Scribd
Upload
55 views21 pages

Regression - Slides and UIP Case-Study Setup

The document discusses testing uncovered interest rate parity (UIP) by running a regression of the change in the exchange rate on the interest rate differential. The regression results show the intercept is not statistically different from zero, but the slope coefficient is statistically different from one, indicating UIP does not hold.

Uploaded by

Ghazi Ben Jaballah
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
55 views21 pages

Regression - Slides and UIP Case-Study Setup

The document discusses testing uncovered interest rate parity (UIP) by running a regression of the change in the exchange rate on the interest rate differential. The regression results show the intercept is not statistically different from zero, but the slope coefficient is statistically different from one, indicating UIP does not hold.

Uploaded by

Ghazi Ben Jaballah
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

International Economics

ESCP Europe

31. August 2022

(ESCP) International Economics 31. August 2022 1 / 20


Outline

Chapter 14
Refresher Regression and UIP case study setup

(ESCP) International Economics 31. August 2022 2 / 20


Uncovered interest rate parity (UIP)

et
Et+1 − Et
(6) Rt − Rt∗ =
Et
et
Et+1 − Et
(6′ ) Rt = Rt∗ +
Et

(ESCP) International Economics 31. August 2022 3 / 20


Testing UIP

Et+1 − Et
= α + β · (Rt − Rt∗ ) + ϵt+1
Et
α = intercept with the vertical axis
β = slope of the regression line
ϵt+1 = error term

(ESCP) International Economics 31. August 2022 4 / 20


Testing UIP: draw a scatter plot

Et+1 −Et
Time R − R∗ Et
Jan 2003 -5 % -5 %
Feb 2003 -10 % -10 %
Mar 2003 0% 0 %
Apr 2003 5% 5 %
Source: Just an hypothetical example.

(ESCP) International Economics 31. August 2022 5 / 20


Before you run the regression

Et+1 − Et
= α + β · (Rt − Rt∗ ) + ϵt+1
Et
Step 1: write down the pair of hypotheses
H0 : α = 0 and Ha : α ̸= 0
H0 : β = 1 and Ha : β ̸= 1
Hence, we want to perform a two sided test!
Step 2: determine the confidence level and the critical values
We want to test on a 95 % confidence level and assume that we have
enough observations so that the normal distribution is the appropriate
choice!
tcritical ≈ 2 or tcritical ≈ −2(in case you want to be precise: 1.96)

(ESCP) International Economics 31. August 2022 6 / 20


Regression results

Et+1 − Et
= α + β · (Rt − Rt∗ ) + ϵt+1
Et
Coeff. SE t-Stat P-val. Low95 Up95
Intercept -0.0619 0.0364 -1.70 0.09 -0.13 0.01
RtYEN − RtAUS -1.1283 0.6383 -1.77 0.08 -2.40 0.14

(ESCP) International Economics 31. August 2022 7 / 20


After running the regression: confidence interval for α

Step 3: How is the lower and upper bound of a confidence interval


computed?
Lower bound: Estimated coefficient MINUS 1.96 * standard error
(SE): -0.0619 - 1.96 · 0.0364 = -0.133
Upper bound: Estimated coefficient PLUS 1.96 * standard error
(SE): -0.0619 + 1.96 · 0.0364 = 0.009

Step 4: Is the value claimed in the hypotheses included in the confidence


interval?
Zero is included. Hence we do not reject H0 : α = 0

(ESCP) International Economics 31. August 2022 8 / 20


After running the regression: confidence interval for β

Step 3: How is the lower and upper bound of a confidence interval


computed?
Lower bound: Estimated coefficient MINUS 1.96 * standard error
(SE): -1.1283 - 1.96 · 0.6383 =-2.379
Upper bound: Estimated coefficient PLUS 1.96 * standard error
(SE): -1.1283 + 1.96 · 0.6383 = 0.1228
Step 4: Is the value claimed in the hypotheses included in the confidence
interval?
1 is not included. Hence we have to reject H0 : β = 1
Conclusion: UIP fails

(ESCP) International Economics 31. August 2022 9 / 20


Alternative for confidence interval: t-test

Step 3a: How is the t-stat for α computed?

estimated coefficient − claimed value −0.0619 − 0


t − stat = = = −1.7
standard error 0.0364
Step 4a: Comparison of t-stat and t-critical. Since the t-test is below the
t-critical, do not reject H0 : α = 0

(ESCP) International Economics 31. August 2022 10 / 20


Alternative for confidence interval: t-test

Step 3a: How is the t-stat for β computed?

estimated coefficient − claimed value −1.1283 − 1


t−stat = = = −3.3343
standard error 0.6383
Step 4a: Comparison of t-stat and t-critical. Since the t-test is above the
t-critical (in absolute terms), reject H0 : β = 1

(ESCP) International Economics 31. August 2022 11 / 20


Interpretation confidence intervals

Claimed value of α = 0 is included in the 95 % confidence interval.


We can NOT reject H0 .
Claimed value of β = 1 is NOT included in the 95 % confidence
interval. We have to reject H0 !
Conclusions
Empirical evidence is NOT in line with UIP. Beta is estimated to be
negative, but should be equal to 1!
Beta is significantly different from 1 on a 95 % confidence interval.
UIP does not hold.
Interest rate differential not an unbiased / good predictor of the
subsequent change in the exchange rate.

(ESCP) International Economics 31. August 2022 12 / 20


Interpretation t-values

2.5% 2.5%
95%
Reject H0 Reject H0
Do not reject H0

-3.40 -3.00 -2.60 -2.20 -1.80 -1.40 -1.00 -0.60 -0.20 0.20 0.60 1.00 1.40 1.80 2.20 2.60 3.00 3.40

-1.96 1.96

Georg Stadtmann - Statistics - Slide No. 45

1
(ESCP) International Economics 31. August 2022 13 / 20
(ESCP) International Economics 31. August 2022 14 / 20
(ESCP) International Economics 31. August 2022 15 / 20
Chinn (2006), p. 9.

(ESCP) International Economics 31. August 2022 16 / 20


Froot/Thaler (1990), p. 182

(ESCP) International Economics 31. August 2022 17 / 20


Froot (1990): NBER 3247.

(ESCP) International Economics 31. August 2022 18 / 20


Chinn (2006), p. 10.

(ESCP) International Economics 31. August 2022 19 / 20


Chinn (2006), p. 13.

(ESCP) International Economics 31. August 2022 20 / 20


Running regressions in excel

(ESCP) International Economics 31. August 2022 21 / 20

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy