From
Science
to
Private
Market
Tech
Tech Matters for Private Market Data
The technology stack used by Scientific Infra & Private Assets to produce the privateMetrics indices and data was built from a decade of research at EDHEC.
Our approach is based on a simple fact: private markets and investors are irrevocably starved of transaction data. There will never be enough raw private transaction prices to allow investors to measure risk and value in either a robust or representative manner. From there, technology is the only way to make up for this lack of raw data.
Two challenges needed to be addressed: how to model market prices to evaluate assets when there are too few comparables, and how to collect enough data from disparate and often unstructured sources of information. EDHEC and SIPA resolved these problems and created a fully integrated data collection technology to acquire, process and organise private asset data, and an asset pricing technology that can be implemented at scale for thousands of assets on a monthly basis. Finally, we developed an API technology to distribute the privateMetrics data to our clients in the format and platform they need.
Asset Pricing Tech
Despite the limited availability of raw price data, a signal processing approach can be developed to transform observable transactions into market pricing information that can be used to estimate the average level of market prices and build benchmarks that are both robust and representative. Based on the EDHEC Infrastructure & Private Assets Research Institute fundamental research in private asset pricing, the SIPA Asset Valuation Methodology is at the heart of the technology used to create privateMetrics. Thanks to a multi-factor asset pricing model and asset classification frameworks a.k.a. PECCS and TICCS, that help capture the systematic differences between market prices, it is possible to re-estimate the average exit value of hundreds of thousands of assets monthly based on the most recent transactions, movements in rates, asset characteristics, etc. This approach works at the segment or market level i.e. with a robust model, thanks to the law of large numbers, the average prices and index returns computed are error-free and the benchmarks reflect trie level of market prices.
AI for Private Data Collection
Collecting data about private firms and private market transactions is a time consuming and resource intensive task. After ten years of developing teams, tools and methods to collect this data, SIPA uses a combination of AI-driven data extraction and human validation to create and maintain the largest database of private company investment data in the world. Our AI-driven tech includes in-house trained and natural language processing to read and understand accounting data from company and fund annual and quarterly reports. This in turns allows turning vast quantities of documents and unstructured data into machine-readable information that is systematically organised using the PECCS and TICCS frameworks. AI-driven classification of companies and fund data are also at the heart of the SIPA Tech stack.
The privateMetrics API
To go beyond clunky websites and difficult-to-use reports, the privateMetrics data, which includes indices, benchmarks and comparables for private equities, private infrastructure and infrastructure debt, has been designed to be accessible directly into our clients' systems whether they use MSExcel or programmatic interfaces programmed in Python or R. Thank to a rich library of functions and API endpoints, almost any benchmarking or asset pricing application can be built and integrated in investor's data systems. The privateMetrics API is built and maintained by our team of top developers , with experience in major financial institution financial data applications.