. Answers to Homework Assignment 4 for Unit 12 (30 points) .

Ahmed specifies and estimates the following regression function (known as the Phillips curve in Macroeconomics)
for 1950-2022 (standard errors are in parentheses):

Population regression function (PRF): Yt = β0 + β1 UNt + β2 Xt + ut

Sample regression function (SRF): Yt = b0 + b1 UNt + b2 Xt + et

OLS SRF estimates: Ŷt = 2.367 – .729 UNt + .692 Xt

R2 = .759, Standard error of regression (S) = 2.48, n = 73

where Yt = the inflation rate (in %),

UNt = the unemployment rate (in %),

Xt = Next year’s inflation rate expected by consumers (in %).

Page 1 of 9
 Population regression function (PRF): Yt = β0 + β1 UNt + β2 Xt + ut

Sample regression function (SRF): Yt = b0 + b1 UNt + b2 Xt + et

OLS SRF estimates: Ŷt = 2.367 + .729 UNt + .692 Xt

R2 = .759, Standard error of regression (S) = 2.48, n = 73

Question (a). Clearly interpret the point estimate of β0.

Answer: b0 = 2.367

The inflation rate irrespective of unemployment and expected inflation is 2.367%.

Clearly interpret the point estimate of β1.

Answer. b1 = .729

Holding Xt constant, when the unemployment rate increases (decreases) by 1% point, the inflation rate decreases
(increases) by .729% point. This means that there is a negative relationship between inflation and unemployment,
when holding Xt constant.

Page 2 of 9
 Population regression function (PRF): Yt = β0 + β1 UNt + β2 Xt + ut

Sample regression function (SRF): Yt = b0 + b1 UNt + b2 Xt + et

OLS SRF estimates: Ŷt = 2.367 + .729 UNt + .692 Xt

R2 = .759, Standard error of regression (S) = 2.48, n = 73

Clearly interpret the point estimate of β2.

Answer. b2 = .692

Holding UNt constant, when the expected inflation rate increases (decreases) by 1% point, the inflation rate
increases (decreases) by .692% point. This means that there is a positive relationship between the inflation rate and
the expected inflation rate, when holding UNt constant.

Interpret the coefficient of determination (R2 = .759).

Answer. R2 = .759
75.9% of the variation in the inflation rate (Yt) is explained by the variations in both the unemployment rate (UNt)
and the expected inflation rate (Xt).

Page 3 of 9
Question (b). Test whether the unemployment rate explains the behavior of the inflation rate. In answering, write the

null and alternative hypotheses and the decision rule. Show your calculations. State your conclusion.

Answer. The null and alternative hypotheses are:

H0: 0 Unemployment does not explain the behavior of inflation (holding Xt constant)

H1: 0 Unemployment explains the behavior of inflation (holding Xt constant)

The decision rule is: Reject H0 at the 100α% level of significance,

if the calculated t-value | | ≥ critical t-value, , /

.
 With 0, b1 = .729, Se(b1) = .251, the calculated t-value = | |=| | = 2.904
.

 With (n – 3 ) = (73 – 3) = 70, the 10%, 5%, and 1% critical t-values are:

For α = .10, ,. = 1.667, for α = .05, ,. = 1.994, for α = .01, ,. = 2.648

With the calculated t-value = 2.904 > 2.648, we reject the null hypothesis at the 1% level of significance. Thus,
there is overwhelming sample evidence to reject the null hypothesis. Accordingly, holding Xt constant, the
unemployment rate explains the behavior of the inflation rate.

Page 4 of 9
Question (c). Test whether the expected inflation rate explains the behavior of the inflation rate. In answering, write
the null and alternative hypotheses and the decision rule. Show your calculations. State your conclusion.

Answer. The null and alternative hypotheses are:

H0: 0 Expected inflation does not explain the behavior of inflation (holding UNt constant)

H1: 0 Expected inflation rate explains the behavior of inflation (holding UNt constant)

The decision rule is: Reject H0 at the 100α% level of significance,

if the calculated t-value | | ≥ critical t-value, , /

.
 With 0, b2 = .692, Se(b2) = .403, the calculated t-value = | |=| | = 1.717
.

 With (n – 3 ) = (73 – 3) = 70, the 10%, 5%, and 1% critical t-values are:

For α = .10, ,. = 1.667, for α = .05, ,. = 1.994, for α = .01, ,. = 2.648

With the calculated t-value = 1.717 > 1.667, we reject the null hypothesis at the 10% level of significance.
Thus, there is moderately strong sample evidence to reject the null hypothesis. Accordingly, holding UNt
constant, the expected inflation rate explains the behavior of the inflation rate.

Page 5 of 9
Question (d). Do the independent variables (UNt and Xt) jointly explain the behavior of the inflation rate at the 5%
level of significance? In answering, write the null and alternative hypotheses and the decision rule. Show your
calculations. State your conclusion.

Answer. The joint null and alternative hypotheses are:

H0: β1 = β2 = 0 UNt and Xt do not jointly explain the behavior of inflation

H1: At least one of these β’s ≠ 0 Otherwise

Decision rule: Reject H0 at the 100α% level of significance, if

the calculated F-value = ≥ the critical F-value, , ,

.
 With R2 =.759 and n = 73, the calculated F-value = = = 110.23
.

 With (n – 3) = (73 – 3) = 70, the 5% critical F-value from the F-table is , ,. = 3.13

 With the calculated F-value = 110.23 > 3.13, we reject the null hypothesis at the 5% level of significance.
Thus, the unemployment rate ( ) and the expected inflation rate ( ) jointly explain the behavior of the
inflation rate ( ).

Page 6 of 9
Question (e). Calculate the 95% confidence interval estimate of σ2. Interpret your result.

Answer.

 Thus, the 100(1‒ α)% C.I. estimator of the population error variance (σ2) is: < σ2 <
, ,

 With (1 – α) = .95, α = .05, α/2 = .025, and (1 – α/2) = .975. In addition, (n – 3) = 73 – 3 = 70.

 From the chi-square table, χ2 , = χ2 ,. = 95 and χ2 , = χ2 ,. = 48.8

. .
 With S2 = (2.48)2 = 6.1504, the 95% C.I. estimate of σ2 is: < σ2 < or 4.53 <σ2 < 8.82
.

 Interpretation: We are 95% confident that this interval estimate contains the true error variance σ2.

Page 7 of 9
Question (f). A critic claims that the true standard error of regression σ is 2.25. Test the validity of this claim using

your results in part (e). Show your calculations. State your conclusion.

Answer.

 With the claimed value of σ = 2.25, the claimed value of σ2 = 5.0625

 In answering Question (e), we found the 95% confidence interval estimate of σ2 is (4.53 < σ2 < 8.82)

 This interval estimate includes the claimed value of σ2 = 5.0625

 Thus, we cannot reject the claim at the 5% level of significance.

 Note that, here, the null hypothesis is H0: σ2 = 5.0625

Page 8 of 9
Question (g). State the assumptions that are needed for your answers to parts (b) through (f) to be valid.

Answer.

 The condition set by the Gauss-Markov theorem must be satisfied. Therefore, we need to assume that the
error term (ut) is purely random.

 For the confidence interval estimation and hypothesis testing, we also need to assume that the error term (ut)
is normally distributed.

Page 9 of 9