Factor models - Tony Yates: research, teaching, media, blog ...

Factor models - Tony Yates: research, teaching, media, blog ...

VARs and factors Lecture to Bristol MSc Time Series, Spring 2014 Tony Yates What we will cover Why factor models are useful in VARs Static and dynamic factor models VAR in the factors Factor augmented VAR. Estimation of factors by principal components. Identification in Var in the factors or FAVAR: sign restrictions. Application : Stock and Watsons Disentangling... paper

Some useful references Stock and Watson: the implications of dynam ic factor models for VAR analysis Stock and Watson: Dynamic Factor Models Stock and Watson 'Disentangling the causes of the 2007-2009 recession Bai and Ng survey Wikipedia entry on principal component analy sis! Geweke lecture Dimensionality motivation for factor models Omitting variables from our VAR means our reduced form shocks dont span the structural shocks. Eg Leeper Sims Zha (1996), 13, 18 variable VAR But including more variables mean no of coeffs to be estimated expands by n^2*lags, while number of data points increases by n*T. Central bank tracks 100s of variables. Unless they are wasting time, maybe they should all enter? Exercise: when does the curse of dimensionality bite?

Some interesting research with factor models Quah and Sargent (1992): you can capture many time series with just 2 factors Confirmed by other authors later Echoing early RBC claims, but using much more agnostic framework. contradicting Smets-Wouters (2007) and similar with many shocks. See also Sims and Sargent (1977) More interesting factor model research Stock and Watson: Disentangling the causes of the crisis[sic] There are 8 factors, not 2 or 3! Financial crisis was not a new shock, just larger versions of the old [Disentangling...] Contradicts narrative of the crisis, and other DSGE based work. We will return to this paper in more detail later. Yet more interesting factor model research Stock+Watson

Implications of dynamic factor models for VAR analysis Redoes SVAR identification with factors. Finds? An exercise for you to summarise it. Harrison, Kapetanios, Yates: Estimating TVPDFM models using kernel methods. Rudebusch: survey of macro-finance work on yield curve, including factor modelling. A simple static factor model Yt F t U t F t AF t 1 Z t Y is our vector of observeables, driven by the latent factors F. Factors follow a VAR process as before. Simple dynamic factor model Yt 0 F t 1 F t 1 U t F t AF t 1 Z t Its dynamic because lags of the factors drive the

observeables too. Dynamic factor model written as a static factor model Ft Sometimes convenient to create the enlarged state vector with current and lagged factors. Ft F t 1 And then re-write in terms of current values only of this enlarged state vector. Yt I A Ft Z t

0 1 Ft U t 3 factor, 6 variable example. Yt F Fr Fx t 1t

1 1r 1x U 1 2t 2 2r

2x U 2 r 1t r1 r1 r r1 x r2

r2 r r2 x x 1t x 1 x 1r x 1x

x 2t x 2 x 2r x 2x r 2t A A r A x F A r A rr A rx

Fr A x A xr A xx Fx F Fr Fx t U r1 U r2 U x1 U x2 Z

t 1 Zr Zx t Here we have six variables, including 2 proxies for each of the variables in the simple sticky price model, which we will assume are the factors. Restrictions in the measurement equation of the factor model Yt 1t 1 0

0 2t 2 0 0 r 1t r 2t 0 r1 r

0 0 r2 r 0 x 1t 0 0 r1 r x 2t 0

0 r2 r U 1 U 2 F Fr Fx t U r1 U r2 Ux1 Ux2 Since we have a clear prior about which observables relate to which

economic concept, we might restrict elements of the factor loading matrix lamda. If we didnt, we would get less well determined estimates. Cost is that we have to be confident our restrictions are valid. Factor Augmented VAR [FAVAR] Yt AYt 1 e t it Yt t ft X it i f t u it Imagine [quite realistically] that we thought inflation and interest rates were pretty well measured, but the output gap was not, and we had several alternative proxies for this. We would extract one factor from these output gap measures, then include it in a

vector of observables and estimate a VAR as before. Blessing of dimensionality Obs on Y driven by a single factor F Yit i F t e it n n Now average over both sides of the equation 1/n Yit 1/n i F t e it i 1 i 1 n n 1/n i F t 1/n e it

i 1 n 1/n Yit i 1 p Ft i 1 Invoke assumption that errors uncorrelated with each other, and we get to the result that as n gets large, the average of our observeables uncovers the factor. Estimation Formulation as state-space model suggests estimation using Kalman Filter [putting it in a wide class of estimation problems, eg estimation of a DSGE/RBC model.

KF computes the likelihood for a given parameter value. Then maximise wrt the parameters. Problem: many parameters therefore large dimensional optimisation problem. Can be reduced with priors about loading matrices. Estimation: preliminaries. See Bai and Ng survey. x it i F t e it X t x 1t , x 2t . . . x Nt F F 1 . . . . F T 1 , . . . . N X t F t et X X 1 . . . X N T N X F e, e e 1 , . . . . e N T N Step by step, we stack the entire data set and factor decomposition in matrix form. Variance-decomposition under the factor

model Variance-covariance matrix of the data Contribution of the factors Variance-covariance of the idiosyncratic shocks F , EF t F t I r Terms in the factors disappear, as we are going to use this normalisation to resolve identification. Identification problem in the factor model X F e F FAA 1

F , We are trying to estimate F and lamda on the RHS here. F FA, A 1 But we can see that we can rotate the factors and loadings with any invertible r*r matrix A, and still preserve the equation with the LHS data matrix. X F e F e Identification to resolve the indeterminacy of

the factors and the loadings. F F I rr 1/2 restrictions diagonal rr 1/2 restrictions Estimation by Principal Components T k matrix of factors F This is what we seek in estimation. Note that in finding k factors, we might not search for the true r factors. k N k matrix of loadings k min Sk, s. t. F kF k I k , k k k

,F k D. N Sk NT 1 T x it kF kt 2 i 1 t 1 Like all estimation, finding the factors and loadings is an optimisation problem. By choice of the factors we try to minimise the residual sum of squares!

Recasting the minimisation problem and its solution. k k max trF X XF k F k F F T k F

k Ik T evc 1 XX . . . . evc k XX k k F X T Two ways to proceed. Usually the same. Here we concentrate out lamda from the objective function. We maximise the explained sum of squares. The estimated factors are the k eigenvectors of XX, corresponding to the k largest eigenvalues. Principal components estimation of the factors T

X T 1 X t X t Define sample var cov matrix of observed data t 1 T MinV r , F, V r , F 1/NT X t F t X t F t t 1 Least squares problem subject to N 1 I r 1 Ft N Xt Solution. Lambdahats are scaled

eigenvectors associated with r largest eigenvalues of sigmahat. Principal Components estimation [Bai and Ng, JOE in press] Write our factor model in matrix form. X it i F t e it X F e trX F X F Factor etimation, of the factors and loadings, minimises this objective. Equivalent to the contribution of the idiosyncratic errors. F F I r Dr These are constraints placed on the

estimation. PC estimates of factors and loadings F F1 . . . . Fr ev 1 Z, ev 2 Z. . . . ev r Z/ T Z XX /TN 1 . . . N X F T Factors are the r scaled eigenvectors of the vcov matrix of the data. The loadings are products of the data matrix and the estimated factors. Watch out: papers sometimes use different notation, partly because the procedure only identifies separately the product of the factors and the loadings, and not each element.

Estimation of the full system 2-step procedure. Having estimated the factors by principal components analysis Treat the factors like you would observed data and then estimate the VAR in the factors using your chosen favoured method (MLE, OLS...). Identifying factors using sign restrictions Yt F t U t Assume static factor model, and VAR(1) in the factors. F t AF t 1 Z t EZZ Z P Z P Z P Z CC P Z PZC Just as with VARs in the variables, we can factor the vcov matrix of

shocks to factors, and factor further using an orthonormal matrix C. We draw multiple Cs, and then inspect the sign of the impact on observeables in the same way as before, except now we have to substitute into the measurement equation and premultiply by lamda. Description in words of sign restriction factor identification Example: monetary policy shock. Normal VAR. A mp shock is one that if it drives cb rate up, will drive output down, inflation down. DFM. A mp shock is a shock to the VAR in the factors such that, given the factor loadings estimated in stage 1, if it drives the cb rate up, it also drives the inflation rate down and output down. One point of factor model would be to have many proxies for inflation. So restriction here would be that it would drive all (eg) proxies for inflation down. Or perhaps most of them.

Identification using Cholesky in a factor model 1 Yt 2 x1 x2 t F t AF t 1 Z t 1 0 2

0 F 0 1x Fx 0 2x Ut t Here Im assuming we have four variables we have a prior are two observations each on two different economic concepts, say inflation

and the output gap Cholesky i.d. with factors B0 F t B1 F t 1 E t As before, we seek the elusive B0inv, which now encodes contemp relationship between the factors U t B0 1 E UU U B0 1 EE B0 1 B0 1 chol U P U F hirf A h P U U Yhirf F hirf A h P U U If we are ok with a lower-triangular B0inv connecting the factors then we simply take it to be the cholesky factor of the vcov matrix of residuals in the var in the factors. Impuse response of factors computed as before. But impulse response of

observables requires substituting into the measurement equation. Application: Stock and Watsons Disentangling.. paper Trying to explain the recession. Note all real series show big drop relative to trend. Not surprising therefore that common component [lamdahat*F] explains a lot. Their research question is: Was it bigger versions of old shocks that explain the crisis. ...Or new shocks. 1. Estimate DFM pre 2007 2. Feed in post 2007 factor outturns. 3. Do factors put through old model explain data any worse post 2007? 4. If there was a new factor, you

would expect R^2 to fall. 5. They dont. 6. Conclusion: there was no new factor. 7. Conflicts with narrative that there was a new financial crisis shock. 8. Recall Christiano, Motto Rostagnos risk shock paper. SWs R^2 exercise et X t F t 1 e 2t X 2t Factors and factor loadings estimated over pre-2007 sample R^2 here cant be >1, but it can be <1 (and sometimes is as we will see) If factors do a good job at explaining the series, then should be close to

1. SW: how well do the old factors explain the new data SW: Tests for break in factor loadings Majority of tests accept stability. Tendency to reject caused by change in 1984. That relates to earlier work dating this as the start of the Great Moderation. Implication is that 2007 not responsible for many breaks. SW: indication of existence of new factor et X t F t Compute ratio of first to sum of remaining eigenvalues. Ee t 2

v eig n v1/ i 2 Construct vcov of idiosyncratic shocks, using pre-crisis loadings and factors. vi Large value of this implies more correlation between idiosyncratic shocks. Tests for equality of this ratio before and after crisis. P value of 0.59. SW: evidence of increased factor variance So if it wasnt new factors, then it must be the old ones that increased. This is the sd of lamdahat*F for selected series. You can see it increases during the crisis. Post-script: Stock-Watson and the old two

factor finding They say you need 7 or 8 factors, not 2. The old finding was, they said, based on i) too narrow a set of data, and ii) the early sample period. This is a huge deal in the business cycle literature, but the finding doesnt seem to have attracted all that much attention. Recap Factor models are a way to overcome curse of dimensionality. In fact there is a blessing of dimensionality. Can be combined with VARs: FAVAR, VAR in the factors. Estimated using PCA. Factors and loadings chosen to minimise contribution of idiosyncratic error variance. Stock and Watsons financial crisis application.

Recently Viewed Presentations

  • How To Deal With Demons &amp; Spiritual Oppression

    How To Deal With Demons & Spiritual Oppression

    (1 John 3:8 NKJV) He who sins is of the devil, for the devil has sinned from the beginning. For this purpose the Son of God was manifested, that He might destroy the works of the devil. (1 John 4:4...
  • Strategies used by the Federalists in the struggle for ...

    Strategies used by the Federalists in the struggle for ...

    Civic Virtue cannot be relied upon as the sole support for government. ... Constitutional checks and balances/sep of powers is the best way to promote republicanism. Believed that the way Senators and Representatives were elected would all ow for good...
  • NOTEBOOK SETUP - Paulding County School District

    NOTEBOOK SETUP - Paulding County School District

    NOTEBOOK SETUP. Science 2017-2018. ABOUT ME DECORATED PAGE. T1 TABLE OF CONTENTS. T2. TABLE OF CONTENTS. ODD Left Student Side: EVEN Right Teacher Side ; What's Wrong Sheet. ... What's wrong in the Lab activity. Scientific Methods. 2 TEACHER SIDEDATE-...
  • Cumulative risk of developing VAP with the duration of ...

    Cumulative risk of developing VAP with the duration of ...

    Cumulative risk of developing VAP with the duration of mechanical ventilation. Although length of time with an endotracheal tube in place increases the risk of nosocomial pneumonia, the greatest risk is during the first 2 weeks of intubation.
  • 20th Century Social Changes in America - History Connections

    20th Century Social Changes in America - History Connections

    20th Century Social Changes in America. IB History. SECTION 1: POSTWAR AMERICA. After WWII, returning vets faced a severe housing shortage ... Westberry. v. Sanders (1964), and . Reynolds . v. Sims ... School teacher in Texas. Served as Texas...
  • What are biological catalysts? - TECC Science

    What are biological catalysts? - TECC Science

    What are biological catalysts? Baseline (Flightpath D): To be able to define the terms 'catalyst' and 'enzyme'. Further (Flightpath C&B ): To be able to describe the properties of enzymes. To use the lock and key theory and collision theory...
  • 2008 NSTE ACS 抗血小板治疗 - haoyisheng.com

    2008 NSTE ACS 抗血小板治疗 - haoyisheng.com

    非ST段抬高性ACS抗栓治疗 ——从循证到实践 北京协和医院 心内科 张抒扬 低分子肝素抗凝治疗是NSTEACS抗栓中重要部分 依诺肝素(克赛)是ACC/AHA 指南唯一被推荐使用的低分子肝素 总 结 1、非ST段抬高性ACS治疗的精髓—抗栓 2、抗栓与出血风险 3、循证到实践 Thank you for attention.
  • Introduction to Syntax

    Introduction to Syntax

    Next Time Constituency tests Phrase Structure Rules Introduction to Syntax Linear structure Hierarchical structure Ambiguity Syntax is: The study of sentence formation Subconscious grammatical knowledge Word order Grammaticality Judgments: We went to my grandmother's house. Visiting relatives can be a...