Tuesday, September 29, 2015

Decoupling Financial Indices with Decentralized Bitcoin Fact Generators

Financial indices such as the Dow Jones Industrial Average or the S&P 500 are well known.  In the age of ETFs and ETNs, a core index is a requirement of the investment product.

In the age of decentralized software, this will be further decoupled into networks of fact generators and verified algorithms.

Creating an index such as the S&P 500 requires two primary components:  Input data (stock prices), and an algorithm (criteria for selecting stock X with proportional weight Y).

Using technologies such as cryptographic hash functionsmerkle treesbitcoin blockchain timestamping and bitcoin oracles, a better, more secure, more transparent financial index system may be developed.  Let's call it "Index-NG."

In the Index-NG system, the algorithm - the software - that turns volumes of input data into "The S&P 500 closing price" or "current gold price at 12:01pm" would transform from a clunky Excel spreadsheet (yes, really) or proprietary S&P software into

  • Open source software
  • Written in a smart contract language such as bitcoin scriptMoxie or ethereum.
  • Secured against corruption and tampering via blockchain hash
  • If not entirely in-chain (bitcoin script, ethereum), processed by a network of oracles run by separate businesses/individuals.
This index algorithm architecture increases transparency and reduces the level of trust we place in any one organization or developer.  The level of peer review is greatly increased.  Auditing is a breeze.

To further decentralize and reduce trust required in the index algorithm, the index's input data is now considered.  The collection of raw data becomes a key act in a decentralized world.

Raw field data is collected by data sensors, and securely stored in the blockchain:  Stock price data, climate station weather data, air pollution data, and more.  Hashing and merkle trees are used to aggregate large volumes of data into small, secure blockchain anchors.

The software and actors that collect the raw data and securely store it in the blockchain are fact generators.  Fact generators are the second half of a decentralized financial index.  Their role is best illustrated with some examples, and is central to the security of the entire system.

Creating an index such as that S&P 500 requires building a secure digital loop for publishing its data and algorithms:
  • NYSE and NASDAQ publish digitally-signed intraday or closing prices, hashed into the blockchain.  Publish this hash in the New York Times and Wall Street Journal stock sections, too!  NYSE and NASDAQ play the role of fact generators, here.
  • Standard and Poor's publishes a digitally-signed S&P 500 algorithm, hashed into the blockchain.
  • Any bank, government agency, individual or machine-based agent may then independently generate the S&P500 index at any time, secured against tampering, with two simple pieces of information:  The hash of the algorithm, and the hash of the data summary.
Creating an ETF, then, becomes a second layer of decentralized algorithms which trigger trades in an ETF's primary markets.  ETFs can exist and be run 100% human-free.  With bitcoin as the value token, both the stock price and the value exist on the blockchain as digitally provable values, ensuring an autonomous agent or DAC can prove with 100% certainty that certain trades should/should not be executed.

Another example is measuring air quality or climate data, a dataset perhaps more subject to manipulation (or accusations thereof).  One can imagine
  • 1st layer: A network of Beijing air quality sensors or US-based climate temperature sensors securely timestamps their data into the blockchain.
  • 1st layer: Satellite infrared and smog imagery is securely hashed into the blockchain.
  • 1st layer: Bitcoin/USD exchange rate data is digital signed by each bitcoin exchange, and securely hashed into the blockchain.
  • 2nd layer: 10 governments and NGOs around the world publish their assessments of this data - and the models/algorithms used to achieve the assessments.
  • 3rd layer:  IMF and other agencies run automated agents which transfer bitcoin value based on the pollution/climate assessments, modified by bitcoin/USD exchange rate to eliminate volatility.
In this example, the fact generators - air quality sensors - are mostly untrusted.  A 2nd layer of software - also fact generators, generating derivative facts - achieves a consensus or quorum over untrusted data.  The 3rd layer of software than acts upon that quorum of derived facts.

In a decentralized world, the gathering of raw data, signing, hashing and synthesizing it - fact generation - becomes the key act upon which software will automatically trigger further actions - including real world actions such as hiring humans, moving shipping containers from point A to B, delivering groceries and more.

Decentralized software - using secured digital facts, running on blockchains (bitcoin, ethereum) or networks of oracles - will form an ecosystem that makes the entire world operate on a more transparent, more efficient, less corruptible basis.

The essence of smart contracts is executing a series of actions (and inactions) based on computer processing of digital facts.

1 comment:

  1. Awesome post. Another layer could be a set of oracles that reference a binary option market or prediction market as an additional layer of redundancy.

    This layer would create an additional market incentive for the outcome to be correct. The data can be trusted incrementally as a function over time akin to block confirmations.