Deciding how to invest a billion dollars can be tough. For Harvard Management Company's (HMC) Scott Sperling and Mike Eisenson, it can be very rewarding as well.

Sperling and Eisenson are the youthful managing partners of the Aeneas Group Incorporated, the high-risk, private placement arm of HMC. As such, they control approximately 20 percent of Harvard's $5 billion plus endowment--the 20 percent that is invested in real estate commodities and venture capital.

And as such, they have earned steep salaries and lucrative bonuses-- bonuses that in 1989 placed their earnings at over $1 million dollars each, far higher than any other Harvard official at the time.

But, as Sperling and Eisenson know all too well, managing high-risk investment can be a risky business. During the last fiscal year, their portfolio was devalued by $200 million, nearly 20 percent of its total worth. Severalmulti-million dollar assets were marked down tozero.

HMC President Jack R. Meyer has told theCrimson that is "pleased" with Aeneas' performanceand with that of its managing partners. And whileestimated that Aeneas has averaged a rate ofreturn of approximately 12.5 percent on itsinvestment over the last decade as opposed toHMC's overall average of approximately 13.5percent, he attributed the lower profits to "toughtimes" for private placements. Sperling andEisenson estimated Aeneas' average returns asabout 14 to 15 percent.

But internal HMC documents obtained by TheCrimson suggest that during the five year periodbetween 1986 and 1991, Aeneas' compound annualreturn averaged only 5.4 percent. That would putthe high risk division's returns not only belowHMC's internal benchmark of 11.1 percent for thatperiod, but below the 6.9 percent for that period,but below the 6.9 percent a very safe and highlyconservative investment in 10-year U.S. TreasuryNotes would have brought.

And while a return rate of 5.4 percent wouldnot necessarily invalidate the estimates made byMeyer, Sperling and Eisenson, it would suggestthat Aeneas was most profitable before 1986, theyear Eisenson arrived and just two years afterSperling took over as a full-time managingpartner.

In fact, several Aeneas investments have comeunder recent scrutiny for their financial merit.

Take Avenue Entertainment, for example. Harvardinvested more than $5 million in the Los Angelesfilm production and distribution company in 1987,when it was just being started up Hollywoodproducer and former island Pictures President CaryBrokaw.

Sperling and Eisenson have said that, at thetime, they were interested in investing in therapidly expanding and potentially highlyprofitable movie business. the Avenue deal was oneof a pair of investments that also included amajor stake in Tristar Pictures.

Tristar was highly successful. When Aeneas soldits stake in the company to ColumbiaPictures--later acquired by the SonyCooperation--Several years later, it did so at asubstantial profit.

But the Avenue deal was not as fortunate. Overthe next three years, sources have told TheCrimson, Avenue lost significant amounts of moneyby acquiring the distribution rights to a seriesof low quality, ill-fated films, including one,Cold Feet that sources said grossed a totalof $250,000 nationwide less than the cost ofproducing the film's preview.

Despite the losses, however, the sources saidHarvard neglected to devalue the Avenue investmentsignificantly until fiscal 1991, when it waswritten down to zero. Holding it at an inflatedvalue, they said, could have contributed to largebonuses like those earned by Sperling and Eisensonin 1989.

"Clearly the company was failing when they tooktheir bonuses and was probably worth zero," onesource alleged. "it was held at artificialvaluation for some length of time and then all ofa sudden written down to zero a short time afterthey got their bonuses..... That seemsoutrageous."

Furthermore, the source charged, Aeneas'dealings with Avenue may be indicative of itsoverall management style.

"If they take one company, hold it at whatappears to be an artificial book value for asubstantial period of time, and then, shortlythereafter write it down to zero, what does thatsuggest about the integrity of the generalvaluation method," the source asked. "Certainlyyou have to wonder when you see something likethat."