This Is How You Can Easily Make Things Look Bad For Boeing's Dreamliner

| About: The Boeing (BA)
This article is now exclusive for PRO subscribers.


The method used to analyze data influences the outcome.

Even a 99% accurate interpolation polynomial does not make for a reliable extrapolator.

Results can only be as accurate as the analyst's understanding of data analysis and the industry or product that is being analyzed.

The Dreamliner deferred costs are extensively covered in my articles. In most articles I have a look at how deferred costs developed and how those compare to my expectations. In other articles I check whether the ultimate goal is still achievable.

In the past I have refrained from sharing my own findings for deferred costs to completely flat out. However, following my article Boeing: 1,300 Dreamliners, Will It Be Enough? I received a lot of comments on the things that I do or do not implement and have received some links towards other analyses. Additionally, some people sent me their own calculations.

Although I am a big fan of investors that do their own due diligence and gather knowledge from various sources, I think some make rather big mistakes caused by unreliable interpretation of data and/or flawed analyses by other parties.

Before I do my assessment on the break-even point for the Boeing (NYSE:BA) 787, I think it is important to address these common mistakes and flaws. In the end a similar result might roll out, but it is more important to understand how you got to the end result. It is also important to understand that some methods are being used in the wrong way, thereby sketching either a too grim or too rosy image.

In this article I will deal with a common data analysis mistake that impacts the deferred costs projection tremendously. This projection was made by a reader to support his view that Boeing's deferred costs are likely to grow by a sizable amount from now and will not peak until 2017.

The data

Figure 1: Development of deferred costs per quarter (Source)

First of all, the data points as shown in Figure 1 are used. These simply are the deferred costs per quarter as reported by Boeing. No calculations per unit, just the basic information Boeing reported for the deferred costs and unamortized tooling costs.

The projection

Now, for the projection the reader sent to me a trend line that was being used, supported by the statement that the formula had a highly accurate R-squared value in excess of .99. The R-squared value is an indication on how well the used formula fits the actual data points.

AeroAnalysis has taken the liberty to recreate this data in Figure 2:

Figure 2: Development of deferred costs with flawed estimator (Source: AeroAnalysis)

As you can see my reader was able to generate an estimator for deferred costs. Using this estimator, he concluded that deferred costs would rise to almost $35bn with deferred costs to be flat in total by Q3 2024.

The mistake

Now, looking at the graph things look quite interesting. A peak and break-even point could be indicated, but the used method does not make sense from a data analysis point of view to me and is unlikely to yield an accurate result.

First of all, the biggest mistake is that a trend line was used. Even with a 2nd degree polynomial with a .99 R-squared value this method is not completely valid. What the reader did was an interpolation of the data and used that formula as an extrapolation polynomial. An interpolation polynomial to extrapolate data can be used, of course, but the accuracy of this prediction method is nowhere close to the .99 R-squared value for the interpolation part. Additionally, one can easily see that the actual data is showing signs of a peak (in line with Boeing's expectations) while the estimator sees a peak in 2017.

Not only the used method makes the obtained results highly unlikely, also the fact that you 'look far into the future' decreases the reliability and accuracy.


While using an interpolating polynomial to extrapolate is not wrong to most of us, one has to understand that the accuracy and reliability of the interpolation do not say anything about the reliability and accuracy of the extrapolating polynomial. In fact, this kind of extrapolation is highly unreliable. Although, it gives an estimate of the break-even point and peak in costs this data is not likely to be anywhere close to the actual future data and is no more reliable than a guesstimate, a mixture of calculations and guesswork.

As an aerospace engineer I have something to share that does somewhat relate to data analysis and cost prediction. A thing that often is said about engineers is that they are people that do precision guesswork based on unreliable data provided by those of questionable knowledge. In this case the data is accurate and provided by knowledgeable people, making the estimate or guesswork as reliable as the analyst's capabilities and understanding of data analysis and the industry.

Although this article is not directly related to making an investment decision it shows how making use of a wrong estimation method can lead to unreliable estimates, which subsequently can lead to making the wrong investment decision.

In the end it is important to know how knowledgeable the analyst whose work you read is, but also in case if you do your own due diligence (which I highly recommend) how reliable the method you use to process data is.

If you would like to receive updates for my upcoming articles, please click the "Follow" text at the top of this page next to my profile.

Disclosure: I am/we are long BA.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.