SIP/2009-WP/11 Business case Traffic Forecasts CHAOUKI MUSTAPHA, Economist, International Civil Aviation Organization Workshop on the Development of Business Case for the Implementation of CNS/ATM Systems (Antigua and Barbuda, 28 September 2 October 2009)
Outline Planning parameters When? Why? How? Input Basic forecasts Peak period forecasts Output Alternative forecast techniques 2
Planning parameters Annual number of flights The average daily number of flights The number of flights in the peak day The number of flights in the peak hour Peak instantaneous aircraft count Others 3
When? New Air Navigation Services Facility Existing Operations 4
Physical Planning Why? To define the air navigation services facilities required To determine the scale and timing of implementation Financial Planning To estimate capital and operating expenditure To estimate operating revenues To carry out Cost/Benefit & Cash Flow Analysis 5
Facts to consider Peak demand rather than annual demand must be used in order to evaluate requirements Traffic Peaks by hour of the day, by day of the week, and by week and month of the year The level of detail of the forecast requirements will depend on the planning phase 6
Why study peaking Capacity utilization most critical during daily and hourly traffic peaks Peaking continues as markets grow The distribution of demand over any period is predictable 7
Peak-period analysis examples ASECNA FIRs Jeddah FIR Muscat FIR Available on the CAFEA website 8
Input: Historic data Yearly, monthly and daily aircraft movements Fleet mix and capacity Load factors Peak period parameters Historic data for passenger traffic 9
Basic forecasts Forecast of passenger traffic Assumptions of future trends for fleet mix & average aircraft size Assumptions for future load factors Unconstrained aircraft movements by type 10
Movements forecast development Movements = Passengers (Load factor) * (Average Seat) 11
Peak period forecasts Analysis of time profile of air traffic Ratios of busy periods applied to annual, monthly or weekly traffic Trend projection of these ratios Factors affecting peak period traffic trends: Business & holiday traffic mix Curfews at airports Changing route patterns 12
Output: Planning parameters Annual number of flights The average daily number of flights The number of flights in the peak day The number of flights in the peak hour Peak instantaneous aircraft count Others Average day of the peak month or week traffic Peak day of the average month or week traffic Peak hour of the average day traffic 13
Alternative Forecast Techniques Quantitative Qualitative Decision Analysis Time-Series Analysis Causal Methods Judgement Delphi Technological Market Research System Dynamics Heuristic Probabilistic Ratio Analysis Trend Projection Moving Averages Spectral Analysis Adaptive Filtering Box-Jenkins Regression Econometric Simulation Bayesian Spatial Equilibrium 14
Time series analysis
An example of trend projection World Passenger Traffic (Scheduled Services) 2000 1600 1200 Time period used for model development Ln (Y) = 121.31 + 0.0737 (T) Forecast horizon 800 400 0 1960 64 68 72 76 80 84 88 92 96 2000 2004 16
An example of trend projection World Passenger Traffic (Scheduled Services) 2000 1600 1200 Time period used for model development Y = 22.811 + 27.986(T) Forecast horizon 800 400 0 1960 64 68 72 76 80 84 88 92 96 2000 2004 17
Types of Trends 100 Linear 100 Polynomial 80 80 60 60 40 40 20 20 0 0 8 16 0 0 8 16 18
Types of trends 100 Exponential 6 Logarithmic 80 60 4 40 20 2 0 0 8 16 0 1 5 9 19
World passenger air traffic trend Linear Trend Polynomial Trend 1600000 1400000 1200000 y = 31301x - 6E+07 R 2 = 0.979 1000000 800000 600000 400000 200000 0 1950 1960 1970 1980 1990 2000 2010 1800000 1600000 y = 487.14x 2-2E+06x + 2E+09 R 2 = 0.9941 1400000 1200000 1000000 800000 600000 400000 200000 0 1950 1960 1970 1980 1990 2000 2010 20
World passenger air traffic trend Exponential Trend Logarithmic Trend 4000000 3500000 3000000 y = 3E-60e 0.076x R 2 = 0.9663 2500000 2000000 1500000 1000000 500000 0 1950 1960 1970 1980 1990 2000 2010 1600000 1400000 1200000 y = 61802888.64Ln(x) - 468489080.17 R 2 = 0.98 1000000 800000 600000 400000 200000 0 1950 1960 1970 1980 1990 2000 2010 21
World passenger air traffic trend Power Trend Moving Average 1200000 1000000 800000 600000 400000 200000 0 y = 54142x 0.7926 R 2 = 0.9266 0 10 20 30 40 50 1200000 1000000 800000 600000 400000 200000 0 1955 1960 1965 1970 1975 1980 1985 1990 1995 22
Types of trends Gompertz 100 80 60 40 20 0 0 40 80 23
Non-linear growth trends (S-Curves) Traffic Growth Market Saturation Slower Growth High Growth Initial Growth Time 24
Econometric analysis
Variables impact on traffic growth trends Cause a change in the demand curve Cause a shift in the demand curve 26
Variables impact on traffic growth trends Price Index 6 5 4 3 2 1 0 D1 Change in Demand Q1 Q2 D1 Tr1 Tr2 Tr3 Tr4 Tr5 Tr6 Number of Passengers 1/2 27
Variables impact on traffic growth trends Price Index 6 5 4 3 2 1 0 Q1 D1 Demand Shifts with Time Q2 D2 Tr1 Tr2 Tr3 Tr4 Tr5 Tr6 Tr7 Tr8 Tr9 Number of Passengers 2/2 28
Market Potential 29 Demographics Demographics & Development Development Greater Greater Income Income Lower Costs
References ICAO Airport Planning Manual (Doc 9184-AN/902), part 1, Chapter 3 ICAO Manual on Air Traffic Forecasting (Doc 8991/2) Reports on the Traffic Forecasting Groups (TFGs( TFGs) 30