Mr. CLAGETT. I am, Senator.

Chairman STEVENS. Thank you very much, sir.

TESTIMONY OF ROBERT P. CLAGETT, CHAIRMAN, COMMITTEE ON CONTINUED REVIEW OF THE TAX SYSTEMS MODERNIZATION OF THE INTERNAL REVENUE SERVICE, NATIONAL RESEARCH COUNCIL; ACCOMPANIED BY C.A. IRVINE, MEMBER, COMMITTEE ON CONTINUED REVIEW OF THE TAX SYSTEMS MODERNIZATION OF THE INTERNAL REVENUE SERVICE, NATIONAL RESEARCH COUNCIL

Mr. CLAGETT. Yes, sir. In fact, I will make the only statement and then we can go into questions.

Chairman STEVENS. Yes, sir.

Mr. CLAGETT. Good morning. I am Robert P. Clagett. I am a lecturer at the University of Rhode Island and I am Chairman of the National Research Council's Committee on Continued Review of the Tax Systems Modernization of the IRS. I am also a member of the National Academy of Engineering.

The National Research Council is the operating arm of the National Academy of Sciences, the National Academy of Engineering and the Institute of Medicine. They were chartered in 1863 to advise the government on matters of science and technology. With me today, as you know, is Al Irvine, who is a member of the committee, and he is president of Eclectic Solutions Corporation.

The project I chaired involved an original committee, followed by an extension of some of the membership plus some new members. It produced 6 reports over a period of 5 years. My remarks today will draw on our last report, the February 1996 publication. That publication is entitled "Continuing Review of the Tax Systems Modernization of the IRS." It was released to the IRS in December of 1995. The IRS was the sponsor of the project, but, consistent with NRC policy, did not preview the findings or the recommendations.

The committee-the original and the extension-which I chaired was rich in private sector experience with large systems development and management, as well as technological expertise. My committee and I strongly believe that modernization of the information systems of the IRS is very important. The Service still has a lot of old computer equipment, which is difficult to maintain. Only part of the returns processing today is computerized, with taxpayer paper files the main mode of operations still. Modernization would enable better customer service and better agency and government performance.

In the course of our work, we focused on the ability of the IRS to modernize, while monitoring its progress. We made several recommendations relating to IRS structure and management that have been implemented. Among them, for example, was the establishment of focused units to address needs in architecture and privacy, elevation of positions with responsibility for human resources and for management of the tax systems modernization; that is, the modernization executive position that is now an Associate Commissioner. Through such actions, the IRS has improved its organization.

A central concern of the committee was management, running across the spectrum from a vision statement through planning and implementation. Quoting from our last report, the 1996 report, "The committee believes that success starts with a concise business vision, a well-organized modernization process, a clear systems architecture, a complete development plan, and a strong set of metrics to determine progress," quite similar to what the GAO showed you earlier.

The IRS has created a vision statement, and it has begun a process of reengineering its business processes. However, these efforts remain somewhat disjoint from TSM. Of greatest concern are the enduring deficiencies in technical management. It is in this area that progress must be made if TSM is to succeed. Those deficiencies show up in the IRS' own software and system development organization, its lack of a concise statement of system architecture, its approach to and slow rate of improvement in development processes, and its incomplete attention to system and data security.

The committee frequently expressed concern about the lack of experience in large systems development and implementation within IRS management. This deficiency shows up in the rate of internal progress and in the nature and effectiveness of interactions with contractors involved in TSM development. I would say here, as I have said repeatedly to the IRS, that the Service should not have had the in-house capability to take on such a large project when it started. That condition was no embarrassment, but it should have assessed the complexity of developing such a large project and then acquired the trained, experienced management team needed to direct it.

Instead, it relied on competent in-house managers who were and are generalists who have advanced within IRS for many years. These are good people who know how to run the IRS as it exists today. At the same time, they have not grasped our recommendations as to what it takes to develop a huge information system such as TSM.

Quoting again from our report, "The IRS must implement the following recommendations immediately and forcefully. Otherwise, the goals of the TSM effort are in serious jeopardy." I am going to run through them. The first recommendation is to acquire technical management expertise through both hiring and contracting, starting with a highly qualified CIO to fill the vacant position. Of course, I am pleased to note that position has now been filled, and I sincerely hope the new CIO has the background and experience to help make TSM effective. In addition, I need to add, though, that other qualified, trained, and experienced managers are needed at other levels as well.

The second recommendation is to effectively integrate operational and technical goals by developing a set of metrics to measure performance; that is, what do you want to have happen when it is complete?

Third, develop an adequate architectural definition design, led by the systems architect's office that they have created, and enforced by a strong set of interface specifications for key applications of systems and a process for maintaining them properly.

Fourth, significantly increase the scope, level of effort, management attention, and tools devoted to security development.

Fifth, implement an overall process improvement plan. I won't really go through it. It is the same idea you have just heard from GAO, to bring their internal capability up to Level 2 and then go from there. That includes, of course, training, assessment of where they are today, and so forth.

Sixth, focus efforts on the integrated case processing project. We picked that one because it is key to all of TSM. What we were trying to do was tell them once again, as we have before, in the other reports, pick one, do it well, test it, get it into the field, and then go from there. Furthermore, I think, and we state in there, TSM funding for projects that do not support such a focused effort should be reduced until definite improvement is shown.

Absent follow-through, new approaches to TSM may be necessary, in our opinion. One possibility, of course, would be to assess, justify, and fund each component project individually. Another might be to out-source, as Senator Glenn says-that is, to contract. The committee believes, however, that correcting the deficiencies as identified and strengthening its management is the best course of action for IRS and the country, again just as GAO said.

Our committee has expressed concern that the problems raised here, like some of the specific recommendations, have been raised previously and, in fact, for quite a few years. We therefore suggest that some ongoing mechanism be established that can both advise the commissioner and foster constructive interaction between the IRS and its oversight organizations to assure the best possible focused decisionmaking and communication.

With that, I thank you. Mr. Irvine and I will be glad to try to answer any questions you have.

Chairman STEVENS. Well, I do thank you very much. Your recommendations are quite similar to GAO, which you heard, and I am sure you have worked with them, too.

Mr. CLAGETT. No, we haven't worked with them very much.
Chairman STEVENS. You haven't?

Mr. CLAGETT. They have been independent assessments.

Chairman STEVENS. That is very good. They are quite concise and similar and point in the same direction as far as the concepts of developing the plan.

Do you think we are at the situation where we ought to say, well, let's go back and start all over again and start it from the beginning as you indicated it should have been started in the first place?

Mr. CLAGETT. Well, let me start by saying it is pretty late to do that. On the other hand, after 5 years I have to tell you that the committee is not very optimistic that IRS will or can implement all the recommendations that we have made.

Chairman STEVENS. Have you analyzed why that is the case, Mr. Clagett? Is it, as the other gentleman said

Mr. CLAGETT. I will be glad to give you my speculation. It certainly is somewhat different than the answer you got from GAO. Chairman STEVENS. Is it civil service restrictions?

Mr. CLAGETT. I don't know if it is civil service restrictions so much as it is the way in which the Federal agencies work. This is my own experience now. We have not addressed that specifically in

the committee, so I will give it as my opinion. I have served on other committees looking at similar modernizations. DLA-I earlier chaired that one, similar kinds of problems, and that is the approach that was taken.

Most private organizations, when they have such a problem-I know CEOs of several companies that have done large projects, and I have asked them how they go about it. The first question, of course, is what is it we want done; what kinds of things do we want to accomplish when we finish. So the very first thing is an overview, a vision of what it is they want done. It took the committee a couple of years to convince IRS that that is the first thing that ought to be done, and that was well after they were into development.

Second, then, usually the private firm is just like IRS and would say, well, we don't have the wherewithal, the technical knowledge, the leadership to do such a job, and they will go out and find a top person to lead such an effort. Certainly, if it were the size of TSM, they would do that. Instead, of course, IRS started from within, had a CIO, and apparently decided they could develop TSM internally.

Yes, there are civil service problems in hiring a first-rate person who has had large-development experience and has technical training, but that is the way it would be done outside. That person would then bring in some key people who had good technical experience, good large-project management experience, and then draw from that people to work with them. That was not done. Therefore, the planning wasn't done in that way. There was not the same thing you saw on the board with GAO; there wasn't a full plan set out, from which metrics could be developed. That said, now we will know when it is we have achieved success, then work back in the project to set milestones to get there.

Absent all that, I believe the only thing to do now is what you have heard earlier, and that is to back off from TSM, put TSM as a whole concept on the shelf, and have IRS propose specific projects-we picked ICP and proposed that they throw all of their efforts on to completing that, doing it well, upgrading their software people at the same time, learning from it, testing it, and then putting it into the field, and then make a proposal to move to the next project. That would also give them time to do the two things we still think are urgently needed; that is, a system architecture overview so that you have a complete picture of the whole project, and also time to develop the standards for interfaces.

One of the things that I worry about when they are doing all of these projects simultaneously without standards is that they are not going to achieve the goals of TSM, because the different developments are not going to be able to fully communicate with each other.

So, that is a long-winded answer to your question. I think it is a serious problem and I don't think IRS is alone in the problem. Chairman STEVENS. What you are really saying is you question whether the IRS is capable of implementing TSM as it is now structured.

Mr. CLAGETT. That is very right.

Do you want to add anything, Mr. Irvine?

Mr. IRVINE. Well, I think everybody needs to understand that software is still among the most difficult things we try to put together. I have been in the business 40 years, and look anywhereprivate industry, government, anywhere-and you will discover that we don't perform as well in terms of meeting schedule, meeting budget, and providing capability in software as we do anywhere else. That is because it is really complicated. These are among the most intellectually complicated things that people try to put together.

If you think about a software project that is several billion dollars, then that needs to be approached-the process by which you are going to carry that out must be careful, controlled, disciplined. As was said earlier today, you have to use the very best practices available if you are going to tackle a project that size because even smaller projects, a couple-of-hundred-thousand-dollar projects, are quite risky in the software game.

Chairman STEVENS. Have you gentlemen been through developments in the private sector in terms of the transition to a new solution of this type? What I am going to get to is whether or not it is acceptable in this period that we are only getting 51 percent of the telephone calls answered.

How does private business handle the transition from the old way of doing things to the new way of doing things? Do they go through this period we are going through of just absolute customer rejection of the concept of the transition?

Mr. IRVINE. I think they manage that very carefully. You have to look in some sense at return on investment, and what the return is is often customer satisfaction and all of that sort of stuff. It is not clear to me—and this is my own personal opinion; I am not speaking for the committee now-that I know how to apply those kinds of measures to what the IRS does. How happy can I be as a customer? If they return me three times my refund, I am really happy. We don't want to do that, so it is not clear to me how to apply those kinds of measures to this circumstance.

Mr. CLAGETT. Well, I would add that in most industry efforts that I have seen, the first thing that is done after they decide what it is they want to do and what the end result will be is then to assess whether or not they can do it; that is, how much resources they have, and so forth. The last thing they do is attempt to develop the whole thing simultaneously and start developing all the systems simultaneously.

So I would say that what I would have hoped would have been done was, having set that out and set the plan out, to then pick the most important thing that was doable first and get started on that while they continued to do a better job on serving the current public. Most firms, as you said earlier about your own situation, couldn't afford to only answer half of their customers' inquiries. They wouldn't have enough money or wouldn't be in business long enough to develop the new system. So you can't let the current operation in any way deteriorate while you are developing a better system.

So I would say, in retrospect, now having spent this time looking at them, that it would have been much more preferable to have done it in smaller pieces and be able to manage it better.