Sep 17, 2010

It Will Work Because It Has To Work. We Have No Choice.

As I have stated before, the key assumption of the "Re-Imagining" report, which recommends that the Social Security Administration pretty much eliminate its field operations and try to handle all of its business by computer, is that Social Security has no choice but to do this because there will be no other way for the agency to handle its upcoming workload. I have been hearing this same sort of language for a long, long time. In the past, this sort of talk has always preceded a disaster.

In 1994 Social Security officials began work on a project to "re-engineer" the process for giving disability claimants hearings. High-priced consultants produced reports recommending a dramatic re-organization. The reaction in the field was that the plan was crazy. I can recall going to a conference of the National Organization of Social Security Claimants Representatives (NOSSCR) and hearing a high level Social Security official describe the re-engineering plan. She took questions. I asked her what Social Security was going to do if the re-engineering plan failed. She replied most emphatically that re-engineering would work "because it has to work." She said that Social Security "had no choice."

You can see the obvious problem in this person's thinking. Ideas do not work because we need for them to work. They succeed or fail based upon their merits rather than upon our perceived need for them to work. Once you say that you have no choice, you close off alternatives and are liable to do something foolish. Thinking this way, a man who cannot make enough money to support his family might decide that his only choice is to buy lottery tickets.

Re-engineering was tried out on a limited basis. It failed and was abandoned before it could do much damage. Tens of millions of dollars were squandered on consultants.

Unfortunately, the notion that Social Security could not possibly get its work done without some productivity breakthrough did not go away. The next effort at a "Great Leap Forward", to use the term that Chairman Mao employed in China to describe something vastly larger but motivated by the same belief that there was no choice but to try a "Hail Mary" pass (to mix my metaphors to a cosmic extent), was called Hearing Process Improvement (HPI). This was another effort at a re-organization of Social Security's structure to give hearings to disability claimants. Again, we were told that there was no choice, that this had to be done to meet future needs. This time, Social Security did not let the dismal results of HPI's trials slow them down. They plowed ahead with nationwide implementation but they made sure to do it at the end of 2000 as the Clinton Administration was leaving office. With only an acting Commissioner of Social Security for many months into the Bush Administration, there was no one to pull the plug on HPI as hearing backlogs soared to horrendous, unimaginable levels. HPI has to rank as the most dramatic mistake in Social Security history.

Social Security never recovered from the HPI debacle during the Bush Administration, mostly because the agency lacked operating funds but also because the new Social Security Commissioner, Jo Anne Barnhart, was able to distract everyone with her own plan to end the agency's backlogs. This plan involved electronic files and something else which she really did not want to describe, other than to tell us that it would be wonderful. Again, Barnhart told us that there was no alternative to her plan, whatever it was. The electronic files were implemented at enormous expense. They have still not led to any dramatic improvement in productivity. The rest of Barnhart's plan, which she called Disability Service Improvement (DSI), just got delayed further and further. She kept the details a huge secret until near the end of the Bush Administration. Once the details were released, it was clear why she had kept her plan a secret. It was nothing more than another ill-conceived reorganization plan. Like HPI, DSI's implementation was delayed until Barnhart was nearly out the door. Fortunately, it was not rushed into nationwide implementation. DSI was another failure. The current Commissioner, Michael Astrue, learned that not long after taking office and began to stop it. I am not sure that DSI has been fully wound down even now, some three years later.

The moral to this long story is that we should be extremely wary of anyone with a plan for Social Security who tells us that there is no alternative to his or her plan.

Sep 16, 2010

Legislation Passed To Name Building After Robert Ball

A press release from Social Security:

Legislation to name the Operations Building at Social Security headquarters in Baltimore, Maryland as the “Robert M. Ball Federal Building” cleared a significant hurdle with passage in the House of H.R. 5773 by unanimous consent. It is fitting to name this building after Bob Ball who served as Commissioner of Social Security under three Presidents and who continued to write and speak about Social Security until his death in 2008.

I want to thank the sponsor of this legislation, Congressman Elijah Cummings, the co-sponsors including the Maryland congressional delegation, Acting House Ways and Means Chairman Sander Levin, and Social Security Subcommittee Chairman Earl Pomeroy. This legislation now heads to the Senate where I hope it quickly passes.

Some Ways To Address Skepticism About The Re-Imagining Report

Social Security has a panel looking at its future technology needs. A subcommittee of that panel has issued a report telling Social Security that it can, over time, make its computers do almost all its work. The report is supposed to be a "re-imagining" of Social Security's future. Actually, I see no "re-imagining" here. It is a vision that already existed in the minds of many high level Social Security officials. They have been talking about this for years. You appoint a group of high tech people, some of whom may be salivating at the thought of obtaining huge contracts with the Social Security Administration, and give them only limited information about the agency's workloads, what do you expect? Of course, they are going to recommend a heavy dose of technology to cure all that ails Social Security.

I am deeply skeptical of the vision of Social Security's future that this panel has produced. The quite unscientific poll I ran on this blog indicates that a lot of other people share my skepticism. The report itself says that the panel has heard that many Social Security employees are deeply skeptical. This skepticism is based upon a much deeper knowledge of Social Security's workloads than this panel has. It is also based upon our memories of past information technology failures and shortcomings at Social Security. If Social Security is going to move towards this panel's vision, the agency is going to have to address the bases for our skepticism. Let me suggest some ways that Social Security can work on this:
  • Social Security has a real problem with something called the windfall offset. The windfall offset reduces back Disability Insurance Benefits (DIB) by the amount by which Supplemental Security Income (SSI) benefits should have been reduced if the DIB had been paid when it was due. Got that? Not likely, unless you work at Social Security or work pretty closely with the agency. Trust me, the windfall offset is a big deal. It eats up a lot of work years at Social Security. It accounts for a lot of mistakes and delays. In the end, though, it is a computation, something that a computer system ought to be able to do. Social Security has tried twice to implement a computer system to automate windfall offset computations. Both efforts were total failures. My recollection is that the last effort to automate the windfall offset cost about $80 million. I am pretty sure that the panel was not told about this history. If Social Security management wants a high tech future for Social Security there is no better place to start than here. If this problem continues to defeat a high tech solution, Social Security management should keep its high tech ambitions modest.
  • Social Security recently tried to implement a computer program to "pull" exhibits for Administrative Law Judge (ALJ) hearings. "Pulling" exhibits means selecting the stuff that is significant and organizing it. Lots of people, including me, said in advance that this would fail. It turned out that we were right. Social Security ought to do a post mortem on this and make it public. How much money was spent on this? Why did high level Social Security officials think it would work? Why did it fail? Could it have worked if it had been done in a different way? What does this experience tell us about the possibilities for future information technology systems?
  • Over the last decade, Social Security has been implementing electronic files for Social Security disability claims. There are still bugs but the system mostly works. There were promises that switching to electronic files would lead to dramatic productivity gains. To the best of my knowledge, Social Security has never issued a report on the effects of the electronic files on productivity. I think that if there had been big productivity gains that Social Security would have been shouting it from the rooftops. They ought to tell the world what the results have been from this recent implementation of a system that mostly works. If they are not comfortable doing that, how much confidence should we have in the prospects of dramatic productivity gains from future information technology systems?

Sep 15, 2010

Unexamined Assumptions

This is, I think, the key sentence from the Re-Imagining Social Security report that first came to light recently:
Given the projected workload increases due to the number of individuals retiring over the next two decades and other demographic trends, electronic self-service appears to be the only solution that will enable SSA to process future transaction volumes.
Really? Why?

First, if the members of this subcommittee think that handling retirement claims is a huge challenge for the Social Security Administration, either they were poorly briefed by Social Security or they were not paying attention. Encouraging those who are retiring to file their claims electronically is a good idea but even if every last retirement claim is filed electronically, Social Security's situation is not much better than it is today. Retirement claims are by far the most common type of claim filed with Social Security but they are so easy to process that they account for a surprisingly modest part of Social Security's workload.

The real problems are the survivor claims, the disability claims and appeals and SSI. Disability and survivor claims are mentioned in the report but I get the distinct feeling that the Panel members had no idea of the challenges they present. They simply assumed these problems to be minor and easily addressed. Social Security's history over the last 40 years tells us the problems with disability claims are intractable and that the problems presented by survivor claims are so complex that Social Security has scarcely tried to tackle them. I looked through the report and the term "SSI" is not even mentioned. How do you do a report on Social Security's future and not mention SSI? My experience is that SSI is irreducibly complex. Again, what kind of briefings did the Panel members get? Were they paying attention to the briefings they did receive? Panel members seem to have made the naive mistake of thinking that Social Security's employees are mainly involved in handling retirement claims. This report is directed not at the real Social Security Administration but at a fantasy agency that exists only in their imagination.

Why is it assumed that Social Security cannot get future appropriations sufficient to allow some reasonable degree of personal service to claimants? It is only by making the assumption that this is impossible that you get something like this report. History suggests that Democrats are quite concerned that there be a reasonable degree of personal service to Social Security claimants. While Republicans are more apt to give this only lip service, they have not been unconcerned with the political consequences of poor public service at Social Security. Even President George W. Bush, the most right wing President that we are likely to ever have, was not unconcerned with service at Social Security. He was consistent in giving Social Security a better break on appropriations than just about any domestic agency. Politicians of all stripes are eager to cut government costs by substituting technology for civil servants but I see no basis for a fundamental assumption that appropriations for customer service at Social Security are going to dry up. If the computers cannot be made to do the work, there will be people available to do the work. The real question is the extent to which information technology can be made to substitute for Social Security employees. This Panel had little idea what it is that Social Security's employees actually do so they had little idea of the complexity of the issues presented when trying to substitute information technology for warm bodies in this context.

Getting an outside opinion is a useful exercise but only if the outsiders know what the facts are. These outsiders know so little that they could do no more than muddy the waters.

Sep 14, 2010

Chief ALJ On Receipt Of Unemployment Insurance Benefits


Click on the page once or twice to see it full size.

Prison Hearings

From a memo issued by Social Security's Chief Judge:
Hearing offices often have difficulty scheduling an in-person or video-teleconference (VTC) hearing when an individual is held in a place of confinement (jail or prison). To reduce delay in adjudicating these cases ... a hearing may be held by telephone for confined individuals. ...

A hearing by telephone cannot be scheduled if the confined individual objects in any manner.

Alan Simpson and AARP

From the Columbia Journalism Review:

[While serving in the Senate, Alan Simpson, the co-chair of the President's Deficit Reduction Commission] disagreed with the AARP’s [American Association of Retired Persons] positions on Medicare and Social Security, [and] believed the group was obstructing budget cuts that Republicans needed to make in order to offset a planned round tax cuts. Simpson held hearings on the AARP’s finances. “I’m a chairman. I can have hearings,” he boasted to reporters in the Capitol corridor, dancing a little jig and pumping his arms in the air. A few days before he announced the hearings, Simpson said “People ought to know where their money comes from and what it’s used for.” As I reported at the time, Simpson never produced a smoking gun, but he created plenty of smoke, focusing on irrelevancies like the size of AARP’s new building and its executives’ salaries.

But the AARP recognized what the hearings were really about. At a meeting with AARP’s board and staff, Simpson told them “I want you to know that the intensity of my investigation will be directly related to the intensity of your fight on Medicare.” In an interview then, AARP’s chief lobbyist John Rother told me: “Many people on the right wing realized that AARP was the force to contend with. They realized they wouldn’t get anywhere unless they dealt with us as an institution.”

The nasty streak Simpson has demonstrated lately is nothing new.

Let me hasten to add that AARP should not be above criticism. My opinion is that it is little more than an insurance company posing as a nonprofit grassroots membership group. A genuine grassroots membership group of retired persons would be a far more formidable force than the AARP.

Social Security Hearing Office Average Processing Time Report















From the newsletter of the National Organization of Social Security Claimants Representatives (NOSSCR). Click on each page once or twice to view full size.

Compare the average processing time as it has changed over time:
  • January 25, 2007 -- 508 days
  • May 25, 2007 -- 523 days
  • July 28, 2007 -- 528 days
  • August 31, 2007 -- 523 days
  • November 30, 2007 -- 500 days
  • February 29, 2008 -- 511 days
  • May 30, 2008 -- 523 days
  • June 27, 2008 -- 529 days
  • July 31, 2008 -- 530 days
  • September 3, 2008 -- 532 days
  • November 5, 2008 -- 476 days
  • December 3, 2008 -- 480 days
  • March 8, 2009 -- 499 days
  • April 24, 2009 -- 505 days
  • June 3, 2009 -- 505 days
  • September 29, 2009 -- 472 days
  • July 5, 2010 -- 415 days
  • July 30, 2010 -- 410