., Manfred Deistler, Elisabeth Felsenstein, and Lukas Koelbl. 2016a. The structure of multivariate AR and ARMA systems: Regular and singular systems; the single and the mixed frequency case. Journal of Econometrics 192: 366–73. [Google Scholar] [CrossRef]
- Anderson, Brian D. O., Manfred Deistler, Elisabeth Felsenstein, Bernd Funovits, Lukas Koelbl, and Mohsen Zamani. 2016b. Multivariate AR systems and mixed frequency data: G-identifiability and estimation. Econometric Theory 32: 793–826. [Google Scholar] [CrossRef] [Green Version]
- Anderson, Brian D. O., Manfred Deistler, Elisabeth Felsenstein, Bernd Funovits, Peter Zadrozny, Michael Eichler, Weitian Chen, and Mohsen Zamani. 2012a. Identifiability of regular and singular multivariate autoregressive models from mixed frequency data. Paper presented at the 2012 IEEE 51st IEEE Conference on Decision and Control (CDC), Maui, HI, USA, December 10–13; pp. 184–89. [Google Scholar]
- Anderson, Brian D. O., Manfred Deistler, Weitian Chen, and Alexander Filler. 2012b. Autoregressive models of singular spectral matrices. Automatica 48: 2843–49. [Google Scholar] [CrossRef] [Green Version]
- Bai, Jushan, and Serena Ng. 2002. Determining the number of factors in approximate factor models. Econometrica 70: 191–221. [Google Scholar] [CrossRef] [Green Version]
- Bai, Jushan, and Serena Ng. 2007. Determining the number of primitive shocks in factor models. Journal of Business and Economic Statistics 25: 52–60. [Google Scholar] [CrossRef] [Green Version]
- Barigozzi, Matteo, Marco Lippi, and Matteo Luciani. 2021. Large-dimensional dynamic factor models: Estimation of impulse–response functions with I (1) cointegrated factors. Journal of Econometrics 221: 455–82. [Google Scholar] [CrossRef]
- Bhansali, Raj, and Subir Ghosh. 1999. Asymptotics, Nonparametrics and Time Series. New York: Marcel Dekker. [Google Scholar]
- Bittanti, Sergio, and Patrizio Colaneri. 2009. Periodic Systems: Filtering and Control. London: Springer, vol. 5108985. [Google Scholar]
- Byrnes, Christopher I., and Anders Lindquist. 1997. On the partial stochastic realization problem. IEEE Transactions on Automatic Control 42: 1049–70. [Google Scholar] [CrossRef]
- Caines, Peter E. 2018. Linear Stochastic Systems. Philadeplphia: SIAM. [Google Scholar]
- Chen, Baoline, and Peter A. Zadrozny. 1998. An extended Yule-Walker method for estimating a vector autoregressive model with mixed-frequency data. Advances in Econometrics 13: 47–74. [Google Scholar]
- Chen, Weitian, Brian D. O. Anderson, Manfred Deistler, and Alexander Filler. 2011. Solutions of Yule-Walker equations for singular AR processes. Journal of Time Series Analysis 32: 531–38. [Google Scholar] [CrossRef]
- Chen, Weitian, Brian D. O. Anderson, Manfred Deistler, and Alexander Filler. 2012. Properties of blocked linear systems. Automatica 48: 2520–25. [Google Scholar] [CrossRef] [Green Version]
- Deistler, Manfred, Brian D. O. Anderson, Alexander Filler, Ch Zinner, and Weitan Chen. 2010. Generalized linear dynamic factor models: An approach via singular autoregressions. European Journal of Control 16: 211–24. [Google Scholar] [CrossRef]
- Deistler, Manfred, Lukas Koelbl, and Brian D. O. Anderson. 2017. Non-identifiability of VMA and VARMA systems in the mixed frequency case. Econometrics and Statistics 4: 31–38. [Google Scholar] [CrossRef]
- Doz, Catherine, Domenico Giannone, and Lucrezia Reichlin. 2012. A quasi–maximum likelihood approach for large, approximate dynamic factor models. Review of Economics and Statistics 94: 1014–24. [Google Scholar] [CrossRef] [Green Version]
- Forni, Mario, and Marco Lippi. 2001. The generalized dynamic factor model: Representation theory. Econometric Theory 17: 1113–41. [Google Scholar] [CrossRef] [Green Version]
- Forni, Mario, Domenico Giannone, Marco Lippi, and Lucrezia Reichlin. 2009. Opening the black box: Structural factor models versus structural VARs. Econometric Theory 25: 1319–47. [Google Scholar] [CrossRef] [Green Version]
- Forni, Mario, Luca Gambetti, Marco Lippi, and Luca Sala. 2020. Common Component Structural VARs. Working Paper. Modena, Italy: CEPR, RECent (Center for Economic Research), University of Modena and Reggio Emilia. [Google Scholar]
- Forni, Mario, Marc Hallin, Marco Lippi, and Lucrezia Reichlin. 2000. The generalized dynamic-factor model: Identification and estimation. Review of Economics and Statistics 82: 540–54. [Google Scholar] [CrossRef]
- Forni, Mario, Marc Hallin, Marco Lippi, and Paolo Zaffaroni. 2015. Dynamic factor models with infinite-dimensional factor spaces: One-sided representations. Journal of Econometrics 185: 359–71. [Google Scholar] [CrossRef] [Green Version]
- Hallin, Marc, and Roman Liška. 2007. Determining the number of factors in the general dynamic factor model. Journal of the American Statistical Association 102: 603–17. [Google Scholar] [CrossRef]
- Hannan, Edward James, and Manfred Deistler. 2012. The Statistical Theory of Linear Systems. Philadeplphia: SIAM. [Google Scholar]
- Ho, B. L., and Rudolf E. Kalman. 1966. Effective construction of linear state-variable models from input/output functions. at-Automatisierungstechnik 14: 545–48. [Google Scholar] [CrossRef]
- Lippi, Marco, Manfred Deistler, and Brian Anderson. 2022. High-dimensional dynamic factor models: A selective survey and lines of future research. Econometrics and Statistics. in press. [Google Scholar] [CrossRef]
- McCracken, Michael W., and Serena Ng. 2016. FRED-MD: A Monthly Database for Macroeconomic Research. Journal of Business & Economic Statistics 34: 574–89. [Google Scholar]
- Miranda, Michele, and Paolo Tilli. 2000. Asymptotic spectra of hermitian block toeplitz matrices and preconditioning results. SIAM Journal on Matrix Analysis and Applications 21: 867–81. [Google Scholar] [CrossRef]
- Mitra, Sanjit Kumar. 2011. Digital Signal Processing: A Computer-Based Approach, 4th ed. New York: McGraw-Hill. [Google Scholar]
- Onatski, Alexei. 2009. Testing Hypotheses About the Number of Factors in Large Factor Models. Econometrica 77: 1447–79. [Google Scholar]
- Rabiner, Lawrence R., and Bernard Gold. 1975. Theory and Application of Digital Signal Processing. Englewood Cliffs: Prentice-Hall. [Google Scholar]
- Serra, Stefano. 1999. Spectral and computational analysis of block Toeplitz matrices having nonnegative definite matrix-valued generating functions. BIT Numerical Mathematics 39: 152–75. [Google Scholar] [CrossRef]
- Shibata, Ritei. 1980. Asymptotically efficient selection of the order of the model for estimating parameters of a linear process. The Annals of Statistics 8: 147–64. [Google Scholar] [CrossRef]
- Shumway, Robert H., and David S. Stoffer. 1982. An approach to time series smoothing and forecasting using the EM algorithm. Journal of Time Series Analysis 3: 253–64. [Google Scholar] [CrossRef]
- Sims, Christopher A. 1980. Macroeconomics and reality. Econometrica: Journal of the Econometric Society 48: 1–48. [Google Scholar] [CrossRef] [Green Version]
- Stock, James H., and Mark W. Watson. 2002. Forecasting using principal components from a large number of predictors. Journal of the American Statistical Association 97: 1167–79. [Google Scholar] [CrossRef] [Green Version]
- Stock, James H., and Mark W. Watson. 2005. Implications of Dynamic Factor Models for VAR Analysis. NBER Working Papers 11467. Cambridge, MA, USA: National Bureau of Economic Research. [Google Scholar]
- Stock, James H., and Mark W. Watson. 2016. Dynamic factor models, factor-augmented vector autoregressions, and structural vector autoregressions in macroeconomics. In Handbook of Macroeconomics. Edited by John B. Taylor, Michael Woodford and Harald Uhlig. Amsterdam: Elsevier, vol. 2, pp. 415–25. [Google Scholar]
- Zamani, Mohsen, Wetian Chen, Brian D. O. Anderson, Manfred Deistler, and Alexander Filler. 2011. On the zeros of blocked linear systems with single and mixed frequency data. Paper Presented at the 2011 50th IEEE Conference on Decision and Control and European Control Conference, Orlando, FL, USA, December 12–15; pp. 4312–17. [Google Scholar]
- Zeiger, Paul, and Julia McEwen. 1974. Approximate linear realizations of given dimension via Ho’s algorithm. IEEE Transactions on Automatic Control 19: 153. [Google Scholar] [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |

© 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Anderson, B.D.O.; Deistler, M.; Lippi, M.
Linear System Challenges of Dynamic Factor Models. *Econometrics* **2022**, *10*, 35.
https://doi.org/10.3390/econometrics10040035

**AMA Style**

Anderson BDO, Deistler M, Lippi M.
Linear System Challenges of Dynamic Factor Models. *Econometrics*. 2022; 10(4):35.
https://doi.org/10.3390/econometrics10040035

**Chicago/Turabian Style**

Anderson, Brian D. O., Manfred Deistler, and Marco Lippi.
2022. "Linear System Challenges of Dynamic Factor Models" *Econometrics* 10, no. 4: 35.
https://doi.org/10.3390/econometrics10040035