Professional Documents
Culture Documents
Profits
Kawika Pierson
MIT System Dynamics Group 3nd Year PhD
Fall 2008 Albany-MIT PhD Colloquium
Outline
Relevant Literature
Reference Mode for Airline Profits
Digging Deeper
The Model
Demand
Price
Capacity
Costs
Profit
Results
Relevant Literature
“Cycles in the sky: understanding and managing business cycles in the
airline market” M Liehr, A Groessler, M Klein, PM Milling - System
Dynamics Review, 2001
Made for Lufthansa as a guide for strategy
Very limited scope (only one feedback loop)
Relevant Literature
“System dynamics for market forecasting and structural analysis”
James Lyneis - System Dynamics Review, 2000
Commercial jet aircraft industry
Focused on use of SD models as forecasts for Jet Orders
Proprietary, but potentially similar to our work
"Analysis of Profit Cycles in the Airline Industry" 2004 Helen Jiang,
R. John Hansman
Very simple model, two stocks one feedback loop
Control theory perspective
Reference Mode
The data for US airline industry profits shows some cyclicality
since before deregulation
Taken from a presentation by Prof. R. John Hansman and Helen Jiang Nov. 2004
Digging Deeper
Profit = Revenue – Costs
Revenue = Price * sales in units
Costs = unit cost * production
Sales is Revenue Passenger Miles
Price is the Price of Tickets
Production is Available Seat Miles
This gives us Profit
How does financial reporting effect our modeling?
Capacity – Causal Loop Diagram
Demand
+
Forecasted
Capacity Demand
+
- +
B Capacity
Shortfall + Desired
Capacity Capacity
Control
+
Investment
Capacity - Model Structure
<Airplane
Planned Retirements>
Replacement
Orders
Desired
+ Capacity
Desired Capacity +
Acquisition Rate Adjustment Time to Adjust
For Capacity - - Capacity
<Time Required to
+ Manufacture an
Desired Aircraft Airplane>
Supply Line
Indicated
Capacity Supply Line
Adjustment Adjustment Time
+ -
Order Approval Adjustment for the Weight on Supply
Delay Time Supply Line - Line Adjustment
+
Indicated Orders Mothballed Capacity
<Return to
for Capacity
Service>
<Cancelled
Orders>
Mothballing Off Mothballing
Cancellation
<Into
Storage>
Airline Capacity
Orders of Airplane Airline Capacity
Supply Line Airplane
Airplanes Manufacturing
Completion Retirements
Initial Airline
<Return to Initial Capacity Capacity <Retired>
Service> on Order
Third Order Stocks –Cancellation
and Mothballing
<Indicated Orders
for Capacity>
Time Required to
Time to Cancel Manufacture an
Airplane
+
<Orders of Cancelled
Airplanes> Orders
Average
<Airplane Service Life
Manufacturing
Completion>
<Operating
<Operating Margin>
Margin> Into Storage Margin Threshold to
Return to
<Time to Service Initiate Mothballing
Mothball>
Time to
Mothball
<Indicated Orders
Capacity on Mothball <Indicated Orders
for Capacity> for Capacity>
Forecasting Demand
Time Horizion for
Reference Demand
Time to Percieve
Indicated
Changes in Demand
Growth Rate
Time to Percieve
Trend in Demand
Perceived
Initial Expected Demand
Expected Growth Change in
Growth Rate in
Rate for Demand Demand
Demand
Change in Perception
Expected Growth
Rate
Gap in Demand
Perception
+
-
Order Approval Adjustment for the Weight on Supply
Delay Time Supply Line - Line Adjustment
+
Indicated Orders Mothballed Capacity
<Return to
for Capacity
Service>
<Cancelled
Orders>
Mothballing Off Mothballing
Cancellation
<Into
Storage>
Airline Capacity
Orders of Airplane Airline Capacity
Supply Line Airplane
Airplanes Manufacturing
Completion Retirements
Initial Airline
<Return to Initial Capacity Capacity <Retired>
Service> on Order
Fitting to Data
Get historical data on important stocks
Airlines are great for this
Airlines.org, MIT Airline Data Project, BTS
Set up summary statistics
John Sterman’s Book plus MAE, RMSE, %E, Thiel,
SSE/M^2
Drive each model sector with historical variables
Use Vensim’s model fitting functions
Lets walk through this
Summary Statistics
Payoff Element <r>
dif cov
Um <dif mean>
Sum of Error
<Sy> <Sx>
Squared over Mean
MSE Us <dif var>
Error over Mean Sum AE MAE over dif var
<dt>
Squared Mean <M Y>
<Yi>
<Yi> <Xi> <dt> <Count> <M X> RMSE Uc <dif cov>
Residuals dif mean
<M X>
<Xi> <M X>
Sum APE MAPE RMSE over
Mean
<Xi> <Yi> <Xi>
<Available Seat
Miles>
SumX2 <dt>
Simulated
<dt> SumXY
<Count>
Y MX2
<Count>
<Historical> Sum Yi
Yi <M X>
Start Time <Time> Mxy
Sx
Percent Error <M Y>
MY
End Time Count Sy
pick <Count>
<Simulated> Interval <TIME STEP> dt
MX r MY2
<dt>
SumY2
Sum Xi
<Yi>
X Xi R^2
<Historical Historical
Available Seat
Miles>
Example of Fitting the Model
1. Open Simulation Control
2.
Create a Payoff
Example of Fitting the Model
3. Run
“Policy”
Negative
Example of Fitting the Model
4.
Set Parameters
Example of Fitting the Model
5.
Capacity Fit – Historical Inputs
Historical and Simulated Airline Capacity
1e+012
750 B
Seat*miles/Year
500 B
250 B
0
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Historical Available Seat Miles : Match All
Available Seat Miles : Match All
Price War
+
Load Factor Congestion
- +
B
Delivery Delay - -
GDP
+ Demand +
+
Forecasted
Capacity R Demand
+ Route
Networks
- +
B Capacity
Shortfall + Desired
Capacity Capacity
Control
+
Investment
Demand – Model Structure
Historical Population
Data and Projections
Initial <Time>
Population Historical GDP Data
and Projections
Miles per Person per
Population
Dollar of GDP
Actual Demand
GDP per
For Seat Miles Capita
Seat Miles Desired
from GDP per Capita
<Time>
Sensitivity of Constant
Demand to Congestion Demand
Congestion Perception Time Demand for Seat Adjustment
Miles per Capita
Indicated Effect of Effect of
Congestion on <Ticket Price>
Congestion on
Demand Demand
Effect of New Elasticity of Demand
Effect of Price
Capacity on with Respect to Price
<Indicated Load on Demand
Demand
Factor> <Comfortable
Industry Load
Factor> Strength of New
Capacity Effect on Reference
Demand Ticket Price
One Year Percent
Change in Capacity <Initial Ticket
Price>
<Lag for
<Available Seat Measuring
Miles> Changes>
Demand Fit – Historical Inputs
Historical and Simulated Revenue Passanger Miles
800 B
600 B
Seat*miles/Year
400 B
200 B
0
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Historical Airline Demand : Match All
Actual Demand For Seat Miles : Match All
0.15
dollars/(Seat*mile)
0.1
0.05
0
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Historical Airline Ticket Prices : Match Price
Ticket Price : Match Price
Variable Cost
Cost per Available
from Operations
Seat Mile
<Average Worker
Compensation> Workers Per Seat
Operating Costs
from Passengers of Capacity
Freight as a
Percentage of
Passenger Operations
Cost Fit – Historical Inputs
Historical and Simulated Operating Costs
200 B
150 B
dollars/Year
100 B
50 B
0
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Historical Airline Operating Costs : Match Cost
Total Operating Costs : Match Cost
Historical
Unemployment <Time>
Data
Profits – Model Structure
<Load Factor>
Freight
Revenue
Operating Operating
Revenue Margin
<Time>
Operating
Profit check year
<Total Operating
Costs>
Accumulated
Operating
New Operating Drain Operating
Profit
Profit Profit
<TIME STEP>
Reported
Operating Profit
Wages Fit – Historical Inputs
Historical and Simulated Average Wages
80,000
dollars/(Year*worker)
60,000
40,000
20,000
0
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Historical Airline Salaries : Match All
Average Worker Compensation : Match All
17,500
15,000
12,500
10,000
1970 1974 1978 1982 1986 1990 1994 1998
Time (Year)
Simulated Real Wage : Match All
Historical Real Wage : Match All
0.15
dollars/(Seat*mile)
1.5e+012
Seat*miles/Year
0.1
1e+012
0.05
500 B
0
0
1971 1975 1979 1983 1987 1991 1995 1999
1971 1975 1979 1983 1987 1991 1995 1999
Time (Year)
Time (Year)
Historical Airline Ticket Prices : Match All
Historical Available Seat Miles : Match All
Ticket Price : Match All
Available Seat Miles : Match All
600 B 150 B
Seat*miles/Year
dollars/Year
400 B 100 B
200 B 50 B
0 0
1971 1975 1979 1983 1987 1991 1995 1999 1971 1975 1979 1983 1987 1991 1995 1999
Time (Year) Time (Year)
Historical Airline Demand : Match All Historical Airline Operating Costs : Match All
Actual Demand For Seat Miles : Match All Total Operating Costs : Match All
5B
dollars/Year
-5 B
-10 B
1971 1975 1979 1983 1987 1991 1995 1999
Time (Year)
Historical Airline Operating Profit : Match All
Reported Operating Profit : Match All
Conclusions
Growth Correction
Partial Model Tests with Historical Inputs
Cyclical Nature not alleviated by Cancellations or
Mothballing
Standard SD Structures fit the industry reasonably well
More dynamics exist in the real system
Comments? Questions?