A Critical Look at the University of Michigan's
Role in the 1987 Merit Agreement

APPORTIONING THE BLAME IN THE ROBERT MOORE OVERCHARGES INCIDENT This report is a greatly condensed version of an unpublished manuscript detailing the machinations of the University of Michigan's complex computer research charging scheme. In January 1993, an investigation by the Department of Justice culminated with an out-of-court settlement between the University of Michigan and the government. U-M agreed to repay nearly $3.1 million in computer processing overcharges that involved federally funded research. In exchange, the government accepted U-M's explanation that the charges were "inadvertent," and therefore dropped the case. The case was an offshoot of a False Claims Act qui tam action that Robert Moore, an accountant for the U-M Information Technology Division (ITD), filed in early 1992. Moore's claim was that the University violated OMB Circular A-21's provisions requiring that rates charged to the government be at least as low as the lowest market rate or no more than the actual cost of the service. The False Claims Act allows citizens to file suit on behalf of the government if they have evidence that a false or fraudulent claim against the government has occurred. If the suit is successful, the plaintiff receives a portion of any money returned to the government. Moore therefore received $570,000. Following the January 1993 settlement, Doug Van Houweling, speaking as the head of the Information Technology Division (ITD), issued a press release in which he claimed both that the overcharges were "inadvertent" and that Robert Moore himself had created the system and problems which resulted in the overcharges. The 1993 press release resulted in Moore's filing of a libel suit against the University. As a result of the libel suit, both Moore and the University have been recitent to speak directly to the issues of the matter. Since obtaining comment on the case was close to impossible, several Freedom of Information Act requests filed in early 1993 have revealed what I believe to be the disingenuity of Van Houweling's claim that the charges were "inadvertent." One memo, from Moore to Sam Plice and Greg Marks, who were immediately below Van Houweling as second and third in the chain of command, openly and frankly discussed the overcharge situation at least two years before Moore reported it to the government. In that November 1989 memo, Moore makes recommendations to his superiors based on which policy would be least likely to reveal problems to the auditors. He also makes note of the nearly $6 million in charges for which he believed the government received no benefit. He wrote, "We could be asked to demonstrate that the federal user is actually benefiting and using these services in proportion to the rest of the University's client base, and I seriously doubt whether that is the case." [ed. emphasis ours] The level of detail with which Moore describes how the University's accounting system was made unnecessarily complicated to confuse outsiders can only be understood by reading the memo. It was in passing in that memo that Moore referred to the "multiple hats" that Eric Aupperle and his Merit staff were wearing and the accounting overlap between U-M and Merit that "concerned" him and another accountant. The very existence of the memo proves that senior ITD officials had foreknowledge of the overcharging. It would take tremendous imagination to believe that Van Houweling wasn't informed of the overcharges by his chief financial officers. It becomes indisputable that he knew of the situation when documents from the Bentley Historical Library were reviewed to provide secondary verification of the sequence of events. Other published reports (Ann Arbor Observer) also directly place possession of the key memo in Van Houweling's hands. Furthermore in the minutes of a March 1985 Computing Center budget briefing, all of the key issues involved in the government's claim against U-M are on the agenda: "Circular A-21 and Cost Reimbursement" University Indirect Cost Surcharge "Our Profit" General Fund support vs. MTS [Michigan Terminal System] GFCA [the cost accounting scheme questioned by the gov't] The GFCA and our "profit" Cost recovery vs. subsidized or "free" services Thought was clearly given to the issues, and in this case even before Moore was hired in 1986. Therefore, could Moore then really have been "responsible" as Van Houweling suggested? In 1986, Van Houweling himself wrote to a staff member that "... we are very concerned about the discrepancy between the need for computing services, as expressed to us by faculty and staff, and this pattern of under utilization by most units." Van Houweling was concerned about a failure of departments to use up their "free" computing allocations because failure to do so would presumably affect the end-of-year overhead balances. Van Houweling solves this problem by creating a "special one time allocation" to researchers, who have a higher demand for MTS time due to their statistical and research needs. He also sent out a letter to every department chair encouraging them to use up their entire original allocation. Van Houweling was effectively encouraging departments to use MTS above and beyond what they had felt necessary previously, because in the long run it "would be to everyone's advantage." This amounted to nothing less than encouraging waste so that long term federal funding could be increased. Moore recognized the same problem when he described how ITD had created two operating systems, "ub" and "um," when they essentially performed the same technical functions. The "um" system was restricted to paying commercial or federal users and UM employees, whereas the "ub" system was used to give students, staff, employees, and outsiders "free" computing accounts. Indeed, it was and continues to be a running joke among ITD employees that the value of an "MTS dollar" was equivalent to a "ruble." The more rubles that Van Houweling could electronically "print," the better the situation for ITD. At one point, an MTS dollar was valued at approximately 30 real cents. Although the above statements require a contextual understanding of the sequence of events, which the Moore memos provide, and aren't necessarily 'obvious' ethical problems, two different documents obtained from Bentley are. 'Redirecting' Capital Funds  On 18 March 1985 Sam Plice writes to Van Houweling: "One strategy for use would be not to use operating funds for capital expenditures and to depend on an indirect cost revenue allocation that. This would free up operating funds for additional salary expenditures. In the long run we need to budgeted directly on general funds, but this strategy might be a bridge until that is possible."  On 13 December 1985, Greg Marks writes Van Houweling:  "Money that we borrow must be for capital investments. Further, it is evidently essential that we describe in significant detail [emphasis in original] the specific capital goods to be acquired, the overall functional purpose, and show how income will be generated to repay the debt. I expected we would have to explain such matters, but it appears that the level of specificity needs to be much greater than I had thought. On the face of it, all that means is some added work on our part, but the real problem is that too large a portion of our new income is restricted to capital goods."     "The total budget for the new initiatives is just under $6.4 million. Income sources that can cover operating expenses total only about $1.7 million (student fee income plus internal reallocation.) That leaves about $4.7 million that may be used only for capital investments. My best estimate at present is we need only about $2.9 million capital funding. Thus within the University we evidently will need to find some way to borrow money for other things, freeing some operating dollars for our new initiatives. I do not know of enough capital activities in the base budgets of either the Computing Center or the Office of Administrative Systems, so I believe will [sic] must find capital activities outside our Division."     "But to do that means we will have to describe something quite different to the lenders, [emphasis ours] so at least some of the detail about our capital needs for the initiatives is irrelevant. Furthermore, I suspect that this might involve enough other people and budgets outside Information Technology (and possibly at a wrong time point for finding other capital projects) such that we will be further delayed in getting the income. I hope that I am seeing more problems than are real in all this. But just in case...". Describing "something different [than the truth] to the lenders" is arguably fraud. If you add up the numbers that Marks gives, his conclusion that non-capital activities could not be legitimately funded seems correct. Was the solution to create new numbers? Conclusion Dr. Van Houweling has been a pivotal player in the operation of the NSFnet. A key part of understanding anyone's role in any management activity is to examine their personality. The events relating to the Moore whistle-blowing case and statements that Van Houweling made following its public disclosure are indicative of his personality. Van Houweling was certainly disingenuous when he claimed that the overcharges were "inadvertent." Such a claim implies that either no one knew about it, or if they did learn of it they would have done something to prevent or report it. It is unbelievable, in light of the documentary evidence, that Van Houweling had no awareness of the accounting system he oversaw. For him to then lay the responsibility on a mid-level accountant and to claim he didn't know what was going on is irresponsible in and of itself. Gordon Cook, President  COOK Network Consultants 431Greenway Ave Ewing, NJ 08618, USA Telephone & fax (609) 882-2572  www.cookreport.com

Text highlighted by burgundy are actual documented quotations.  Original publication may vary from this reprint in subtle ways.

Reprinted with permission of the
The COOK Report on Internet and NREN
January 1995 (Vol. 3, No. 10)

(editor's introduction)

    Recent University of Michigan graduate Chetly Zarko has spent much of the last two years digging in the University of Michigan archives and serving the University authorities with FOIA requests for the contents of computer conferences and other University documents.

    He sheds much light on the way in which Dr. Doug van Houweling was brought in to shepherd a major technology push at the University one that including ousting Amdahl as the University mainframe and bringing in IBM who quickly became a joint study partner in the 1987 Merit Cooperative Agreement for the NSF backbone.  Zarko describes the complex series of accounting arrangements between Merit and the University of Michigan as tempting University administrators to identify all possible loop-holes in the cooperative agreement and manipulate them to the University's advantage.

    In early 1992 Robert Moore an accountant in the UM Information Technology Division filed a False Claims Act against the University that resulted in  an out-of-court settlement between the University of Michigan and the government. U-M agreed to repay nearly $3.1 million in computer processing overcharges that involved federally funded research. In exchange, the government accepted U-M's explanation that the charges were "inadvertent," and therefore dropped the case.  While this affair was not apart of the Merit NSF cooperate agreement, Zarko shows it as part of the same institutional system that appears not to have enough checks and balances in it on behalf of the public interest.

    It is in this respect that this primarily historical research we believe has meaning for the present and future.  Anything that the NSF in the future can do exert more effective oversight with regard to the way its money is spent is desirable.  Zarko's article helps to underline the importance of DNCRI's movement toward a fixed unit price arrangement and away for payment for costs incurred with the vBNS Cooperative Agreement.  We publish the University's rebuttal.

A CRITICAL LOOK AT THE UNIVERSITY OF MICHIGAN'S ROLE IN THE NSFNET 1987 COOPERATIVE AGREEMENT

by Chetly Zarko, 
Original Copyright 1994-1995

    Editors Introduction: We present here an article that is at this point of primarily historical significance. Nevertheless we think it is worth presenting for the general insights it shows into the ways in which the interests of large corporations, and major academic institutions have inter-acted when federal money and national prestige are at stake. It also gives a useful context for judging the
complex foundation on which the seven and one half year Merit cooperative agreement was built. For after all, this became the single most important effort underlying the current takeoff of what has become the commercial Internet. Some would call it selfless service for the national good. Others would call it empire building. A close to final draft of this article was sent to Doug Van Houweling and Walter Harrison at the University of Michigan at the beginning of October. We offered them an opportunity to rebut Chetly Zarko's research. We received a reply saying only that they disagreed with the article and pointing out that he had lawsuits in effect against the university. We sent the final draft to them on December 3rd.

    Dr. Douglas Van Houweling, Vice Provost and Director of the University of Michigan's (U-M) Information Technology Division (ITD), has certainly played a pivotal role in the expansion of the NSFNET over the last decade. In December 1984, Van Houweling was recruited from Carnegie Mellon University (CMU) to fill a newly created position. The purpose of this position was to restore U-M's national leadership role in academic information technology applications. In the decade since, he has demonstrated an impressive ability to expand the size of the ITD administrative bureaucracy. In 1980, the Computing Center, ITD's precursor, had approximately 200 employees. By 1993, that number had increased to over 600. The largest increases came during Van Houweling's first six years at U-M, from 1984 to 1990. To accomplish that, Van Houweling followed a specific set of plans to attain goals that have, on the whole, been largely realized. Although he met with initial resistance and often vehement disagreement from career Computing Center employees in his first years, many of those same managers now consider him to be a "visionary" despite their early differences.
    Perhaps his only major setback was a negotiated settlement with the Department of Justice in which ITD agreed to repay nearly $3.1 million in general computer-related overhead overcharges involving a whole class of federally-funded research projects. He partially deflected this potential problem by attempting to blame the problem on lower management (the charges were proclaimed to be "inadvertent") and the whistle blower himself, and the case never received much publicity after an initial short burst of interest.
    The important question of this analysis is not whether Van Houweling has lead his organization to new heights, but whether his very success in doing so was appropriate and responsible, both from the perspective of the U-M and the public at large. The central conclusion of this report is that Van Houweling's leadership was motivated by pork barrel institutional aspirations more so than it was motivated by a belief that he was promoting the public good. This is not my vision of responsible leadership.

    Perhaps without much thought, the University of Michigan's role in the controversy over the management of the NSFNET has often been overlooked, both in Ann Arbor and nationally. Although Merit Network, Inc.'s (Merit) role in the emerging controversy is well known through its cooperative agreements with the NSF, very little is heard about the underpinnings of that role. Despite the fact that Merit is officially a consortium of nine public universities in the state of Michigan, Merit's ties to U-M are greater than those to any of the other participating schools.
    Every Merit employee is legally first and foremost an employee of U-M. Physically, Merit had for many years relied entirely on U-M office space and staff. Ethereally, Merit receives considerable free access to U-M computer time and facilities. The organization is further funded and assisted, in a variety of ways, by U-M. In fact this is a point of contention among some member schools, and as we shall see later, some accountants. Eric Aupperle, President of Merit and senior manager in ITD at U-M, is responsible to both Doug Van Houweling, member of the Merit Board of Directors; and Doug Van Houweling, Vice Provost for Information Technology at the University of Michigan. Aupperle also deals as a contractor with Doug Van Houweling, member of the Board of Directors for Advanced Network Services (ANS). As is evidenced by the U-M Freedom of Information officer's characterization that Van Houweling's ANS records are his "private and personal" affairs, it is difficult to know in what capacity Aupperle and Van Houweling communicate at any given instant.
    In 1985 NSF began a process of solicitations for building what would become known as NSFNET. Van Houweling, with his recent arrival at U-M, was interested in pursuing several of these NSF grants in order for Merit to increase U-M's prestige and leverage in the national policy making process. It is likely though that his key motive was the tremendous funding it would bring his way, helping to expand the scope and power of the ITD bureaucracy (see Bentley Historical Library archive documents mentioned below). With the technical requirements and range of competition for NSF's solicitation being significant, Van Houweling and others also realized the need to secure support from corporate partners. He actively sought out that support and cooperation using his numerous contacts obtained prior to his arrival at U-M.
    By early 1987 Van Houweling had in place a team which consisted of IBM, MCI, and Merit. The team's proposal was accepted by NSF in November 1987, with the resulting grant award. As a result of the Michigan connection in the grant, the State of Michigan Strategic Fund also partially matched award funds. After the initial cooperative agreements expired in 1992, they were temporarily extended and follow-up solicitations were later won by members of the same team.
    After the surprise creation of ANS in 1990, speculation abounded that Van Houweling and others had worked out a deal with the director of NSF, Stephen Wolff, behind closed doors. Those dealings, most of which are still not public, that resulted in this sequence of award grants that for a long time were questioned by competitors and critics. In 1991, following ANS's 1990 launch, ANS CO&RE was created as a for-profit subsidiary. ANS CO&RE collects the profit obtained from the sale to private entities of unused excess data transfer capacity.
    A for-profit organization's restrictions on how it could assist other organizations, like IBM and MCI, are much less than those of a non-profit. Theoretically then, ANS should have maintained a strict separation from ANS CO&RE, just as Merit needed to separate itself adequately from purely private business dealings. The NSF was also criticized for a "cooperative-agreement" format rather than a traditional bidding "contract." The legal difference is significant since the latter obligates, with sanction, the contractor to perform certain spelled out duties, whereas the former allows failure and lack of performance, without sanction. Audit determinations are also held to a lower standard.
    To further their defense, Van Houweling and Aupperle both basically recited the IBM and MCI claim that IBM and MCI contributed money to the NSFNET for purely "philanthropic" reasons. Their contributions would benefit the academic community, but legally they could see no benefit at all to themselves. In addition, Van Houweling has loftily testified before Congress suggesting that NSF's small contribution to the infrastructure had almost single-handedly "energized a nation." The believability of these arguments will be left to the reader's discretion. A more useful and relevant answer though, would come from a different question. Was this the best deal the taxpayer could get? This report attempts to shed a little bit of light on the historical origins of that debate.

    This author's investigation into these activities began with a marginally related incident. In January 1993, Robert Moore, an accountant for U-M's Information Technology Division, in conjunction with the Departments of Health and Human Services (HHS) and Justice (DOJ), reached a $3.1 million out-of-court settlement in a suit filed regarding alleged overcharges by U-M for research using University computers (see sidebar). That settlement, pursuant to the False Claims Act, resulted from accusations by Moore, supported by documentation he had retained, that the University had improperly charged the Federal Government for computer related research processing costs which weren't actually related to the research in question. Almost all of these overcharges involved research requiring use of the University's academic mainframe, known as the Michigan Terminal System (MTS).
    Information regarding those events is and was not readily available due to University silence resulting from a related libel suit Moore has pending against the University. As a result, the Moore memos were released to this author following state Freedom of Information Act requests filed with the University in early 1993. The memos, after the government learned that a high- level U-M official ordered them destroyed, became a central part of the government's case against U-M. The one key memo is mostly a lengthy expose of the issues leading to the research overcharges. Buried on page 14 almost in the style of a footnote is a description of unusual accounting practices between Merit, U-M, and the NSFNET. On 7 November 1989, Moore confides: 

    An interesting finding is that there is some tension and dispute among Merit member schools. One memo relayed Michigan State University's (MSU) concerns about the "dubious" nature of some of Merit's claims with reference to supercomputer policy. There is also other evidence of friction and competition between members. Therefore, when referring to Merit, one must be careful to realize that their activities are not necessarily representative or reflective of all of the involved Michigan schools.
    An important tertiary confirmation that there is something to Robert Moore's 1989 concerns of accounting overlap between U- M and Merit is Linda Charlesworth's 1985 statement to Van Houweling. Charlesworth is Van Houweling's executive secretary.This was at least one year before Moore was hired at U-M, and was during a time in which U-M accounting structures were undergoing major changes, which the government later determined were illegal. She asks, "Will the Merit accounts remain separate? (I would expect that to be necessary)".
    Necessary, but not necessarily so? According to state and Federal tax laws, a non-profit organization such as Merit, must be operated exclusively for the tax-exempt purposes for which it was created and no part of earnings of the organization may "inure to the benefit of any private shareholder or individual." In memos from Virginia Rezmierski, one of Van Houweling's key "ethics policy" assistants, to Merit employees these restrictions are discussed explicitly. According to the memos, "Merit must demonstrate that remaining services [after Merit decided to sell services to an unrelated for-profit corporation in the late eighties] are its primary activity," and must "avoid adding outside sales that will cumulatively exceed the insubstantial limitation."
    Although these memos are discussing the sale of Merit services that don't pertain to the ANS controversy, they demonstrate simply that high level University and Merit officials are aware of their restrictions. The memos also help confirm a probable explanation for the creation of ANS CO&RE. Both ANS and Merit were constrained by the "insubstantial limitation" of federal non-profit tax laws and therefore neither could safely sell surplus network capacity to the business community. Creating a for-profit subsidiary, a common tactic among non-profit agencies, was the logical choice. It would allow the creators (IBM, MCI, Van Houweling, and ANS) to create the appearance of a certain amount of distance and yet maintain control over the actions of the new organization. But, according to principles of both conflict of interest and tax laws, there must be more than an "appearance" of separation. Unfortunately, there has been no legal precedent set in this area of non-profit law.
    It was still not clear whether U-M and Merit funds have benefited private organizations like IBM and MCI through ANS. Do activities like Merit's subcontracts to ANS and ANS's subcontracts for the same service back to Merit, along with activities that might benefit ANS CO&RE, violate these rules? In the Bentley documents, I found an interesting e-mail printout addressed from oberst@educom.mailnet to Van Houweling on 19 July 1985 which discusses a meeting with AT&T Bell Labs in New Jersey.
   Presumably, this message was from Dan Oberst of Princeton, who has had long-standing contacts with Bitnet and Educom. The message, which was preserved by Bentley without much context, gives a summary of the NSFNET solicitation and "Doug's [Van Houweling] candid behind the proposal" look at that proposal.
    Oberst describes that look as detailing a "window of opportunity" requiring the "leveraging" of NSF's money. After the NSF grants in 1987, it was clear that what NSF wanted was to get a larger investment in the network from private organizations that would partially match its own commitment. While this is a laudable strategy, known as leveraging, it can work in reverse as well. Competitors and critics have alleged that the cost-sharing arrangement among the Merit team's partners has been skewed in IBM and MCI's favor. Arguably, IBM and MCI "leveraged" NSF money by making relatively smaller commitments which essentially bought the rights to technology developed mostly at the expense of the taxpayer.
    A discussion draft by John McCredie of Digital Equipment Corporation was retained in Bentley by a computing center employee. That paper answers the question of why large computer firms have given "millions (in free computers and technical equipment) to a few campuses." The answer is that they are "huge computer experiments," presumably designed to serve as a tax-deductible test-bed for new systems. McCredie's discussion draft goes on to quote Derek Bok, former President of Harvard University, when he questioned the appropriateness of vendor - University partnerships in general. McCredie cites Bok as basically stating that one does not get "something for nothing." In the context of the University of Michigan's involvement in any particular partnership, the possession of this discussion draft simply shows that U-M employees were aware of what they were getting into and were aware of some of the ethical and policy questions that might arise.
    In another record retained by CC employees, commentary by the employees focused around the validity of this comment which they cited from an unlabeled House hearing testimony during the mid- eighties: "Already rich universities have developed superior sensors for pinpointing new bundles of money in Washington." Keeping that seemingly true observation in mind, let's analyze the following two statements from the Bentley records.
    The first is: "I do expect NSF to provide some funding for us to develop TCP/IP support within Merit. One way to use a portion of these funds is to reduce this year's member fees (each Merit university pays a membership fee to maintain Merit). Another is to hire additional staff." [1986] Perhaps someone overlooked the other possible use of that kind of grant. Why not develop TCP/IP; rather than kick it back to the schools or to expand the bureaucracy? Obviously the purpose of getting these grants is not
to actually accomplish the goal of the grant, but rather it is to "pinpoint" those "bundles of money in Washington." In the second, created at about the same time, Van Houweling is apparently quoted in an unknown CC employee's handwritten notes of a high level meeting: "Doug [presumably Van Houweling] -- we have to do that to seem as national leaders & to get $, contact C leaders"
    Visionary?
   The meaning of this passage seems pretty clear with respect to Doug's intent to get dollars, but it is not fully clear what it means by contacting "C leaders." One might suppose that if the university seemed to be a national leader that it would get more support from other national "[C]omputing" leaders in industry. The Vice Provost for Information Technology at U-M seems to have played his role well in the acquisition game. His intent was to locate money sources. In this game, as long as you seem to be the leader, then you get the money. Appearances, not reality, are the most important part of the grant process. One who did not wish to be 'charitable' might say that the U-M efforts were motivated more by the desire to get grant money for the sake of getting a piece of the pie, for expanding the bureaucracy and power of U-M, than by the laudable goals that Van Houweling proclaimed in an interview with the author in June 1993.

    The quid pro quo nature of University-vendor relations that McCredie and Bok cite is demonstrated by a specific sequence of events and some well placed speculation by mid-level computing employees at U-M. In 1986, U-M bid out a contract to provide a new mainframe computer to house its Michigan Terminal System and other services. The purpose of this new computer was to replace two Amdahl machines which, according to Van Houweling's press statements at the time, were outdated and overburdened.
    The competition for this replacement boiled down to two bids. One from Amdahl and another from IBM. The decision process by which U-M selected the winning bid is illustrative of the political processes at work. The Computing Center (CC), which was the old organization that Van Houweling had inherited upon his arrival, had recommended that the Amdahl be chosen because it was technically superior to IBM according to most of its benchmark speed and capacity tests and the cost of the two machines was similar. Many CC employees also believed from their past experiences with both companies that Amdahl would give much better technical service if their machine had any problems. Van Houweling and the executive officers however presented to the Board of Regents a recommendation that the IBM machine be purchased.
    As a result of seeing material released pursuant to a FOIA lawsuit seeking a restricted access "computer conference" the U- M Board of Regents was using in 1986-87, this author was pointed in a surprising direction. One Regent discussed privately in that conference how he was influenced in his decision process by the public debate in a separate publicly accessible computer conference discussion used by U-M computing center employees, staff, and students interested in computing at the U-M. According to the September 1986 evidence from that public computer conference, that decision was made in part (this was the third of three possible political reasons cited) because "the University hopes to patch up its relationship with IBM by buying this machine and once this is done it hopes that millions of dollars of IBM money will flow to the University in the form of gifts and grants over several years."
    The source of that quotation, Jeff Ogden, a CC employee at the time, is now a network administrator with Merit. Later in that public computer conference, Maya Bernstein, who at one time worked as a CC employee and is now an information policy analyst for the OMB in Washington D.C., corroborated Ogden. She emphasized that Ogden's third point was the key. She stated, "It sounds to me like reason #3 of Jeff's above is the main issue. After proposing a Japanese make in U-M's bid for a supercomputer last year, what's the big deal about a Japanese computer now?? It's hard to believe UM would go with IBM without knowing for sure that gifts and grants could be expected, but then if they did know, it might be considered bribery. Politics indeed."
    Others on the conference were asking themselves the same type of question. A research associate asked, "Can the University really treat vendors like that and get away with it? I would (naively) think that Amdahl would be in a position to sue. (They apparently offered a superior product for the same cost and lost the bid for political reasons." The response from an attorney, "Yeah, I don't think there's much there. Now, if Amdahl could show that IBM had bribed Regents there might be a "tortious interference with business relations" claim, but I don't know much about that."
    Instead of continuing to discuss it, an ITD research systems employee suggested, "We're in bed with IBM for the next several years so we may as well stop asking why and complaining."
    Bernstein and Ogden later go on to address the question of what impact moving from two Amdahl's to one IBM would have on the accounting and rate structure and how that move could be justified in terms of the rates U-M charged the Federal government (see sidebar). Previously, the University had used the existence of two separate Amdahl machines to justify charging lower rates to staff than to Federally funded researchers. With only one IBM machine, UM had set itself up for the eventual problems it would face with the Robert Moore/Department of Justice overcharges case.
    Why did Van Houweling or his officers feel a need to "patch" up relations with IBM? Ogden provides a partial answer when he stated, "Over the year's U-M chose to buy several machines from Amdahl rather than IBM (470V/6, V/7, V/8 and a 5860). "Amdahl's existence has changed the large computing environment in which IBM lives. IBM doesn't like Amdahl and as a result IBM doesn't like the U-M for helping Amdahl in its early years when there were lots of questions about Amdahl's continued existence. I think this is the main thing to be patched up from IBM's point of view." But why would Van Houweling or U-M care about that relationship so much that he would take the management risks of overriding the recommendations of his technical people? The answer is left to speculation, but the U-M and Merit only one year later entered into the 1987 NSFNET cooperative agreement with IBM and MCI, U-M later received a gift of a free second supercomputer from IBM, and became one of the pilot test programs for the Institutional File System (IFS) that IBM offered to U-M.
    When I queried Ogden about his 1986 statements, Ogden delicately attempted to maneuver away from them, and claimed that his initial thoughts on the problem were somewhat off-base. When asked why he almost quit his job (at least two other employees did resign) at U-M in December 1986, and why he had a high-level discussion with then President Harold Shapiro, his response was that he had confronted Shapiro with his concerns and that Shapiro assured him that there was no quid pro quo between IBM and U-M. Although Ogden still believes that UM officials may have "hoped" for reciprocation, he claimed to have no knowledge that IBM officially agreed to such a deal and that there was no explicit quid pro quo.
    Beyond U-M's "official" relationship with IBM, Van Houweling himself had done outside consulting for the firm, according to his own recent (1992-93) curriculum vitae. When directly queried earlier in May 1994 about the time frame of that consulting experience, Van Houweling did not respond at all. After more than two weeks of waiting for an answer, FOIA requests were filed for any documents Van Houweling may have filed with U- M or the State of Michigan to satisfy state conflict of interest reporting laws with respect to IBM. The response to that request was that although their were no responsive documents for IBM consulting related conflicts, the Regents had made the appropriate public notification (under state law) with respect the potential conflict between Van Houweling's Merit Board of Directors membership and his position at U-M.
    Since this release made no reference to his IBM ties, which were clear to have existed at some time, I made a formal in-house request with the Regents to investigate a possible failure to report an outside consulting engagement. Richard Kennedy, the Secretary of the University, speaking on behalf of the Regents, formally replied that he had been "assured" that Van Houweling's IBM consulting was prior to his employment with the University in 1984. A follow-up FOIA revealed that Kennedy didn't collect any documentation that would demonstrate that "assurance" either way. Regardless of whether Van Houweling's consulting occurred before or after he was employed by U-M, which is a vital legal question, IBM was certainly a convenient company for Van Houweling to ally with. In fact, it was an often stated maxim in the local and national writings of the time, that Van Houweling was hired precisely for the national information technology contacts he possessed. McCredie and Bok may have hit a bull's-eye.

    In a May 1993 taped interview between the author and Eric Aupperle, several interesting claims were made. Perhaps the most central result of the interview was the point at which I showed the quotation from the Robert Moore memo in 1989 which indicated his concern about "Aupperle's multiple hats" and the "cross-subsidization" between U-M accounts and staffing and Merit accounts and staffing. Aupperle's response was, "I read it, so what's your question .... [pause indicated] I disagree with the premise. [laughter from Aupperle]" When I followed up with the question of "what kinds of things would cause that level of concern with both Hankins and Moore," Aupperle responded unequivocally that "I have no idea. No idea." He went on to explain that all accounting is kept strictly separate with special standards for NSFNET related expenses, and stated that, "I have no idea how Bob [Robert Moore] arrived at that conclusion."
    Recall from earlier in this article, that just recently, in a 1994 FOIA release from the 1986 Merit Director's Advisory Committee meeting minutes, Aupperle is cited as expressing a concern over the fact that Merit was benefiting from U-M staff funding. Recall that he himself questioned, before that committee, the "fairness" of that funding benefit. Certainly in 1986, seven years before his interview with the author, Aupperle had an idea, even an explicit ethical concern, in relation to the same thought that Moore would eventually express in his 1989 memo. Was Aupperle caught in a lie in his 1993 interview, or shall we be "charitable" enough to attribute this to poor memory on behalf of the President of Merit?
    Even in other parts of my 1993 interview, Aupperle confirms some very strange accounting mechanisms which might lend credence to the Moore memo. When asked where Merit funding flows from and to, Aupperle noted that "I don't work for Merit, I work for U-M," and that "all Merit staff are U-M employees." Essentially, Merit is an organization where "there are no Merit employees." He rationalizes this by noting that Merit staff are only paid with a U-M check and that the Merit fiscal agent, which is another university on a rotating basis, is charged back by U-M for Merit related expenses including staff. According to Aupperle, he is a U-M employee in name only. Some might conclude however that he is rather a Merit "employee" in name only. Through a complex web involving U-M and Merit's fiscal agent which is either (at any given time) Wayne State University or Western Michigan University, NSF and Michigan Strategic Fund dollars are sent to the fiscal agent and then redistributed to either U-M, IBM, MCI, or now ANS depending on the invoices they submit. The possibility for abuse, waste, or even simple error is significant.
    Of other significant interest, Aupperle, when asked about Van Houweling's role in the 1987 solicitation process, referred to Van Houweling as a "visionary" [Aupperle was at least the third source who used that word] who could see through the national policy issues. He identified Van Houweling as the instrumental player in securing the support and cooperation of IBM and in securing the grants from the Michigan Strategic Fund. Aupperle also described the reasons for the creation of ANS and the reasons IBM and MCI were selected as partners.
    ANS was created in 1990 because, as a result of the growth of the national network, Merit no longer saw within its primary corporate mission the need to carry out the majority of NSFNET's work. Merit's original purpose was primarily to serve the networking interests among its' nine Michigan member universities, and although NSFNET indirectly benefited them by increasing their national connectivity, it grew beyond the regional purpose of Merit. Therefore, ANS was created singularly to replace Merit in its role to promote national network interests. However, ANS took over mostly the operational functions of the network while Merit retained the management functions and Network Operations Center.
    IBM was brought in mostly for its hardware and its experience in technical services which could benefit Merit in maintaining the network. MCI was brought in for its ability to produce data switching circuitry and its ability to coordinate the administrative problems (Aupperle emphasized the administrative end) associated with building a national network across a wide number of LATA's created in the wake of the AT&T breakup.

    [Note, 2003: While writing in 1995, the Defense Department was caught in an excessively priced hammer story, hence the reference].

    I tried to gain a little broader perspective on the questions of this report by speaking with several informed University sources. One person's initial reaction to a draft version of this report was that although it may be substantively accurate, it was difficult to imagine a pattern of "mal-intent" on the part of IBM, MCI, or the University. These organizations simply were "so large" that it would have been difficult to organize such an undertaking. Without dismissing the possibility, I tend to agree that this report hasn't unearthed evidence of any grand conspiracy on the part of the grantees, but I do believe it sheds light on some of the pork- barrel political motivations at play specifically in the NSFNET grant process, and in the university-research process generally.
    It should be clear that Van Houweling and the U-M in this case, and universities in general, may too often be only trying to "pinpoint bundles" of taxpayer money for bureaucratic expansion, rather than accomplish the sometimes legitimate purposes of publicly funded research. The public can not automatically count on the interests of a university matching its' own. Robert Moore's memo and the resulting $3.1 million settlement between the University and Department of Justice, does however lend credibility to the notion that there may be problems deeper than just waste and fat in appropriation process.
    That case highlights an important area for improvement in government oversight generally. The central issue there was that the U-M incorporated the costs of non-related services in its overhead calculation and that it shifted money it received from the government to pay for services not related to government research. The same picture of shifting costs and growing budgets is beginning to emerge within the NSF grant process. The cooperative-agreement format doesn't make a grantee's role and responsibilities very clear.  In essence, the government needs to either exercise the strict control over fungibility that it would ordinarily exercise in a traditional contract arrangement, or it needs to make it very clear to both winners and losers of a "cooperative agreement" that the only criteria it will judge the grantee by is the "outcome." As long as there is confusion about the legal responsibilities of the "cooperative-agreement," there will continue to be legal challenges and bitter competitors who perceive the process as arbitrary.
    According to Van Houweling, an internal NSF/Merit audit revealednothing unusual. I examined this 13 page audit by Bollam, Sheedy, Torani and Co. As Van Houweling indicated, there were no questioned costs and the auditing firm came to no negative conclusions. However, Bollam, et al seem to rely exclusively on summary aggregate data provided by Merit and make no detailed analysis of how they arrived at their onclusions. The audit essentially checked the Merit accountants' ability to add and subtract, something that has never been questioned. One can question the reliability and neutrality of such an audit. Such a routine audit examines "random" portions of a budget, and not being exhaustive or not looking in the right places might bias the report or make it incomplete. By the audit's own admission, it even used a "non-statistical sampling" and noted that the "report may not represent total costs that may have been questioned had
all expenditures been tested.
    These financial questions could be avoided in the future if Merit separated itself from the U-M. Looking a little bit deeper into the national controversy, competitive bitterness, and the public records Van Houweling and Merit have left behind leads one to ask some pretty hard questions. As of yet, the University has yet to come up with good answers. Let's hope they have some.

    Editor's Note:  Walter Harrison, director of university relations, sent us the following response on December 6, 1994.  The U-M had nearly two months to compose its response.

   "Chetly Zarko's article is long on innuendo and supposition and short on facts and proper context. Mr. Zarko repeats allegations that have been raised in the past and never been found, by the University of Michigan or outside agencies, including the Office of Inspector General of the National Science Foundation, to have had factual merit.

    We are proud to have played a role in the creation of the National Science Foundation computer network. Mr. Zarko is correct in writing that this is a network born of relationships between government, industry and higher education. It is a successful collaboration that has served, and will continue to serve, as a model for computer network infrastructures. Readers should know Mr. Zarko is hardly an impartial author.  He has several outstanding lawsuits against the University of Michigan. Objectivity is not a driving philosophy behind his reporting of these particular issues. We specifically object to the attacks on the personal and professional integrity of University of Michigan officials. These attacks transcend the boundaries of responsible journalism.  While we are always concerned that our policies and procedures are correct, we are confident that in this case there is no substance to Mr. Zarko's allegations."

[the author] replies that he has two Freedom of Information Act lawsuits against the U-M, not "several."

Author's Biography (1995)

    Mr. Zarko is a freelance writer currently living in Arlington, Virginia. He graduated with honors in political science from the University of Michigan in 1993. While at U-M, he had written a variety of reports about university research overhead rates, university administrative actions, and allegations of misuse of the university computer conferencing system.

    After graduating, he resided in the Ann Arbor, Michigan area for more than a year. During this time he learned that the Board of Regents of the University had been operating a restricted closed access computer conference. Although he had filed and been denied similar requests for other closed conferences, as an elected body, the Regents were subject not only to the Freedom of Information Act, but also subject to a state Open Meetings Act.

    The University denied him access in the interests of preserving the privacy of the Regental participants, despite the fact that they had funded the operation with their "Regents' expense account." In December 1993, Mr. Zarko filed suit seeking to compel the release of the conferences under both FOIA and OMA and alleging that even if these were somehow "private" communications, that the Regents had misappropriated public moneys for private use in violation of Michigan common law. In March 1994, the Ann Arbor News and Detroit Free Press threatened to join Mr. Zarko as co-parties in the litigation if the University didn't release the records to them. The two newspapers had only one month earlier won a 6 year $500,000 Michigan Supreme Court battle seeking to enforce greater openness in the selection process that yielded President James Duderstadt.

    In April, admittedly recognizing the new prospect of "expensive litigation," the University promptly allowed the newspapers first access to paper transcripts of the conferences. Until he retained an attorney, the University refused even to pay Mr. Zarko $82 filing fees. He is continuing his lawsuit, seeking access to the electronic version of the conferences and alleging capricious and discriminatory treatment in the disparate treatment he received relative to the newspapers. In March and April 1994, Mr. Zarko also requested access to the records the University had been ordered to release in the newspapers' 6 year presidential selection process battle, on the grounds that newspaper stories cited original sources pertaining to the use of e-mail in their decision- process. Mr. Zarko received highly edited (blacked-out) copies of those records, despite complete unedited disclosure to the newspapers. The University reasoned that the newspapers were entitled to some special 'spoils of victory' as the prevailing party in the suit, but that other citizens were not. In April 1994, Mr. Zarko filed a second lawsuit alleging capricious and discriminatory treatment, violation of the original court order compelling the University to release the records in order to "inform the public," that the 14th Amendment would bar interpretations of the law that did not protect all citizens equally, and a second violation of both the FOIA and OMA. In the only matter directly related to this article, Mr. Zarko has been denied access to Merit conferences like NSF:STAFF, on the ironic grounds that they are also the "private and personal" affairs of the participants.  He has not yet filed suit on that issue.