foreign countries with limited resources, fellowships for graduate students and fellowships for M.D.s to get postdoctoral research training after they complete their M.D. degree. However, we are phasing out the latter program because the NIH has instituted awards that do the same thing. We are delighted to let the NIH take over this activity while we pursue other approaches to funding physician scientists. If the NIH imitates us, it is the sincerest form of flattery.
With regard to funding for medical research, 80 percent goes directly to HHMI Investigators. There are now 325 investigators at 72 institutions in the United States. The rest is controlled from HHMI headquarters in the form of space and overhead payments to the host institutions, and for major equipment items, which are given out separately to investigators.
As the endowment increased from about $5 billion back in 1986 to a little over $11 billion today, the number of investigators has increased from 96 to a planned plateau of 330, however, there have been as many as 348. Some investigators leave to assume administrative jobs, or jobs in biotech or academia, which requires them to give up their appointment. In addition, we review our own investigators every 5 years, and about 20 percent are not renewed. These two sources of attrition are about the same magnitude. We have periodic competitions where we appoint new investigators.
This was the situation two and a half years ago when Tom Cech took over as president of HHMI, and recruited David Clayton and me as vice presidents. We began to evaluate how the Institute spends its money to determine what changes would be appropriate for the future. When we started this examination the endowment had just gone through a period of rapid growth. We agreed that even if the endowment continued to grow, it would not make sense to increase the number of investigators to 400. There are two reasons for this conclusion.
The first was that we know our investigators as individuals because the bureaucracy is not large. Our organization had already been strained with 330 investigators. To increase to 400 would mean that we could not maintain our style of review, and personal interaction. Moreover, at this time, the NIH was providing generous funding for many scientists. We felt that if we increased the number of investigators to 400, the best we could hope for was a 15 percent increase in the output. That did not seem like the best way to spend money. The primary mission of HHMI is to fund and advance basic biomedical research in the United States. Our challenge was to determine the best way to use this additional money.
We anticipated additional funds of $50 million to $60 million. The cost of an investigator, if you include the money that is given to the host institution in payments for rent and utilities, is about $1.2 million per year. Not all the investigators are in this range, but that is the average.