Tennis greats at the top of their game
The time of bird seeds and chart tuneups

Pulling the multi-national story out, step by step

Reader Aleksander B. found this Economist chart difficult to understand.


Given the chart title, the reader is looking for a story about multinationals producing lower return on equity than local firms. The first item displayed indicates that multinationals out-performed local firms in the technology sector.

The pie charts on the right column provide additional information about the share of each sector by the type of firms. Is there a correlation between the share of multinationals, and their performance differential relative to local firms?


We can clean up the presentation. The first changes include using dots in place of pipes, removing the vertical gridlines, and pushing the zero line to the background:


The horizontal gridlines attached to the zero line can also be removed:


Now, we re-order the rows. Start with the aggregate "All sectors". Then, order sectors from the largest under-performance by multinationals to the smallest.


The pie charts focus only on the share of multinationals. Taking away the remainders speeds up our perception:


Help the reader understand the data by dividing the sectors into groups, organized by the performance differential:


For what it's worth, re-sort the sectors from largest to smallest share of multinationals:


Having created groups of sectors by share of multinationals, I simplify further by showing the average pie chart within each group:



To recap all the edits, here is an animated gif: (if it doesn't play automatically, click on it)



Judging from the last graphic, I am not sure there is much correlation between share of multinationals and the performance differentials. It's interesting that in aggregate, local firms and multinationals performed the same. The average hides the variability by sector: in some sectors, local firms out-performed multinationals, as the original chart title asserted.



Very interesting read!

The one thing I would also consider is... reversing the colors.
Since the focus of the chart is on the multinational companies, it would make sense to highlight them in red. Even the chart's title mentioning "global" is misfortunatelly placed next to a red rectangle, color-coded for "local" elswhere.

Actually, this single change could make the original chart much more readable by making the order of the list (by decreasing return of global companies) much more prominent.

The comments to this entry are closed.