IRS ANSWER FOR SENATOR AKAKA

Toll-free calls from Hawaii are answered by our assistors in Customer Service sites in Portland, Seattle, and Fresno. Taxpayers in Hawaii who have received a notice with the 1-800-829-8815 telephone number can receive assistance between 3 a.m. and 8:30 p.m. Taxpayers who call the 1-800-829-1040 telephone number with an account related question receive service from 5 a.m. to 8:30 p.m. If the taxpayer has a tax law question, assistors are available from 5 a.m. to 4 p.m. After 4 p.m., taxpayers may leave a message and we will return the call within 24 hours.

Senator GLENN. Thank you, Senator Akaka.

We are going to have to conclude this panel. The time has gotten away from us this morning. Thank you very much, gentlemen. We hope you will be responsive to any additional questions that we may have to submit on behalf of individuals or the Committee staff so we can publish them as part of the record. Mr. DOLAN. Thank you, Senator.

Senator GLENN. Our next panel includes Bob Clagett, chairman of the Committee on Continued Review of the Tax Systems Modernization of the Internal Revenue Service of the National Research Council; John Gioia, president and CEO, Robbins-Gioia, Inc., and Robert Tobias, national president of the National Treasury Employees Union, and he is accompanied by Steve Herrington, president of NTEU, Chapter 27, in Columbus, OH.

Gentlemen, if you would all come to the table, we would appreciate it, and obviously, time being what it is, I always hate to have people come in-I know you have prepared testimony and are prepared to give a 3-hour dissertation, and we ask you to summarize a lifetime of experience in 3 minutes, but if you could summarize your statements this morning, it would obviously be appreciated. Mr. Clagett?

TESTIMONY OF ROBERT P. CLAGETT,1 CHAIRMAN, COMMITTEE ON CONTINUED REVIEW OF THE TAX SYSTEMS MODERNIZATION OF THE INTERNAL REVENUE SERVICE, NATIONAL RESEARCH COUNCIL

Mr. CLAGETT. Senator, I would appreciate it if my statement could be put into the record.

Senator GLENN. All statements will be included in their entirety in the record.

Mr. CLAGETT. Anticipating this, I first want to say that I was very pleased to hear Mr. Gross say that they are (1) creating a systems architecture plan for TSM; (2) that they are hiring more technical expertise-I am not clear how they are going to do that with the RIFs, but I am pleased to hear it-and (3) that they are developing the detailed plans that GAO spoke about.

I would only say that our Committee met for 5 years, which is less than yours has been working on this, and we have been frustrated, like you, during all of that time that these things have not occurred, and we have been saying those same things for those 5 years.

My statement is in two parts. The first part of it I prepared earlier, after the Treasury had submitted their response to you, so I will summarize that first and then get into a second part, with a few comments and summary.

1 The prepared statement of Mr. Clagett appears on page 342.

The first testimony that I prepared talked about the Treasury plan, and if you will recall that sharp turn that they had proposed for revising TSM and the way in which IRS operates, my fundamental comment on it was only that the sharp turn was a good idea, but if you recall, it said nothing about reducing the budget for IRS back then. I think that was in March. And I strongly felt that indeed the TSM part of the IRS budget should be cut back and that they should focus on a single project; that has been the recommendation. It was the recommendation in the NRC final report, and I made that recommendation when I appeared here before.

I do disagree, by the way, I should add, with GAO this morning. They said that their recommendation was to cut back to a few projects, but they suggested that IRS pick some technically simple projects on which to go forward commensurate with their technical capability. Our Committee took quite a different tack.

In the first place, of the large number of technical people in IRS, there are indeed many who are very expert. So our recommendation was to pick a technically challenging project that is currently underway-the ICP project, which is a large one and technically challenging—and pull their best people together on that and focus on completing it, putting it into the field and being able to report to you that it is complete, using it as a learning vehicle to increase their own capabilities and as an example for the rest of the IRS internal people in pulling them up to a current technical level. That summarizes the first part.

On the second one, I am not prepared to comment about Cyberfile, which is new, but I am going to talk about probably treats the same thing. Recently, the Committee staff sent to the National Research Council a copy of a proposal for a bill, and I would like to comment on that briefly.

First, I certainly would like to congratulate the Committee on identifying one of TSM's most serious problems, and that is the software development capability inside IRS. It is weak; their own review of their level and the CMM levels was at Level 1, as you have heard both from me and from GAO before. And we had recommended in our final report that they set as their target to get in quick order to Level 2. The ability once you get to Level 2 in increased speed, accuracy and efficiency will more than pay for that effort.

As I understand it, the proposal was-and I again congratulate the Committee on the way they put it, which was to set a Level 3 target and to withhold funds for any further TSM development until they got there. I have a reaction to that that I will take a couple of minutes to comment on.

First, the levels of the software CMM are fairly arbitrary. The development, as you can imagine, of increased skills or improved equipment is a continuum from the lowest level to the highest. So I backed that up with a survey that I was just given by Roger Bate, who is a member of my committee and also a senior member of the Software Engineering Institute at Carnegie Mellon. He sent me a 1994 survey summary done by Software Institute of a large body of the numbers of companies, both private companies and in the Federal system, that have reached Level 3.

First, from that survey, they found that over those large numbers, it took on average 25 months to move one level-that is, from Level 1 to Level 2, or from Level 2 to Level 3. Most significant to me were the gains during the time they were increasing their capability, the continuum. Productivity in software development increased 35 percent per year in the time they were working on improving their capability. Reduction in completion or, for the private sector, time to market, shortened by 19 percent pear year of the time they were in that development. And to me, the most significant number, the post-release that is, before release defect report, that is, the number of defects they found, dropped by 39 percent per year of the time they were doing the development. Those are big numbers, and in fact that last one, of decreasing the bugs in software by 39 percent a year, is an astonishing one and one that would qualify doing the whole effort just on that basis.

From that, you see that even with these skilled groups in a large number of organizations, it took them 2 years to get from Level 1 to Level 2, and indeed it took another 2 years to get to Level 3. That means that if you have set a target for the IRS to get to Level 3, it is at minimum a 4-year project if they go at it as hard as they

can.

In summary, we would suggest that you give them a target of Level 2. That is the level that they are now requiring their contractors to be. That has been one of our major concerns, and I know it is yours as well, that they have Level 1 capability people trying to administer to people who are already at a Level 2, a much more sophisticated level. There is no way they can manage them because they cannot even understand what it is the contractor is doing, and that is a very serious problem.

So I would suggest that you consider going to Level 2, and when they get to Level 2, set a new target, set the same kinds of stringent goals that you are proposing for getting to Level 3, let us say by the year 2000 by then, and move from there.

Finally, I should comment also that I learned from Mr. Bate that the software acquisition CMM that was referred to has just been published this year, so his recommendation is that you not make that part of the target because there has been no study and in fact no experience yet in its use, but that IRS indeed use that new CMM in developing their own acquisition capability and that maybe at the time you start the target for software Level 3, you would also introduce Level 3 of the acquisition CMM.

That is about it, except that I would like to make a couple of points. First, if you are going to set targets and set funding based on achieving those targets, our recommendation, or certainly mine, would be that you have some kind of outside body outside the IRS certify that they have gotten to that level. I do not think it is proper to have IRS say we are there. I think somebody outside should certify to that.

And second, in understanding that a new bill is a comment or a creation of the control board for management, I would like to suggest that the Committee also consider insisting that there be some kind of outside technical equivalent of a control board to make sure that the kind of advice that was given, for example, by the Com

mittee I had could continue to be given to the IRS and could continue to recommend to this Committee further technical items.

With that, I will stop and try to answer any questions you may have.

Senator GLENN. Thank you. We will wait until everyone has had a chance to testify.

Mr. Gioia?

TESTIMONY OF JOHN GIOIA,1 PRESIDENT AND CHIEF EXECUTIVE OFFICER, ROBBINS-GIOIA, INCORPORATED Mr. GIOIA. I also will submit my comments for the record, and summarize what I think are the high points. You mentioned some of them. I have read some of the best plans written and not implemented well—not only in IRS, but in many other agencies. And I have asked myself the question over the years-why is that so. The IRS does understand PERT charts-although they are more complicated now. They do understand change process, earned value, and all those things, and they do have the plans to execute thembut they do not get it done. So you wonder why, and I reached a conclusion, and I think I can validate it. It is that a decision process is lacking; a good, strong, disciplined and controlled decision process does not exist. It does not exist for a lot of reasons; mostly, it is because of the culture and the core competencies within the agencies. And I am not picking on IRS. As I said, I can make the same case for the FAA, Social Security, and others.

Agencies must have a discipline and control process and a successful program environment. I am talking now about a program environment for managing programs, not for managing the total IRS. Such as managing a program within the IRS like TSM, that you can bound, that has a set of objectives that you can define, that can be measured to achieve success. In that kind of an environment, you must have a discipline and control process. You must be able to minimize micro management. You have got to have a streamlined decision process but also some checks and balances. You have also got to have minimum but effective oversight and, most important, you must have effective communications with all the stakeholders involved so they help you rather than try to crucify you.

Industry succeeds in making a lot of these things happen. You heard testimony this morning that in industry, VISA and companies like that, seem to get these jobs done, and the government does not.

I suggest that in order to build this kind of environment and process that I have just described, which is really very simple-I mean, we are not talking about magic bullets or things that are esoteric; we are talking about something that is basic in management that most people understand-but you have to implement it. As you said, the Congress cannot direct the implementation, but you should be able to make sure it is being accommodated.

In the implementation process, to ensure that the agency has a disciplined and controlled environment, I suggest a control board, similar to what Mr. Clagett just referred to, very similar to a board

1 The prepared statement of Mr. Gioia appears on page 347.

of directors in a big corporation; an independent, strong board with the ability to make the decisions, approve the requirements, redirect the program if it is necessary to be redirected, not be influenced by any time or political change, hold changes to a minimum, and redirect or cancel the program if it requires that.

The board should be independently manned. It should be manned by people who do not have any skin in the game. A member of the government should not chair the board. Nor should members of industry who are suppliers to the particular agency executing the program.

That board should be in place prior to program implementation and that board should exist under some kind of a charter that allows them to continue through a political environment such as a reelection or a change of administrators or a redirection of the program by any other external influence.

That board should have the authority to validate requirements and control changes, which is critical. It is because of requirement creep, lack of change control, and poor program management and implementation that these programs ultimately fail.

The next thing below the board organizationally should be a good program management environment with a good discipline and control process that allows the kinds of things that you were referring to, Senator Glenn-good PERT charts, good understanding of interdependency, the ability to integrate various systems together as a whole, the ability to measure progress against a set of metrics, to know when you have variances and to have the ability to adjust for those metrics or variances when they occur.

That is, again, nothing that is uncommon, but it is very, very hard to implement. It is hard to implement if you do not have a control process on board that forces decisions to be made, allows people to say "no," minimizes changes, has the authority and the responsibility to execute and act as a corporate entity.

I think, in my humble opinion, that is the way to improve the process for fielding programs that we are facing throughout the Federal Government, particularly within the IRS.

With that, I will conclude and let you ask any questions.

Senator GLENN. Thank you very much.

Robert Tobias, national president of the National Treasury Employees Union. Bob, we are glad to welcome you again. You have appeared before this Committee before.

TESTIMONY OF ROBERT M. TOBIAS,1 NATIONAL PRESIDENT, NATIONAL TREASURY EMPLOYEES UNION

Mr. TOBIAS. Thank you very much, Senator Glenn.

Senator GLENN. We welcome your accompanist from Ohio as well. Welcome, Steve.

Mr. HERRINGTON. Thank you, Senator.

Mr. TOBIAS. I am pleased that this Committee in general and you in particular, Senator, are interested in the views of employees on the most important issues that are before this Committee today.

My written testimony discusses funding for tax systems modernization and the associated issues of buyouts, but I would like to

1 The prepared statement of Mr. Tobias appears on page 362.