Outline

1. Some basic ideas

2. The problem of estimation: OLS method
3. Classical Normal Linear Regression Model
(CNLRM)
4. Interval estimation and hypothesis testing
5. Extensions of the two variable linear regression
model

10/24/2017 Mai VU-FIE-FTU 2

5. Extension of the two variable linear
regression model
 In this section, we consider some commonly used
regression models that may be nonlinear in the variables
but are linear in the parameters or that can be made so by
suitable transformations of the variables. In particular, we
discuss the following regression models:
 The log-linear model
 Semi-log models
 Reciprocal models
 The logarithmic reciprocal model
 We discuss the special features of each model, when they
are appropriate, and how they are estimated.
10/24/2017 Mai VU-FIE-FTU 3

5.1. How to measure elasticity: the log-

linear model
 Consider the following model, known as the exponential
regression model:
𝛽
𝑌𝑖 = 𝛽1 𝑋𝑖 2 𝑒 𝑢𝑖 (5.1.1)
Which may be expressed alternatively as:
𝑙𝑛𝑌𝑖 = 𝑙𝑛𝛽1 + 𝛽2 𝑙𝑛𝑋𝑖 + 𝑢𝑖 (5.1.2)
Where ln= natural log. We can rewrite as:
𝑙𝑛𝑌𝑖 = 𝛼 + 𝛽2 𝑙𝑛𝑋𝑖 + 𝑢𝑖 (5.1.3)
 Where 𝛼 = 𝑙𝑛𝛽1 . This model is linear in the parameters α and
β2, linear in the logarithms of the variables Y and X, and can be
estimated by OLS regression.
 Because of this linearity, such models are called log-log,
double-log, or log-linear models.

10/24/2017 Mai VU-FIE-FTU 4

5.1. How to measure elasticity: the log-
linear model
 If the assumptions of the classical linear regression
model are fulfilled, the parameters can be estimated
by the OLS method by letting:
𝑌𝑖∗ = 𝛼 + 𝛽2 𝑋𝑖∗ + 𝑢𝑖 (5.1.4)
 Where 𝑌𝑖∗ = 𝑙𝑛𝑌𝑖 and 𝑋𝑖∗ = 𝑙𝑛𝑋𝑖 .
 The OLS estimators 𝛼ො and 𝛽መ2 obtained will be best
linear unbiased estimators of α and β2, respectively.

10/24/2017 Mai VU-FIE-FTU 5

5.1. How to measure elasticity: the log-

linear model
 One attractive feature of the log-log model, which has
made it popular in applied work, is that the slope
coefficient β2 measures the elasticity of Y with
respect to X, that is, the percentage change in Y for a
given (small) percentage change in X.
 Thus, if Y represents the quantity of a commodity
demanded and X its unit price, β2 measures the price
elasticity of demand, a parameter of considerable
economic interest.

10/24/2017 Mai VU-FIE-FTU 6

 If the relationship between quantity demanded and price is as shown in
the Figure (a), the double-log transformation as shown in Figure (b) will
then give the estimate of the price elasticity (−β2).

10/24/2017 Mai VU-FIE-FTU 7

5.1. How to measure elasticity: the log-

linear model
 Two special features of the log-linear model:
 The model assumes that the elasticity coefficient between Y
and X, β2, remains constant throughout, hence the
alternative name constant elasticity model.
 Although 𝛼 ො and 𝛽መ2 are unbiased estimates of α and β2, β1
(the parameter entering the original model) when estimated
as 𝛽መ1 = antilog (𝛼)
ො is itself a biased estimator.
 In the two-variable model, the simplest way to decide whether
the log-linear model fits the data is to plot the scatter-gram of
lnYi against lnXi and see if the scatter points lie approximately on
a straight line.

10/24/2017 Mai VU-FIE-FTU 8

5.2. Semi-log models: log-lin and lin-log
models
 The Log–Lin Model
 The Lin–Log Model

10/24/2017 Mai VU-FIE-FTU 9

The Log-lin model: how to measure the

growth rate
 Economists, businesspeople, and governments are
often interested in finding out the rate of growth of
certain economic variables, such as population, GNP,
money supply, employment, productivity, and trade
deficit.
 Suppose we want to find out the growth rate of
personal consumption expenditure on services. Let Yt
denote real expenditure on services at time t and Y0
the initial value of the expenditure on services. Then
we have:
𝑌𝑡 = 𝑌0 (1 + 𝑟)𝑡 (5.2.1)

10/24/2017 Mai VU-FIE-FTU 10

The Log-lin model: how to measure the
growth rate
 Where r is the compound rate of growth of Y. Taking the
natural logarithm, we can write:
𝑙𝑛𝑌𝑡 = 𝑙𝑛𝑌0 + 𝑡𝑙𝑛(1 + 𝑟) (5.2.2)
 Now letting: 𝛽1 = 𝑙𝑛𝑌0 and 𝛽2 = ln(1 + 𝑟), we have:
lnYt = β1 + β2t (5.2.3)
 Adding the disturbance term, we obtain:
lnYt = β1 + β2t+ ut (5.2.4)
 This model is like any other linear regression model in that
the parameters β1 and β2 are linear. The only difference is
that the regressand is the logarithm of Y and the regressor
is “time,” which will take values of 1, 2, 3, etc.

10/24/2017 Mai VU-FIE-FTU 11

The Log-lin model: how to measure the

growth rate
 Models like this are called semilog models because only
one variable (in this case the regressand) appears in the
logarithmic form.
 For descriptive purposes a model in which the regressand is
logarithmic will be called a log-lin model.
 In this model the slope coefficient measures the constant
proportional or relative change in Y for a given absolute
change in the value of the regressor (in this case the variable
t), that is:
𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑟𝑒𝑔𝑟𝑒𝑠𝑠𝑎𝑛𝑑
𝛽2 = (5.2.5)
𝑎𝑏𝑠𝑜𝑙𝑢𝑡𝑒 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑟𝑒𝑔𝑟𝑒𝑠𝑠𝑜𝑟

10/24/2017 Mai VU-FIE-FTU 12

The Log-lin model: how to measure the
growth rate
 Linear Trend Model. Instead of estimating model (5.2.4),
researchers sometimes estimate the following model:
Yt = β1 + β2t + ut (5.2.6)
 That is, instead of regressing the log of Y on time, they
regress Y on time, where Y is the regressand under
consideration.
 Such a model is called a linear trend model and the time
variable t is known as the trend variable.
 If the slope coefficient in (5.2.6) is positive, there is an
upward trend in Y, whereas if it is negative, there is a
downward trend in Y.

10/24/2017 Mai VU-FIE-FTU 13

The Log-lin model: how to measure the

growth rate
 The choice between the growth rate model and the
linear trend model will depend upon whether one is
interested in the relative or absolute change in the
expenditure on services, although for comparative
purposes it is the relative change that is generally more
relevant.

10/24/2017 Mai VU-FIE-FTU 14

The Lin-Log Model
 Unlike the growth model just discussed, in which we
were interested in finding the percent growth in Y for
an absolute change in X, suppose we now want to find
the absolute change in Y for a percent change in X. A
model that can accomplish this purpose can be written
as:
Yi = β1 + β2 ln Xi + ui (5.2.7)
 For descriptive purposes we call such a model a lin–
log model.

10/24/2017 Mai VU-FIE-FTU 15

The Lin-Log Model

 We can interpret the slope coefficient β2 as follows:
𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑌 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑌
𝛽2 = = (5.2.8)
𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 ln 𝑋 𝑟𝑒𝑙𝑎𝑡𝑖𝑣𝑒 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑋
 Symbolically, we have:
∆𝑌
𝛽2 = (5.2.9)
∆𝑋/𝑋
 Where ∆ denotes a small change. Equation (5.2.9) can
be written equivalently as:
∆Y = β2(∆X/X) (5.2.10)
 This equation states that the absolute change in Y ( =
∆ Y) is equal to slope times the relative change in X.
10/24/2017 Mai VU-FIE-FTU 16
The Lin-Log Model
 When is a lin–log model like useful?
 An interesting application has been found in the so-
called Engel expenditure models, named after the
German statistician Ernst Engel, 1821–1896.
 Engel postulated that “the total expenditure that is
devoted to food tends to increase in arithmetic
progression as total expenditure increases in
geometric progression.”

10/24/2017 Mai VU-FIE-FTU 17

5.3. Reciprocal models

 Models of the following type are known as reciprocal
models:
1
𝑌𝑖 = 𝛽1 + 𝛽2 + 𝑢𝑖 (5.2.11)
𝑋𝑖
 Although this model is nonlinear in the variable X because
it enters inversely or reciprocally, the model is linear in β1
and β2 and is therefore a linear regression model.
 As X increases indefinitely, the term β2(l/X) approaches
zero (note: β2 is a constant) and Y approaches the limiting
or asymptotic value β1.
 Therefore, models like (5.2.11) have built in them an
asymptote or limit value that the dependent variable will
take when the value of the X variable increases indefinitely.

10/24/2017 Mai VU-FIE-FTU 18

5.3. Reciprocal models
 One of the important
applications of this model is the
celebrated Phillips curve of
macroeconomics.
 There is an asymmetry in the
response of wage changes to the
level of the unemployment rate:
Wages rise faster for a unit
change in unemployment if the
unemployment rate is below Un,
which is called the natural rate
of unemployment and then they
fall for an equivalent change
when the unemployment rate is
above the natural rate, β1,
indicating the asymptotic floor
for wage change.

10/24/2017 Mai VU-FIE-FTU 19

5.4. Log hyperbola or logarithmic reciprocal

model
 Y increases at an increasing rate 1
(i.e., the curve is initially convex) 𝑙𝑛𝑌𝑖 = 𝛽1 − 𝛽2 + 𝑢𝑖
and then it increases at a
𝑋𝑖
decreasing rate (i.e., the curve
becomes concave).
 Such a model may then be
appropriate to model a short-
run production function.
 Recall from microeconomics
that if labor and capital are the
inputsin a production function
and if we keep the capital input
constant but increase the labor
input, the short-run output–
labor relationship will resemble
as showed in this Figure.

10/24/2017 Mai VU-FIE-FTU 20

5.5. Choice of functional form
 The choice of a particular functional form may be
comparatively easy in the two-variable case, because
we can plot the variables and get some rough idea
about the appropriate model.
 The choice becomes much harder when we consider
the multiple regression model involving more than
one regressor.
 There is no denying that a great deal of skill and
experience are required in choosing an appropriate
model for empirical estimation. But some guidelines
can be offered:

10/24/2017 Mai VU-FIE-FTU 21

5.5. Choice of functional form

1. The underlying theory (e.g., the Phillips curve) may
suggest a particular functional form.
2. It is good practice to find out the rate of change (i.e.,
the slope) of the regressand with respect to the
regressor as well as to find out the elasticity of the
regressand with respect to the regressor. The
knowledge of these formulas will help us to compare
the various models:

10/24/2017 Mai VU-FIE-FTU 22

5.5. Choice of functional form

10/24/2017 Mai VU-FIE-FTU 23

5.5. Choice of functional form

3. The coefficients of the model chosen should satisfy
certain a priori expectations. For example, if we are
considering the demand for automobiles as a function of
price and some other variables, we should expect a
negative coefficient for the price variable.
4. Sometime more than one model may fit a given set of
data reasonably well. One major difference was that the r2
value of the linear model was larger than that of the
reciprocal model. One may therefore give a slight edge to
the linear model over the reciprocal model. But make
sure that in comparing two r2 values the dependent
variable, or the regressand, of the two models is the same;
the regressor(s) can take any form.

10/24/2017 Mai VU-FIE-FTU 24

5.5. Choice of functional form
 In general one should not overemphasize the r2
measure in the sense that the higher the r2 the better
the model.
 As we will discuss in the next chapter, r2 increases as
we add more regressors to the model.
 What is of greater importance is the theoretical
underpinning of the chosen model, the signs of the
estimated coefficients and their statistical significance.
If a model is good on these criteria, a model with a
lower r2 may be quite acceptable.

10/24/2017 Mai VU-FIE-FTU 25