The landscape of digital assets and blockchain technology has rapidly evolved, introducing concepts such as smart contracts, stablecoins, tokens, non-fungible tokens (NFTs), and other web3 technologies. These advancements have expanded the definition of assets, markets, and economies. Digital assets are no longer limited to payment tokens or stores of value; they now encompass works of art, virtual real estate, and entertainment assets.
These new assets pose challenges for financial reporting, both conceptual and methodological. One such challenge is fair value pricing, the standard required by US and international accounting standards (US GAAP and IFRS) for reporting value in financial reports.
Fair value pricing is a crucial aspect of the mark-to-market accounting method, which mandates that asset valuations in financial reporting should accurately reflect current market prices, with regular adjustments made to account for prevailing market conditions. Under this accounting standard, the value of an asset is determined by its price on a specified principal market, which is typically the exchange where the asset is officially listed and traded (e.g., NYSE, Nasdaq, CBOT). In other cases such as globally traded commodities or foreign exchange, the principal market is determined to be the market with dominating volumes, e.g. EBS and Reuters for foreign exchange, ICE and NYMEX for oil. These markets are local to some geographic region and are active a limited number of hours a day. Assets are exchanged against a sovereign fiat currency, and apart from potential currency conversions, fair value is derived from the price of the most recent transactions on the designated principal market. This ensures that assets are reported at their current market values, providing transparency and accuracy in financial reporting.
In the universe of digital assets, trading is global and happens across many different venues without a specific venue dominating any of the others. Moreover, most transactions are swaps of digital assets against other digital assets, this is true even if one of these assets is a stablecoin. Even for assets that trade against fiat currencies it is often the case that a small fraction of the volume is converted to fiat (on/off ramping) while the real drivers are crypto to crypto swaps. With such fragmented markets, it becomes impossible to apply existing standards to reflect market realities.
The Principal Path Method presents an alternative approach to pricing assets, considering the unique structure of the market and the diversity of assets within it.
When an asset doesn't directly trade against USD or trades with low volumes, it can be indirectly converted to USD through various asset sequences. PPM aims to establish a methodology for selecting a pricing path, similar to how accounting guidelines determine principal markets for pricing assets. However, due to the dynamic nature of the market, PPM ensures the designation of the principal path remains adaptable to changing conditions, meeting the criteria of fair value including availability, reliability, and relevance.
To illustrate, given assets like BTC, ETH, USDT, USDC, MANA, GLT, APE, etc., we define exchangeable pairs as those traded on compliant public markets with established exchange rates. A pair like (USDT, ETH) is exchangeable across multiple exchanges, while (DOGE, APE) is not. These assets and pairs form the exchangeability graph. A pricing path connects two assets, possibly across multiple exchanges and blockchains, considering fees, risks, and other factors.
The PPM assesses each path's quality by considering asset traits, exchange adherence, trading volumes, data recency, transitions, and other factors. These path scores are derived from corresponding scores linked to exchanges or blockchains. These scores are categorized into compliance, volume, decay, and latency. A path's compliance, volume, and decay scores are determined by their respective minimum scores along the path, while latency accumulates across the path. The principal path, having the highest score, is selected for pricing the asset in an ephemeral manner.
The library is an implementation of the PPM method in python, designed to compute historical PPM fair value prices from historical data. The different components of the score have very different time scales during which changes and updates take place. The compliance score is assessed quarterly, volume and latency on a monthly bases and decay either tick by tick or in short intervals. As such the libraries are designed to use pre-computation and cache intermediate computations as much as possible. Historical data used in the demo and stress test covers ten assets and four exchanges adding up to 30G of compressed tick data covering the period from November 2nd to November 30th 2022. In addition another 5G of disk space are required to store cache data.