Sunday, January 13, 2013

Data Update 2013: The Dark Side of Numbers

For the last two decades, I have dedicated the first two weeks of each new year to a ritual. I obtain/collect/download data on all publicly traded companies listed globally, using a variety of data sources, and then analyze and present the data, aggregated at a number of different levels: by country, by region (US, Europe, Emerging Markets, Japan, Australia & Canada) and by industry. I report on measures of operations (profit margins, turnover ratios, working capital), measures of leverage (debt ratios), measures of risk (beta, standard deviation, equity risk premiums, country risk premiums) and pricing measures (earnings multiples, book value multiples, revenue multiples). I just completed my 2013 update and you can find it by clicking here.

I start with a belief that all data should be accessible and available to all investors at low or no cost, but my motives for providing my reading of the data are far from altruistic. I draw on the numbers that I estimate through the rest of the year for my teaching, analysis (valuation or corporate finance) and writing (blogs, books). In other words, I would have analyzed all of this data anyway and having completed the work, I see little benefit in keeping it behind a pay wall or passwords. Let me hasten to add that nothing that I do is particularly original nor is it path breaking and my task is made easier by the easy access that we have to raw data. I do hope, though, that while I do make mistakes, that I have not let my personal biases or views color the data, and that that nothing that I do is opaque.

Each year, I also try to add something new to the dataset to keep it fresh and this year, I have added company-specific estimates of  costs of equity and capital (in US dollar terms) in the individual company data sets (look to the top of the linked data page). In making these estimates, though, I had to make very broad assumptions about country risk.  For instance, I used the risk premium of the country of incorporation to the company, though it is preferable to use the risk premium based on operations. So, take these cost of capital estimates with a grain of salt, and if you prefer a more precise estimate for a company, you should do in more detail.

When I finished my update a year ago, I posted on it here, and talked about one of my favorite movies/books, Moneyball, in the context of arguing that intuition & experience were vastly overrated in business. Much of what we think we have learned or think we know about investing and corporate finance is skewed by psychological flaws that we all share: faulty framing, hindsight bias and selective memory, and good data can play a cleansing role. That post represented the “good” that I see in data/numbers, and I thought that this year’s post, for balance, should offer the other side of the argument. I know that data can be misused and manipulated, and that some of my own data has been used to back up specious arguments in multiple settings. In particular, here are three practices relating to data that I find distasteful and suggestions on how you can counter them.

1. Data to intimidate: An article in the Wall Street Journal  pointed to fact that people who are unfamiliar with numbers tend to give them too much weight to them and are particularly swayed by "mathematical" arguments, even if they are nonsensical. It is this weakness that is used by some number crunchers to intimidate those that may not have the same degree of facility with numbers. I have seen corporate financial analyses and valuations where analysts use table after table of numbers, to bludgeon others into submission, using acronyms, jargon and greek alphabets to further the rout.
The counter: The best weapons against number intimidation are common sense and a focus on the big picture. I hope that having access to my data will give you some ammunition in this endeavor but having a solid grounding in first principles of valuation and corporate finance alway helps.

2. Data to mislead: If you have access to a great deal of data, you can parse the data and choose pieces to back up a preconception or argument that you want to advance. A couple of years ago, the effective tax rates that I publish on my site, for US companies, were used by some to advance the argument that US companies were not paying enough in taxes. Looking at the 2013 update on tax rates, that number is low (14.93%), but it is the average effective tax rate across all US companies, including those that are money losing (and thus paid no taxes). Looking only at money-making companies, the average effective tax rate is 28.37%, and the weighted average tax rate is even higher at 30.05%. So, if you have an agenda, you can take your pick to make the argument that US companies pay too little, just enough or even too much in taxes.
The counter: While there is little that you can do to stop people from using data selectively, you can counter their arguments by presenting them with the numbers that they are ignoring. In fact, it was in response to the tax rate debate that I started reporting the average tax rates for money-making companies and aggregated tax rates in my datasets.

3. Data to deflect and evade responsibility: Many analysts use data to avoid making tough judgments about businesses or dealing with uncertainty. Thus, assuming that a company will earn a profit margin typical of the industry is much easier to do than analyzing its competitive advantages and estimating a margin, based on your assessment. Similarly, using a historical or a service supplied equity risk premium in valuation is far simpler than estimating one, based upon the macroeconomic risks that we face in markets today. In fact, using an expert or a service estimate of these numbers (using an equity risk premium from a data service like Ibbotson or even my website) allows analysts to claim immunity from errors and to pass the buck, if the numbers turn out to be wrong in hindsight.
The counter: I have absolutely no concerns about you borrowing data and spreadsheets from my website but please make them your own by adapting and modifying them to not only fit your needs to but also to reflect your points of view.

I hope that you find my data useful in your work or research. If you do, that is more than sufficient return on any time that I have invested in putting it together. If you can think of ways in which it can be more useful or complete, please do let me know and I will try my best to incorporate those suggestions into next year's update.

37 comments:

  1. Very generous of you, Professor. Thank you!

    ReplyDelete
  2. Prof, what is your thought on the next generation of investors. I just saw in one of the yahoo message boards..where there were few people urging not to vote for certain people and also disallow them with any benefit..
    since more and more people, who invest in stocks..are going to message boards for whatever reason..

    What do you think about kind of activity in the internet by people to encourage/discourage others to take certain action on a particular stock.

    I am just thinking as 20 years ago, these things didn't happen..I wonder how the social structure of the investment world will change 5 years from now. For someone like me, it is extremely helpful. Thanks to the internet..that I'm following people like you.

    I hope my question makes sense and apologize that it is off-topic to you original discussion.

    ReplyDelete
  3. Dear Professor,

    In your sheet

    Earnings. Book Value and Sales Multiple Averages by Country

    the aggregated (not average) values are all 0's. Is this an error or will they no longer be provided?

    ReplyDelete
  4. Dear Professor
    Thanks for continuing to provide such useful data at no cost. It is invaluable to individual investors like me for whom access to (free) data is usually not available.

    I was looking at your betaIndia and Indiacompfirm spreadsheets and find that the beta figures in Indiacompfirm spreadsheet in columns G and H are different from the beta numbers in the betaIndia spreadsheet columns C and F.

    I would think the beta numbers (column H in Indiacomp and column C in betaindia) should be the same but I am probably reading the spreadsheets improperly and would appreciate your help / clarification.

    Loganathan

    ReplyDelete
  5. very interesting and insightful. I am adding you to my FAVORITES! A high honor, that if you add 5>50$, gets you a small coffee at Starbucks...

    ReplyDelete
  6. Great information, thanks for providing! As for the cost of capital calculation (specifically referencing the US data compfirm.xls, columns CB-CE), can you explain the cost of debt calculation? E.g. the pretax cost of debt for Ford is 1.00% based on the spreadsheet, a lookup from the 3-yr standard deviation in stock price (column AK) to the table in G1:I9.

    ReplyDelete
  7. Anonymous,
    Thank you for noticing the cost of debt. That was a screw up. It should be fixed now.

    ReplyDelete
  8. Loganathan,
    A good catch. My mistake again. Should be fixed now.

    ReplyDelete
  9. Thank you for this historical summary.
    When instead, you take all Yahoo downloads of publicly available historical data of stocks publicly listed on Nasdaq, the NYSE and the ASE, you end presently up with 5630 stocks, a little less than the number you quote for listed US companies. As it only takes a few minutes to collect this General Watch List, and not even an hour to collect all available historical exchange (non-fundamental) data back to 1962, it becomes an almost trivial exercise to back test any set of selection rules. You may add historical fundamental data and historical forward looking statements like analysts’ estimates and revisions thereof and even the entire collection of delisted stocks. One of the peculiar things with a back test on forward-looking statements is that if you would have known them a few days to a few weeks early, you really start to make money, and that does not appear to be the case with quarterly reports.

    When you weekly randomly pick 50, 100, 500 or 1000 stocks from a collection of 5000 stocks over the past 50 years after eliminating stocks with an average daily trading liquidity of less than $1 Million, your CAGR of the equally weighted picks becomes somewhat larger than 17%/year independent of the number of picks. Hence, over the past 50 years, these random picks collected 100x more money than the S&P500 over the same period. On a semi-logarithmic scale, this performance approaches a straight line, even up to today.

    The question that I have concerns the rigorous division into Sectors or Industries you use. Yahoo identifies 8 market sectors and 215 industry categories. My findings show that any limitation in market sectors and industries lowers the annual returns of the random picks. I understand that profitability is different in different industries, and I also understand various valuation models. But do you think that share prices really reflect such valuation models?

    ReplyDelete
  10. The marketplaces are the same now as they were 5 or 10 decades ago because they keep changing-just like they did then.

    ReplyDelete
  11. We are a network of short sale agent of Real Estate agents throughout the United States who specialize in helping home owners who are in a difficult financial situation by offering free home owner consultation services as well as free short sale processing and brokerage services.so being with american destiny realty.com.

    ReplyDelete
  12. Thank you professor! Your data have been the most helpful to my investment operations. I can't thank you enough!

    ReplyDelete
  13. Thank you professor we are all better investor reading your thoughts.
    Greetings from Bosnia.
    Saša

    ReplyDelete
  14. Thank you Professor for all the valuable support you provide to us.

    Utsav Jha

    ReplyDelete
  15. I touch really precise version these articles I stingy there are writers that can create moral stuff.voip

    ReplyDelete
  16. I love your blog and ideas - particularly when focused on business - please keep it up. However, the contamination of your ideas by MPT, "beta", and similar nonsense (primarily focused on short-term price movements) is regrettable.

    ReplyDelete
  17. Professor,

    The names of some industries in the Total Beta By Industry Sector for CHINA are doubled - there's two each for Healthcare Equipment, Facilities and Services but there is no data on Heavy Construction, Homebuilding, etc.

    Also, it seems that Debt Ratio Trade Off Variables by Industry for CHINA were not updated for 2013.

    Thanks for your efforts and all the best!

    ReplyDelete
  18. I don't know why it happened, but I think it is fixed now.

    ReplyDelete
  19. Thanks Prof. Btw i just drop by to know what is the different of market risk premium calculated using implied risk premium and market risk premium from service provider by bloomberg ( they are using market cap * avarage ddm index). And like you said, many people still let go their responsibility to service provider for the data eventhough they dont know what the principle behind the formula and how service provider derive the number.

    ReplyDelete
  20. This is regarding beta information for India. Pl let me know if my comments are correct on them.

    1) Bottom up Beta for sector - This is unlevered beta for the sector
    2) Bottom up levered beta - This is levered beta for the sector
    3) Beta (column AC) - This is levered beta for the firm (business)

    Your comments will be useful in interpreting the information supplied.

    Thanks



    ReplyDelete
  21. Professor,

    When are the spring corporate finance and valuation classes starting?

    Thanks

    ReplyDelete
  22. On the betas for Indian companies, you are almost right. The only difference is that the levered bottom up beta is the sector beta levered up at the company's debt to equity ratio.
    And my spring classes start a week from Monday (February 4, 2013).

    ReplyDelete
  23. Professor
    For calculating the cost of equity and cost of capital you have used the global betas (indiacompfirm file) and not the purely Indian betas so to say from the betainda file. Is there any particular reason for this? I ask this because there is a substantial difference in the two beta numbers as is to be expected and for my own valuation exercises I am not sure now which beta to use. Your clarification would greatly help.

    ReplyDelete
  24. thank you professor

    ReplyDelete
  25. Nice post!!your post is very informative and impressive.

    ReplyDelete
  26. "For calculating the cost of equity and cost of capital you have used the global betas (indiacompfirm file) and not the purely Indian betas so to say from the betainda file. Is there any particular reason for this? I ask this because there is a substantial difference in the two beta numbers as is to be expected and for my own valuation exercises I am not sure now which beta to use. Your clarification would greatly help. "

    Pls. do reply.

    Its not a ritual bt a useful tool for anyone followin u.Keep it up sir.

    ReplyDelete
  27. The odds associated to numbers are not a taboo for me anymore. That’s what I thought, but I stand by corrected with this post. There are many other negatives on numbers I didn’t realize they are until I read it here.

    ReplyDelete
  28. I'm not good actually on numbers that caused on me sometimes to hate it. It seems my head aches when I think numbers. But, I believe it cannot be avoided. And, it's good to know someting related about numbers that I found here.

    ReplyDelete
  29. We're looking for 10,000 individuals who want to receive closing stock quotes every 24 hours. We have over 35 countries to choose from. WE DON'T WANT YOUR CHARGE CARD INFO. There is NO charge for this data as we are testing a new stock delivery system and need 10,000 individuals to stress test the system.

    If you are interested, please send me an email providing me with the following information:

    First & Last Name
    Zipcode
    Email Address where stock data files are to be sent.

    Tell us which stock exchanges (up to four) you wish to receive on a daily basis.

    What's the catch??? No catch!

    Please help us spread the word.

    Tom - thomasfbuchanan@gmail.com

    ReplyDelete
  30. On Indian data update:
    Cost of equity (column J) and Cost of capital (column N) are calculated on dollar terms.

    However (ROE - Cost of Equity) and (ROIC - Cost of Capital), i.e columns O and P - are based on different currencies.

    Due to this companies appear to have excess returns, but actually are not.

    Am I missing something? Please advice. Thanks for help.

    Sunny

    ReplyDelete
  31. Well said, the post is really the freshest on this valuable topic. I fit in with your conclusions and will thirstily look forward to your next updates.best web hosting Bangalore

    ReplyDelete
  32. Remarkable blog! I have no words to praise, it has really allured me.
    disability living allowance contact number

    ReplyDelete
  33. That’s really a nice one, I have seen many blogs but they are outdated so I’m pleased to see this blog now.
    Bevis

    ReplyDelete
  34. Hi to everybody, here everyone is sharing such knowledge, so it’s fastidious to see this site, and I used to visit this blog daily.working tax credits phone number

    ReplyDelete

Given the amount of spam that I seem to be attracting, I have turned on comment moderation. I have to okay your comment for it to appear. I apologize for this intermediate oversight, but the legitimate comments are being drowned out by the sales pitches and spam.