SACRAMENTO, Calif. — In the late 1990s, equities delivered double-digit returns while stock market volatility declined. Fed this data, the typical optimizer told pension executives to keep loading up on stocks.
That answer quickly proved to be wrong with the subsequent collapse of the tech-stock bubble
Officials at the $187 billion California Public Employees' Retirement System in Sacramento want to avoid any future irrational exuberance. In an unusual move for a pension fund, CalPERS' officials have devised a way to impose a "reality check" on the total portfolio by running a "reverse optimization," said Mark Anson, chief investment officer.
While investors pump historical data, interest-rate and return assumptions into a traditional mean-variance optimizer to create a choice of optimal asset mixes, reverse optimization starts with the asset weightings and works backwards.
A reverse optimization generates the implied excess rates of return within the total portfolio, based on the risk associated with each asset class, the covariance matrix and the allocations to each asset class. If the results are out of whack with a pension executive's beliefs about risk and return, the reverse optimization would signal that the portfolio should be rebalanced, Mr. Anson explained.