Thursday, March 6, 2014

Do We Need Ratings Agencies?


TED.com recently published a talk by Annette Heuser about ratings agencies. She argues that the current ratings agencies need to be replaced with a new, non-profit one. This, she feels, will resolve current potential conflicts of interest and will, presumably, lead to better ratings. Specifically, her focus is on giving sovereign debts higher ratings in order to reduce borrowing costs for countries which have a poor credit history.

Unfortunately, her proposal suffers from a misunderstanding of basic economics. The issue is not public vs. private. The issue is who pays. Currently, as she says, the issuers (who are being rated) pay, causing the potential for conflict of interest. The potential conflict obviously doesn't go away if the rating agency is non-profit. A non-profit still needs to cover the costs of their research and rating process. Thus, the very very simple fix is to switch the payer from the issuer (country or company) to the lenders and buyers of the securities (banks or investors). This could take the form of a transaction fee, or it could take the form of a consortium whose budget is covered annually by buyers/investors. I would strongly suggest that Annette refocus her attention on this aspect and ensure that INCRA is funded thus. Her focus on non-profitism, transparency, breaking the grip is just a set of side-shows.

But wait, there's more. An achilles heel of any ratings agency is their reliance on their chosen rating methodology/algorithm. As we saw in 2007, securities with AAA ratings crashed. This objectively demonstrated that the prevailing rating methodologies were inadequate because they did not accurately factor in all risks and probabilities and/or did not accurately represent them to consumers. Ms. Heuser's proposal to clone the current (flawed) rating-agency structure would no correct this problem. A far better solution would be to empower investors.

How?

Replace ratings (the conclusion of one entity's analysis) with the ability to rate. In other words, give investors the ability to formulate their own conclusion, and to see what others have concluded. To do so, empower investors with powerful analytical tools that can access rich, diverse, open-source data.

Now is the time. All of the component pieces are proliferating today. OpenGovernment info can be combined with open-source historical markets and financial data, open-source regression models, data analytics, data visualization tools.

Of course, there are issues to be sorted out. The first to get these right will have a massively winning business model:

  • Making money - who pays? for what? how is it priced?
  • Messy data - to be valuable, the platform needs data from a multitude of sources. Necessarily, data will thus be heterogenous in terms of structure, quality, accuracy, completeness, encoding, and definitions. To be useful, the platform will need to provide ways of using a heterogeny of data without degrading quality of outcomes
  • Source Neutrality - are certain data sources "better" than others? Should certain data providers be made preferential? Should opinions or interpretations or results of analytics be included in the data? If so, how should they be differentiated from raw facts?
  • User Neutrality - should all users have access to all data? Should users have to subscribe to specific data sources/quality levels/time periods?
  • Modeling skills - Will users have the ability to create their own valid ratings? Will they have the requisite knowledge of data, probability concepts, risk concepts, details of financial products? How can the platform abstract these concepts? 

Bottom line, ratings agencies are relics of a prior age, just like Encyclopedias and farriers.

Sunday, March 2, 2014

How Your Customers Should Think About Bitcoin

Bitcoin reminds me of the Gold Standard or UN SDRs. They're great in concept, but probably needs to weather a good crisis or two before I'll trust it.

The first thing to understand is that it is a bearer-owned stored-value mechanism like gold, dollars, or bearer bonds. It's not a payment network like paypal. It's not a payment mechanism like a credit card. It's not an account. Whoever physically "holds" the bitcoins owns their value.

This leads to one great benefit: Unlike the USD or other national currencies, no politician can use them as a political tool by using inflation to ruin the value or using executive orders to inflate the value of Bitcoins. So, like gold, they should be safe stores of value, immune to a specific country's situation.

But the reality is far less sanguine or simple. There are problems:

  • The notion that it is outside the influence of governments hasn't been tested. I'm sure government lawyers are working on this. Every single time a bitcoin changes hands, its authenticity is verified at the central database. This means every transaction and transfer can be monitored. The amount of data is too juicy for the NSA to ignore. But will other parts of the government (executive branch, for example) also get ahold of the info? 
  • The way they determine the value of the currency. Bitcoin is based on supply and demand. Because it is a small market with relatively few traders, it is easily manipulated. The issuing company says they are increasing the supply steadily based on the size of the economy. The good news is that, since 
  • Counterfeit. Every computer system has weaknesses. Even the most secure networks have been attacked, and most of them have been compromised at some point. I don't trust that Bitcoin is secure enough, if for no other reason than because it hasn't been attacked hard enough. 
  • Fraud. There are system-administrators in the Bitcoin foundation who must have special system access in order to do their jobs. These people could certainly be bought off for the right price. We haven't invented a really technologically fool-proof way to avoid this yet. Mt. Gox is the obvious example of what can go wrong here.
  • and, of course, not many places accept bitcoins as payment (currently)