I've been helping out with a new music blog, Federal Audio Reserve. The plan is to provide a little bit of background into how electronic music gets made. The dirty guts that doesn't quite make it in the material you hear once it gets released. A range of artists are invovled, each doing something slightly different with electronic music - but sharing a common interest in music technology.

I shamelessly stole the above photo of Mr. Volcker from the NY Times, and photoshopped it a little to add some more gravitas to the man. I like to think it now says:

Yeah, I'm smoking, wearing bifocals and listening to tunes. You have a problem with that? Stagflation had a problem with me. See how far that went.

Click on the picture for the video.

---

for the bandwidth deprived, there is a youtube version.

No one writes an annual letter like Warren:*

A story I told you some years back illustrates our problem in accurately estimating our loss liability: A fellow was on an important business trip in Europe when his sister called to tell him that their dad had died. Her brother explained that he couldn.t get back but said to spare nothing on the funeral, whose cost he would cover. When he returned, his sister told him that the service had been beautiful and presented him with bills totaling $8,000. He paid up but a month later received a bill from the mortuary for $10. He paid that, too . and still another $10 charge he received a month later. When a third $10 invoice was sent to him the following month, the perplexed man called his sister to ask what was going on. "Oh," she replied, "I forgot to tell you. We buried Dad in a rented suit."

---

*: Bill said it was OK if I refered to him on a first name basis.
ht: Tanta

I can't sleep, and I feel the need to rant, so please excuse the indulgence... Comments represent the will of my loud cats keeping awake more than those of my firm; caveat emptor, etc.

---

Each quarter, companies around America file their regulatory statements with the SEC. The SEC wants to ensure that the investing public, both institutional and mom and pops, have access the fundamental information that is relevant to understanding the essentials of corporate performance. This information is published in a very loosely structured format. It is an absolute nightmare to parse this information into a database free of errors. Plenty of services will sell you parsed SEC data in cleaner formats. At our firm, we get access to 10K and Q information via at least ten different channels. In some cases, the same parent company provides the same data via 4 subsidiary channels, each of which are totally un-interoperable with each other.

Take the simple problem of tickers. Anyone who has dealt with equities trading technology knows about this nightmare. The problem is that there, despite the existence of a few quasi-standards, are no true standards that specify how to deal with the concept of 'a single company' with 'a single equity issue' trading on 'a single exchange'. Each of our providers of SEC data has come up with their own unique, and incompatible, way of solving this problem. CRSP has a solution. Reuters/Bridge has another. Reuters Knowledge Direct has yet another. Bloomberg has one. Yahoo Finance and Google each have their own. CapitalIQ has another. Our prime broker has one and our IT team has come up with their own too. And our traders often just make up new symbols when it all gets too much. And the SEC has their own too. In addition, there are various (flawed) standards, CUSIP, ISIN, SEDOL...

This means that each trading firm, somewhere in their backend systems, a big cross-referencing table. Ok. Not a pain to deal with. The cross-reference table that I use works for 99% of the companies we follow, and that's close enough. But the only reason why this simple problem hasn't been solved is that none of the institutions selling their data have any incentive for interoperability. Despite what we heard today around the halls at Money:Tech. If we can't get to a point whereby trivial company identifiers interoperate, there is little hope for grander 'mashups.'

On a panel today, a company made the case that they embraced openness, evidenced by the fact that they freely publish a small subset of their data on Yahoo Finance. Great. So what if a quant savvy small investor wants to put together a database, albeit only containing the 10 line items that this company choses to share online for 'free'? Well, they would have to write a bot to scrape this information to compile the data. As someone who has done this type of stuff, I can tell you that Yahoo is not a fan of this type of behavior. Likewise for Google. If you write a simple bot to scrape this stuff, you will find yourself blacklisted from the service pretty damn quick. In the end, the small investor will end up spending more time trying to hide their botting activity (which violates Yahoo's terms of service) than actually collecting and working with the data.

Even worse is that if our savvy investor spent the time to write software that could competently parse the raw files as published on SEC's EDGAR, they will find that even the SEC, whose purpose of late is solely ensuring that price sensitive information is shared freely to all, takes the same attitude as Yahoo and Google and will shut down any bot that tries to grab data from their site automatically. So even with the introduction of XBRL, which should mitigate many problems with the current EDGAR file format, it wouldn't surprise me if the small investor is no better off in the end.

For all the promise of openness and the virtues of information sharing, primary investment research finds itself operating with a business model more akin to BMG and EMI than represented at today's conference. Through the heavy-handed use of legal, and ultimately anti-technical, constraints, the primary research providers as inflicting their own DRM on information that wants to be free. And whereas commercial music is sponsored through merchandising and tours, there is no clear adjunct market for raw numbers. In a way that is more pressing than for the music industry, these 'research' providers need to wake up and find a better model lest risk irrelevance.