Time Series Models and Applications

Part 10: Time Series Applications [ 1/62]
Econometric Analysis of Panel Data
William Greene
Department of Economics
University of South Florida
Time Series Applications

 Panel Data Time Series Models
 Univariate Time Series
 Autocorrelation in Regression
 Vector Autoregression
 ARCH and GARCH Models
 Macroeconomic Data
 Nonstationarity and Integrated Series
 Unit Roots
 Cointegration
Cross Country Growth Convergence

Solow - Swan Growth Model

Yi,t  K i,t (A i,tL i,t )1
A i,t  index of technology
K i,t  capital stock = Ii,t-1 + (1-)K i,t-1
Ii,t 1  s i Yi,t 1  Investment  Savings
L i,t  labor  (1  ni )L i,t 1
Equilibrium Model for Steady State Income
logYi,t  i  i t   i log Yi,t 1  i,t
i  (1-i )gi , gi =technological growth rate
i  "convergence" parameter; 1- i = rate of convergence to steady state
Cross country comparisons of convergence rates are the focus of study.
A Heterogeneous Dynamic Model
logYi,t  i  i log Yi,t 1  i x it  i,t

i
Long run effect of interest is i 
1  i

Average (over countries) effect:  or
1
(1) "Fixed Effects:" Separate regressions, then average results.
(2) Country means (over time) - can be manipulated to produce
consistent estimators of desired parameters
(3) Time series of means across countries (does not work at all)
(4) Pooled - no way to obtain consistent estimates.
(5) Mixed Fixed (Weinhold - Hsiao): Build separate i into the
equation with "fixed effects," treat i    ui as random.
“Fixed Effects” Approach

i 
i  , =
1  i 1
(1) Separate regressions;  î , ˆ
î ,  i
ˆ 1 N ˆ i 1 î or
(2) Average estimates =   i1  Ni1
N 1 î N
ˆ= (1 / N)Ni1ˆ
i
Function of averages: 
î
1  (1 / N)Ni1
In each case, each term i has variance O(1/Ti )

N
Each average has variance O(1/N)  i=1
(1/N)O(1/Ti )
Expect consistency of estimates of long run effects.
Country Means
(Average the T observations for each i)
i 
i  , =
1  i 1
logYi  i  i log Yi,1  i x i  i
i contains log Yi,1
Estimates are inconsistent. But,
 tT1 log y it
logYi  ,
Ti
 tT2 log y i,t 1 y i,T  y i,0
log Yi,1   logYi   logYi   T (y) / Ti
Ti Ti
logYi  i  i log Yi,1  i x i  i  i  i (logYi   T (y) / Ti )  i x i  i
i i i i
  xi   T (y) / Ti 
1  i 1  i 1  i 1  i
Country Means (cont.)

logYi  i  i log Yi,1  i x i  i
 T (y) y i,T  y i,0
 i  i (logYi   T (y) / Ti )  i x i  i   0
Ti Ti
i i i i
  xi   T (y) / Ti 
1   i 1  i 1  i 1  i
(1) Let = i0  i / (1  i ), expect i0  0 to be random
i =i /(1  i ) and i   to be random
(2) Level variable x i should be uncorrelated with change (logYi   T (y) / Ti ).
Regression of logYi on 1, x i should give consistent estimates of 0 and .
Time Series of Means

(Average across countries)
log y t  (1 / N)Ni1 log y i,t
=  + (1/N)Ni1i y i,t 1  (1/N)Ni1i x i,t  (1/N)Ni1 i,t
Use =(1/N)Ni1i so i    (i  ) then
(1/N)Ni1i log y i,t 1   log y 1,t  (1/N)Ni1 (i  ) log y i,t 1
Likewise for (1/N)Ni1i x i,t
log y t     log y 1,t  x t  ( t  (1/N)Ni1 (i  ) log y i,t 1 ...)
Disturbance is correlated with the regressor. There is
no way out, and by construction no instrumental variable
that could be correlated with the regressor and not the
disturbance. This is hopeless.
Pooling
Essentially the same as the time series case.

OLS and GLS are inconsistent
There could be no instrument that would work
(by construction)
A Mixed/Fixed Approach
log y i,t  i  Ni1idi,t log y i,t 1  i x i,t  i,t

di,t = country specific dummy variable.
Treat i and i as random, i is a 'fixed effect.'
This model can be fit consistently by OLS and
efficiently by GLS.
A Mixed Fixed Model Estimator
log y i,t  i  Ni1 idi,t log y i,t 1  x i,t  (w i x i,t  i,t )

i    w i
Heteroscedastic : Var[w i x i,t  i,t ]=2w x i,t
2
 2
Use two step least squares.
(1) Linear regression of logy i,t on dummy variables, dummy
variables times logy i,t-1 and x i,t .
(2) Regress squares of OLS residuals on x i,t2 and 1 to
estimate 2w and 2 .
(3) Return to (1) but now use weighted least squares.
Nair-Reichert and Weinhold on Growth

Weinhold (1996) and Nair–Reichert and Weinhold (2001) analyzed growth and
development in a panel of 24 developing countries observed for 25 years, 1971–1995. The
model they employed was a variant of the mixed-fixed model proposed by Hsiao (1986,
2003). In their specification,
GGDPi,t = αi + γi dit GGDPi,t-1 + β1i GGDIi,t-1 + β2i GFDIi,t-1

+ β3i GEXPi,t-1 + β4 INFLi,t-1 + εi,t
GGDP = Growth rate of gross domestic product,

GGDI = Growth rate of gross domestic investment,
GFDI = Growth rate of foreign direct investment (inflows),
GEXP = Growth rate of exports of goods and services,
INFL = Inflation rate.
The constant terms and coefficients on the lagged dependent variable are country specific.
The remaining coefficients are treated as random, normally distributed, with means βk and
unrestricted variances. They are modeled as uncorrelated. The model was estimated using
a modification of the Hildreth–Houck–Swamy method
ki  k   k wki
wki ~ N [0,1]
0.757
5.67
6.45
22.18
Heterogeneous Dynamic Models

i
long run effect of interest is i 
1  i
See :
Pesaran,H., Smith,R., Im,K.," Estimating Long-Run Relationships
From Dynamic Heterogeneous Panels," Journal of Econometrics, 1995.
(Repeated with further study in Matyas and Sevestre, The
Econometrics of Panel Data.
Smith, J., notes, Applied Econometrics, Dynamic Panel Data Models,
University of Warwick.
http://www2.warwick.ac.uk/fac/soc/economics/staff/faculty/jennifersmith/panel/
Weinhold, D., "A Dynamic "Fixed Effects" Model for Heterogeneous
Panel Data," London School of Economics, 1999.
TIME SERIES DATA

Modeling an Economic
Time Series
 Observed y0, y1, …, yt,…
 What is the “sample”
 Random sampling?
 The “observation window”
Estimators
 Functions of sums of observations
 Law of large numbers?
 Nonindependent observations
 What does “increasing sample size” mean?
 Asymptotic properties? (There are no finite
sample properties.)
Interpreting a Time Series

 Time domain: A “process”
 y(t) = ax(t) + by(t-1) + …
 Regression like approach/interpretation
 Frequency domain: A sum of terms
 y(t) =
 j  j Cos( j t )  (t )
 Contribution of different frequencies to the observed
series.
 (“High frequency data and financial econometrics
– “frequency” is used slightly differently here.)
For example,…
Decomposed
Studying the Frequency Domain

 Cannot identify the number of terms
 Cannot identify frequencies from the time series
 Deconstructing the variance, autocovariances
and autocorrelations
 Contributions at different frequencies
 Apparent large weights at different frequencies
 Using Fourier transforms of the data
 Does this provide “new” information about the
series?
Stationary Time Series

 zt = b1yt-1 + b2yt-2 + … + bPyt-P + et
 Autocovariance: γk = Cov[yt,yt-k]
 Autocorrelation: k = γk / γ0
 Stationary series: γk depends only on k, not on t
 Weak stationarity: E[yt] is not a function of t, E[yt * yt-s] is not a
function of t or s, only of |t-s|
 Strong stationarity: The joint distribution of [yt,yt-1,…,yt-s] for any
window of length s periods, is not a function of t or s.
 A condition for weak stationarity: The smallest root of the
characteristic polynomial: 1 - b1z1 - b2z2 - … - bPzP = 0, is greater
than one.
 The unit circle
 Complex roots
 Example: yt = yt-1 + ee, 1 - z = 0 has root z = 1/ ,
| z | > 1 => |  | < 1.
The characteristic polynomial is 1 - 1.220175z - (-0.262198)z 2 = 0

Side Issue
How does y(t) = 1.220175 y(t-1) - 0.262198 y(t-2) + a behave?
y(t) = 1.220175 y(t-1) + a is obviously explosive.
1.220175 0.262198
How to tell: A =  
 1 0 
Smallest (possibly complex) root must be greater than 1.0.
Stationary (et) vs.

Nonstationary (yt) Series
1 2 .5 9
8 .0 1
3 .4 3
V a ri a b l e
-1 . 1 5
-5 . 7 3
-1 0 . 3 1
0 20 40 60 80 100
T
YT ET
The Lag Operator

 Lc = c when c is a constant
 Lxt = xt-1
 L2 xt = xt-2
 LPxt + LQxt = xt-P + xt-Q
 Polynomials in L: yt = B(L)yt + et
e.g., B(L) = (1.22L – 0.262L2)
 A(L) yt = et
 Invertibility: yt = [A(L)]-1 et
Inverting a Stationary Series

 yt= yt-1 + et  (1- L)yt = et
 yt = [1- L]-1 et = et + et-1 + 2et-2 + …
1
 1  (L)  (L)  (L)  ...
2 3
1  L
 Stationary series can be inverted
 Autoregressive vs. moving average form of series
VECTOR
AUTOREGRESSION
Vector Autoregression
The vector autoregression (VAR) model is one of the most successful, flexible,
and easy to use models for the analysis of multivariate time series. It is
a natural extension of the univariate autoregressive model to dynamic multivariate
time series. The VAR model has proven to be especially useful for
describing the dynamic behavior of economic and financial time series and
for forecasting. It often provides superior forecasts to those from univariate
time series models and elaborate theory-based simultaneous equations
models. Forecasts from VAR models are quite flexible because they can be
made conditional on the potential future paths of specified variables in the
model.
In addition to data description and forecasting, the VAR model is also
used for structural inference and policy analysis. In structural analysis, certain
assumptions about the causal structure of the data under investigation
are imposed, and the resulting causal impacts of unexpected shocks or
innovations to specified variables on the variables in the model are summarized.
These causal impacts are usually summarized with impulse response
functions and forecast error variance decompositions.
Eric Zivot: http://faculty.washington.edu/ezivot/econ584/notes/varModels.pdf
VAR
y1 (t )  11 y1 (t  1)  12 y2 (t  1)  13 y3 (t  1)  1 x(t )  1 (t )
y2 (t )   21 y1 (t  1)   22 y2 (t  1)   23 y3 (t  1)   2 x(t )   2 (t )
y3 (t )   31 y1 (t  1)   32 y2 (t  1)   33 y3 (t  1)   3 x(t )   3 (t )
(In Zivot's examples,
1. Exchange rates
2. y(t)=stock returns, interest rates, indexes of industrial production,
rate of inflation
VAR Formulation
y (t) = y (t-1) + x(t) + (t)
SUR with identical regressors.
Granger Causality: Nonzero off diagonal elements in 
y1 (t )  11 y1 (t  1)  12 y2 (t  1)  13 y3 (t  1)  1 x(t )  1 (t )
y2 (t )   21 y1 (t  1)   22 y2 (t  1)   23 y3 (t  1)  2 x(t )   2 (t )
y3 (t )   31 y1 (t  1)   32 y2 (t  1)   33 y3 (t  1)  3 x(t )   3 (t )
Hypothesis: y2 does not Granger cause y1: 12 =0
Impulse Response
y (t) = y (t-1) + x(t) + (t)
By backward substitution or using the lag operator (text, 943)
y (t)  x(t)  x(t-1)   2 x(t-2) +... (ad infinitum)
+ (t)  (t-1)   2 (t-2) + ...
[ P must converge to 0 as P increases. Roots inside unit circle.]
Consider a one time shock (impulse) in the system,  =  2 in period t
Consider the effect of the impulse on y1 ( s ), s=t, t+1,...
Effect in period t is 0.  2 is not in the y1 equation.
 2 affects y2 in period t, which affects y1 in period t+1. Effect is 12  
In period t+2, the effect from 2 periods back is ( 2 )12  
... and so on.
Zivot’s Data
Impulse Responses
ARCH AND GARCH

MODELS
GARCH Models: A Model for Time Series

with Latent Heteroscedasticity
Bollerslev/Ghysel, 1974
ARCH Model
GARCH Model
Estimated GARCH Model

----------------------------------------------------------------------
GARCH MODEL
Dependent variable Y
Log likelihood function -1106.60788
Restricted log likelihood -1311.09637
Chi squared [ 2 d.f.] 408.97699
Significance level .00000
McFadden Pseudo R-squared .1559676
Estimation based on N = 1974, K = 4
GARCH Model, P = 1, Q = 1
Wald statistic for GARCH = 3727.503
--------+-------------------------------------------------------------
Variable| Coefficient Standard Error b/St.Er. P[|Z|>z] Mean of X
--------+-------------------------------------------------------------
|Regression parameters
Constant| -.00619 .00873 -.709 .4783
|Unconditional Variance
Alpha(0)| .01076*** .00312 3.445 .0006
|Lagged Variance Terms
Delta(1)| .80597*** .03015 26.731 .0000
|Lagged Squared Disturbance Terms
Alpha(1)| .15313*** .02732 5.605 .0000
|Equilibrium variance, a0/[1-D(1)-A(1)]
EquilVar| .26316 .59402 .443 .6577
--------+-------------------------------------------------------------
MACROECONOMIC DATA
Analysis of Macroeconomic Data

 Integrated series
 The problem with regressions involving nonstationary
series
 Spurious regressions
 Unit roots and misleading relationships
 Solutions to the “problem”
 Random walks and first differencing
 Removing common trends
 Cointegration: Formal solutions to regression models
involving nonstationary data
 Extending these results to panels
 Large T and small T cases.
 Parameter heterogeneity across countries
Nonstationary Data
Integrated Series
Stationary Data
Integrated Processes
 Integration of order (P) when the P’th differenced
series is stationary
 Stationary series are I(0)
 Trending series are often I(1). Then yt – yt-1 = yt is
I(0). [Most macroeconomic data series.]
 Accelerating series might be I(2). Then
(yt – yt-1)- (yt – yt-1) = 2yt is I(0)
Historic Hyperinflations
Interwar Germany, Hungary 1946, Zimbabwe 2007-2008
Money stock in hyperinflationary economies.
Price level in Venezuela in 2016 - 2017
A Unit Root?
 How to test for  = 1?
 By construction: εt – εt-1 = ( - 1)εt-1 + ut
 Test for  = ( - 1) = 0 using regression?
 Variance goes to 0 faster than 1/T. Need a new
table; can’t use standard t tables.
 Dickey – Fuller tests
 This invokes the possibility of unit roots in
economic data. (Are there?)
 Nonstationary series
 Implications for conventional analysis
Unit Root Tests
y t   0 y t 1    t  Ll1 l y t l   t
Different restrictions on parameters produce the model.
The parameter of interest is  0.
Augmented Dickey-Fuller Tests:
Unit root test H0 :  < 1 vs. H1 :   1.
KPSS Tests:
Null hypothesis is stationarity. Alternative is broadly
defined as nonstationary.
KPSS Test-1
KPSS Test-2
Panel Unit Root Tests

Purchasing Power Parity

Cross country purchasing power parity hypothesis:
Ei,t  i  iPi,t  i,t
Ei,t  log of exchange rate country i with U.S.
Pi,t  log of aggregate consumer expenditure price ratio
Hypothesis of PPP is i  1.
Data on numerous countries (large N) and many periods (large T)
Standard simple regressions based on SUR model?
(See Pedroni, P., "Purchasing Power Parity Tests in Cointegrated
Panels," ReStat, Nov, 2001, p. 727 and related cited papers.)
Application
“Some international evidence on price determination: a non-stationary

panel Approach,” Paul Ashworth, Joseph P. Byrne, Economic Modelling,
20, 2003, p. 809-838.
80 quarters, 13 OECD countries
log pi,t = β0 + β1log(unit labor costi,t)

+ β2 log(world price,t)
+ β3 log(intermediate goods pricei,t)
+ β4 (log-output gapi,t) + εi,t
Various tests for unit roots and cointegration
Implications
 Separate analyses by country

 How to combine data and test statistics
 Cointegrating relationships across countries
COINTEGRATION
Cointegrated Processes:
Real DPI and Real Consumption
6967
5775
4582
Va ria b le
3390
2198
1006
0 41 82 123 164 205
Ob serv.#
REALDPI REALCONS
Lack of Cointegration
Divergent Series?
2 1 .0 4
1 6 .8 5
1 2 .6 7
Va ria b le
8 .4 8
4 .3 0
.1 2
0 20 40 60 80 100
Ob serv.#
YT XT
Cointegration
 X(t) and y(t) are obviously I(1)
 Looks like any linear combination of x(t) and y(t) will
also be I(1)
 Does a model y(t) = x(t) + u(t) where u(t) is I(0)
make any sense? How can u(t) be I(0)?
 In fact, there is a linear combination, [1,-] that is I(0).
 y(t) = .1*t + noise, x(t) = .2*t + noise
 y(t) and x(t) have a common trend
 y(t) and x(t) are cointegrated.
Cointegrated Variables?
Cointegration and I(0) Residuals

2 .0 0
1 .5 0
1 .0 0
.5 0
ET
.0 0
-. 5 0
-1 . 0 0
-1 . 5 0
0 17 34 51 68 85 102
Ob serv.#
Cointegrating Relationships
 Implications:
 Long run vs. short run relationships
 Problems of spurious regressions (as
usual)
 Problem for existing empirical
studies: Regressions involving
variables of different integration.
E.g., regressions of flows on stocks
Money demand example

APPENDIX
Autocorrelation in Regression
 Yt = b’xt + εt
 Cov(εt, εt-1) ≠ 0
 Ex. RealConst = a + bRealIncome + εt U.S. Data, quarterly, 1950-2000
Autocorrelation
 How does it arise?
 What does it mean?
 Modeling approaches
 Classical – direct: corrective
 Estimation that accounts for autocorrelation
 Inference in the presence of autocorrelation
 Contemporary – structural
 Model the source
 Incorporate the time series aspect in the model
Regression with Autocorrelation
 yt = xt’b + et, et = et-1 + ut

 (1- L)et = ut  et = (1- L)-1ut
 E[et] = E[ (1- L)-1ut] = (1- L)-1E[ut] = 0
 Var[et] = (1- L)-2Var[ut] = 1+ 2u2 + …
= u2/(1- 2)
 Cov[et,et-1] = Cov[et-1 + ut, et-1] =
= Cov[et-1,et-1]+Cov[ut,et-1]
=  u2/(1- 2) + 0
OLS vs. GLS with Autocorrelation

 OLS
 Unbiased?
 Consistent: (Except in the presence of a lagged
dependent variable)
 Inefficient
 GLS
 Consistent and efficient
+----------------------------------------------------+
| Ordinary least squares regression |
| LHS=REALCONS Mean = 2999.436 |
| Autocorrel Durbin-Watson Stat. = .0920480 |
| Rho = cor[e,e(-1)] = .9539760 |
+----------------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient | Standard Error |t-ratio |P[|T|>t] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
Constant -80.3547488 14.3058515 -5.617 .0000
REALDPI .92168567 .00387175 238.054 .0000 3341.47598
| Robust VC Newey-West, Periods = 10 |
Constant -80.3547488 41.7239214 -1.926 .0555
REALDPI .92168567 .01503516 61.302 .0000 3341.47598
+---------------------------------------------+
| AR(1) Model: e(t) = rho * e(t-1) + u(t) |
| Final value of Rho = .998782 |
| Iter= 6, SS= 118367.007, Log-L=-941.371914 |
| Durbin-Watson: e(t) = .002436 |
| Std. Deviation: e(t) = 490.567910 |
| Std. Deviation: u(t) = 24.206926 |
| Durbin-Watson: u(t) = 1.994957 |
| Autocorrelation: u(t) = .002521 |
| N[0,1] used for significance levels |
+---------------------------------------------+
+---------+--------------+----------------+--------+---------+----------+
|Variable | Coefficient | Standard Error |b/St.Er.|P[|Z|>z] | Mean of X|
+---------+--------------+----------------+--------+---------+----------+
Constant 1019.32680 411.177156 2.479 .0132
REALDPI .67342731 .03972593 16.952 .0000 3341.47598
RHO .99878181 .00346332 288.389 .0000
Detecting Autocorrelation
 Use residuals directly (Durbin and Watson)
 d= Tt  2 (e t et 1 ) 2  / Tt 1et2   2(1  r )
 Assumes normally distributed disturbances, strictly
exogenous regressors
 Variable addition (LM test) (Godfrey)
 Based on yt = ’xt + (εt-1 + ut)
 Use regression residuals et and test  = 0 in
regression of et on xt and et-1
 Assumes consistency of b.
Reinterpreting Autocorrelation
Regression form
yt   ' xt  t , t  t 1  ut
Error Correction Form
yt  yt 1   '(xt  xt 1 )  ( yt 1   ' xt 1 )  ut , (    1)
 ' xt  the equilibrium
The model describes adjustment of y t to equilibrium when x t changes.

Time Series Models and Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Time Series Models and Applications

Uploaded by

Copyright:

Available Formats

Part 10: Time Series Applications [ 1/62]

Econometric Analysis of Panel Data

Time Series Applications

Cross Country Growth Convergence

A Heterogeneous Dynamic Model

logYi,t  i  i log Yi,t 1  i x it  i,t

“Fixed Effects” Approach

Country Means (cont.)

Time Series of Means

Essentially the same as the time series case.

log y i,t  i  Ni1idi,t log y i,t 1  i x i,t  i,t

A Mixed Fixed Model Estimator

log y i,t  i  Ni1 idi,t log y i,t 1  x i,t  (w i x i,t  i,t )

Nair-Reichert and Weinhold on Growth

GGDPi,t = αi + γi dit GGDPi,t-1 + β1i GGDIi,t-1 + β2i GFDIi,t-1

GGDP = Growth rate of gross domestic product,

Heterogeneous Dynamic Models

TIME SERIES DATA

Interpreting a Time Series

Studying the Frequency Domain

Stationary Time Series

The characteristic polynomial is 1 - 1.220175z - (-0.262198)z 2 = 0

Stationary (et) vs.

The Lag Operator

Inverting a Stationary Series

ARCH AND GARCH

GARCH Models: A Model for Time Series

Estimated GARCH Model

Analysis of Macroeconomic Data

Unit Root Tests

Panel Unit Root Tests

Purchasing Power Parity

“Some international evidence on price determination: a non-stationary

80 quarters, 13 OECD countries

log pi,t = β0 + β1log(unit labor costi,t)

 Separate analyses by country

Cointegration and I(0) Residuals

Money demand example

Regression with Autocorrelation

 yt = xt’b + et, et = et-1 + ut

OLS vs. GLS with Autocorrelation

You might also like