Modern Times at the Postal Service: Network Rationalization gives more work to idle machines


Last Tuesday, David Williams, USPS VP of Network Operations and the man running the Network Rationalization plan, was cross-examined before the Postal Regulatory Commission, which is working on its Advisory Opinion about the mail-processing consolidations and the change in service standards they depend on.  Williams explained “the whole concept” behind the plan like this:

“If you go into anyone of our plants today at this hour, we’ve got about 6,000 delivery bar code sorters.  The vast majority of those machines are sitting on the workroom floor and they’re not doing anything.  They’re just sitting there.  And those machines were purchased to process delivery point sequencing volume, and that operation typically starts at say 11:30 at night and goes to 6 or 7 o’clock in the morning, and it’s that idle time in terms of mail processing equipment for delivery point sequencing that we need to address. 

“Those machines take up a lot of square feet, those machines are sitting idle, and the concept of this network change is to get a full operating use of that equipment so instead of a five-to-six-hour operating window on about 6,000 machines we want to get to a 20-to-16 hour operating window….  It’s that lengthening of the operating window of that equipment which drives a significant reduction in the number of pieces of equipment that we need and therefore the number of square feet that we need, which allows us to reduce the fixed overhead that we have in our network.” (PRC webcast at 1:18:00; written transcript, p. 246)

From Williams’ point of view, then, the Postal Service has invested heavily in machines that are not operating enough hours per day.   Relaxing service standards will enable the Postal Service to make better use of the machines and the facilities that house them, and thereby reduce fixed operating costs.  It’s all about machines and square footage and fixed overhead.

In a few weeks, the Postal Service will begin implementing a plan that will cost 35,000 postal jobs and probably just as many non-postal jobs, displace tens of thousands of workers, cause long commutes, break up families, hurt the economies of over 200 communities, slow down the mail, and drive business away — all so that the Postal Service can give more work to idle machines.

The leaders of the Postal Service call this a “rationalization” plan.  They value the efficient use of machines over things like jobs, families, and community.  They think feeding a machine is more important than feeding a family.  


The incredible shrinking cost-saving numbers

Let's say that the Postal Service has a point.  Let’s say that the postal deficit is real (and not a manufactured crisis being caused by payments to the retiree health care fund), and let’s say mail volumes will continue to drop significantly over the coming years (even while the rest of the economy recovers from the recession and resumes growth).  Let’s say that automation in the long run is a good thing (even though tens of thousands of mail-processing jobs may be lost in the short run).   Let’s say that the Postal Service, if it’s to act like a business, should make the most efficient use of its operations (even if that means more work for machines and less work for people).

To justify this line of thinking, the Postal Service should be required to demonstrate that its Network Rationalization plan will in fact save money, and not just a little, but enough to make all the pain worthwhile.  

It's becoming increasingly clear, however, that the Postal Service cannot prove the consolidation plan will realize any significant amount of savings. 

When the Network Rationalization plan was first announced in September, the Postal Service said it would save $3 billion.  When the Postal Service presented its case to the Postal Regulatory Commission in December, they said it would save $2.1 billion ($2.6 billion in lower operating costs, minus $500 million in lost revenue).

Now that the AMP (Area Mail Processing) studies have been completed, we see that 223 of the original 252 facilities have been approved for consolidation.  Nine studies are still underway, but even if most of them are approved, about 10% of the plants will have been removed from the plan.  That presumably means the savings should also be reduced by 10%, so the $2.6 billion becomes $2.34 billion, minus the $500 million in lost revenue, for a net savings of $1.84 billion.

As we learned last week from the secret marketing survey, however, that calculation may seriously underestimate the revenue losses that Network Rationalization will cause.  In order to estimate how much business changing service standards and slowing down the mail would drive away, the Postal Service commissioned a market research study.   When the results came in, the Postal Service didn’t like the numbers, and it commissioned a second round of research to produce more "reliable" data.  (That research, by the way, cost nearly one million dollars.  The rest of the story is here.)

The phase-2 survey became part of the Postal Service’s testimony when it made the Request for an Advisory Opinion with the PRC.  It estimated a $1.3 billion revenue loss, which translates into a $500 million net contribution loss. (The contribution loss reflects the lower costs of processing less mail.)  As noted above, that $500 million is subtracted from the total operating-cost savings of $2.6 billion, for a net savings of $2.1 billion.

The phase-1 survey, kept secret for many months, showed that the revenue loss would be about $5 billion, and contribution loss, $2 billion.  Losses of that magnitude would wipe out nearly all of the $2.34 billion that consolidating about 230 plants would achieve.   No wonder the Postal Service commissioned a second survey.

The Postal Service now says that the phase-1 survey was “seriously flawed” and the reason it shows such large revenue losses is that the “concept statement” used to initiate the interviews made the mistake of mentioning two other big initiatives — closing post offices and eliminating Saturday delivery — and this caused participants to overestimate their reactions to the change in service standards.  

For the phase-2 survey, the Postal Service removed the passage referring to closing post offices and ending Saturday delivery, but it did a lot more.  It completely re-wrote the concept statement and added passages that made it sound as if the Network Rationalization plan would actually improve service in some cases.  It also reconstructed the way questions were posed, elminating, for example, options that allowed customers to indicate how much mail they might divert to email, FedEx, and UPS.  (There's more about the changes in the survey here.)

The phase-1 survey results are not really "flawed."  They just need to be examined in the context of the other two initiatives the concept statement mentions.  

In its case for ending Saturday delivery, the Postal Service hired the same market research firm and determined that it would suffer a revenue loss of $456 million and a net contribution loss of $200 million if it went to five-day delivery. 

The Postal Service didn’t bother doing a market research study about closing post offices.   In its case for closing 3,600 post offices under the Retail Access Optimization Initiative (RAOI), the Postal Service simply assumed that the revenue lost at a closing post office would migrate to another post office or "alternate retail access point."

So closing post offices and ending Saturday delivery don't add up to much in terms of potential revenue losses.  The phase-1 survey shows a net revenue loss of $2 billion.  Of that, only $200 million can be attributed to eliminating Saturday delivery, and none of it, according to the Postal Service’s own reasoning in the RAOI case, is associated with closing post offices.  That means the remaining $1.8 billion in contribution losses must be due to the change in service standards.

Now, it is possible that concern about the “cumulative effect” of all three initiatives caused participants in the survey to overestimate their reactions.  So if we want to focus solely on the effects of the change in service standards, perhaps that $1.8 billion should be reduced a bit.  Let’s say phase-1 showed a $1.5 billion contribution loss due to change in service standards.  That would turn the $2.34 billion in cost savings for 230 consolidatioins into a net savings of about $840 million.  That's a far cry from the $3 billion announced in September or the $2.1 billion that’s the basis of the Network Rationalization case before the PRC.


AMPs vs. Network Rationalization: Go figure

The Postal Service’s cost-savings analysis requires high-level mathematics to compute and understand, and it’s based on a simulation, a hypothetical model built by computers.  The testimony by the George Washington University economist who did the calculations contains equations like this:

It may be difficult to understand the computations, but the bottom line is easy to grasp: The Postal Service plans to eliminate 35,000 positions from the mail-processing network.  That’s over 20% of the 160,000 workers now employed in the network.  Figured at about $75,000 per position, including benefits, that comes to $2.6 billion — the total cost savings the Postal Service claims the Network Rationalization will achieve.  (You can read more about how the 35,000 shake out here.)

The problem, however, is that the AMP studies don’t show anything like that.  Though they obviously have their own flaws, the AMP studies look at operations much closer to the ground and they deal with real-world scenarios.  In February, the Postal Service released the AMP studies on the 183 facilities approved for consolidation.  We added up the numbers and they showed about 12,000 lost positions and a total cost savings of $962 million.

That estimate was confirmed this past week by the Postal Service.  In an interrogatory posed to Dave Williams, the APWU asked what the dollar value of savings estimated for those 183 facilities would be.  “The AMPs estimated approximately $1 billion in savings associated with those facilities studied in the AMP process,” responded Williams.  [APWU/USPS-T1-46]

Asked under cross-examination on Tuesday about the large discrepancy between what the AMP studies say and what the Network Rationalization will supposedly save, Williams explained that the AMP studies are “conservative” in their estimates.  He went into more detail a couple of weeks ago, when he was asked the same question in an interrogatory posed by the APWU.  [APWU/USPS-T1-26]

“There are major areas of savings that the AMP process does not examine and hence were not taken into account,” Williams explained.  The AMPs don’t add up to the projected $2.6 billion in savings for several reasons: The number of consolidations approved for implementation is smaller than originally being considered; the AMP packages don’t estimate increased productivities at facilities other than the gaining facility; the AMP process takes a short-term view of savings while the Network Rationalization plan takes the long view; and there are many areas of savings simply not captured by the AMP process.

Let’s take these explanations one by one.

1. Smaller number of facilities: When the Postal Service originally presented the Network Rationalization plan, it was looking at consolidating 252 facilities, but only 223 were approved.  That means 90% of the plants initially part of the Rationalization have been approved, so this explanation can only account for 10% of the missing cost savings. 

2. Increased productivity: Williams says, “An AMP package does not assess any estimated increase in productivities for any operations that remain behind in the consolidated site.”  He adds that facilities not even included in the AMP study process (a site that neither gained nor lost workload) may experience increases in productivity. 

But “increased productivity” doesn’t save anything in itself.  It just means each worker is processing more mail.  The only way that this productivity can be converted into savings is by eliminating positions, and the AMP studies already include the cost savings from the staffing changes at both the losing and gaining sites.  It’s hard to imagine that there are going to be many positions eliminated at the few facilities not directly involved in the consolidation plan.

3. Short-term versus long-term: Williams offers another reason for the discrepancy between the AMP studies and the total Network Rationalization savings.  The AMP studies offer a “conservative estimate” because they focus on the short-term (one year), while the Network Rationalization focuses on the long term. 

As an example, the Postal Service says that in the short-term, a facility may remain open for local transportation purposes, but in the long-term, it might be closed down completely.  This would mean, then, that the AMP study is showing some positions remaining at the losing facility, but these positions would eventually be eliminated. 

In other words, the AMP studies are not really showing how many positions are ultimately going to be lost.  Many more positions will eventually be eliminated, and that’s where the rest of the savings will come from.

4. Other areas of savings: Williams says there are other areas of savings not included in the AMP process, and he lists them as follows: “The Postal Service does not include the savings associated with premium pay reductions, rents or rental opportunity savings, additional DPS sorting, or service-wide benefits as part of the wage rates utilized in the AMP packages. In addition, the Postal Service has not included the additional air cost into the AMP packages. There are also areas where an estimate of savings is made.  However, the Postal Service is persuaded that the vast majority of these savings have not been captured through the AMP process. Examples include utilities, supplier and contractor costs, parts and supplies, reductions in outgoing secondary sortation and the productivity improvements associated with the Upgraded Flats Sorting Machine 100 and Carrier Sequence Barcode Sorter consolidation.”

Sort through all that if you like, but there’s nothing in the explanation to account for over a billion dollars in savings not captured by the AMP studies. 

Williams can’t say where all the savings are coming from.  He can only make the vague remark, “the Postal Service is persuaded that the vast majority of these savings have not been captured through the AMP process.”  The Postal Service may be persuaded, but that’s not the kind of reasoning that’s likely to persuade others.


AMPs versus PIRs: "Concurrent savings" make the difference

As evidence that AMP studies provide “conservative estimates” of cost savings, the Postal Service likes to cite the Post-Implementation Reviews (PIRs) conducted six months and then one year after a consolidation is implemented.  These PIRs consistently show that the actual cost savings turns out to be far more — three, four, even five times greater — than the AMP study predicted.

In his testimony, Williams illustrated this phenomenon by pointing to 24 final PIRs that showed an actual savings of $345.3 million, as opposed to the AMP estimate of $71.6 million.

We can see the same thing is this sampling of AMPs compared to first and second PIRs.  The following table shows the cost savings and number of postions lost for 20 randomly selected facilities, according to the AMP study, the first PIR, and the second PIR.

AMP pos.
PIR 1 pos.
PIR 2 pos.

As we can see in the table, the final PIR shows an average annual savings of $11 million — almost four times the average of $3 million predicted by the AMP studies.

As yet further evidence along these same lines, the Postal Service has been citing an OIG audit (published Jan. 9, 2012) that also showed PIR savings far greater than AMP savings. 

This OIG report came up on Tuesday when Williams was cross-examined by the attorney for the APWU.  The attorney wanted to call Williams’ attention to the OIG’s explanation for the discrepancy between AMP savings and PIR savings: “This variance occurred," says the OIG, "because other concurrent initiatives were added to the AMP consolidation savings.  They included the concurrent savings in the PIRs, but these savings were not directly related to the AMPs.  We anticipate conducting an audit of the PIR process and related savings calculations in the future.” 

These “concurrent savings,” as we learn in a footnote in the OIG report, include “workhour savings from other mail processing initiatives and the retirement incentive in 2009.”  Asked if he agreed with this analysis, Williams replied yes.  (written transcript, p. 228)

One of the other concurrent initiatives is the 2010 APWU national agreement that made it possible for the Postal Service to use more temporary contract workers.  In his cross-examination, the APWU attorney talked about this with Williams.  That agreement, which went into effect during the summer of 2011, said that up to 20% of the workforce could be temporary workers (PSEs, or Postal Support Employees), and in some contexts, even more than 20% would be possible.  The Postmaster General said the plan would save $3.7 billion, and a large part of that savings are captured in the PIRs.  But those savings have nothing to do with the consolidations per se. 

We’re learning pretty much the same thing with the Network Rationalization plan as a whole.  As interrogatories and cross-examinations have revealed, much of the savings supposedly associated with the consolidations would be achieved by aspects of the plan that have nothing to do with benefits from the consolidations themselves — like using more contract workers who receive lower wages and fewer benefits than career employees.


Coming ahead: Revised testimony and an independent analysis

The table comparing AMP studies to PIRs shows something besides cost savings.  It also shows that the AMP studies seriously underestimated the number of positions that would eventually be eliminated from the mail-processing network.  That’s why there’s such a big discrepancy in the cost savings.

The average AMP shows 40 lost positions, but a year later, when the second PIR was done, an average of 151 positions had been eliminated — some as a result of the consolidation, but many as a consequence of other initiatives, like retirement incentives and using more contract workers.

The 183 approved AMPs show where 12,000 positions will be eliminated.  If 230 plants are ultimately consolidated, that would account for about 15,000 lost positions.  Somehow, an additional 20,000 positions will need to be eliminated to get to the 35,000 called for by the Network Rationalization plan.  AMP studies, in other words, identify only about 40% of the positions that will be lost.

This means that when the Postal Service held public meetings across the country and gave those PowerPoint presentations showing how many positions would be eliminated at the “losing” facility, it did not tell the whole story.

Just to take one example.  When the Postal Service gave its presentation about consolidating the plant in Buffalo, NY, where there are currently 734 employees, it said that there would be a net loss of 199 positions.  Five weeks later, the AMP study said there would be a net loss of 216 positions. 

But for the Postal Service to realize $2.6 billion in cost savings and eliminate 35,000 positions, Buffalo would actually need to lose more like 500 positions.  And a year from the implementation date, that’s probably what the PIR will show.

It’s not clear why the Postal Service did not explain all this to people at those public hearings.  Perhaps the AMP studies only show how many positions the consolidation itself would eliminate, and not those related to other concurrent initiatives.  Perhaps the Postal Service just wanted to be “conservative” in its estimates.  Or perhaps the Postal Service feared that telling people the whole story would provoke too much alarm, anger, and protest.

Whatever the explanation, low balling the number of lost positions makes those meetings a sham.  Postal employees, their families, and the communities that stand to lose a processing facility deserved better.

Hopefully the PRC will be able to do better than the Postal Service in providing an objective analysis of how much the Network Rationalization plan will actually save.  The Postal Service witnesses are now responding to interrogatories asking them to submit revised testimony and new calculations based on the 183 AMP studies.  

The PRC has also hired its own consulltants "to provide an independent evaluation of cost and staffing changes resulting from network consolidation." The results of that work should be available by April 23.

Not that it will mean anything to the Postal Service.  In his testimony last week, Dave Williams said the Postal Service would issue a Final Rule on the change in service standards in mid-April and immediatley begin the process of helping customers to adjust  to the new system.  The actual consolidations will begin early this the summer.  The Advisory Opinion won't be ready until September.   By then, the Network Rationalization plan will be a reality.

(Photo credits: worker operating a bar code sorter; the Incredible Shrinking Man; New Yorker cover; Charlie Chaplin caught in the gears of Modern Times; productivity cartoonworker at the Buffalo plant)