Financial Tests Report

American Airline group lnc.

Submitted to: Sir Iftikhar-Ul- Hassan

Submitted by: Asad Ghaffar

Roll No. 8315

Class: B.B.A 7TH Finance (Evening)

Session: 2018- 2022

Government College University Faisalabad

Sahiwal Campus
 Introduction
In this report, I have to explain the fluctuations of volatility of American Airline Group
lnc of assets returns through different financials tests or analysis from the last three years. I have
to collect the time series data of AAL from Jan 1st, 2019 to Dec 31st, 20121 for analysis like to
check the stationarity of data, unit root tests, correlation analysis and regression analysis. These
analyses are performed by using Stata.

In this report, I have to analyze the data normality, to conduct stationary tests,
autocorrelation, heteroskedasticity and a detailed description of the univariate ARMA/ARIMA
model and the notion of conditional volatility with the popular family of Arch and Garch that
specially design to capture the autoregressive nature of the volatility of assets returns.

In our case, I have a non-stationary variable called “log y”. firstly, I must convert this
non-stationary variable into stationary variable by first differencing in the shape of “X” variable.
Afterwards, I can use this stationary variable to run various time series models such as unit root
tests, correlation and regression analysis etc. I shall be appointing graphs like line plot,
correlogram and histogram for better understanding.

Source of data for Analysis:

Company Name: American Airline Group lnc (AAL)

Website: www.yahoofinance.com

Link: https://finance.yahoo.com/quote/AAL?p=AAL&.tsrc=fin-srch
Financial Tests Reports Saturday January 1 05:05:51 2022 Page 1

___ ____ ____ ____ ____(R)

/__ / ____/ / ____/
___/ / /___/ / /___/
Statistics/Data Analysis

User: Asad Ghaffar

Project: American Airline Group Inc.

___ ____ ____ ____ ____ (R)

/__ / ____/ / ____/
___/ / /___/ / /___/ 14.2 Copyright 1985-2015 StataCorp LP
Statistics/Data Analysis StataCorp
4905 Lakeway Drive
Special Edition College Station, Texas 77845 USA
800-STATA-PC http://www.stata.com
979-696-4600 [email protected]
979-696-4601 (fax)

Single-user Stata perpetual license:

Serial number: 10699393
Licensed to: Andrey

Notes:
1. Unicode is supported; see help unicode_advice.
2. Maximum number of variables is set to 5000; see help set_maxvar.

. *(2 variables, 36 observations pasted into data editor)

. generate t = tm(2019m1) + _n -1

. format t %tm

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month

. twoway (tsline logy)

. ac logy

. histogram logy
(bin=6, start=2.3513753, width=.20161561)

. corrgram logy

-1 0 1 -1 0 1
LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor]

1 0.8856 0.8887 30.652 0.0000

2 0.7624 -0.0753 54.042 0.0000
3 0.6151 -0.2787 69.724 0.0000
4 0.4539 -0.0994 78.531 0.0000
5 0.3271 0.0479 83.252 0.0000
6 0.1983 0.0650 85.046 0.0000
7 0.0657 -0.2121 85.25 0.0000
8 -0.0548 -0.2400 85.396 0.0000
9 -0.1631 0.1117 86.745 0.0000
10 -0.2602 0.1605 90.307 0.0000
11 -0.3347 -0.0305 96.438 0.0000
12 -0.3930 -0.1465 105.24 0.0000
13 -0.4253 -0.0442 116 0.0000
14 -0.4092 0.1413 126.41 0.0000
15 -0.3675 0.0992 135.21 0.0000
16 -0.3065 0.0130 141.64 0.0000
Financial Tests Reports Saturday January 1 05:05:52 2022 Page 2

. twoway (tsline logy) (tsline x)

. ac x

. histogram x
(bin=5, start=-.44276237, width=.13360626)

. corrgram x

-1 0 1 -1 0 1
LAG AC PAC Q Prob>Q [Autocorrelation] [Partial Autocor]

1 0.0362 0.0363 .04991 0.8232

2 0.2322 0.2327 2.1655 0.3387
3 0.0299 0.0131 2.2016 0.5316
4 -0.0647 -0.1248 2.3765 0.6669
5 -0.1307 -0.1431 3.1143 0.6824
6 0.0667 0.1438 3.3131 0.7686
7 0.0476 0.1331 3.4178 0.8439
8 -0.1234 -0.2517 4.1483 0.8435
9 -0.1433 -0.2807 5.1717 0.8191
10 -0.1420 -0.0156 6.2167 0.7967
11 -0.1512 0.1041 7.4508 0.7615
12 -0.1472 -0.1050 8.6715 0.7307
13 -0.0862 -0.2431 9.109 0.7647
14 0.1158 0.1928 9.9367 0.7668
15 -0.1033 0.1967 10.628 0.7785

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month

. dfuller logy, trend regress lags(1)

Augmented Dickey-Fuller test for unit root Number of obs = 34

Interpolated Dickey-Fuller
Test 1% Critical 5% Critical 10% Critical
Statistic Value Value Value

Z(t) -1.467 -4.297 -3.564 -3.218

MacKinnon approximate p-value for Z(t) = 0.8402

D.logy Coef. Std. Err. t P>|t| [95% Conf. Interval]

logy
L1. -.1214762 .0828079 -1.47 0.153 -.2905924 .0476401
LD. .0718011 .1803432 0.40 0.693 -.2965087 .440111
_trend .0002644 .0030493 0.09 0.931 -.0059631 .006492
_cons .3415998 .2820299 1.21 0.235 -.2343821 .9175818

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month
Financial Tests Reports Saturday January 1 05:05:53 2022 Page 3

. wntestb logy, table

Cumulative periodogram white-noise test

Bartlett's (B) statistic = 3.0779

Prob > B = 0.0000

. wntestq logy, lags(1)

Portmanteau test for white noise

Portmanteau (Q) statistic = 30.6523

Prob > chi2(1) = 0.0000

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month

. correlate logy x
(obs=35)

logy x

logy 1.0000
x 0.1267 1.0000

. pwcorr logy x, obs sig star(5)

logy x

logy 1.0000

36

x 0.1267 1.0000
0.4682
35 35

. spearman logy x, stats(rho obs p) star(0.05)

Number of obs = 35
Spearman's rho = 0.1162

Test of Ho: logy and x are independent

Prob > |t| = 0.5060

. ktau logy x, stats(taub obs p) star(0.05)

Number of obs = 35
Kendall's tau-a = 0.0655
Kendall's tau-b = 0.0655
Kendall's score = 39
SE of score = 70.415

Test of Ho: logy and x are independent

Prob > |z| = 0.5894 (continuity corrected)
Financial Tests Reports Saturday January 1 05:05:53 2022 Page 4

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month

. regress logy x

Source SS df MS Number of obs = 35

F(1, 33) = 0.54
Model .071231413 1 .071231413 Prob > F = 0.4682
Residual 4.36350266 33 .132227353 R-squared = 0.0161
Adj R-squared = -0.0138
Total 4.43473407 34 .130433355 Root MSE = .36363

logy Coef. Std. Err. t P>|t| [95% Conf. Interval]

x .3063964 .4174539 0.73 0.468 -.54292 1.155713

_cons 3.003674 .0619772 48.46 0.000 2.87758 3.129767

. dwstat

Durbin-Watson d-statistic( 2, 35) = .1056223

. arima D.logy, ar(1) ma(1)

(setting optimization to BHHH)

Iteration 0: log likelihood = 17.327345
Iteration 1: log likelihood = 17.4329
Iteration 2: log likelihood = 17.456678
Iteration 3: log likelihood = 17.525773
Iteration 4: log likelihood = 17.576407
(switching optimization to BFGS)
Iteration 5: log likelihood = 17.583651
Iteration 6: log likelihood = 17.611155
Iteration 7: log likelihood = 17.618431
Iteration 8: log likelihood = 17.619267
Iteration 9: log likelihood = 17.619339
Iteration 10: log likelihood = 17.619347
Iteration 11: log likelihood = 17.619347

ARIMA regression

Sample: 2019m2 - 2021m12 Number of obs = 35

Wald chi2(2) = 1.19
Log likelihood = 17.61935 Prob > chi2 = 0.5514

OPG
D.logy Coef. Std. Err. z P>|z| [95% Conf. Interval]

logy
_cons -.0192147 .0332725 -0.58 0.564 -.0844277 .0459982

ARMA
ar
L1. .6054437 1.083556 0.56 0.576 -1.518288 2.729175

ma
L1. -.5156457 1.176413 -0.44 0.661 -2.821372 1.790081

/sigma .1462276 .0208161 7.02 0.000 .1054288 .1870263

Note: The test of the variance against zero is one sided, and the two-sided
confidence interval is truncated at zero.
Financial Tests Reports Saturday January 1 05:05:53 2022 Page 5

. tsset t
time variable: t, 2019m1 to 2021m12
delta: 1 month

. regress D.logy

Source SS df MS Number of obs = 35

F(0, 34) = 0.00
Model 0 0 . Prob > F = .
Residual .758759527 34 .022316457 R-squared = 0.0000
Adj R-squared = 0.0000
Total .758759527 34 .022316457 Root MSE = .14939

D.logy Coef. Std. Err. t P>|t| [95% Conf. Interval]

_cons -.0190519 .025251 -0.75 0.456 -.0703681 .0322643

. estat archlm, lags(1)

LM test for autoregressive conditional heteroskedasticity (ARCH)

lags(p) chi2 df Prob > chi2

1 2.149 1 0.1426

H0: no ARCH effects vs. H1: ARCH(p) disturbance

. arch D.logy, arch(1)

(setting optimization to BHHH)

Iteration 0: log likelihood = 18.573569
Iteration 1: log likelihood = 18.700473
Iteration 2: log likelihood = 18.707599
Iteration 3: log likelihood = 18.708395
Iteration 4: log likelihood = 18.708474
(switching optimization to BFGS)
Iteration 5: log likelihood = 18.708483
Iteration 6: log likelihood = 18.708484

ARCH family regression

Sample: 2019m2 - 2021m12 Number of obs = 35

Distribution: Gaussian Wald chi2(.) = .
Log likelihood = 18.70848 Prob > chi2 = .

OPG
D.logy Coef. Std. Err. z P>|z| [95% Conf. Interval]

logy
_cons -.0094948 .0241092 -0.39 0.694 -.0567479 .0377583

ARCH
arch
L1. .292099 .2675825 1.09 0.275 -.2323531 .8165511

_cons .015052 .0046631 3.23 0.001 .0059125 .0241915

.
 3.5
3
log y
2.5
2

2019m1 2020m1 2021m1 2022m1

t
.2
0
x
-.2
-.4

2019m1 2020m1 2021m1 2022m1

t
 3.5
3
log y
2.5
2

2019m1 2020m1 2021m1 2022m1

t
.2
0
x
-.2
-.4

2019m1 2020m1 2021m1 2022m1

t
 Interpretation
Firstly, the times series data shows that the normality between all variables like time, dependent
and Independent variables.

Stationarity:
 Null Hypothesis: Stationary Data
 Alt. Hypothesis: Non Stationary Data

According to the Correlogram or Stationary tests, the corresponding probability value of log y variable is
zero that is too much less than 5%. The variable of log y is non-stationary. It means I reject Null
Hypothesis and to select Alternative Hypothesis. On the other hand, the corresponding probability value
of “X” variable is greater than 5%. The variable of “X” is stationary. It means I cannot reject Null
Hypothesis rather I accept Alternative Hypothesis.

Unit Root:
 Null Hypothesis (H0): variable is not stationary or got unit root
 Alt. Hypothesis( H1): variable is stationary or no unit root

According to the Augmented Dickey Fuller tests, the value of absolute test statistics is less than the
critical value, it means I cannot reject Null Hypothesis because the variable of log y has unit root and also
the coefficient of L1 has negative value. The model is unit root or valid.

White Noise:
According to the Bartlett’s and Portmanteau white noise test, the probability value of white noise is
less than 5% and is equal to zero. The test allows running further analysis. It means I can accept Null
Hypothesis rather than Alternative Hypothesis.

Correlation Analysis:
 Null Hypothesis (H0): P = 0, reject null hypothesis
 Alt. Hypothesis( H1): P ≠ 0, accept null hypothesis

According to the Pearson’s correlation, the probability of list wise and pair wise is greater than zero it
means I accept Null Hypothesis that shows the high correlation relationships among the variables log y
and X.

 Null Hypothesis (H0): P = 0, no monotonic relation

 Alt. Hypothesis( H1): P ≠ 0, some monotonic relation

According to the Spearman’s correlation, the probability value is greater than zero. It means the variables
have high monotonic relationship between them.
Regression Analysis:
The probability of regression analysis is greater than 5%. It means there is no relationship among
variables. The coefficient of X variable is also positive. The R-Squared value is too small which is less
than 5%. It indicates that the model is not fit for process.

According to the Durban Watson, the value is less than 2%. It indicates there is no autocorrelation
detected in the sample. It means there is a positive autocorrelation among variables.

ARIMA Model:
The value of likelihood in ARIMA Model is positive but not in high that not show the
significance level. The coefficient value of AR and MA is greater than 5%. These values are not show the
high level of significance.