Will the Airline Industry Recover? Econ 240C Group E: Daniel Grund Daniel Jiang You Ren David Rhodes Catherine Wohletz James Young
Table of Contents Motivation Data Collection Identifying the Model Intervention Variables ForecastConclusions
Motivation Will the airline industry be able to recover after the September 11 th terrorist attack? What events significantly effect the airline industry post deregulation? What are forecasted revenue passenger miles?
Data Collection Bureau of Transportation Statistics Final Scheduled and Non-Scheduled Revenue Passenger Miles Monthly Data, January 1981-December –
The Raw Data
Difference of Logs
AutocorrelationPartial CorrelationAC PAC Q-Stat Prob **|. | **|. | |. |.|. | |. |.|. | *|. | *|. | |** |.|* | *****|. | *****|. | |** | *|. | *|. | **|. | |. | **|. | |. | ***|. | **|. | ****|. | |****** |.|*** | **|. |.|* | |* |.|* | |. | *|. | |. |.|* | |* |.|. | *****|. |.|. | |** |.|. | *|. |.|. | Clearly there is a lot of structure to this data.
Basic Observations Clearly the series is highly regular, following a more or less constant cycle in the difference of logs until September After that time there are distinctive disturbances to that pattern. Obviously September 11 th was a major factor, but we found two other significant disturbances.
CoefficientStd. Errorz-StatisticProb. JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC AR(1) AR(2) SAR(12) MA(2) SMA(12) Variance Equation C E ARCH(1) GARCH(1) Pre-9/11 Regression
AutocorrelationPartial CorrelationAC PAC Q-Stat Prob.|* |.|* | |* |.|* | |. |.|. | |. |.|. | |. |.|. | *|. | *|. | |* |.|* | |. |.|. | *|. | *|. | |. |.|. | |* |.|* | |. |.|. | |. |.|. | |* |.|* | |. |.|. | |. |.|. | *|. |.|. | |. | *|. | |. |.|. | |. |.|. | |. |.|. | |. |.|. | Correllelogram of Residuals: further ARMA terms could be used, but unnecessary.
AutocorrelationPartial CorrelationAC PAC Q-Stat Prob.|. |.|. | |. |.|. | |. |.|. | |. |.|. | *|. | *|. | |. |.|. | |. |.|. | *|. | *|. | |* |.|* | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |* |.|* | |* |.|* | |. | *|. | |. |.|. | |. |.|. | |. |.|. | *|. | *|. | Correllelogram of Squared Residuals: Appears to be adequately clean, due to GARCH.
Residuals are fairly normal.
Actual Growth versus Pre-9/11 forecasting.
Difference in recolored forecast versus actual (in 1000s of RPMs) 9/11 Troop deployment War begins
Disruptions to the forecast By observing the data we found that there were three obvious events which significantly shocked the airline industry, each followed by a one or two month recovery with rates of growth far from predicted values. These events were Sept 2001 (9/11), Dec 2002 (Deployment of troops) and March 2003 (Commencement of war with Iraq). During the deployment of troops, US commercial airlines were used to move troops and equipment. The following two months were a return to equilibrium. The commencement of war with Iraq brought additional concerns of terrorist attacks, but these concerns faded.
CoefficientStd. Errorz-StatisticProb. EVENT AFTEREVENT PREEVENT EVENT AFTEREVENT EVENT AFTEREVENT JAN FEB …………… NOV DEC AR(1) AR(2) SAR(12) MA(2) SMA(12) Variance Equation C ARCH(1) GARCH(1) December 2002 (PREEVENT2) was not statistically significant but had some explanatory power. All other events were significant.
Event variables capture the effects of major events in recent years.
Residuals appear to be adequately clean. AutocorrelationPartial CorrelationAC PAC Q-Stat Prob.|* |.|* | |* |.|* | |. |.|. | |. |.|. | |. |.|. | *|. | *|. | |. |.|. | |. |.|. | *|. | *|. | *|. | *|. | |* |.|* | |. |.|. | |. |.|. | |* |.|* | |. | *|. | |. |.|. | *|. |.|. | |. |.|. | |. |.|. | |. |.|. |
Residuals Squared also clean. AutocorrelationPartial CorrelationAC PAC Q-Stat Prob.|. |.|. | |. |.|. | |. |.|. | |. |.|. | *|. | *|. | |. |.|. | |. |.|. | |. |.|. | |* |.|* | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |. |.|. | |* |.|* | |* |.|* | |. |.|. | |. |.|. | |. |.|. |
Residuals are normal.
The Pre-9/11 forecast and event model forecast give almost identical results for 2004 growth rates. This suggests that no further recovery from 9/11 is expected.
Conclusion Airline industry not expected to recover to pre-September 11 th, 2001, growth path by the end of Airline growth rates appear to have returned to their original pattern except for additional shocks due to Middle-East current events, but the industry appears to have permanently lost at least 5 million RPMs per month.
QUESTIONS?