One of the central arguments in Trump's attacks on Amazon is all wrong - and it's Citigroup's fault

Nicholas Kamm/AFP/Getty Images

• President Donald Trump has been relying on a 2017 Citigroup equity research note to justify his claim the post office loses about $1.50 per Amazon package it delivers.
• But Citigroup greatly underestimated how lucrative services like package delivery are for the post office.
• Using the right numbers, Citi would have said rates need to go up by 11 cents, not $1.46.
• Beyond the math error, Citi's pricing ideas - which were cribbed from UPS - have been rejected by the Postal Regulatory Commission, which says USPS package rates are already high enough.

On Monday, Donald Trump tweeted that the post office loses $1.50 for every package it delivers for Amazon. On Tuesday, he said it was $1.47.

Where did he get these figures? He didn't make them up. He's relying on an equity research report from Citigroup, which doesn't quite say what the president says, but it says something pretty close.

The Citi note, from last April, proposes that the post office might need to raise the average price per piece for all of its "competitive products" by $1.46 - that is, not just deliveries for Amazon, and not just parcels, but also priority and express mail.

But Trump's garbling of the precise figure isn't the important thing.

The important thing is that the Citi note itself is wrong. The bank made a key math error in coming up with the figure the president has seized on.

Essentially, Citi assumed competitive products were less lucrative for the post office than they really are, and so they greatly overestimated how much prices would have to rise to meet a specific revenue benchmark. If Citi had the right numbers, they would have called for an 11-cent increase in price-per-piece for competitive products, which is a lot less than $1.46.

Plus, Citi's underlying analysis of how the post office ought to price package delivery makes little sense, even if you correct the math. It leads to a useful talking point for a president with a vendetta against Amazon, but it's not good business analysis.

In what industry would you let your competitors tell you how to price your products?

Robert Couse-Baker/flickr




The Citi note was released shortly after a Postal Regulatory Commission rulemaking process on postal prices began, and it closely echoes arguments UPS made in those proceedings.

UPS called for allocating more of the post office's institutional costs away from first-class mail and toward competitive services like package delivery, which would have had the effect of forcing the post office to raise prices on packages.

Obviously, UPS has an interest in getting the post office to raise prices, so it can raise its own prices. Citi's note estimates UPS's earnings per share would rise nearly 50% if the post office followed Citi's guidance on parcel pricing.

But contrary to Citi's argument that current package pricing is "uneconomic," the commission rejected the pricing approach advocated by UPS this February, saying the UPS proposal is "inherently arbitrary" and "fails to maximize economic efficiency because it is not based on marginal cost and does not yield prices reflecting market demand."

The flaw here - both for UPS and Citi's related analysis - comes from the fact that the post office is in a lot of businesses at once.

Yes, it competes with FedEx and UPS on packages and express letters. It also has a legislative mandate to deliver first-class mail all over the country, six days a week and for a fixed price, which is often a money-losing endeavor. And it has enormous legacy costs from a time when it used to employ a lot more people and its real estate needs were quite different than they are today.

The UPS proposal, as endorsed in the Citi note, would have obligated the post office to cover more and more of those legacy costs by raising prices on packages and other products where it competes with UPS.

Such a move would have been good for UPS, but it would have been bad for consumers, bad for shippers, and quite possibly bad for the post office and taxpayers, if the higher prices caused package volumes to decline.

Instead of going this direction, the PRC has proposed a rule that gives the post office flexibility and does not require any increase in prices for competitive products - the post office could even cut prices and remain in compliance.

UPS, for its part, still objects to the direction PRC has taken.

"UPS believes that the Postal Service's explosive package delivery growth is driving additional costs that are not being accounted for in the Postal Regulatory Commission's existing and proposed formulas," said Kara Ross, a spokeswoman for UPS, in an email to Business Insider.

How Citi got the math wrong

Associated Press/Evan Vucci

President Donald Trump speaks during a roundtable discussion on tax policy, Thursday, April 5, 2018, in White Sulphur Springs, W.Va.


You can break most of the lines of business at the US Postal Service into two categories.

There are "market-dominant products," like first-class mail, where the post office has a legal monopoly and a legal obligation to provide universal service. And there are the competitive products, like parcel delivery and priority mail, where the post office competes with private companies.

The Postal Accountability and Enhancement Act of 2006 requires that competitive products be, well, competitive. The post office is required to price them so they bring in at least enough revenue to cover the costs directly associated with providing them, plus an "appropriate share" of costs that can't be readily allocated, such as certain real estate expenses or the salary of the Postmaster General.

For the last 10 years, this minimum "appropriate share" was set at 5.5%. But package delivery has been growing as a share of USPS business, and UPS has been urging the PRC to raise this share.

The Citi note proposed an increase in the appropriate share from 5.5% to 24.6%, relying on prior proposals from UPS. By Citi's estimate, this change would have required the post office to cover an additional $7 billion of operating expenses through higher prices on competitive products, which would have required charging an extra $1.46 per piece delivered.

The big problem in Citi's analysis is associated with the phrase I italicized a few paragraphs above: "at least."

The 5.5% regulation is a floor. And lately, the post office has been far above the floor.

You can see it here in this table - Citi assumed revenues from competitive products were only covering the regulatory minimum of 5.5% of non-allocated costs ($2 billion) as of 2017. But PRC figures show they actually covered 23.2% of those costs ($8.5 billion).

As such, raising the floor by imposing what Citi calls a "corrected" cost methodology would only have required the post office to come up with an additional $500 million a year in profits from competitive products, not $7 billion. USPS could have met that goal by raising competitive product prices by 11 cents per piece, not $1.46.


In fairness to Citi, 2017 financial data from the postal service wasn't available when they made projections for 2017 in this research note, because 2017 wasn't over.

But figures from 2015 and 2016 were available, showing the postal service had achieved ratios of 13.3% and 16.5% in those two years, respectively. Using an average of those figures would have cut Citi's proposed price increase roughly in half.

I asked Citi's lead analyst on the note, Christian Wetherbee, whether it was simply a mistake that they hadn't used figures on the post office's actual performance to calculate how much it would have to raise prices to meet the revenue benchmarks they proposed.

"I'd love to say it wasn't a mistake," he said. "It's something that should have been in there."

Citi and UPS want to use package pricing to fix unrelated financial problems at the postal service

Even after fixing the math error, the Citi note doesn't make a good substantive argument that the post office's pricing model for competitive products is "unsustainable," or that prices need to go up.

In recent years, the post office has been losing several billion dollars a year, if you account for legally-mandated payments to shore up a fund for retiree health benefits (which the service has shirked making.)

Citi sought to address those financial troubles through the other pricing scenario it considered in its note: Raise competitive products prices by $1.74 per piece and the postal service's annual multi-billion dollar losses get wiped out, making it possible to make the retiree benefits fund payment.

"Inherently in our analysis there was an assumption that you should be able to generate enough revenue from the competitive products for the whole enterprise to break even," Wetherbee told me.

There are a few big problems with that idea.

First, it assumes raising prices by more than 50% would not lead to any reduction in package and express-mail volume, which doesn't seem reasonable.

Second, it treats the legally-mandated retiree healthcare payment as a current-period business expense, when in truth only part of it is.

The retirement benefits situation at the post office is very complicated, but essentially the service spent decades failing to pre-fund benefits it was promising workers, and then Congress passed the PAEA, which mandated the service close the funding gap within 10 years. This seemed reasonable in 2006, when the economy was strong and first-class mail volumes were not yet totally cratering.

Payments to close the retiree benefit funding gap have to come from somewhere, but they're not properly counted as business expenses for the current period. It's illogical to say a payment to cover a retirement benefit that a worker earned in 2004 is part of the cost to deliver a package for Amazon today, because the retiree will be owed the benefit whether any more packages are delivered or not.

But the biggest problem with Citi's break-even analysis has to do with first-class mail and standard mail, which is the technical term for junk mail.

Why would the assumption inherent in Citi's analysis be that the post office should rise prices on competitive products to a level so that its entire operation runs at a break-even? Why not raise the price of stamps, or increase the rates that are charged to junk-mail senders?

Calls for higher USPS prices are special pleading from private shippers

Brendan McDermid/Reuters

Jeff Bezos.




In July 2017, the Wall Street Journal ran an op-ed based on Citi's note, called "Why the post office gives Amazon special delivery."

"An April analysis from Citigroup estimates that if costs were fairly allocated, on average parcels would cost $1.46 more to deliver," wrote Joshua Sandblute. "It is as if every Amazon box comes with a dollar or two stapled to the packing slip - a gift card from Uncle Sam."

Who is Sandblute? Per the Journal, he is "co-president of Greenhaven Associates, a money-management firm that owns FedEx common stock."

The truth about the postal service is that it has enormous fixed costs, some associated with those retiree benefit obligations, and some associated with real estate and other facilities it acquired decades ago when its business model looked very different, and which it faces political and practical barriers to shedding today.

As the volume of first-class mail has declined, growing its package delivery business has given the USPS a way to reuse some of its fixed assets, service some of those fixed costs, and limit the losses it would otherwise impose on taxpayers.

That's why the post office has fought efforts to restrict the way it prices; if it couldn't grow this business, it would lose more money. As the PRC correctly noted, the correct question to ask when seeking to determine whether competitive product pricing is competitive and fair is whether the post office makes money at the margin when it sells those products, and they conclude it does.

It's hard to see how forcing USPS parcel price increases could benefit consumers by fostering competition - after all, there is an existing, robust competitive market in package delivery, as demonstrated by the fact that UPS and FedEx already exist and make profits.

A free hand on package pricing at the post office has been good for consumers, good for the post office, and good for online sellers for Amazon. It's not great for shippers that compete with the post office, and it's not great for presidents who hate Jeff Bezos because he owns the Washington Post. But they'll live.

NOW WATCH: This incredible animation shows how humans evolved from early life

function receiveMessage(event){ console.log(event.data.fullscreen); Android.fullscreen(event.data.fullscreen); }window.addEventListener("message", receiveMessage, false);