This is the second part of a two part series where I interview John Schroy, author of Capital Flow Analysis and the founder of Capital Markets Wiki. From the takeaways from our discussions, it’s clear John has put a lot of thought and analysis into thinking about collaborative, crowdsourced investment research.  Just wanted to start the discussion.  I’d love to hear from others who are focused on this space.  Interested readers can find Part I here.

Can you describe CMW, its genesis, any usage statistics/anecdotes that provide some color? What’s the future for the site?

John Schroy: CMW is a “semantic wiki” organized around a strict informational structure called “Capital Market Taxonomy”. It was opened to the public in March 2009, without promotion, and has few editors at this time. However, it has a robust policy base and extensive documentation of its taxonomy.

Today, it has 2,136 articles in the main encyclopedia, almost all of which are “stubs” but which serve to demonstrate the “semantic” features of the site.

The advantage of a “semantic wiki” is that it is possible to quickly select articles based on highly specific criteria, such as “a list of all derivatives settled in dollars or euros relative to soybeans, wheat, or corn” with a linked table showing contract terms, the market where traded, and other data.

On opening the wiki to the public, I found that people were having a hard time understanding the purpose. This led to a complete revamping of the front pages and help sections.

Two months ago, my daughter, who is a portfolio administrator, offered to try to set up an article and found this difficult to do. Again, it was back to the drawing board and an “easy setup” feature was added.

In mid 2008, I made a pre-launch presentation to a group of financial consultants in Southeast Asia, which resulted in creating special templates for the use of college professors.

Issues regarding a more appropriate server for the wiki are still being resolved.

The future of this wiki is expected to depend upon the willingness of university professors to use it as a teaching tool for open source financial research. With this as a basis, it is hoped that the “tipping point” will be reached, with enough articles and editors that the wiki will grow on its own.

Once this point is passed, it is expected that institutional sponsors will support a non-profit foundation, providing the modest funding needed to cover server expenses as the wiki grows.

In order to ensure impartiality and lack of bias, there is no intent to admit advertising to the wiki, nor to allow any institution to dominate the supporting non-profit foundation.

In a recent article on Buffett in The Atlantic, What Would Warren Do?, the reporter cited that so much free information has taken away the value investor’s advantage. In your article opportunities in OSINT, you mention that this free open source information is largely unexploited. Can you explain and do you think more info/data makes it harder to outperform?

JS: Although there is an abundance of “free” information on the Internet — more than anyone can hope to digest in a lifetime — this information has a “time cost” to turn it into investment intelligence. Only a tiny portion of this information is published in Standard & Poor’s and Moody’s. Capital Markets have become complex and most information is “buried” in official documents, covered with boilerplate, disclaimers, and useless “fluff”.

For example, the closed-end fund NRO was rated five stars by Morningstar and its preferred shares were given an AAA rating by Moody’s, but the “toxic” features of its preferred shares were not revealed in its annual reports to shareholders and the fund was not even reported by Standard & Poors. Like similar funds, it fell dramatically with the Crash of 2008. I did a research article in CMW on this issue and it took 75 hours to dig out the relevant facts ( I could have expended ten times this effort in finding out all that was available about this relatively simple company.)

The Crash of 2008 showed that the market is far from efficient and that the “Efficient Market Hypothesis” is nonsense. Billions are invested in “index funds” without any research at all. The information base available to Benjamin Graham was tiny compared to the tsunami of free information available today. However, in Graham’s day, most information was easily available through “Standard Statistics”, the ancestor of S&P. Because companies were so much simpler, this information carried the analyst a long way to getting an advantage over those who didn’t know how to interpret the data.

When I started in investment analysis, in Rio de Janeiro in 1959, although there was no equivalent to S&P (until I set one up), companies were exceedingly simple and easy to understand. The terms and conditions of issuers were more or less standard. Companies had few, if any, subsidiaries. Most had a single line of business, like brewing beer or making cigarettes. A few ratios from the financial statement and a trip to visit the factory could give one a good understanding of the business and its prospects.

Today, the problem is exceeding complexity and the flood of non-standard terms and conditions that few have the patience to digest. Technical analysis is far more popular today than fundamental analysis and dominates most investment decisions.

Without help in collaborative research, it is difficult to attack the information tsunami alone. Because “toxic” details can be well buried among irrelevancies, it takes a lot of time to understand today’s opportunities. Some companies (like the big international banks) are so complicated that even top management doesn’t understand them. I know, because I’ve been in that position.

The idea that markets are efficient because of so much “free” information is simply not true. The information is not “free” at all — there is an enormous “time cost” of digging it out.

You mention the inner dynamics of leaderless organization must show that the ROI is positive for contributors. What Is the qualitative ROI for contributors to CMW? What about free riding? It’s a lot easier to sit back and wait for others to do the work. How can i incentivize others to do the work that’s important to me?

JS: This is really the key to whether CMW will be successful or not.

First of all, an analyst need not be afraid of “free riders”. (See: http://www.capital-market-wiki.org/wiki/index.php?title=Help:Free_Riders). CMW does not provide “analysis”, opinions, or recommendations — only facts. To be able to judge the reliability of these facts, one has to be quite familiar with the workings of the encyclopedia. Also, information is spread over many articles, connected by a semantic network. Most casual readers won’t fully understand how this works unless they become editors. Finally, any “free rider” who really understands the encyclopedia will be tempted, now and then, to make a correction or addition — and will no longer be a “free rider”. Its not necessary to be a heavy-hitter analyst to make valuable contributions to a wiki. Just removing spam or vandalism, or correcting grammar or style is useful.

CMW also needs editors with many different skills — no one is expected to know everything. For example, one contributer might be a lawyer in Spain, another can translate from Mandarin to English, still another has information about the tax laws in New Zealand. Most editors of Wikipedia contribute only one article. Others only do little things, like clean up typos. However, with millions of editors, the results are astounding.

Furthermore, experience with Wikipedia indicates that active editors are more likely to cooperate and help other active editors. For example, if you are interested in a stock traded on the Jakarta Stock Exchange and there is an active editor on that topic, he or she is more likely to be willing to help you answer a specific question if you already have a “wiki reputation” as an active editor. CMW has special procedures to show one’s “wiki reputation”.

With regards to motivation, the CMW model presumes three target editors:

• Educators: Academic professors and students.
• Investors: People looking for information to guide investment decisions.
• Entrepreneurs: People looking for opportunities to make money.

If you go to the wiki and follow the links from the first page, you’ll eventually get to a series of sub-systems that support these different target markets.

For example, most information in CMW is open source and can be republished and reconfigured without royalties. An entrepreneur may take this information, add investment advice and advertisements, and publish or sell the feed. The wiki provides a number of suggestions a how entrepreneurs can benefit from the wiki.

However, for entrepreneurs to benefit, they will need to help the wiki grow, either by helping to recruit academic or institutional patrons, or by recruiting editors. No wiki, no benefit to entrepreneurs.

CMW can also be used by financial institutions to outsource or home-source investment research, for a dramatic cost reduction. The basic idea is that is costs a lot less to hire a stay at home mom with a CFA in Boise, Montana than a Harvard MBA in New York City. CMW provides the templates needed to manage such out-sourced jobs (in fact, the same templates used by college professors to give assignments to students)

See: http://www.capital-market-wiki.org/wiki/index.php?title=Help:Editor_benefits

For information on possible institutional use of CMW, see the article “Modern Technology for institutional investment research” (http://capital-flow-analysis.com/capital-flow-watch/modern-technology-for-institutional-investment-research.html)

You mention some parallel motivational worlds like finding jobs and promoting research firms. Can you elaborate? Seeking Alpha which is not collaborative but aggregates 3rd part content sees a lot of people marketing their own advisory firms via content, as well as creating a portfolio of work to raise money off of or find a job. Do you see this playing out as well on CMW?

JS: CMW is designed specifically to provide economic benefits to editors in a number of ways.

Because it is a semantic wiki, tied into Capital Market Taxonomy, it has been possible to set up a formal “job opportunity network”.

For example, if an editor specializes in translating articles from English into Japanese (CMW is multi-lingual), he or she can “advertise” this skill on his or her user page, placing the information in the semantic database. Then if someone else is looking for a person to translate a financial article from English to Japanese, they simple input this information on a form on their own user page and the name of the translator will appear. The employer may then check out the work of the translator on the wiki to see if its satisfactory and then contact the translator directly to offer an assignment.

There is also a formal system of “merit awards”. If someone thought that the translator above did a good job of translating from English to Japanese, they can go to that user’s “talk” page and give them a “merit award” which will be entered in the database. “Merit awards” are also a way to thank other editors for helping you on a job. On every article, you can see who did the most work and who, if anyone, earned “merit awards” for this work.

It is expected and hoped that editors of CMW gain economic advantage from participating in the wiki. In fact, here are some of the special systems (software modifications) that are specifically for this purpose: “merit awards”, “job opportunity network”, “patrons”, “RSS sub-pages”, “assignment templates”, “repackaging services”, and two sidebar links: “main contributors” and “prizes, services, jobs”.

The reference to “parallel worlds” is relative to the global scope of the encyclopedia. It is possible to have a highly developed encyclopedia with reference to the market in Hong Kong, with almost no entries with regards to the markets in Belgium. There can also be parallel worlds with regards to languages. CMW works off a single database and taxonomy, allowing different language versions of the same article to be referenced through the same database.

I’d like to know about governance and because you’re focused on investing with huge incentives to game the system, how your governance differs from Wikipedia’s.

In many ways, the governance of CMW is similar to that of Wikipedia. However, here are some of the major differences:

• CMW formally encourages self-promotion (within limits), whereas Wikipedia actively discourages the practice.
• CMW has formal, hard-wired systems to provide economic advantages to editors (like the job opportunity network) whereas Wikipedia has none.
• CMW has formal, hard-wired “merit awards” that serve not only to help editors get jobs, but also as a peer review system. Wikipedia has informal “barn stars” unconnected to the database or any opportunity network.
• If a small accounting firm in Chile were to place an article in Wikipedia, it almost certainly would be deleted within minutes. In CMW, as long as the information is factual, related to capital markets, and in the proper style and format, the article would be encouraged and allowed to enter the database.
• A consulting firm specializing in, say, cotton derivatives, could write an encyclopedia article on this topic in CMW, placing a relevant RSS feed published by the firm on the subpage of the article. In Wikipedia, this is not possible nor permitted.
• CMW has a policy of “Ethical Promotion”. Wikipedia’s policy is “no blowing your own horn.”

I’d like what you’re doing with taxonomy. Can you just explain why it’s so important and how this works?

JS: Articles in CMW are divided into four categories: Markets, Institutions, Operations, and Instruments. These groups are called “namespaces”.

For examples, there might be articles on Markets; United States; Institutions: NYSE; Operations: Insider Trading Rules (Indonesia); and Instruments: General Motors Common Stock (US)

Capital Market Taxonomy consists of formal, written definitions of what goes in each category.

For example, the definition for the namespace “Markets” is “legal jurisdictions within which capital markets exist.”

Each namespace is further broken down into sub-categories.

For example, the namespace “Markets” is broken down into “nation-state”, “sub-national entity”, “territory”, “monetary union”, “monetary and economic union”, and “status disputed”.

Capital Market Taxonomy consists of several thousand defined categories and other semantic forms.

However, an editor doesn’t have to know any more than the top four categories. When setting up a new article, there is a form on the sub-page that automatically classifies the article into the proper categories.

This makes it possible to find articles without knowing the exact name, simply by concept.

For example, you can make a list of all securities exchanges in Southeast Asia by conducting a semantic search.

All editors have a “private sandbox” (something than Wikipedia doesn’t have) where they can work on articles or set up private “tables of contents” or “working lists”. If you are researching articles about securities exchanges in Southeast Asia, a few lines of code in your “private sandbox” gives you a working table of contents with links to all articles in this category.

Of course, as the encyclopedia grows, far more sophisticated searches can be set up just as easily. For example, you might wish to make a list of all beer companies in Europe that have issued preferred stock or convertible debentures. Capital Market Taxonomy allows this to be done quickly and securely.

Furthermore, lists created in this way are automatically updated. If a new beer company appears in Europe with convertible debentures, it would automatically appear on your list.

Capital Market Taxonomy makes it possible for editors who do not know each other or perhaps even speak the same language, to find articles of mutual interest simply by concept.

Although Wikipedia uses categories, there is no taxonomy or documentation and the list of categories is in the several hundred thousands and virtually useless. CMW enforces taxonomy discipline in creating new categories and also has two other semantic forms: “attributes” and “relations”.

An important feature of Capital Market Taxonomy are the built-in outlines for over seventy types of articles. These serve as research guides, suggesting questions to be answered and the format of articles. Editors can set up an outline in a few minutes, appropriate to a specific topic. For example, one type of outline is available for “preferred shares” while another is available for “financial market regulators”. (See: “Recommended Formats at http://www.capital-market-wiki.org/wiki/index.php?title=Help:Recommended_Format).

The best way to learn about Capital Market Taxonomy is to read the help files on the wiki (which are extensive) and then test out the system in your private sandbox.

Note: Although any reader can use the features of Capital Market Taxonomy, casual “free riders” may not even realize that the system exists. This suggests why active editors are more likely to benefit from CMW than the casual “free rider”.

Very interesting, but again, why is this so important to the future of collaborative, crowdsourced investment research?

JS: The primary purpose of Capital Market Taxonomy is to make the job of collaborative research easier and more efficient by introducing the concept of “every fact in its proper place”, reducing the “time-cost” of fact gathering.   In other words, the benefit is for editors primarily, rather than readers.

Taxonomy is not needed if an analyst is working alone, but if many analysts (who don’t know each other) are to work together efficiently, they need to have an prior understanding as to the purpose of the research and where the results are to be published.

Capital Market Taxonomy makes this more or less automatic in the process of setting up new articles, in filling out the form that creates the “InfoBox”, and in using the recommended formats for each type of article.

Wikipedia, for example, has no taxonomy whatsoever.  The result is duplication of articles, lack of standard formats and uncertainty as to whether someone has already written something about a certain topic.

See: “Articles to be merged” in Wikipedia, (http://en.wikipedia.org/wiki/Category:Articles_to_be_merged), which shows over 14,000 articles in this category as of August 2009.  There is also another category for articles to be split up.