May 21: This post has been updated.

Before there was a much-maligned investment strategy for the agency charged with safeguarding the pensions of 44 million American workers, there was a consultant who devised that strategy. New information obtained by ProPublica reveals how the consultant, Rocaton Investment Advisors, reached conclusions that have now been criticized by two federal agencies and one inspector general.

Back in September 2007, the Pension Benefit Guaranty Corporation paid Connecticut-based Rocaton $395,000 to study the agency's finances and suggest how best to allocate its money, according to its contract (PDF). What the PBGC ultimately approved -- reducing its investment in bonds to increase the money it holds in stocks -- differed little from the recommendation made by Rocaton.

Fortunately, the PBGC never took its planned plunge into the stock market, avoiding catastrophic losses from buying stocks when they were near an all-time high. The agency announced last Thursday that while it had engaged Wall Street firms in November 2008, it had not made any investments based on the new strategy. Given the congressional investigation into former PBGC Director Charles Millard's handling of the investment contracting, it appears likely that the new policy will be severely modified, if not scrapped.

Nonetheless, Rocaton's work for the PBGC holds interest. Just as Millard shaped who would receive the contracts to implement the new strategy, he also served as the senior PBGC official on a three-member evaluation panel that selected Rocaton, according to a report by the PBGC inspector general (PDF). The speed and zeal with which Millard, a former Lehman Brothers executive, proceeded from hiring a consultant to signing contracts with investment banks created the appearance that he was not surprised by the findings of his handpicked consultant. One question that senators on the U.S. Senate Special Committee on Aging might want to ask when Millard testifies under subpoena this Wednesday: Did he give any instructions to Rocaton beyond the bland "statement of work" in the contract?

Millard and Rocaton did not return calls seeking comment.

Prior to February 2008, the PBGC had about 75 percent of its investment portfolio in bonds, with the rest in equities, according to the Congressional Budget Office. The new strategy would have reduced the allocation in fixed-income assets to 45 percent, raised the amount dedicated to equities to 45 percent, and devoted 10 percent to alternative asset classes like real estate.

Both a CBO report in April 2008 and a GAO report in July 2008 criticized the consultant's analysis and conclusions. In particular, the agencies faulted the new strategy for "reducing the timing match" between the PBGC's portfolio and its need for cash in the short-term to pay benefits. Yet neither agency provided more than cursory descriptions of Rocaton's work. Through a Freedom of Information Act request, ProPublica has obtained the 112-page report that Rocaton presented to the PBGC. It lays out the guts of the investment strategy Millard adopted.

The rationale can be boiled down to three basic tenets: Stocks perform well over time. The PBGC, which pays out pensions over a lifetime, has the luxury of investing for the long term. And diversification of assets, e.g. equities and real estate, spreads risk and is thus safer. Millard was adamant that the only chance the PBGC had to earn its way out of a deficit -- now well over $11 billion -- was through betting on stocks and other risky assets. As recently as last month, he made the case in an opinion piece in the Austin American-Statesman, when he said, "a low-risk strategy would simply lock in the existing deficit." (See update.)

As support for its conclusions, Rocaton looked at the U.S. stock market and the U.S. bond market from January 1927 to December 2006. "For investors who were able to wait 20 years, there has never been a period when stocks have underperformed cash," Rocaton concluded. The consultant even took the agency to task for so-called "opportunity costs" for not investing in bull markets.

If Rocaton had been able to include the current financial crisis in its analysis, it might have reached different conclusions. The bookending of two major financial meltdowns in 1929 and 2008 would also have underscored that stocks can be too volatile an asset class to be able to accurately predict their behavior, according to Jason Hsu, chief investment officer for California investment group Research Affiliates.

Hsu's colleague at Research Affiliates, Rob Arnott, released a paper last month in the Journal of Indexes showing that from 1979 through 2008, an investor in 20-year Treasury bills would beat the S&P 500.

"Depending on how much data you are using, depending on the window you are looking at, you are going to get very different conclusions," said Hsu, who also teaches at UCLA's business school. "When you are investing in a risky asset class, even if you expect to do better, there is always a significant probability that the actual result will be much worse."

Hsu sees danger in the diversification mantra, where again, history might not be the best guide. The stated reason for diversification is that different asset classes rarely move in concert. So, for example, when stocks are up, bonds are often down. Yet every once and a while, an outlier event, like a global financial meltdown, drags down all assets together.

Rocaton seems to suggest that even such volatility can be surmounted. "Shorting the stocks of companies that present PBGC with the greatest risk may seem impolitic, but nevertheless is one of many similar derivatives strategies which should be considered," Rocaton suggested.

The consultant said the time horizon is on the agency's side. "In addition, the modeling illustrated that even the most drastic short-term declines in PBGC's funded status were overcome in the long-term by the additional return gained from equity investments."

But did Rocaton's models anticipate the perfect storm currently buffeting the PBGC? The stock market has plunged, as have most other asset classes in the agency's portfolio. Interest rates are low, which increases the agency's liability. Specific industries that carry the defined benefit pension plans backed by the agency -- carmakers, media and finance, to name a few -- are reeling. If they all went bankrupt at once, the PBGC could be forced to take on their underfunded plans, ballooning the agency's deficit.

The financial crisis has stripped bare historical models like Rocaton's. "When you structure a portfolio just blindly taking historical averages and assume the future will look exactly like before, you are not going to get good results that way," Hsu said.

Ironically, Hsu thinks PBGC's new investment policy might make sense at the moment since the price of most stocks has dropped. "If you had asked me this question 18 months ago, I would decidedly tell you, no, being heavily invested in equities is not a good idea given how fragile the market," he said. "Now ... risky assets are beginning to look attractive."

Update: After ProPublica filed this report, the PBGC announced that its deficit had grown to $33.5 billion.