Financial Forecasting for Airline Prices

This study analyzes the impact of explanatory variables on the consumer price index of airline fares in the United States. The multiple regression analysis results show that the selected independent variables are not good predictors of the consumer price index. A what-if-analysis can be conducted by considering other variables that may have a significant impact on the consumer price index.

Running Head: FINANCIAL FORECASTING FOR AIRLINE PRICES
Financial Forecasting for Airline Prices
Name of the Student
Name of the University
Author Note

1FINANCIAL FORECASTING FOR AIRLINE PRICES
Table of Contents
Introduction......................................................................................................................................2
Impact of Explanatory Variables.....................................................................................................2
Multiple Regression Analysis Results.............................................................................................6
Interpretation of Multiple Regression Results:............................................................................6
Residuals......................................................................................................................................7
Back-Testing and Errors..............................................................................................................8
Conclusion and What-If-Analysis...................................................................................................9
References......................................................................................................................................11

2FINANCIAL FORECASTING FOR AIRLINE PRICES
Introduction
Air transport is extremely important for the economy of the country. There are several
factors in airline business which contribute to the economy of a country. Airline provides a
network to the whole world helping the global tourism and business (Belobaba, Odoni &
Barnhart, 2015). Exports of goods have become much easier with the presence of airlines.
Exports can be made at a faster rate than previous with the presence of airlines. Nowadays
everybody is busy with their daily routine. Thus travelling abroad is always preferred by air
rather than by water which is extremely time consuming. The air fares vary a lot with domestic
and international segregation and distance as well (Williams, 2017). Thus, to study the reasons
affecting the air fares, consumer price index has been considered as the independent variable.
The original air fares were not available. Thus, the study was conducted using the price indices
(Wensveen, 2016). The data for the analysis has been collected from the website of Economic
Data Series and Research Guides (fred.stlouisfed.org, library.uhd.edu).
Impact of Explanatory Variables
Out of the 5 selected independent variables which are supposed to affect the consumer
price index of the airline fair, at first a correlation analysis has been performed. From the
correlation analysis, it has been observed that the independent variables producer price index by
industry and the international producer price index by commodity has been eliminated from the
model as these two variables had a very high correlation with other independent variables such
as domestic producer price index of the commodity. Although there are a lot other factors on
which the airlines price can be affected. But all the data were not available for the United states.
The other independent factors such as air revenue ton miles of freight and mail, and air revenue

3FINANCIAL FORECASTING FOR AIRLINE PRICES
passenger miles have very weak correlation with each other. Thus these variables are significant
enough to be considered into the study. If the variables considered as independent variables had
inter-dependence, then studying one variable would be able to predict the effect of the other on
the dependent variable. Thus, these eliminations have been made. Now, the variables selected
have a very correlation with each other and thus there is not much chance that the influence of
one variable will coincide the other.
Producer
Price
Index by
Industry
Producer Price
Index by
Commodity
(international)
Producer Price
Index by
Commodity
(Domestic)
Air Revenue Ton
Miles of Freight
and Mail (Price
Index)
Air Revenue
Passenger
Miles (Price
Index)
Producer Price
Index by Industry 1
Producer Price
Index by
Commodity
(international)
0.789328
429 1
Producer Price
Index by
Commodity
(Domestic)
0.981067
421 0.660125271 1
Air Revenue Ton
Miles of Freight
and Mail (Price
Index)
0.444472
157 0.241018696 0.459360699 1
Air Revenue
Passenger Miles
(Price Index)
0.364925
684 0.09705588 0.412585727 0.405488512 1
In this study, the consumer price index of Airline fares is considered for the country the
United States. A lot of flights ply all day from the US airports to various parts of the world.
Thus, it is important to identify the factors that are responsible for the variations in the flight
fares. One of the factors that have been considered in this study the air revenue that the airline
company earns from transporting the freights and mails via air. It can be seen clearly from the
figure that the airlines company is only interested to earn high revenue. They do not have much
interest to analyze the factors when the revenue is low. Thus, all the data points for revenue for

4FINANCIAL FORECASTING FOR AIRLINE PRICES
freights and mails are high. It also has been seen that the consumer price index is high when the
revenue is high. This diagram also shows that all the values are concentrated at one particular
place. Hence nothing can be commented on the nature of change of consumer price index with
the change in the revenue. This indicates that is a very weal correlation or association between
the airline prices and the revenue earned from the transport of freights and mails. The
relationship is only obtained for very high revenues and not for low revenues.
90.0 100.0 110.0 120.0 130.0 140.0 150.0
0.000
50.000
100.000
150.000
200.000
250.000
300.000
350.000
Air Revenue Ton Miles of Freight and Mail
Air Revenue Ton Miles of Freight and Mail
Consumer Price Index
Figure 1: Relation between Revenue from Freight and Mail and consumer Price Index
The consumer price index is supposed to depend on the revenue earned from the
travelling of passengers. It can be seen clearly from the graph that the values of price index do
not much change with the changes in the revenue. The price indices of the consumers are more
or less constant. However, the price indices of the customers are high when the revenue earned
from customer travel is high. Thus, it can be said that when high revenue is earned, the price
index of the airline fares are also high. Thus, in order, it can also be said that the airline fares will
be high when the revenue earned is high. Nothing much can be said about the airline prices or
price indices when the airline companies earn low revenue. Thus, it can be said that the relation

5FINANCIAL FORECASTING FOR AIRLINE PRICES
between air revenue from passengers and the price index are not much strong. If the revenue
earned is low, then nothing can be said about the price indices of the airline prices.
100.0 110.0 120.0 130.0 140.0 150.0 160.0 170.0 180.0 190.0 200.0
0.000
50.000
100.000
150.000
200.000
250.000
300.000
350.000
Air Revenue Passenger Miles
Air Revenue Passenger Miles
Consumer Price Index
Figure 2: Relation between Revenue from Passengers and consumer Price Index
The next figure shows the relationship between the producer price index of the
commodity and the consumer price index. It can be seen from the figure clearly that there is a
relationship between the producer and the consumer price indices. In case of the producer price
indices, the producers will be giving priority to the business class tickets sold as the earning from
those types of tickets will be much higher than the economy class tickets. On the other hands, the
consumers will be giving priority to the tickets which are of cheaper rate, such as economy class.
Thus, the consumers will more prone to buying the economic class tickets and the producers will
be much more interested in selling the business class tickets. Selling of more economy class
tickets will give the producers a good revenue but that same amount of revenue can be earned by
selling lesser amounts of business class tickets. This creates a direct relationship between the
producers and the consumers price indices. In this case also, there is no relation between the
variables for lower price indices. Thus, when the price indices are high, a positive association

6FINANCIAL FORECASTING FOR AIRLINE PRICES
exists between the consumers and the producers price indices. No relationship exists between the
variables when the price indices are low.
90.0 95.0 100.0 105.0 110.0 115.0 120.0 125.0
0.000
50.000
100.000
150.000
200.000
250.000
300.000
350.000
Producer Price Index by Commodity (Domestic)
Producer Price Index by Commodity (Domestic)
Consumer Price Index
Figure 3: Relation between producer price index and consumer Price Index
Multiple Regression Analysis Results
Consumer Price Indices (Yt) = 0.0002 – (0.001 * X1) + (0.0002 * X2) + (0.072 * X3)
Coefficient
s P-value
Intercept 0.0002
0.91179
5
Air Revenue Ton Miles of Freight and Mail (Price Index)(Percentage Change)
(X1) -0.001
0.98595
5
Air Revenue Passenger Miles (Price Index)(Percentage Change)(X2) 0.0002
0.99003
2
Producer Price Index by Commodity (Domestic)(Percentage Change)(X3) 0.072
0.35073
1
Interpretation of Multiple Regression Results:
 For every 1% increase in the air revenue ton miles of freights and mails, the price index
of consumers are expected to decrease by 0.001%.

7FINANCIAL FORECASTING FOR AIRLINE PRICES
 For every 1% increase in the Air revenue of passenger miles, the consumer price index is
expected to increase by 0.0002 percent.
 For every 1% increase in the producer price index by commodity, the consumer price
index is expected to increase by 0.072 percent.
 The p-value indicates that the three independent variables such as air revenue ton miles of
freight and mail, air revenue of passenger miles and price index of the producer by
commodity are not good predictors of the price index of the consumers. Thus, the model
is not at all a good model in predicting the consumer price indices due to the high p-
values of the independent variables at both 5% and 1% level of significance.
Residuals
The figure below, which is the residual chart, gives a clear picture of the actual values of
the consumer price indices at different time points (t) and the forecasted values of the consumer
price indices at the time t which has been obtained from the regression model Yt. From this
residuals graph it can be seen that no missing values from the data can be identified from here
and thus the errors cannot be predicted. Hence the forecasted value might not be a good fit.

8FINANCIAL FORECASTING FOR AIRLINE PRICES
0 2 4 6 8 10 12
0
2
4
6
8
10
12
Residuals
Figure 4: Residuals of Consumer Price Index
Back-Testing and Errors
The MAE and MAPE for the four independent variables considered in this study such as
Consumer price index, air revenue ton miles of freight and mail, air revenue passenger miles and
producer price index by commodity. For the variable producer price index by commodity, the
percent changes were 0.00% for some months. Thus, it is mathematically impossible to calculate
the average values. Thus, the APE for the forecasted value already calculated for the last month
that is August 2017 has been considered in place of the MAPE. The errors of prediction are
usually lowered in case of moving averages. Since the variables are not much good predictors of
Consumer Price Index, the forecast will not give a satisfied result.
Consumer Price Indices (Percentage Changes)
MA (5) MA (10) MA (20) MA (50)
Forecasted Value -0.0125 -0.0017 -0.0031 -0.0018
MAE 0.0161 0.0147 0.0146 0.0125
MAPE 12.18% 25.42% -15.74% -30.68%

9FINANCIAL FORECASTING FOR AIRLINE PRICES
Air Revenue Ton Miles of Freight and Mail (Price Index) (Percentage Change)
MA (5) MA (10) MA (20) MA (50)
Forecasted Value 0.0159 0.0114 0.0085 0.0043
MAE 0.0219 0.0212 0.0206 0.0187
MAPE 21.15% 11.23% 2.90% 9.43%
Air Revenue Passenger Miles (Price Index) (Percentage Change)
MA (5) MA (10) MA (20) MA (50)
Forecasted Value 0.0237 0.0203 0.0139 0.0080
MAE 0.0846 0.0897 0.0854 0.0861
MAPE -28.08% -16.41% -14.77% -9.51%
Producer Price Index by Commodity (Domestic)(Percentage Change)
MA (5) MA (10) MA (20) MA (50)
Forecasted Value -0.0021 0.0027 0.0015 -0.0006
MAE 0.0234 0.0226 0.0225 0.0233
APE* -66.31% -151.96% -107.31% -102.24%
Conclusion and What-If-Analysis
As the p-values of all the dependent variables are high, this indicate insignificance of the
variables in predicting the consumer price index. The regression table shows that only 1.06
percent of the variability in the independent variables can be explained by the independent
variables that are considered in the model for forecasting.
Regression Statistics
Multiple R
0.10336
3
R Square 1.06%

10FINANCIAL FORECASTING FOR AIRLINE PRICES
Adjusted R
Square -0.019
Standard Error
0.01849
8
Observations 104
Thus, in order to increase the explanatory percentage of variability, several what-if-
analysis will be performed. The four different types of moving averages that has been obtained
for all the variables are used in the regression model to increase the explanation of variability of
consumer price index to a certain percentage by the insignificant independent variables. The
percentage is supposed to rise a little as in case of moving averages, the error is minimum.
 When the producer price index by commodity is -0.0021, the consumer price index for
September 2017 will be:
Consumer Price Indices (Yt) = 0.0002 + (0.001 * 0.0159) + (0.0002 * 0.0237) + (0.072 *
-0.0021) = 0.000069%. Thus, the price index will rise by 0.000069%
 When the producer price index by commodity is 0.0027, the consumer price index for
September 2017 will be:
Consumer Price Indices (Yt) = 0.0002 - (0.001 * 0.0114) + (0.0002 * 0.0203) + (0.072 *
0.0027) = 0.00039%. Thus, the price index will rise by 0.00039%.
 When the producer price index by commodity is 0.0015, the consumer price index for
September 2017 will be:
Consumer Price Indices (Yt) = 0.0002 - (0.001 * 0.0085) + (0.0002 * 0.0139) + (0.072 *
0.0015) = 0.00030%. Thus, the price index will rise by 0.00030%.
 When the producer price index by commodity is 0.0080, the consumer price index for
September 2017 will be:
Consumer Price Indices (Yt) = 0.0002 - (0.001 * -0.0018) + (0.0002 * 0.0043) + (0.072 *
0.0080) = 0.00078%. Thus, the price index will rise by 0.00078%.

11FINANCIAL FORECASTING FOR AIRLINE PRICES

12FINANCIAL FORECASTING FOR AIRLINE PRICES
References
Belobaba, P., Odoni, A., & Barnhart, C. (Eds.). (2015). The global airline industry. John Wiley
& Sons.
Economic Data Series by Tag | FRED | St. Louis Fed. (2017). Fred.stlouisfed.org. Retrieved 6
December 2017, from https://fred.stlouisfed.org/tags/series
Research Guides: Economics: Statistics & Data. (2017). Library.uhd.edu. Retrieved 6 December
2017, from https://library.uhd.edu/c.php?g=356697&p=2405882
Wensveen, J. G. (2016). Air transportation: A management perspective. Routledge.
Williams, G. (2017). The airline industry and the impact of deregulation. Routledge.