President Trump is continuing his Twitter attack on Amazon over its deal with the Postal Service. Yesterday Trump tweeted, “Only fools, or worse, are saying that our money losing Post Office makes money with Amazon. THEY LOSE A FORTUNE, and this will be changed. Also, our fully tax paying retailers are closing stores all over the country… not a level playing field!”
A couple of days ago, his tweets were more specific: “While we are on the subject, it is reported that the U.S. Post Office will lose $1.50 on average for each package it delivers for Amazon. That amounts to Billions of Dollars…. If the P.O. ‘increased its parcel rates, Amazon’s shipping costs would rise by $2.6 Billion.’ This Post Office scam must stop.”
The attacks were in the same vein as his earlier tweets back in December. They are apparently based on an op-ed in the Wall Street Journal by Josh Sandbulte published last July. Sandbulte claimed that each Amazon parcel was getting a $1.46 subsidy. “It’s like a gift card from Uncle Sam,” said the WSJ headline.
As Vox has noted, Sandbulte is a money manager who works for a firm that owns FedEx stock, but that may or not be relevant. Anyone invested in mutual funds probably owns some FedEx. In any case, he didn’t invent the subsidy idea. It came from an analysis done by Citigroup published in April 2017.
The thesis of the Citigroup report is that taxpayers are essentially paying for the free shipping offered by Amazon. As the Citi analysts write, “In this note, we examine the true profitability of the Post Office and show that by charging below market rates on parcel volume (mainly eCommerce) the Post Office has essentially turned free shipping into a future tax payers’ burden.”
The Citigroup report, it should be noted, is intended to give advice to investors in the stock market. It claims that “a day of reckoning” is coming when the Postal Service will have to implement a significant increase in shipping rates, and this will provide a large “revenue opportunity” for the Postal Service’s competitors, FedEx and UPS — something on the order of $15 to $19 billion a year in additional revenue. This, they say, “supports upside for both stocks.”
The Citigroup report is somewhat less bullish on Amazon because it will have to bear the brunt of rate increases by the Postal Service and also FedEx and UPS, who will be in a better position to raise rates themselves. According to the analysts’ “worst case scenario,” Amazon will have to pay $2.6 billion a year in additional shipping costs.
As a closer look at the Citigroup report reveals, the case for a huge Postal Service rate hike on parcels is seriously flawed, and the report provides no evidence for Trump’s tweets that the Post Office is losing a fortune on the Amazon deal.
Before we get to the Citigroup report, it will be helpful to lay out a few basic facts about the types of U.S. mail, the way postal accounting works, and the particular service Amazon is using. If you’re familiar with all this, you can cut to the chase and go to the section below on the Citigroup analysis.
A Postal Primer
In 2006, the Postal Accountability and Enhancement Act (PAEA) divided postal products and services into two categories, Market Dominant and Competitive.
Market Dominant products and services are those in which the Postal Service “dominates the marketplace” as a result of its two monopolies — the letter monopoly, which gives the Postal Service a monopoly on non-urgent First Class mail, and the mailbox monopoly, which gives the Postal Service exclusive right to put mail in mailboxes. Market Dominant includes First Class, Standard (mostly ad mail), and Periodicals, as well as certain types of international mail. According to the USPS 2017 10-K report, Market Dominant accounts for about 70 percent ($50 billion) of total revenues ($70 billion) and about 95 percent of total volumes (150 billion pieces)
Competitive mail includes shipping services like Priority Mail, parcels, and some other types of international mail. Its name derives from the fact that there are competitors in the private sector for these types of products. Competitive products account for 29 percent ($20 billion) of total USPS revenues and about 5 percent of volumes.
The rates for Market Dominant mail are constrained by the “price cap” regulation, which limits rate increases in a class of mail to the Consumer Price Index. The rates on Competitive products are essentially constrained by the marketplace, but the Postal Service is free to set prices as it sees fit, subject to approval by the Postal Regulatory Commission, so long as the products cover their costs, aren’t cross-subsidized by Market Dominant products, and make an appropriate contribution to institutional costs. In other words, the rates can’t be too low.
The phrases “covering their costs” and “contribution to institutional costs” refer to the way the Postal Service analyzes the costs of each product and service. As in every business, for every product there are two types of costs, variable and fixed.
The variable costs are those associated directly and indirectly with a specific product or service, and they change relative to volume. The Postal Service calls these attributable costs.
The fixed costs include wages, rent on leased post offices, and all the other overhead expenses that stay the same regardless of how much volume the Postal Service is handling. The Postal Service calls these institutional costs.
For each product and service, the Postal Service figures out the variable costs that can be attributed to that product; whatever revenue is brought in beyond that is considered contribution to the institutional costs.
The cost coverage for a product or service is determined by dividing the unit’s revenue by the attributable cost. A cost coverage of 100 percent means that the product has covered all of its attributable costs but contributed nothing to institutional costs. Ideally, a product will therefore have a cost coverage greater than 100 percent, so it can contribute something to the Postal Service’s fixed overhead costs.
Finally, there are the Negotiated Service Agreements, i.e., contracts between the Postal Service and individual mailers that provide customized pricing and other arrangements. For business reasons, the details of these deals are withheld from the public, but each NSA must be approved by the PRC before it goes into effect, and then again annually as part of a review to ensure that the deal is still in compliance with the law.
There are several hundred NSAs in effect right now, including 846 Competitive domestic agreements, ranging widely in size and scope. At least one, perhaps several of them, cover the deals between Amazon and the Postal Service.
The PRC’s Annual Compliance Review
The PRC conducts an Annual Compliance Determination Review (ACDR) to ensure sure all the products and services provided by the Postal Service are in compliance with the laws governing postal matters. Generally speaking, the review determines if the costs incurred by each product and service are covered by the revenue generated by that product or service.
The compliance review thus looks at how each type of mail — including each NSA — is doing with respect to cost coverage, i.e., the extent to which it is covering attributable costs and how much it is contributing (or failing to contribute) to institutional costs.
The Commission also examines how much Competitive mail as a whole is contributing to institutional costs to ensure that the Postal Service isn’t using the products over which it has a monopoly to unfairly subsidize products for which there’s competition from the private sector.
As it happens, the PRC issued the 2017 Annual Compliance Determination Report just last week. Noting that the law requires “each Competitive domestic NSA product to cover its attributable cost,” the Commission found that “all but four Competitive domestic NSAs covered their attributable costs and complied with this statutory requirement” (p. 84). Three of these NSAs expired or were terminated, and the fourth (a Priority Mail Contract that is almost definitely not the Amazon NSA) is being monitored pending reevaluation.
The PRC’s compliance report means that the PRC has reviewed the Amazon contract or contracts and determined that they are indeed covering their attributable costs. They are not losing money for the Postal Service.
The Amazon NSAs and Parcel Select
While we know very little about the details of Amazon’s contract or contracts with the Postal Service, we do know that most of the parcels delivered by the Postal Service for Amazon fall under the category called Parcel Select.
Back in 2013, when the Postal Service announced that it was doing Sunday delivery for Amazon Prime, we were able to locate the NSA in PRC docket CP2014-1 and confirm that it was a Parcel Select product. You can see the agreement here, but it’s almost entirely redacted.
According to the USPS description, “Parcel Select service provides very competitive pricing. It is often used by other private parcel companies to complete delivery of the ‘last mile’ for their shipments — particularly for deliveries in non-metropolitan or rural areas because the Postal Service is the only carrier that offers delivery to every door 6 days a week.”
The biggest users of Parcel Select are Amazon, FedEx, and UPS. They’ve determined that using the Postal Service for the “last mile” (from the post office to the home or business) is much more cost-effective than trying to deliver to millions of addresses themselves.
In fiscal year 2016, according to an article in DC Velocity, about 2.5 billion packages moved under Parcel Select. Amazon was responsible for about 1 billion packages; FedEx (through its “SmartPost” product) used USPS for 600 million pieces, and UPS (through “SurePost”) had USPS deliver about 275 million pieces of Parcel Select. The balance came from several parcel consolidators that aggregate packages from multiple smaller shippers.
Parcel Select generally takes two to nine days, but the big mailers, consolidators, and private shippers prepare and presort the parcels and deliver them to the Destination Delivery Unit (DDU) — usually your local post office — or a regional processing facility, thus saving a lot of time and expense.
If the shippers get the parcels to the DDU by a certain time — Early Bird DDU — the Postal Service can often provide same-day delivery. Regular DDU — dropping off the parcels after the carriers have left the facility — usually means next working day delivery.
The Postal Service’s ability to deliver packages in two days or less has been an important factor in growing its parcel business, and it points to the synergy between the Postal Service and its larger customers who help shuffle packages to the right place in the network, labeled and sorted in the right way.
Because the users of Parcel Select do some of the work themselves, they’re entitled to “workshare” discounts based on the costs the Postal Service is avoiding. These discounts are arranged through the nonpublic NSAs, so we don’t know how much Amazon is paying, but the public pricing for Parcel Select is shown here. As you can see, the prices start at $2.85 for a parcel weighing a pound or more, dropped off at a DDU. Prices go up from there, depending on the weight and how close to the destination it’s dropped off. For parcels of less than a pound, prices start at $1.42.
Due to its volume discount, Amazon is paying much less than these published prices. According to a Bloomberg article, David Vernon, an analyst at Bernstein Research who tracks the shipping industry, estimates that Amazon is probably paying, on average, about $2 per parcel. That estimate can also be derived from this USPS financial report. It shows that in 2017 Parcel Select brought in $5.66 billion on 2.8 billion pieces, an average of $2 per piece. Amazon’s deal is probably comparable to the agreement the Postal Service has with UPS and FedEx to deliver the last mile for them.
As the chart for Parcel Select prices shows, there’s a big range in pricing, and the pricing on Amazon packages probably varies significantly, depending on geography, time of year (holidays), and who’s doing the delivery (e.g., union workers or lesser-paid non-union workers like City Carrier Assistants).
If Amazon is now sending about a billion parcels through the Postal Service, and if the average is about $2 per piece, the relationship is bringing the Postal Service about $2 billion a year. That strictly a ballpark guess — the Citigroup report puts the estimate at $3 billion a year — but it shows that Amazon has become a big part of the Postal Service’s business.
The Citigroup Analysis
Citigroup is bullish on UPS and FedEx because, it says, they will enjoy “a better pricing environment in the future” since the Postal Service is soon going to need “to raise rates meaningfully to capture its true costs.” (Of course, if USPS shipping rates were to go up, UPS and FedEx would need to pay more for “last mile” deliveries by the Postal Service, but with higher rates or a bigger market share for their own deliveries, the net result would be to their advantage.)
If these “true costs” are not eventually covered by higher rates, says Citigroup, the taxpayer is going to be on hook for the Postal Service’s financial problems and, by extension, for the “free shipping” Amazon is offering.
In order to identify these “true costs,” the Citi analysts outline two scenarios that illustrate how the Postal Service is charging below market prices for shipping services.
Scenario #1: Paying off the prefunding
In the first scenario, Citigroup focuses on the so-called “prefunding mandate” established in 2006 by Congress with the Postal Accountability and Enhancement Act. Under PAEA, the Postal Service was required to pay $5.5 billion a year, for ten years, into a fund to cover the costs of retiree health care for decades down the road.
How this Retiree Healthcare Benefit Fund (RHBF) came to be is a long story, as discussed in this previous post. Much of the story remains shrouded in mystery, though, because the U.S. Senate decided to withhold publication of the Committee report that described its rationales for the Act’s provisions. But the long and short of it is that the decision to mandate annual contributions of $5.5 billion turned out to be a disaster.
In 2006, the Postal Service was doing great and coming up with over $5 billion a year didn’t seem like a big problem, but then the Great Recession hit, postal revenues took a dive, and it wasn’t long before the Postal Service had to start borrowing from the Treasury to make the payments. When it reached the borrowing limit, the Postal Service just started defaulting on the payments.
This prefunding mandate is responsible for almost all the deficits you read about in the news every time a USPS fiscal report is published. Were it not for the prefunding requirement, the Postal Service would be doing just fine over the past few years, usually breaking even or showing a relatively small profit or loss. According to the FY 2017 10-K report (p. 17), if one looks solely at “controllable” costs and income (i.e., excluding the RHBF payments and other actuarial issues), the Postal Service lost $814 million in 2017 and made a profit of $610 million in 2016 and $1.188 billion in 2015.
In any case, for their first scenario, the Citigroup analysis assumes that “a day of reckoning is approaching” when the Postal Service “must resume making annual prefunding payments to the PSRHBF as well as meeting the amortization schedule on the accumulated $33.9B it owes, as laid out in its FY16 annual report to Congress.”
In addition, says Citi, the Postal Service will need to contend with rising operating costs and declining Market Dominant revenues. According to Citigroup, in fiscal year 2017 the Postal Service will therefore need to bring in an additional $8.3 billion, growing to $9.6 billion in 2019.
For some reason, the Citi analysts then assume that parcel mail would need to cover all of this additional revenue. This, they estimate, would require a 50 percent hike in the price of an average parcel, an increase from $3.50 to $5.25 — or about $1.75 per piece.
The problems with Scenario #1
This scenario is seriously flawed for several reasons.
First of all, there’s no reason to believe that such “a day of reckoning” is coming. If anything, we’re heading for a day when Congress recognizes the mess it created with the prefunding mandate and does something to fix it.
As reported by Government Executive, there’s currently a bill in Congress (the text isn’t available yet) under which “Outstanding payments [to the PSRHBF] would be wiped clean and USPS would make actuarial payments toward the remaining liabilities over the next 40 years.”
A 40-year amortization schedule was actually being recommended back in 2006, but the Bush administration pushed for the 10-year schedule that caused all the problems. Such a relaxed schedule would mean annual payments on the order of $1 billion rather than $6 billion, and essentially eliminate the premise on which Citi’s Scenario #1 is based.
A second problem with this scenario is that it assumes the entire $8 billion in increased revenue would have to come from shipping services. But if the Postal Service had to bring in an additional $8.3 billion, it would surely spread the hurt out over all types of mail and not put it all on parcel shippers.
Annual revenues in 2017 were about $70 billion. To bring in another $8.3 billion would require an across-the-board rate increase of about 12 percent. Since such an increase exceeds the price cap on Market Dominant products, the Postal Service would need to request what’s called an exigent rate increase from the PRC due to the “extraordinary” circumstances of being required to pay off the prefunding mandate. That’s happened once before. In 2013, the PRC granted a 4.3 percent increase for about two years to help the Postal Service make up losses due to the Great Recession.
An across-the-board increase of 12 percent would be a serious matter, and mailers, large and small, would go ballistic. One of the effects would be that the rates Citigroup pays to send credit card bills and solicitations would also go up, which would cost the corporation millions of dollars. Maybe that’s why the Citi analysts assume that parcels would need to bear the entire burden of producing the additional revenue.
Anyway, a 12 percent increase on an average piece of Parcel Select ($2) might mean about 24 cents more on a typical Amazon package — not the kind of thing to generate headlines or presidential tweets.
Citigroup actually acknowledges the problems with this scenario later in its report. It notes that the Postal Service could increase postage on other types of mail, Congress might permit the partial reinstatement of the exigent surcharges previously approved by the PRC, and Congress could also provide relief from statutory obligations to prefund certain benefits obligations. Plus, if the Postal Service were to increase parcel pricing significantly, it would do so gradually.
Citigroup calls these “caveats” but they are really much more. Each of these possibilities is much more likely to happen than a 50 percent increase on parcel rates just to cover the costs of retiree health care. Scenario #1 is never going to happen.
Scenario #2: Changing the contribution to institutional costs
In the second scenario, Citigroup notes that when PAEA became law in 2006, Competitive products were assigned a 5.5% share of institutional costs. At the time, parcels represented a relatively small part of the Postal Service’s business, but with the boom in e-Commerce, parcels have become a much larger part. Competitive products now account for about a third of the Postal Service’s revenues.
Over the past few years, UPS has filed several analytic studies and legal briefs with the PRC arguing that Competitive products are not paying their “appropriate share” of institutional costs. According to UPS, the appropriate share would be 24.6% (or even more), as opposed to the 5.5% determined over a decade ago. UPS’ motives are not ambiguous. An increase in contribution to institutional costs would mean an increase in the Postal Service’s parcel prices, which would improve UPS’ position in the market — it could raise its own rates and/or grab a larger piece of the pie.
For its Scenario #2, the Citigroup analysts “work through the impact if the UPS’s suggested 24.6% estimate is instituted.” To do this, they “calculate competitive products’ share at both the current understated 5.5% rate and the updated 24.6% rate proposed by UPS. The difference between the two rates represents the incremental institutional costs that need to be allocated to competitive products.” Here’s how that works out.
For 2017, total USPS operating expenses (i.e., not including the retiree health care expense) were about $70 billion, about half of which were categorized as variable (attributable) costs and half were fixed (institutional) costs. Figured at the 5.5 percent set by PAEA and Citigroup’s estimate of institutional costs (52 percent of total costs), the appropriate share for Competitive products would come to about $2 billion (that’s $70 billion x 52% x 5.5%).
Citi then looks at what would happen if Competitive products had to cover their appropriate share of the fixed costs by using the 24.5 percent figure. The result would be that Competitive products would have to contribute about $9 billion (i.e., $70 billion x 52% x 24.3%).
According to Scenario #2, Competitive products would therefore need to pay $7 billion more for Institutional costs. That would require a rate increase on shipping services comparable to the one determined by Scenario #1 — almost 50 percent. Average parcel rates would go from $3.51 to $4.97, an increase of $1.46. That is the origin of the claim made in the WSJ article, and it’s the basis of Trump’s tweets.
The fatal flaw in Scenario #2
The fatal flaw in Scenario #2 is the assumption that competitive products are currently contributing only 5.5 percent to institutional costs. This 5.5 percent share is actually just the floor set after PAEA was enacted back in 2006, when the PRC determined (as stated in the ACDR) “that if Competitive products contribute at least 5.5 percent toward the Postal Service’s total institutional costs, then, as a whole, they will cover an appropriate share of the Postal Service’s total institutional costs.” (p.92)
But competitive products contribute much more than 5.5 percent. According to the 2017 Compliance report, “In FY 2017, the total institutional costs of the Postal Service were $29.700 billion. To comply with 39 U.S.C. § 3633(a)(3) for FY 2017, Competitive products must have contributed at least $1.634 billion toward the Postal Service’s institutional costs. In FY 2017, the total Competitive products contribution was $6.806 billion (approximately 23 percent), which exceeds the minimum contribution requirement.” (italics added, p. 92).
While the numbers differ slightly from Citigroup’s estimates for 2017, the point is clear enough. Competitive products are not contributing 5.5 percent to institutional costs; they are contributing more than four times that amount, and they are contributing almost exactly what Citigroup and UPS say they should be contributing, about 23 or 24 percent.
The Citigroup report observes in a footnote that UPS has subsequently argued that competitive products’ appropriate share should be even higher, like 29 percent. Even at that level, the impact on the prices of Competitive products would be modest compared to what Scenario #2 envisions.
For the past couple of years, there’s been a lively debate at the PRC about how much Competitive products should contribute to institutional costs. (See, for example, PRC Docket No. RM2017-1 on “Institutional Cost Contribution Requirement for Competitive Products.”) While UPS has argued for a higher minimum floor, others have argued that there doesn’t need to be a minimum at all. In its comments on the issue (January 23, 2017), for example, Amazon states this:
“Competitive products are covering almost are four times the share of the Postal Service’s institutional costs that the Commission mandated five years ago. The contribution and cost coverage of competitive products are now far too high to support any credible allegation that a binding minimum contribution requirement is needed to preserve a ‘level playing field’ for the Postal Service’s competitors, let alone to avoid cross-subsidy, predatory pricing, or any other alleged form of unfair price competition, or provide a margin of safety.”
Whatever the Commission ultimately decides to do about the minimum contribution level, the key point here is that Citigroup’s Scenario #2 is based on the mistaken assumption that Competitive products are currently covering only 5.5 percent of the Postal Service’s institutional costs, when in fact they are contributing more than four times that amount. There’s no shortfall of $7 billion that needs to be made up by a big increase in parcel prices.
Considering how sophisticated the Citigroup analysts are when it comes to making projections, it’s perplexing that they missed this basic fact.
The $2.6 billion in additional costs for Amazon
The other number that Trump has tweeted also comes from the Citigroup report. In this tweet, Trump says, “If the P.O. ‘increased its parcel rates, Amazon’s shipping costs would rise by $2.6 Billion.’ This Post Office scam must stop. Amazon must pay real costs (and taxes) now!”
Here’s where that number $2.6 billion comes from.
The Citi analysts outline a “worst case scenario” in which Amazon must bear a 50 percent increase in its costs with the Postal Service, plus a 20 percent price jump with FedEx and UPS, which will be able to raise their prices thanks to the price hike by their competitor.
The analysis gets pretty complicated, but the bottom line, says Citigroup, is that Amazon would incur an additional $2.6 billion in shipping costs.
Trump’s tweet makes it sound as if the Postal Service itself is missing out on $2.6 billion a year in revenue from Amazon, but that’s not the case, and it’s really misleading, to say the least, to make such a claim.
Cost coverage on the Amazon deal
If you want to get even deeper into the weeds on all this, let’s take a look at the cost coverage for the type of mail Amazon is using.
While the Amazon NSAs are nonpublic, one can get a good sense of just how successfully the products are covering costs by looking at a couple of USPS financial reports.
This USPS report shows that in 2017 Parcel Select brought in $5.67 billion on 2.8 billion pieces. This second USPS report shows the cost coverage for each type of mail. It doesn’t mention Parcel Select, but it provides data for Total Ground mail, which includes Parcel Select, Standard Post, and Parcel Return Service. It shows that in 2017, Ground brought in about $6.2 billion in revenue on about 2.88 billion pieces. Parcel Select thus accounts for almost all Ground mail, so the data on cost coverage for Ground can give us a ballpark view of the cost coverage for Parcel Select and, by extension, the Amazon NSAs.
The second report shows that the average Ground piece brought in $2.148 in revenue. The variable cost was $1.221; the rest — $0.927 — was contribution to institutional costs. The cost coverage was 175.86 percent.
If you compare this to the cost coverage for other types of mail, you’ll see it’s typical. First Class letters, for example, have a cost coverage of 164 percent, and First Class mail overall has a cost overage of 210 percent.
A few types of mail have a cost coverage below 100 percent. Standard Mail flats, for example, have a cost coverage of about 74 percent, and Periodicals, about 70 percent. While these types of mail do not cover their attributable costs, there are legal and policy reasons why some of that failure may be considered excusable. In any case, the issue is regularly discussed in the PRC’s compliance reviews, and all the stakeholders are well aware of the issues.
If the cost coverage on Amazon’s Parcel Select is anything like 175 percent, there’s clearly no cause for concern that the Postal Service or the taxpayer is somehow subsidizing Amazon delivery.
Perhaps an upside
There are many reasons to criticize Amazon — its success is hurting brick-and-mortar businesses, it exploits its workers with low pay and poor working conditions, and it’s “no fan of labor unions” — and there are plenty of reasons for criticizing the Postal Service as well, like the way it underpays its non-union employees and abuses the emergency suspension provision to close post offices without due process.
There are also reasons to raise questions about the relationship between Amazon and the Postal Service, as we’ve done on this website since the deal to deliver for Amazon on Sundays was first announced in 2013. STPO contributor Mark Jamison’s posts on Amazon and the lack of transparency in the NSAs are still worth reading. (Our Amazon posts are archived here.)
Criticisms aside, there’s little reason to believe that the Postal Service is significantly undercharging Amazon for delivering its parcels. The cost coverage data show that parcels are more than covering both their attributable and institutional costs, the PRC has reviewed the NSAs and found them in compliance with the relevant statutes, and there are serious flaws with the Citi analysis on which the claims for underpaying are based. There’s absolutely no evidence that the Postal Service is losing a fortune on the Amazon deal.
While Trump’s tweets on the post office are bogus, perhaps we should at least be thankful to the President for raising some important questions about the Postal Service. Perhaps his tweets will lead to some thoughtful Congressional hearings on postal topics. Perhaps the pension and health care cost issues will finally get clarified. Perhaps the Postal Service will address public concerns by providing more transparency about the deals it strikes with companies like Amazon. Perhaps, perhaps.
(For previous posts about Amazon, start here.)
(Photo: Wall Street Journal)