Investors get less while risking more, reports say

Institutional investors might be taking on more risk and earning less return from their venture capital and buyout investments than they think.

Two researchers, in separate reports, show that not only are investors probably not getting the returns they think they are, but also the investments are more correlated to stocks than previously believed. That, academics say, could play havoc with their asset allocation models.

One of the reasons for the disconnect is that smoothing — or averaging several quarters of returns into a single quarter — done by fund managers makes the funds appear to be doing better than they really are, and be a lot less risky, said Susan Woodward, founder of Sand Hill Econometrics Inc., Menlo Park, Calif. On average, venture capital fund managers go back six quarters for portfolio company valuations; buyout managers' portfolio company values might cover as many as five quarters, she said.

For example, venture capital investments earned 20% per year, gross, between 1995 and 2005, with a beta of 2.0, a standard deviation of about 60% per year and a correlation with the stock market of 0.7, according to Sand Hill, a provider of analytical tools for private equity.

Buyout exposure is like stock market exposure, with beta of 0.9 and 15% standard deviation. Buyout return correlation with the S&P 500 is 0.9 to 0.95. (By definition, the beta of the stock market is 1; the S&P 500 has a standard deviation of 17%.)

"What this means is that venture capital and buyouts are avenues of diversification, but they are not gold mines," Ms. Woodward said. "Venture capital returns have been about equal to the Wilshire 5000."

Ms. Woodward, a former chief economist of the U.S. Securities and Exchange Commission, undertook the review of some 6,500 "value events" in an effort to create a new venture capital benchmark. Value events are either financing rounds, sales of companies or bankruptcies.

Many consultants are aware of the problem.

"Many people talk about private equity as a diversifier, but most serious consultants understand that it is more accounting diversification than economic diversification," said Stephen L. Nesbitt, chief executive officer of Santa Monica, Calif.-based consulting firm Cliffwater LLC. "Private equity correlations and risk are understated tremendously."

Instead of using historical private equity data, many consultants use the public markets as a substitute in asset allocation modeling. "We realized the numbers were garbage," he said.

A separate study by John H. Cochrane, the Myron S. Scholes Professor of Finance at the University of Chicago, concurred with Ms. Woodward that venture capital returns are lower than what is reported.

"We were trying to get behind the smoothing," Mr. Cochrane said in an interview. "On average, venture capital acts very similar to small growth stocks."

His paper, published last January in the Journal of Financial Economics, shows that between January 1987 and December 2001, the average annualized return of the smallest Nasdaq stocks was 62%, about the same as the 59% mean return of venture capital funds during the same period.

"Overall, with all the venture capital projects, I could not find much of an illiquidity premium," he said. In other words, venture capital funds do not appear to provide much extra return over the stock market to compensate investors for locking away their money for 10 years.

"Venture capital is very risky," Ms. Woodward said. She noted half of all venture capital-backed companies fail and some 50% of the money invested in venture capital investments is lost. However, the failure rate is not reflected in existing venture capital benchmarks, she added. The buyout fund failure rate is less than 3%, she added.

Lack of accounting for survivor bias is a big problem with most venture capital benchmarks, Mr. Cochrane said.

"They collect the returns for everybody that is around," he said. "It is like collecting data from everyone still in the casino: They're not asking the people on the bus … who are on their way home."

But survivorship bias is not unique to private equity, Cliffwater's Mr. Nesbitt said. "There are some survivorship biases present in all indexes," he added.

Not only does the smoothing affect private equity returns but the data can alter institutional investors' overall asset allocation modeling, said Josh Lerner, Jacob H. Schiff Professor of Investment Banking at Harvard Business School, Boston. Using stale prices — even a one- or two-day delay — as the basis of valuing a fund's portfolio affects its correlation with the public markets. Correlations are much more important to investors than overall volatility, Mr. Lerner said. If private equity investments are more correlated to the stock market than institutional investors think, it makes their asset allocation models inaccurate.

"The risk numbers in the asset allocation models are potentially problematic," Mr. Lerner said. "Risk and correlation are very challenging things to handle in private equity."

Mr. Cochrane said the venture capital and buyout numbers would suggest a high level of correlation between those sectors and the stock market.

"They (institutional investors) are kidding themselves that buyout and venture capital are more of an alternative asset class than they really are," he said.

The problem is more severe with buyout funds than with venture capital, Mr. Lerner said. Private equity prices are "sticky," and general partners are holding their portfolio company investments at cost, which might not reflect their current value. Venture capital fund managers revalue their portfolio companies more frequently than buyout fund managers because there are more rounds of financing. Buyout funds may end up keeping their company investments at cost for five to 10 years, he said.

Ms. Woodward said buyout managers are starting to report values closer to current values or "fair market" as is recommended by Private Equity Industry Guidelines Group, an industry group formed to create reporting guidelines.

When looking at her data for the five years ended June 30, 60% of the portfolio companies were carried on the fund managers' books at current or "fair market" value, up from 50% for the 10 years ended June 30, Ms. Woodward said.