Whatever you think about Amazon‘s reporting $0 R&D expenses and going for a $51 million R&D tax credit, this is a post about the danger’s of data as filtered and interpreted for you by infomediaries.
Amazon’s recent 10-Q reported USD 1.6 billion (appoximately) in “technology and content” expenses. This particular line item is a creation of Amazon, and they use their own XBRL tag for it. Looking back through their filings, they have done this consistently. As I said previously, it is clear to me from Amazon’s definition that these are not R&D expenses.
When we look at how infomediaries handle the situation, clearly someone disagrees with me, and they work at Morningstar. MSN.com sources its financial data through Morningstar, according to the fine print on their web page, and the “technology and content” line has been mapped to research and development for years. Contrast this with Yahoo Finance. That site sources their data from Edgar Online (again, according to the fine print), and that source merges Amazon’s fulfillment, marketing, and T&C lines into SG&A. There is explicitly no amount for R&D shown on the Yahoo Finance site.
Who should we trust? The answer is easy – Amazon. It is their data, their filing. They are breaking out numbers to help investors understand their company better. If they use custom XBRL tags, so be it. That is one of the great values of the extensible taxonomy functionality of XBRL, and the allowance for that functionality in the SEC’s implementation of XBRL-enabled financial reporting.
In contrast, the mapping strategies of the data consolidators are not transparent, and lead to clear anomalies as shown above. It doesn’t have to be that way. Infomediaries can also use XBRL to publish their mappings from company specific tags to standard tags or specific fields in their databases. The capital market would benefit from that openness.