In this article, the authors propose a risk estimation model that addresses both smoothness and idiosyncratic risk dynamics of narrow private equity portfolio returns. The authors subsequently apply the model to a broad set of private equity return streams with tightly controlled diversification properties. They show that increased model complexity is detrimental to the model’s forecasting power and that idiosyncratic risks increase as one goes between hypothetical investment in a broad (albeit noninvestable) private equity index and a real-life, narrowly diversified portfolio of private equity funds. The authors also find that private equity returns have demonstrated a level of exposure to public equity markets that may be considered surprisingly low and offer a possible qualitative justification for this phenomenon. Finally, the authors incorporate the newly obtained private equity risk model into a general portfolio construction framework and demonstrate how to study the inevitable trade-offs between the costs and benefits of increased portfolio diversification. <b>TOPICS:</b>Portfolio theory, portfolio construction, equity portfolio management <b>Key Findings</b> • By applying a somewhat novel econometric technique to a proprietary dataset, we offer a way to separately estimate systematic and idiosyncratic risks of private equity (PE) programs. Our methodology controls for data overfitting and program concentration. • Observed volatility and equity beta of PE are much lower than that of public equity. We believe this is not an aberration but instead a fundamental consequence of PE being less exposed to the excess volatility of public equity markets. • With the exception of the tumultuous period of the global financial crisis (GFC), risk properties of private equity have been surprisingly persistent; if anything, systematic risks of private equities decreased slightly after the GFC.