Panel Threshold Regression Models 2018

Chapter 3.
Panel Threshold Regression Models

School of Economics and Management - University of Geneva
Christophe Hurlin (Université of Orléans)
University of Orléans
May 2018
C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 1 / 86

1. Introduction
In econometrics, threshold regression models are a category of

regime-switching models in which
The slope parameters vary according to a "regime" switching

mecanism that depends on a threshold variable,
The regime is observable ex-post, contrary to the Markovian regime

switching models.

1. Introduction
De…nition (threshold regression model)

A typical threshold regression model is given by
yt = α + β00 xt + β10 xt h (qt ; θ ) + εt
where β0 and β1 are K 1 vectors, qt is a threshold variable, θ a vector

of parameters and h (qt ; θ ) a transition function.

1. Introduction
Example (threshold regression model)

If the transition function is a binary function such that
(
1 if qt c
h (qt ; c ) = ,
0 if qt < c
then the model is simply de…ned by
yt = α + β00 xt + β10 xt I(qt c) + εt
where I(.) is the indicator function and c is a location parameter.

1. Introduction
Example (threshold regression model)

If the transition function is a binary function, we have a regime-switching
mechanism for the slope parameters that depends on the threshold
variable and a location parameter c
(
α + β00 + β10 xt + εt if qt c
yt = ,
α + β00 xt + εt if qt < c
In this simple case, we have two "regimes" for the slope parameters, i.e.
β0 and β0 + β1 .

1. Introduction
Remarks
The transition function h (qt ; θ ) may be smoother than a binary

function.
In general, the transition function h (qt ; θ ) is assumed to veri…ed
0 h (qt ; θ ) 1
Sometimes, we have
lim h (qt ; θ ) = 1 lim h (qt ; θ ) = 0

q t !+∞ qt ! ∞

1. Introduction
Example (Logistic transition function)

A logistic transition function is de…ned as:
1
h (qt ; γ, c ) = , γ>0
1 + exp ( γ (qt c ))

1. Introduction
Figure: Logistic transition function with c = 0
0.9
0.8
)
0.7
transition function h(q,
0.6
0.5
0.4
0.3
=0.5
=1
0.2
=5
0.1
0
-6 -4 -2 0 2 4 6
threshold variable q

1. Introduction
Panel data
The threshold regressions in panel data models allow to model the

heterogeneity of the slope parameters.
These models give a parametric approach of the heterogeneity

which is associated to an economic "story" (interpretation).

Introduction
The outline of this chapter is the following:

Section 1: Introduction
Section 2: Panel Threshold Regression (PTR) Model
Section 3: Panel Smooth Threshold Regression (PSTR) Model

Section 2
The Panel Threshold Regression (PTR) Model

2. The panel threshold regression model
Objectives
1 Introduce the panel threshold regression (PTR) model.

2 Understand the link with heterogeneous panel models.
3 Understand the link with time varying parameters panel models.
4 Understand the di¤erences with random coe¢ cient models.

The Panel Threshold Regression (PTR) model has been

introduced by Hansen (1999).
In this paper, threshold regression methods are developed for

non-dynamic panels with individual …xed e¤ects.
Hansen, B. E. 1999. Threshold e¤ects in non-dynamic panels: estimation,

testing, and inference, Journal of Econometrics, 93, 334-368

De…nition (panel threshold regression model)

The panel threshold regression (PTR) model is de…ned as
yit = αi + β10 xit I(qit c) + β20 xit I(qit >c ) + εit
where the dependent variable yit is scalar, αi is a …xed e¤ect, the threshold
variable qit is scalar, the regressor xit is a k vector, and I(.) is the indicator
function and c is a threshold parameter.

Assumptions
The threshold variable is exogeneous or at least predetermined

(qit = yi ,t d with d 1).
For the identi…cation of β1 and β3 , it is required that the elements of

xit are not time invariant.
The threshold variable qit is not time invariant.
The error εit is assumed to be i.i.d. with E (εit ) = 0 and V (εit ) = σ2ε .

Panel Threshold Regression model

An alternative speci…cation of the PTR is
yit = αi + β00 xit + β30 xit I(qit c) + εit
or equivalently
yit = αi + β00 xit + β30 xit h (qit ; c ) + εit
whith h (qit ; c ) = I(qit c) the binary transition function

(
αi + β00 + β30 xit + εit if qit c
yit = ,
αi + β30 xit + εit if qit > c

Figure: Transition function h (qit ; c ) with c = 0
0.9
0.8
transition function h(q,c)
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-6 -4 -2 0 2 4 6

De…nition (heterogeneous panel data model )

The PTR model can be viewed as an heterogeneous and time-varying
parameters panel data model
(
0
αi + β10 xit + εit if qit c
yit = αi + βit xit + εit = ,
where the marginal e¤ect (slope parameters) satisfy

(
∂yit β1 if qit c
= βit = ,
∂xit β if qit > c 2

Heterogeneous panel data model

(
yit = 0
,
At a given time t, two cross-section units i and j may have two di¤erent
slope parameters
∂yit ∂yjt
= β1 6 = = β2 if qit c and qjt > c.
∂xit ∂xjt

Time-varying parameter panel data model

(
yit = ,
A given cross-section unit i may have di¤erent slope parameters at

di¤erent dates t and s
∂yit ∂yis
= β1 6 = = β2 if qit c and qis > c.
∂xit ∂xis

Example (Marginal e¤ect and PTR)

Consider a PTR model with K = 1 regressor such that
yit = αi + 0.5xit I(qit 3) 0.8xit I(qit >3 ) + εit
Then, the marginal e¤ect of the regressor xit on yit is equal to

(
∂yit 0.5 if qit 3
= 0.5I(qit 3 ) 0.8I(qit >3 ) = ,
∂xit 0.8 if qit > 3

Figure: Marginal e¤ect with K = 1, β1 = 0.5, β2 = 0.8 and c = 3
6 1
0.8
5
0.6
0.4
t/dx t
i
3
0.2
i
Threshold variables
Marginal effect dy
2 0
-0.2
1
-0.4
-0.6
-1
-0.8
-2 -1
0 10 20 30 40 50 0 10 20 30 40 50
Time Time

Panel Threshold Regression model

The PTR model is an alternative to the random coe¢ cient model
Panel Threshold Regression βit = f (qit ; c ) βit is a constant term

Economic interpretation (through qit )
Random Coe¢ cient Model βi i.i.d. β, ∆ βi is a random variable
No economic interpretation

Estimation by NLS
The PTR model can be rewritten as:
yit = αi + β0 xit (c ) + εit
xit I(qit c ) β1
xit (c ) = β =
(2K ,1 )
xit I(qit >c ) (2K ,1 )
β2

Estimation by NLS
In order to eliminate the individual e¤ects, we apply the Within
transformation
yit = β0 xit (c ) + εit
T
yit = yit yi yi = T 1
∑ yit
t =1
T
xit (c ) = xit (c ) x i (c ) x i (c ) = T 1
∑ xit (c )
t =1
T
εit = εit εi εi = T 1
∑ εit
t =1

Let us de…ne
0 1 0 1 0 1
yi ,1 x1,i ,1 (c )0 εi ,1
B y C B x1,i ,2 (c )0 C B ε C
yi = B i ,2 C
@ ... A Xi (c ) = B
@ ...
C
A εi = B i ,2 C
@ ... A
(T ,1 ) (T ,2K ) (T ,1 )
0
yit x1,it (c ) εit
0
1 0 1 0 1
y1 X1 (c ) ε1
B y C B X (c ) C B ε C
Y =B 2 C
@ ... A X (c ) = B 2
@ ...
C
A ε =B 2 C
@ ... A
(Tn,1 ) (Tn,2K ) (Tn,1 )
yn Xn (c ) εn

De…nition (given threshold)

For any given threshold c, the slope coe¤cient β can be estimated by
ordinary least squares (OLS).
1
β (c ) = X (c ) 0 X (c )
b X (c ) 0 Y
The vector or residuals is given by
bε (c ) = Y X (c ) b
β (c )
and the sum of squared errors is
SSR (c ) = bε (c )0 bε (c )

De…nition (Threshold estimation)

The estimation if the threshold parameter c is obtained by minimization of
the concentrated sum of squared
b
c = arg min SSR (c )
c 2Θ

Remarks
It is undesirable for a threshold c to be selected which sorts too few

observations into one or the other regime.
This possibility can be excluded by restricting the search in to values

of c such that a minimal percentage of the observations (say, 1% or
5%) lie in each regime.
h i
c 2 Θ = quantile fqit gN ,T
i =1,t =1 , 0.05 , quantile f q g N ,T
it i =1,t =1 , 0.95
Given b
c , we can compute the estimates for β as
b β1 (b
c)
β = β (b
c) =
β2 (b
c)







De…nition (extension to r + 1 regimes)

A general speci…cation of the PTR can be proposed with r + 1 > 2
regimes or r threshold parameters c1 , ..cr .
r
yit = αi + ∑ βj0 xit I(c j 1 <q it cj ) + εit
j =1
with c0 = ∞ and cr +1 = +∞.

Example (PTR with three regimes)

For instance, a three-regimes model (r = 3) with 2 threshold parameters
c1 and c2 is given by
yit = αi + β10 xit I(qit c1 ) + β20 xit I(c1 <qit c2 )

+ β30 xit I(qit >c2 ) + εit

Example (PTR with four regimes)

For instance, a three-regimes model (r = 4) with 3 threshold parameters
c1 , c2 and c3 is given by
yit = αi + β10 xit I(qit c1 ) + β20 xit I(c1 <qit c2 )

+ β30 xit I(c2 <qit c3 ) + β40 xit I(qit >c3 ) + εit

Inference
If one comes to test whether the threshold e¤ect is statically
signi…cant in the model with two regimes, the null hypothesis is
H0 : β 1 = β 2
This null hypothesis corresponds to the hypothesis of no threshold

e¤ect.
Under H0 the model is then equivalent to a linear model.
yit = αi + β0 xit + εit
with β = β1 = β2 .

Inference
The null hypothesis H0 : β1 = β2 can be tested by a standard test.
If we note S0 the sum of squared of the linear model, the approximate

likelihood ratio test of H0 is based on:
SSR0 SRR1 (b
c)
F1 = 2
b
σ
b2 denotes a convergent estimate of σ2 .

where σ

Inference
The main problem is that under the null, the threshold parameter c is
not identi…ed (nuisance parameter).
Consequently, the asymptotic distribution of F1 is not standard and,

in particular, does not correspond to a chi-squared distribution.
This issue has been largely studied in literature devoted to threshold

models, notably since the seminal paper by Davies (1977, 1987).
One solution is to use a bootstrap procedure to determine the

asymptotic distribution of the statistic F1 .

Number of regimes (thresholds)

The same kind of inference procedure can be applied in order to
determining the number of thresholds.
A likelihood ratio test of one threshold versus two thresholds is based

on the statistic
SSR1 (b
c) SSR2 (b
c1 , b
c2 )
F2 = 2
b
σ
where bc1 and b
c2 denote the threshold estimates of the model with
three regimes , and SSR2 (b
c1 , b
c2 ) denotes the corresponding residual
sum of squares.


The hypothesis of one threshold is rejected in favor of two thresholds
if F2 is larger than its critical value.
The corresponding asymptotic p-value can be approximated by

bootstrap simulations (Hansen, 1999).
If the model with two thresholds (three regimes) is not rejected

accepted, we to test the hypothesis of two thresholds (three regimes)
against the alternative of three thresholds (four regimes).


The corresponding likelihood ratio statistic, denoted F3. , is de…ned as:
SSR2 (b
c1 , b
c2 ) SSR3 (b
c1 , b
c2 , b
c3 )
F3 = 2
b
σ
where SSR3 (bc1 , b
c2 , b
c3 ) denotes the residual sum of squares of the
model with four regimes and three threshold parameters.
Thus, a sequential procedure based on F1 , F2 , F3 , etc. allows to

determining the number of regimes

Example
Candelon, Colletaz, Hurlin (2012) investigate the threshold e¤ects in the
productivity of infrastructure investment in developing countries.
(
ai + α1 kit + β1 hit + γ1 xit + εit if qit λ
yit =
ai + α2 kit + β2 hit + γ2 xit + εit if qit > λ.
where yit is the aggregate added value, kit is physical capital, hit is human
capital, xit is infrastructure stock.
Candelon B., Colletaz G., Hurlin C. (2013), Network E¤ects and

Infrastructure Productivity in Developing Countries, Oxford Bulletin of
Economics and Statistics, 75(6), 887-913.





Key Concepts Section 2
1 Panel Threshold Regression Model

2 Heterogeneous and time-varying parameters
3 Non Linear Least Squares
4 Inference for the number of regimes
5 Davies problem

Section 2
The Panel Smooth Threshold Regression

(PSTR) Model

3. The panel smooth threshold regression model
Objectives
1 Introduce the panel smooth threshold regression (PSTR) model.

2 Understand the link with heterogeneous panel models.
3 Understand the link with time varying parameters panel models.

The Panel Smooth Threshold Regression (PSTR) model has been

introduced by Gonzalez, Teräsvirta and van Dijk (2005).
This model can be viewed as a generalization of the PTR model.
González, A., Teräsvirta, T., van Dijk, D., 2005. Panel smooth transition
regression model. Working Paper Series in Economics and Finance, vol. 604.

De…nition (PSTR model)

The PSTR with two extreme regimes can be de…ned as
yit = αi + β00 xit + β10 xit g (qit ; γ, c ) + εit
where g (qit ; γ, c ) is a transition function, qit a threshold variable, c a

location parameter and γ a slope parameter.

Transition function
The transition function g (qit ; γ, c ) is a continuous function of the

observed variable qit
The transition function is normalized to be bounded between 0 and 1.
0 g (qit ; γ, c ) 1

De…nition
Gonzalez, Teräsvirta and van Dijk (2005) consider a logistic transition
function
1
g (qit ; γ, c ) = , γ>0
1 + exp ( γ (qit c ))

Figure: Logistic transition function with c = 0
0.9
0.8
)
0.7
transition function h(q,
0.6
0.5
0.4
0.3
=0.5
=1
0.2
=5
0.1
0
-6 -4 -2 0 2 4 6

Remark 1
If γ ! ∞, the logistic function tends to an indicator function and the
PSTR corresponds to a PTR model
(
1 if qit c
lim g (qit ; γ, c ) = I(qit c ) = ,
γ!∞ 0 if qit < c
yit = αi + β00 xit + β10 xit I(qit c) + εit

Remark 2
If γ ! 0, the logistic function tends to an indicator function and the
PSTR corresponds to a linear panel model
1
lim g (qit ; γ, c ) = 8qit
γ!∞ 2
β0 + β1
yit = αi + β0 xit + εit β=
2

De…nition (heterogeneous panel data model )

The PSTR model can be viewed as an heterogeneous and time-varying
parameters panel data mode since the marginal e¤ect (slope parameters)
satisfy
∂yit
= βit = β0 + β1 g (qit ; γ, c )
∂xit
as soon as qit does not belong to xit , with by convention
β0 βit β0 + β1

Remark 3
Even there are two "extreme" regimes/values for the slope parameters, in
fact there is an in…nity of regimes (possible values) for the slope
parameters βit given the observed values for qit :
βit = β0 + β1 g (qit ; γ, c )
β0 βit β0 + β1

De…nition (generalization to r+1 regimes)

The PSTR model can be generalised to r + 1 extreme regimes as follows:
r
yit = αi + β00 xit + ∑ βj0 xit gj (qit ; γj , cj ) + εit
j =1
where the r transition functions gj (qit ; γj , cj ) depend on the slope

parameters γj and on the location parameters cj .

Marginal e¤ects
If the transition variable does not belong to the set of regressors, the slope
parameters become:
r
δyit
βit = = β0 + ∑ βj gj (qit ; γj , cj )
δxit j =1

Marginal e¤ects
If the transition variable does not belong to the set of regressors, the slope
parameters become:
r r δgj (qit ; γj , cj )
δyit
βit = = β0 + ∑ βj gj (qit ; γj , cj ) + ∑ βj qit
δxit j =1 j =1 δqit

Estimation
The estimation of the parameters of the PSTR model consists of
eliminating the individual e¤ects αi and then in applying NLS to the
transformed model
See Gonzàlez et al. (2005) or Colletaz and Hurlin (2006), for more
details.
González, A., Teräsvirta, T., van Dijk, D., 2005. Panel smooth transition
regression model. Working Paper Series in Economics and Finance, vol. 604.
Colletaz, G., Hurlin, C., 2006. Threshold e¤ects in the public capital
productivity: an international panel smooth transition approach. Document
de Recherche du Laboratoire d’Economie d’Orléans. 2006-1.

Inference
Gonzàlez et al. (2005) propose a testing procedure in order (i ) to test
the linearity against the PSTR model and (ii ) to determine the
number, r , of transition functions, i.e. the number of extreme regimes
which is equal to r + 1.
Testing the linearity in a PSTR model can be done by testing
H0 : γ = 0 or H0 : β0 = β1
But in both cases, the test will be non standard since under H0 the
PSTR model contains unidenti…ed nuisance parameters.

Inference
A possible solution is to replace the transition function gj (qit , γj , cj )

by its …rst-order Taylor expansion around γ = 0 and to test an
equivalent hypothesis in the auxiliary regression
yit = αi + θ 0 xit + θ 1 xit qit + εit
The parameters θ i are proportional to the slope parameter γ of the

transition function.
Thus, testing the linearity of the model against the PSTR simply
consists of testing
H0 : θ 1 = 0

De…nition (linearity/homogeneity test)

The F-statistic for the homogeneity assumption H0 : θ 1 = 0 is then
de…ned by:
LM F = (SSR 0 SSR 1 )/ [SSR 0 /(TN N 1)]
with SSR0 the panel sum of squared residuals under H0 (linear panel
model with individual e¤ects) and SSR1 the panel sum of squared residuals
under H1 (PSTR model with two regimes). Under the null hypothesis, the
F-statistic has an approximate F (1, TN N 1) distribution.

Choice of number of transitions

The logic is similar when it comes to testing the number of transition
functions in the model or equivalently the number of extreme regimes.
We use a sequential approach by testing the null hypothesis of no

remaining nonlinearity in the transition function.

We want to test whether there is one transition function (H0 : r = 1)
or whether there are at least two transition functions (H0 : r = 2).
Let us assume that the model with r = 2 is de…ned as:
yit = αi + β0 xit + β1 xit g1 (qit ; γ1 , c1 ) + β2 xit g2 (qit ; γ2 , c2 ) + εit
The logic of the test consists in replacing the second transition

function by its …rst-order Taylor expansion around γ2 = 0 and then in
testing linear constraints on the parameters in
yit = αi + β0 xit + β1 xit g1 (qit ; γ1 , c1 ) + θ 1 xit qit + eit
and the test of no remaining nonlinearity is simply de…ned by

H0 : θ 1 = 0.


Let us denote SSR0 the panel sum of squared residuals under H0 , i.e.
in a PSTR model with one transition function.
yit = αi + β0 xit + β1 xit g1 (qit ; γ1 , c1 ) + εit
Let us denote SSR1 the sum of squared residuals of the transformed

model.
yit = αi + β0 xit + β1 xit g1 (qit ; γ1 , c1 ) + θ 1 xit qit + eit
The F-statistic LMf can be calculated in the same way by adjusting

the number of degrees of freedom.

Testing procedure
1 Given a PSTR model with r = r , we test the null H0 : r = r
against H1 : r = r + 1.
2 If H0 is not rejected the procedure ends.
3 Otherwise, the null hypothesis H0 : r = r + 1 is tested against
H1 : r = r + 2.
4 The testing procedure continues until the …rst acceptance of H0 .
5 Given the sequential aspect of this testing procedure, at each step of
the procedure the signi…cance level must be reduced by a constant
factor 0 < ρ < 1 in order to avoid excessively large models. Gonzàlez
et al. (2005) suggest ρ = 0.5.








Example
Hurlin, Rabaud and Fouquau (2008) considers a PSTR model for
determining the relative in‡uence of …ve factors on the Feldstein and
Horioka result for OECD countries.
Iit = αi + β0 Sit + β1 Sit g (qit ; c ) + eit
They consider …ve main factors (as potential threshold variable) generally
considered in this literature: (i) economic growth, (ii) demography, (iii)
degree of openness, (iv) country size and (v) current account balance.
Hurlin C., Rabaud I., and Fouquau J. (2008) The Feldstein-Horioka Puzzle:
a PSTR Approach. Economic Modelling, 25, 284-299.





Key Concepts Section 3
1 Panel Smooth Threshold Regression Model

2 Marginal e¤ects (slope parameters)
3 Non Linear Least Squares
4 Inference for the number of regimes

End of Chapter 3
Christophe Hurlin (University of Orléans)

Panel Threshold Regression Models 2018

Uploaded by

Document Information

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

Panel Threshold Regression Models 2018

Uploaded by

Copyright:

Available Formats

Chapter 3.

Panel Threshold Regression Models

Christophe Hurlin (Université of Orléans)

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 1 / 86

In econometrics, threshold regression models are a category of

The slope parameters vary according to a "regime" switching

The regime is observable ex-post, contrary to the Markovian regime

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 2 / 86

De…nition (threshold regression model)

yt = α + β00 xt + β10 xt h (qt ; θ ) + εt

where β0 and β1 are K 1 vectors, qt is a threshold variable, θ a vector

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 3 / 86

Example (threshold regression model)

then the model is simply de…ned by

yt = α + β00 xt + β10 xt I(qt c) + εt

where I(.) is the indicator function and c is a location parameter.

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 4 / 86

Example (threshold regression model)

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 5 / 86

The transition function h (qt ; θ ) may be smoother than a binary

In general, the transition function h (qt ; θ ) is assumed to veri…ed

lim h (qt ; θ ) = 1 lim h (qt ; θ ) = 0

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 6 / 86

Example (Logistic transition function)

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 7 / 86

Figure: Logistic transition function with c = 0

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 8 / 86

The threshold regressions in panel data models allow to model the

These models give a parametric approach of the heterogeneity

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 9 / 86

The outline of this chapter is the following:

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 10 / 86

The Panel Threshold Regression (PTR) Model

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 11 / 86

1 Introduce the panel threshold regression (PTR) model.

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 12 / 86

The Panel Threshold Regression (PTR) model has been

In this paper, threshold regression methods are developed for

Hansen, B. E. 1999. Threshold e¤ects in non-dynamic panels: estimation,

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 13 / 86

De…nition (panel threshold regression model)

yit = αi + β10 xit I(qit c) + β20 xit I(qit >c ) + εit

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 14 / 86

The threshold variable is exogeneous or at least predetermined

For the identi…cation of β1 and β3 , it is required that the elements of

The threshold variable qit is not time invariant.

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 15 / 86

Panel Threshold Regression model

yit = αi + β00 xit + β30 xit I(qit c) + εit

yit = αi + β00 xit + β30 xit h (qit ; c ) + εit

whith h (qit ; c ) = I(qit c) the binary transition function

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 16 / 86

Figure: Transition function h (qit ; c ) with c = 0

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 17 / 86

De…nition (heterogeneous panel data model )

where the marginal e¤ect (slope parameters) satisfy

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 18 / 86

Heterogeneous panel data model

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 19 / 86

Time-varying parameter panel data model

A given cross-section unit i may have di¤erent slope parameters at

C. Hurlin (University of Orléans) Advanced Econometrics II May 2018 20 / 86

Example (Marginal e¤ect and PTR)

yit = αi + 0.5xit I(qit 3) 0.8xit I(qit >3 ) + εit