Research
Work in progress
Project description
Private-market funds are exposed to risk dynamics that are difficult to capture using standard linear-Gaussian models. Their valuations evolve over the fund lifecycle, exhibit time-varying uncertainty, and may display skewness and heavy tails. This paper develops a nonlinear, non-Gaussian state-space framework that accommodates these features. The model is evaluated through particle filtering, using both simulated data and a sample of private equity and venture capital funds. Relative to Kalman-filtering approaches, particle filtering provides a more flexible characterization of the latent risk dynamics and performs better with departures from Gaussianty. The model also provides a criterion for selecting, among multiple candidates, the public-market return that is closer to the fund’s observed valuation dynamics.
Project description
Government defaults typically have devastating consequences for a financial system: they can absorb much of the risk-bearing capacity of domestic financial intermediaries, thus constraining funding to firms. Local financial architecture, however, can act as a buffer by moderating how government defaults transmit to the real economy. While government default can plunge banks into crisis, a well-developed corporate bond market could act as a "spare tire" for financial intermediation.