Behind the scenes

Behind the scenes: Dow 30 tax burden

Business reporter Jia Lynn Yang got her hands on SEC filings for the 30 companies in the Dow Jones Index and paired up with graphics editor Darla Cameron to see how much each company pays in taxes compared to how much it makes. Here’s their graphic.

final graphic

How much data did you start with?

Darla: Digitized SEC files only go back to the early 1990s, so Jia Lynn had to track down more data from a research firm called S&P Capital IQ, which generously opened up their databases to us. Their data went back as far as 1969 for twelve of the companies; for almost all the rest, the data went back to the 1970s and 1980s. For each company, we focused on two data points: something called the “current tax provision,” which approximated how much the company paid in federal taxes in a given year, and also the company’s total income (both international and U.S.). We then divided those two figures to get a sense of the proportion of income going towards federal taxes. (Interestingly, researchers say is basically the only way to see what a company pays in taxes since they don’t have to report an actual tax figure. Jia Lynn called each company to ask, in case. None of them would share the information with us.) That resulting proportion tells a story of what they have paid over time.

How did you arrive at your final visualization? What other presentation options did you consider?

Darla: We started making fever charts of each company’s tax proportion over the last 40-odd years. The result was a bit messy:

step 1

Thanks to tax credits, in some years the companies reported negative income or 200+ percent taxes, and those outliers made the charts difficult to compare. We noticed a pattern, though: the overall slope for most companies trended downward. Post graphic artist Todd Lindeman looked at the charts and suggested just using the first and last number.

step 2

That tells the story (which is that these companies report paying less in taxes than they did 40 years ago) in a much simpler way.

What tools did you use and what resources did you find useful in developing your project?

Darla: Jia Lynn obtained some of the data from the SEC’s EDGAR database, a beloved research tool for business reporters and document junkies since it contains all the reports and disclosures that publicly traded companies are required by law to submit. We also relied heavily on data from S&P Capital IQ, which helped fill in some of the digital gaps in EDGAR. For the final visualization, I used Highcharts charting library with custom styles and JS.

What difficulties did you have with cleaning up the data?

Darla: Once we tracked down 40 years of data from Capital IQ (and narrowed it down to first and last tax years only), cleaning the data up wasn’t too difficult. We worked in a shared Google doc and then I used Mr. Data Converter to convert part of the data to a JavaScript object which populates the chart.

What was the biggest programming challenge?

Darla: Customizing the Highcharts styles and interaction was definitely the biggest programming challenge for me, since I am relatively new to programming. I looked into the documentation to tweak the chart colors and styles, then added elements like the hovering infoboxes and combobox/search selector separately.

Anything you’d do differently?

Darla: Our original plan included small charts for each company showing their detailed tax data for each year to help tell a more specific story. Those would be nice to add, but it’s time to move on to the next story.