The union throws the Commission a Hail Mary on the change in service standards
June 25, 2012
Sometime this week, the Postal Regulatory Commission will issue a ruling on the APWU’s Motion for an Emergency Order, which seeks to prevent the Postal Service from implementing a change in service standards for First Class mail on July 1.
If the PRC does not issue an order to stop the Postal Service, overnight delivery will come to an end for Inter-SCF First-Class Mail and periodicals. That’s mail that originates or destinates outside of the geographic area served by a sectional center facility (SCF). It’s basically regional mail, in between local and long distance, and it represents about 20 percent of First Class mail. The other 80% of overnight delivery is Intra-SCF mail — the “turnaround” mail that originates and destinates within the same geographic area served by an SCF — it's basically local mail that "stays in the building" — and it will be preserved until 2014, then that will be the end of overnight delivery for First Class mail.
The APWU says the Postal Service should wait until the PRC issues its Advisory Opinion before implementing the changes, sometime around Labor Day. The Postal Service says it won’t wait, and it’s filed an Opposition brief to the union’s request, arguing that the PRC doesn’t have the authority to issue an injunction, and even if it did, there’s no justification for it.
The APWU argues that the 1970 Postal Reorganization Act (PRA) — the legislation that created the Postal Service, the PRC and Advisory Opinions — says the Postal Service must request an Advisory Opinion when it’s planning a change in service with nationwide impacts. The Postal Service replies that the Act says only that it must request an Opinion; the Act doesn’t say that the Postal Service must wait until the Opinion is issued.
The APWU says that if the Postal Service implements the changes before the Opinion is out, it would render the legal requirement for an Opinion “nugatory” (of no value). The Postal Service responds that if it were expected to wait for an Opinion, the regulations would not specify that the request for an Advisory Opinion must be submitted “not less than 90 days in advance of the date on which the Postal Service proposes to make effective the change” (39 CFR § 3001.72).
The APWU argues that if the service changes have already been implemented when the Opinion comes out and the Commission makes a recommendation against the changes, the Postal Service will say that it would impose too great a cost on itself and its customers to undo the changes. The Postal Service replies that if it proves desirable to reverse the changes, mailers and the Postal Service can adjust. In the meantime, says the Postal Service, a delay would lead to additional expenses and require it to revise its consolidation plans yet again. (All the documents concerning the APWU Complaint and motion for an emergency order are in docket C2012-2.)
No stopping summer consolidations
As good as the arguments may be for an emergency order, it’s not likely the Commission will rule in favor of the AWPU’s request. Such an order would say, in effect, that the Postal Service must wait until the Opinion is issued. As the Postal Service notes in its Opposition brief (p. 14), PRC Chairman Ruth Goldway has already expressed her view that the Postal Service needn’t wait for an Advisory Opinion to be issued before it implements the change in service standards.
The APWU knows this, and it has known for a long time that the change in service standards would most likely be implemented before the Opinion was issued. In March, USPS VP David Williams told the Commission that the Final Rule on the changes would be published in the Federal Register in mid-April and the Postal Service would immediately begin the process of helping customers to adjust to the new system. Everyone knew that the Opinion would not be out until months later.
The Final Rule was not published until May 25. The delay of over five weeks was possibly due to developments in Congress and complaints from mailers, but most likely because the Postal Service was busy revising the plan and developing a two-phase approach. So rather than an implementation date in May — about four months before the expected date of the Opinion — the changes will be implemented on July 1, around two months before the Opinion.
That may not seem like a significant difference, but the legal battle waging between the APWU and the Postal Service suggests that there’s more to the matter, even though the legal arguments don’t make clear exactly what’s at stake.
At first glance, one might think that the Postal Service needs to make the service standard changes on July 1 so that it can proceed with 48 plant consolidations in July and August. That would be a strong argument in favor of implementing the service standard changes now, but that appears not to be the case.
The APWU brought forth an expert witness, Pierre Kacha, who provided supplemental testimony about his analysis of the two-phase approach announced by the Postal Service on May 17. Mr. Kacha ran a network simulation model on the 48 plants slated for consolidation in July and August. It showed “that the impact on inter-SCF overnight delivery will be minimal. This suggests that the Postal Service can implement the consolidations planned for this summer while maintaining the current service standards.”
The Postal Service did not provide any testimony challenging this assessment, and it does not contest the claim in its opposition to the emergency order. In its Opposition, the Postal Service only says that given the suspension on consolidations between September and January, “a delay in the July implementation of the Final Rule could mean a delay in implementation until 2013” (p. 24).
The Postal Service does not directly address the APWU argument that the 48 summer consolidations could take place with current service standards in place. And that raises a good question: If the Postal Service can proceed with the summer consolidations without changing service standards, why not wait until the Opinion comes out before changing the standards?
The ever-evolving Rationalization plan
One of the arguments put forward by the APWU is that the Postal Service's latest plan is so different from what it started with, it ought to be submitting a new request for an Advisory Opinion. That’s not going to happen, but it points to one of the main problems in the whole process.
While the Postal Service has provided voluminous documentation about its mail processing system, along with plenty of details about operating costs and cost savings, it’s extremely difficult to predict what will happen when the Postal Service changes service standards and consolidates plants. And one of the reasons it’s difficult to make good projections is that the Postal Service keeps changing the plan, and the numbers it provides always seem to be a step behind the latest version.
The Network Rationalization plan has been “evolving” almost since it was announced, and that makes it a moving target for opponents. It’s also plain frustrating for the Commission, which is simply trying to get a grasp of the plan. Without a good understanding of how much the plan would save, it’s difficult to do a cost-benefit analysis and render a useful Advisory Opinion. It’s still not clear if the plan will do more harm than good.
When the plan was first announced last year, the Postal Service said it would consolidate nearly 300 plants and save $3 billion. That number, by the way, still appears on the USPS website.
When the Postal Service presented the plan to the PRC in December, it encompassed about 260 consolidations, yielding an operational savings of $2.6 billion. After deducting $500 million in lost revenues due to slowing down the mail, the net savings would be $2.1 billion. The whole plan hinged on eliminating overnight delivery for almost all First Class mail as well as slowing a lot of 2-day delivery to 3-day. The idea was that the overnight standard necessitated using sorting machines for just eight hours a day; eliminating overnight delivery would enable the Postal Service to use those machines 20 hours a day. Fewer machines would be needed, hence a smaller amount of floor space, so plants could be consolidated and tens of thousands of jobs eliminated.
After the plan was submitted to the PRC, not all of the AMP studies were approved, so in late April the Postal Service submitted new testimony revising the cost savings for about 230 plants. The new numbers showed an operating cost savings of $2.1 billion, minus the $500 million in lost revenue, for a net savings of $1.6 billion.
About three weeks later, even before the PRC and participants in the Advisory Opinion process could evaluate the revised cost savings, the Postal Service announced a new plan — a two-phase approach that would break up the consolidations into three periods. Phase 1 begins this summer, with 48 consolidations, and resumes in January, with another 92 consolidations. Phase 2 takes place in 2014, with another 89 consolidations. During phase 1, overnight delivery for Inter-SCF mail will end; with phase 2, nearly all overnight delivery will end.
In announcing the new version of the plan, the Postal Service said that about $1.2 billion of the savings would be realized during the “interim” phase (Phase 1). When asked by the PRC about this, the Postal Service revealed that the $1.2 billion was calculated using the original plan with 260 consolidations, rather than the trimmed down version of 230. That confused matters, of course, and left it to the PRC to figure out how much each phase would save using 230 consolidations. As for how much revenue loss each phase might cause, the Postal Service had no answer for that at all — it never did a market research study on the two-phase approach.
The lack of good cost-savings analysis on phase 1 versus phase 2 is important because it’s possible that the PRC will recommend sticking with something like phase 1. This interim phase is similar to what the Senate bill (S.1789) calls for, and it’s similar to what the PRC’s expert witnesses recommend. But without cost-saving numbers on the interim phase, it’s harder to make the case that this is the way to go.
Savings & Losses
Getting a clear sense of how much the consolidation plan will save is made difficult not only because the Postal Service keeps changing the plan. It’s also a problem because the Postal Service tends to mix savings that are due specifically to the consolidations with savings that can be realized independently of consolidating plants.
As several witnesses have pointed out, much of the cost savings shown in the Post-Implementation Reviews (PIRs) conducted after consolidations take place can be credited to the use of PSEs, declining volumes, increased use of automation and newer, better machines, and other factors that don’t have anything to do with consolidating plants. The Postal Service could realize substantial savings without consolidations.
There’s still another problem with projecting net savings. The change in service standards is definitely going to drive away business, but who knows how much? To answer that question, the Postal Service employed a market research company to interview customers. But there are two versions of the research, with drastically different forecasts. The first one was kept hidden for months because the results argued against the Postal Service’s plan. It projected a contribution loss of $2 billion, whereas the second version projected only a $500 million loss.
The Postal Service says it abandoned the first research study because the “concept statement” used to initiate interviews mentioned closing post offices and ending Saturday delivery, which caused respondents to inflate their estimates of how much less mail they would send. This problem was corrected in the concept statement for the second round of market research, so the Postal Service says it's more reliable and that's what they used in testimony.
That, however, is not an accurate picture of the differences between the two phases of the market research. The concept statement for the phase-1 research may mention closing post offices and ending Saturday delivery, but the interview questions themselves are clearly about how customers will respond to the change in service standards, so it's not likely that respondents were overly influenced by references to these other initiatives.
The revised concept statement used in the phase-2 research doesn't simply avoid mentioning post offices and Saturday delivery. It does something else that's important. It adds passages that make it sound as if the new service standards might actually be beneficial to mailers.
That may explain why the results indicated that some mailers might actually increase their volumes when the mail slows down. The NALC introduced Michael Crew, Professor of Regulatory Economics at Rutgers, as an expert witness, and he pointed out this anomaly as evidence that the study was seriously flawed. Why, after all, would mailers increase volumes when service is going down? It makes no sense.
The Postal Service had the witness who conducted the survey submit a reply. Rebecca Elmore-Yalch explains that mailers care about other things besides speed of delivery, like reliability and convenience. That’s true, of course, but the quantitative research she conducted asked mailers how they would respond to slower delivery, not to a broad range of other changes, such as improvements in reliability.
Another problem with the market research is that it introduces a “probability factor” which allows the Postal Service to decrease the lost volume. Respondents in the survey were asked not only how much less mail they would send if service standards changed, but also, using a probability range called the Juster scale, how likely they would be to send that much less mail. Those two numbers were then multiplied.
If you said you would send 20% less mail, and you were 60% likely to do so, the total volume reduction would be 12%, rather than 20%. Mr. Crew thought this made no sense. The same issue came up with the Five-Day delivery case, for which Mr. Crew testified, and the PRC agreed that it was inappropriate to use the Juster scale this way. It’s not like asking someone the likelihood they would vote for a particular candidate or buy a particular product. The volume reduction number implicitly contains a likelihood estimate: why introduce another? The Postal Service and its market research firm continue to maintain that it’s appropriate to downsize the estimates of lost volume this way. (More details about the two surveys are here.)
Savings scenarios: From rosy to worst case
Given all the difficulties in making projections about cost-savings, it’s dangerous to depend on any single estimate. The Postal Service should be in the best position to make reliable estimates — it has all the data — but it has a history of understating losses and overstating savings. Even Senator Susan Collins has had to chastise the Postal Service for its “fuzzy math": "By cutting service and raising prices and not calculating the resulting disastrous revenue losses," says the Senator, "we have to ask ourselves if the savings estimates under the Postal Service plan are pure fantasy."
Rather than depending solely on the Postal Service’s estimates, perhaps it would be helpful to consider a variety of scenarios, just to get a sense of the range of possibilities. Here’s a table showing a half dozen scenarios, as described below. (By the way, there's a much more sophisticated analysis of various scenarios for processing costs in the testimony of William Weed, but it doesn't encompass all costs or any revenue losses.)
1. The best-case scenario is the one the Postal Service has presented in its testimony. According to the Postal Service’s top-down computer modeling, the full-out plan will save $2.1 billion in operating costs and lose $500 million in contribution losses for a net savings of $1.6 billion. As noted above, when it announced the two-phase approach in May, the Postal Service said phase 1 would save $1.2 billion, but that estimate was made using the original plan, with 260 plants and $2.6 billion is operational savings. Scenario 1 adjusts for 230 plants by distributing the $2.1 billion proportionally.
The Postal Service has not done market research on each phase, so for this scenario, we’ve distributed the $500 million between the two phases in proportion to how the mail would be affected. Since 20 percent of overnight delivery will end with phase 1, we’ve assigned 20 percent of the revenue loss to phase 1, 80 percent to phase 2. (Actually, more than 20 percent of the mail will be affected, but we’re talking customer perception and behavior.)
2. The cost savings for this scenario are projections based on the AMP studies, which add up to about half the Postal Service’s estimate. The AMPs are closer to the ground than the top-down modeling used by the Postal Service, and the unions argue they are much more reliable. There’s been a lot of testimony about why the AMPs don’t add up to the Postal Service top-down estimates, and the Postal Service has repeatedly pointed to the types of savings the AMPs don’t capture. But the gap between the two estimates is huge, and the Postal Service hasn’t provided hard numbers to explain the difference.
The numbers on 203 AMP studies are here; 180 AMP studies are separated into phase 1 and phase 2, here. The estimates in this scenario require extrapolation from these tables, since there are no AMP studies for about 30 of the plants. Note that while the Postal Service says phase 1 will save less than phase 2, the AMPs suggest phase 2 might save more. There are fewer plants in phase 2, but they tend to be larger.
This scenario uses the second round of market research, as in scenario #1.
3. This scenario goes back to the top-down modeling estimate for cost savings, but switches to the first round of market research, the one with the more drastic revenue losses. It distributes those losses in the same way as in scenario #1: 20 percent for phase 1, 80% for phase 2.
4. This worst-case scenario uses the cost savings from the AMP studies and the revenue losses from the first round of market research.
5. This scenario is a modification of #4. It uses the AMP studies, but changes the distribution of the revenue losses. Rather than assuming the losses will be distributed proportionally to how much mail will be impacted by each phase, it splits the losses evenly over the two phases. That’s based on the assumption that mailers will begin mailing less sooner rather than later. It’s probable, in fact, that mailers have already begun mailing less just in anticipation of the change in service standards. This scenario leads to the same bottom line as scenario 4, but it changes the total savings for each phase.
6. This scenario averages the other five, but it’s not a simple mean. Using the Juster scale, it weights each scenario according to our shot-in-the-dark view of the probability that it will occur. A score of 3 means “some possibility,” 4 means “fair possibility,” 6 means “good possibility,” etc. Using this method of averaging gives the more likely scenarios more weight in the average. The results are fairly reasonable: The savings will be somewhere in between the top-down estimate and the AMPs, closer to the AMPs, and the losses will be somewhere in between the two rounds of market research, closer to the first, less manipulated round one.
You can download the spreadsheet by going to Google docs. You can change the Juster numbers or modify the scenarios and see how your own estimates affect the average scenario.
What to make of it all
While the Postal Service would undoubtedly challenge the validity of all but its best-case scenario, each of these scenarios is within the realm of possibility. What’s striking is just how widely they vary. The plan could save $1.6 billion or lose $800 million, or anywhere in between. From best to worst represents a difference of nearly $2.5 billion. The scenarios make it seem as if coming up with projected savings is a big guessing game. And it is.
Something else stands out in this table. Phase 1 will save some money in most scenarios, and phase 2 will save something only in the good scenarios. Phase 2 could easily prove a big loser.
While the Advisory Opinion is weeks away, there’s a good chance that the PRC will recommend something along the lines of Phase 1. That’s pretty much what’s in the Senate bill, and it’s very similar to what the PRC’s expert witnesses, William Weed and Harold Matz, came up with. They describe how eliminating overnight delivery for Inter-SCF mail would expand the DPS window from four to eight hours and make it possible to use far fewer sorting machines. They recommend an "incremental" approach to consolidations rather than the "all-or-nothing" approach of the Postal Service's plan.
Common to all of these versions of network rationalization is the elimination of overnight delivery for Inter-SCF First-Class mail, while preserving it for Intra-SCF mail. Such an approach limits ending overnight delivery to 20 percent of the mail, and it would presumably have a much less negative impact on mail volumes. It would still permit a significant number of consolidations to take place, and it would allow all plants — not just the gaining plants — to expand the hours the machines can operate and thereby realize some savings.
That’s not to say proceeding with the 140 consolidations in phase 1 is a good idea. According to our most likely scenarios, #3 and #6, the net savings for the plan will be about $100 million. That represents a very small savings for the Postal Service, and it comes at great cost to 230 communities that will be losing a postal facility, and to tens of thousands of postal workers and their families.
It may be hard to believe that the plan would end up saving so little, but that very well might happen. Reducing service in order to downsize the network drives business away, and chasing falling revenues can be a losing game. Plus, smaller plants tend to show higher rates of productivity, so it’s very difficult to consolidate a small plant into a large one and come out ahead.
Consolidating plants may or may not save money, but as witness Max Heath of the National Newspaper Association observed in his testimony (p. 3035), the consolidations rarely go seamlessly and there are always delays: “Frankly, in my view I think the Area Mail Processing system has been quite a disaster for most mailers.”
Just ask nicely
While it’s unlikely that the PRC will issue an order telling the Postal Service that it cannot implement the change in service standards on July 1, the Commission could take a different approach. It may not be an appropriate legal response to the APWU's motion for an emergency order, but the PRC could simply ask the Postal Service to wait until the Advisory Opinion is done.
The PRC could explain to the Postal Service why implementing the changes on July 1 is a bad idea — not just because of the effects the changes may have on the Postal Service and its customers, but because of the precedent it sets and the message it sends about the relevancy of the PRC and its Advisory Opinion.
It’s simply not healthy for the Postal Service to implement a change in service without waiting to hear the Advisory Opinion. As the APWU says, that renders the Advisory Opinion “nugatory.” One of the main objectives of an Advisory Opinion is that it provides due process to all those affected by the change in service. Where's the due process if the Postal Service can implement the change before the Opinion is completed?
It’s fine for the Postal Service to tussle with the PRC — the tension between a business and a regulator can be beneficial — but when one government agency tells another government agency it's irrelevant, there’s a problem, especially when one of those agencies is charged with protecting the interests of stakeholders and the general public.
That may be exactly the message the Postal Service wants to send. Perhaps it is insisting on making the changes on July 1 just to show that it can do whatever it wants, and nothing can stop it — not Congress, not the unions, certainly not the PRC. Isn't that what it means to run the Postal Service "like a business"?