Like lipstick on a pig, fancy graphs don’t change bad data!

A headline, from Bloomberg News, in my daily newsfeed, attracted my attention.

It said “Truck Cargo Numbers Point To Growing Economy”. Now that was an interesting departure from the recent crescendo of recessionary doom and gloom articles. With my interest being piqued, I went to the Bloomberg News site to read more about this and get the actual story. There I was met with very large and very bold headlines (see below) proclaiming the fact that a “key US recession predictor was missing”. The message was clear. We should stop the hand-wringing about an upcoming recession and instead look at the prospects of an expanding economy. OK – now I was hooked. I had to find out more.

I learned that Bradley Jacob, the expert whose opinion was being represented in this news report, believes that “truck cargo data has proven to be a reliable predictor of future economic activity, with a steady string of monthly freight declines preceding each of the past five U.S. recessions. There’s no real sign of that kind of downturn now. Since early 2013, the tonnage hauled by trucks hasn’t fallen even one month, when compared with the year-earlier period, according to the American Trucking Associations.”

The graph visible in the background of the speaker was the one shown below and supposedly was the data that substantiated the statement above.

Observe that this graph shows several downturns, almost equally spaced. That did not seem to make any sense. US Recessions haven’t been that precisely cyclical. On closer examination, I found that the month and year data on the x-axis was not readable. Hence, I could not ascertain if the downturns corresponded to known recessionary periods. Fortunately, there was a graph on the web version of the report. I have included that below.

I think that you’ll agree that this one does not look anything like the one shown in the video footage. Anyways, the speaker on the news program had stated that the Truck Freight Tonnage had dropped steadily prior to every one of the past five recessions. Looking at this particular graph, I have to assume that he was referring to the downturns (represented by the orange arrows that I added to the graph).

The first problem I have with such a statement is that the downward trend for the past three recessions was followed by an upward trend (blue arrows added by me) just prior to the recessionary troughs (shown in the blue line on the graph). Imagine being present in that moment of time where the upward swing begins. If you had access to a crystal ball that showed you the portion of the graph representing the future, then it would be a different story. Without the benefit of that crystal ball, it would be very difficult for you to ascertain if the downward trend preceding the upward swing in the graph was an actual indicator of a recession or not. After all, there were similar downward swings followed by upward swings (for example, the period between 1983 through 1986 on the graph) which were not related to any recessions.

Secondly, the graph seems to totally contradict the premise of the story. Let me explain. Look at the last two orange arrows that I have shown in the graph. (The last one being the one leading into Dec 2015) How do those downturns fit the point being made about no indication of an upcoming recession? Assume for a moment that the premise of a decreasing truck tonnage as a leading indicator of a recession is true. Then by all accounts, the graph presented actually indicates an upcoming recession. Doesn’t it? Isn’t that quite contrary to the statement – “There’s no real sign of that kind of downturn now.” Huh? Am I missing something?

I was by now completely confused. Surely a respected business information source as Bloomberg would not present something that is internally contradictory. So I did some more research to take a closer look at the Truck Tonnage data. Here are a couple of graphs that I replotted with data that I found from other sources.

The first one (shown above) is a plot of the ATA Monthly Tonnage index drawn from data at the U.S. Department of Transportation, Bureau of Transportation Statistics (BTS) site. I really don’t see how this index can be considered as a proven indicator at accurately predicting an upcoming recession. The index does not show a steady decline leading into the recessionary periods. For example, take a look at the time period preceding the “Great Recession”. It actually shows an increase through the second half of 2007, then remains more or less steady (normal ups and downs) through the first half of 2008 and then starts plummeting in the second half of 2008. However, by then the recession was already well underway. Now that cannot be called a leading indicator!

The second graph (shown above), is a replot of the YOY data that is discussed in the TV segment as well as the web article. I wanted to check the claim – “…since early 2013, the tonnage hauled by trucks hasn’t fallen even one month, when compared with the year-earlier period, according to the American Trucking Association.”

Looking at this graph it is true that the Year on Year tonnage has been positive, except for the last data point which is for Nov 2015. Now that November contraction contradicts the statement. Furthermore, shouldn’t we be more concerned about the decrease in the magnitude of increase of volume over prior year?

By this time, I had determined that the ATA Truck Tonnage Index, while interesting, is not a sure-fire predictor of a recession or lack thereof. Also, aren’t the Consumer Spending and Durable Goods Orders (See graph below, From Gardner Research) more of a leading predictor of a recession than a decreasing Truck Tonnage Index?

So by now I have wasted a good part of a day investigating the effectiveness or lack thereof of truck tonnage hauled as an indicator of a recession. Wasted? Yes! After all, people don’t pay me to assess the effectiveness of recessionary indicators! I could have just taken the article at face value, as many readers have certainly done and been comforted by the fact that we have strong evidence to suggest that we are not heading into a recession in the US.

However, I would have been clearly misinformed. This is what gets my goat! Throwing up graphs and talking about them on TV is cool, fashionable and exudes an air of expertise. However, if the facts presented are wrong or misleading, then that is a big problem. If done knowingly, for some other ulterior motive or gain, then that is simply wrong! People are affected by the data that is presented to them. People make decisions that involve significant financial risks based on data that they receive. Making sure that presented data is accurate should be a fundamental responsibility of the presenter/speaker/author.

The point of this article and the reason I spent some more of my time writing this (although I don’t consider this portion of it wasted, for even if one person is benefited….), is to make a point about data.

The power of data is awesome! Data, when used intelligently and with some analytical thinking and interpretation, can provide tremendous insights and provide clear guides to action. However, inaccurate, misleading, skewed data (typically done by special interests), when bandied about by professionals and presented as facts, is garbage!

Besides, being a statistical practitioner myself, I seldom get perturbed by a few data points showing good or bad trends. Variation is normal! Now, if I see 6 or 7 points in a row, all ascending or all descending, then I start to pay attention. That is a trend! All else is just routine ups and downs.

I don’t believe everything that I come across without adding a bit of critical thinking. The need for this is increasingly becoming essential every day. There are innumerable sources of data and information without any editorial or peer-review system. Data is coming at us in flooding waves. Data and misinformation can spread around the world in a matter of seconds. We need to filter data and information through the lens of critical thinking. Or else we may be numbed into making incorrect decisions, whether it be in business, personal finance, electing the next president or simply choosing a restaurant for a romantic Valentines Day dinner!

At this point, I have to bring up a quote that is typically attributed to Deming. I have lived by that rule from the day that I heard it. “In God we trust, all others bring data.”

Caveat emptor!


About the author:

Rahul Sarkar is a Professional Engineer and entrepreneur working with manufacturing companies (with a focus on the small, mid-sized companies) in increasing top-line sales and bottom line savings.

His company, Chiketa–Phoenix, works on lowering manufacturing costs (materials and labor), developing Lean Manufacturing strategies, identifying energy savings opportunities, advising on automation systems, managing Low-Cost Country (LCC) sourcing, developing purchasing management systems and more…

To learn more, e-mail: