I'm using free Quandl financial data to reproduce a view of S&P 500 Dividend Aristocrats. What is important for me is not the result (we can get the list on the internet for free) but I'd like to do this by myself and make sure I understand the methodology.

Quandl data is pretty good, but there is some unexpected dividends I don't know how to explain.

For example :

  • Mac Donald's (MCD) has a dividend of 22¢ on 2000-11-13 and 23¢ on 2001-11-13, which is true and in the Quandl data. But in there we also find a dividend of 47¢ on 2000-09-13.
  • Kimberly Clark (KMB) usually has 4 dividends a year (less than $1). But on 2014-11-03, there is an additional super-high dividend (4.77$).

So my questions are :

  • Can we explain this phenomenon ? Is the data legit ?
  • Is there a data source I could use to find and remove these extra dividends ?

For MCD, the 47¢ is a regular dividend on preferred stock (see SEC filing here). Common stock holders are not eligible for this amount, so you need to exclude this amount.

For KMB, there was a spin-off of Halyard Health. From their IR page on the spin-off:

Kimberly-Clark will distribute one share of Halyard common stock for every eight shares of Kimberly-Clark common stock you own as of the close of business on the record date.

The deal closed on 2014-11-03. At the time HYH was worth $37.97 per share, so with a 1:8 ratio this is worth about $4.75. Assuming you were able to sell your HYH shares at this price, the "dividend" in the data is something you want to keep.

With all the different types of corporate actions, this data is extremely hard to keep clean. It looks like the Quandl source is lacking here, so you may need to consider looking at other vendors.

Not the answer you're looking for? Browse other questions tagged or ask your own question.