D. J. Bernstein
What it's like to work at UIC

MSCS grant mismanagement, part 1: hiding cost information (2005)

A sponsor---the National Science Foundation, for example, or the Sloan Foundation---makes a grant to UIC: it sends UIC some money to spend in support of my research. I can then make specific requests to UIC; I might request that UIC use some of the grant money to hire a graduate student to help with my research, for example, or that UIC spend $1000 of the grant money purchasing a computer for the research assistant to use.

Suppose the university imposes a $550 surcharge on the $1000 computer; as a result of this surcharge, the grant runs out of money the next month; the university slashes the research assistant's paycheck; it's too late for the research assistant to find another job. That's not an acceptable situation. To sensibly plan my requests, I have to know how much the requested items will cost.

Unfortunately, at UIC, cost information is unnecessarily difficult to find. This web page gives three examples. The overall effect of many examples is that I've had to waste quite a bit of time acquiring numbers that should be readily available on the university web pages. In many cases, I still don't have the numbers I need.

Equipment overhead charges

When an NSF grant pays (e.g.) $1438 salary for a student for a month, it also pays nearly $800 to the university for ``facilities and administration.'' Almost all ``direct'' charges are accompanied by these ``overhead'' charges. (The exact overhead rate is negotiated by NSF and UIC years in advance.)

However, some equipment payments have no overhead charges. Obviously this makes a big difference in the amount of equipment that a grant can purchase.

The lucky equipment, as far as I can tell, is equipment that lasts for at least a year and costs at least $500; or components for building such equipment; or accessories necessary to make such equipment usable. For example, buying ten components from four different companies and assembling them into a single computer, at a total cost of $855, is charged no overhead, even if each component is below $500. Similarly, if a computer costs more than $500, then there's no overhead for adding or replacing a keyboard, network cable, or any other necessary attachment.

But I'm not sure about this. So I've asked for confirmation, with detailed quotes from the policies that I believe are applicable. After a year, after repeated email requests to several levels of university administration, I still don't have an answer.

You might wonder why I'm not sure about the rules. One answer is that I've seen the university take overhead money from equipment that---according to my understanding of the rules---has no overhead charges. My guesses are that I have the rules right, and that the university has been stealing quite a lot of grant money in this way; but this theft is a separate issue, and is discussed on a separate web page. The point of this web page is that it's unnecessarily difficult to find out the costs.

Course buyouts

The Sloan Foundation named me a Sloan Fellow early in 2002, and sent UIC $40000 to use in support of my research. In March 2002, the UIC MSCS department head agreed to relieve me of three courses in exchange for $35000 ($11666.67/course). I started making plans on this basis.

The department head did not obtain the required approvals from the college before making this agreement. He also did not communicate the agreement to the UIC MSCS grant accountant.

Many months later, unaware of the department agreement, the grant accountant asked me to sign a document specifying a rate of $18004/course. On 10 December 2002, I carefully explained to the grant accountant that the department had agreed on a total of $35000 for three courses. She continued trying to convince me to accept a larger amount, first saying that the increase was ``allowable'' under ``Sloan guidelines,'' then saying ``don't argue with me about it,'' then saying that the department head's agreement needed college approval that had not yet been obtained.

The department head later obtained college approval for $11666.67/course. But this incident still illustrates the difficulty of finding cost information at UIC. Months after I thought I knew the costs, I was suddenly being told very different numbers, and had to waste time resolving the discrepancy.

Research assistants

There are several costs in paying a graduate research assistant:

In April 2004, I asked the department to confirm that it would cost $10791.15 for one of my grant accounts to hire two research assistants for the summer (the usual two months, half time), if both of them were registered for 6 hours of classes (3 being the cutoff, I believe).

The department replied that it would cost $11303.77. I understand a small part of the discrepancy: graduate students, like faculty, had recently received a 1.5% raise; I should have said $10952.78 instead of $10791.15. But the remaining gap is worrisome.

I explained in detail where my numbers came from and what I believed the department's mistake to be. After a year, after repeated email requests to several levels of university administration, I still don't have an explanation of the discrepancy.

Of course, this $351 discrepancy is a much smaller relative error than a $550 discrepancy in the cost of a $1000 computer. But I don't want to have to set aside hundreds of dollars for a ``UIC incompetence'' line in my budget. The same money could be spent providing books for my research assistants, for example, or sending someone to a nearby conference.