Using the Distribution of Returns to Guide Portfolio Construction

Equity returns don’t follow a neat bell curve, they are right-skewed, meaning a small number of extreme winners pull the average higher. In a normal distribution, outcomes are symmetric and the mean equals the median. In equities, however, the upside istheoretically unlimited while the downside is capped at a total loss. That asymmetry produces a long right tail where a handful of outliers drive most of the gains.

This is why the mean return often exceeds the median return in equity indices. In public markets, most stocks underperform the average, while a few mega-winners such as NVDA or TSLA carry the index. To put a finer point on this, since the launch of ChatGPT in November 2022, NVDA has increased 9x, contributing a full 3.9% of the S&P500’s total earnings, and in 2024 alone, was responsible for 22% of the total return of the index (RBC, Jan 2025).

Owning the S&P 500 exposes you to those winners, but the “typical” stock in the index delivers far less. The gap between mean and median highlights how skew dominates equity outcomes. According to Dimensional Advisors, in an analysis of rolling 5-year returns, only one third of companies in the S&P 500 outperformed the index. If you extend the timeframe to 20-year rolling periods, only 20% outperformed the index.

This dynamic holds across public equity, private equity, and especially venture capital, where return distributions follow a power law—an extreme right-skew in the distribution. This explains why according to PitchBook, from 2005–2019 top-quartile venture fund-of-funds averaged ~0.27x higher TVPI than top-quartile primary VC funds. The outperformance not only persists but becomes increasingly more dramatic at lower return bands highlighting that indexation in extremely skewed returns tend to yield more consistent returns.

Fixed income (i.e. debt) investments work differently by design: the upside is generally a capped contractual coupon while the downside remains total loss. Both tails are bounded, but in the opposite direction of equities. The result is a negatively skewed return profile: steady coupons most of the time, punctuated by occasional severe losses. According to the Federal Reserve, defaults in sponsor-backed direct lending run about 5% with loss-given-default up to 67% (Cliffwater puts it closer to 2.5% defaults and 65% LGD). In other words, when defaults occur, you may recover only ~33 cents on the dollar.

The point isn’t to disparage bonds, but to clarify how investors should think about cross-asset portfolio construction. Equities, with their right-skewed distributions, are best captured through broad indexing, which ensures exposure to the few big winners. Credit, with its left-skew, may benefit from more concentrated selection, where underwriting skill and collateral discipline can mitigate asymmetric downside. Unfortunately, most advisors promote portfolios that are the opposite of this logic. Clients wind up with concentrated equity portfolios that underperform (especially in PE and VC), and index-esque credit portfolios that capture the market’s risk of loss given default (again, especially in private credit).

There’s a logical way to construct a durable portfolio that blends the arithmetic and the intuition required to create consistency in returns. It involves putting the allocation at the top of the pyramid (Brinson et al., 1986) and then building a portfolio of sensible concentrations and indexations where the data supports it.

‍