Investors face tremendous problems when attempting to value investments in unlisted private companies. Such entities lack reliable valuation histories, preventing the construction of meaningful, accurate, and representative indices in private markets. In addition, currently available benchmarks typically rely on appraised valuations, which have many drawbacks. They are subject to several well known biases, such as smoothing, staleness, and poor representativeness, and do not reflect all private companies or all the information available in private markets.

Due to this lack of robust benchmarks, many investors perform comparable analyses (or comps) using the available transaction data on private companies. However, such data is sparse, noisy, and biased because few private companies transact regularly and each valuation is subject to idiosyncratic (or individual) factors. Therefore, comps based on such data are not very useful to investors.

In this paper, we propose a solution that is neither based on appraised valuations nor subject to these biases in transaction data. The factor model based solution when calibrated with transaction data and novel risk factors proposed in the paper, can transform the sparse, noisy, and biased transaction data into meaningful information that aids asset allocation, benchmarking, and monitoring of private investments.