Monday, July 6, 2015

Seasonality, The Heart Of A Boeing Forecast

When making statistical projections seasonality skews the projection according to what may occur if the statistical influences are added to the math model for projecting outcomes. The reason both Airbus and Boeing come up with different results is based on different seasonal variables. These can include many affects on the not so straight line assumption of how many aircraft will be ordered. Below are a few seasonal indices affecting a twenty year forecast.
  • Fuel Market
  • Economic Stability 
  • Labor Market
  • Financial Market 
  • Demand Constraint
  • Supply Constraint
These are just to name a few where a statistical model factors in each of these seasonal influences, smoothing the outcome, and  apportioning its influences in the forecast. With all these variables the manufacturer who makes its proclamations about the airplane market futures; comes up with a plausible outcome, which reasonably affixes a demand for just a general prediction for customers who will need an airplane through til 2034.  Each years projection updates the date range. During 2016 the date range will go through 2035. Each succeeding year's forecast will include the "n"th year twenty years out. If going longer than twenty the forecast suppositions starts to lose accuracy and relevance. In fact a ten year forecast is a better predictability reliance than twenty years. The narrower the time range, the better the forecast will match what actually occurs. 

It's just like the weather forecast. Having a 72 hour weather window is an excellent method for predictability, and usually a most accurate forecast. Adding seasonality factors into the weather forecast can extend a reasonable reliance 30 days out with high confidence. The Confidence component is another statistical tweak which continues to enhance a more accurate outcome by defining a range of accuracy expected. A ninety day weather forecast becomes closer to a guess with some  seasonality influences and is less reliable. The further a statistical forecast goes out the more unstable the outcome accuracy becomes compared to what will really happen.

A statistical Airplane model would have to include some seasonal factors, including the affects of  War or Depression.  It would include additions with more extensive seasonality factoring with any 20 year model. The 20 year prediction  reels its long range seasonality influences when adding those factors to an forecast outcome. A complexity of factors become so unstable and whimsical that it becomes a useless model if too many seasonal consideration are applied. The only sanity in the process is doing a year to year with a twenty year range. War and depression forecasts are similar to predicting the next earthquake. Therefore, why predict the airplane market dynamics during the next twenty years using "earthquake data" in its forecast model? It is more solid to use straight forward industry growth projections and social and economic indicators in the model, which are a manageable, and a reasonable resource for projecting the next twenty years from year to year reports.

A reasonable answer from forecasting should be simple and should be part of every participant from the industry. Remember in the weather forecast example above, the first 72 hours are usually a golden prediction,  with only having a few degrees of temperature variance from what actually is recorded when it happens. A week long weather report becomes plausible and fairly reliable. A month long weather outlook is a best information outcome with all the seasonal bells and whistles included. It even has its own forecast within the forecast. How reliable and how accurate will the outcome actual match the forecast is another forecast in itself. That is known as the confidence factor. A twenty year forecast has a broader spectrum for its confidence when calculating what the market will become. Hence an annual forecast keeps updating the Airplane industry's "weather report", adding on a year. It is similar to what the news contemplates when conducting your own local weather reporting. The importance of the Airplane forecast is those first five years and how Boeing sees it. Those years are the most accurate and are within a customers own five year financial plans. 

It then becomes important to a customer, supplier, or financial markets, to take into consideration the first five year window of a Boeing 20 year market prediction.

It also important for development teams, innovators and industry planners to take into account the 5-10 year segment of what Boeing predicts. It takes that much time for completing innovation steps, allowing for maturing, and plant facility growth to occur. The last ten years of the prediction, is an over-arching case-maker when organizations are proposing a quantified change affected by predicted demand quotient.

Boeing doesn't claim absolute accuracy, but does claim all things considered you can expect 22,000 units from here to 2034. Next year it will be from then to 2035. Airbus smooths it out a little less perhaps accommodating a seasonal decline on its own narrow body demand during the next twenty years.