Doug Garnett’s Blog


Law of the Jungle: Does Amazon Avoid Antitrust Using Goodhart’s Law?

Law of the Jungle:  Does Amazon Avoid Antitrust Using Goodhart’s Law?

The following is a “thought experiment” considering new ways to view Amazon with respect to antitrust. Thoughts and responses which further explore or further challenge what I’ve written are quite welcome — we’re all just trying to sort this out.

In 1997, anthropologist Marilyn Strathern generalized what has become known as Goodhart’s Law into its most common statement:

When a measure becomes a target, it ceases to be a good measure. 

What Goodhart saw was that once metrics become a goal they lose their value because they are easily gamed such that numerical goals are reached without the intended result. Metrics are far more susceptible to gaming than descriptive goals because numbers measure only a small part of the total — they do not measure a “coherent whole.”

Modern business is plagued by this gaming. Bonus systems? Gamed. Corporate share price? Gamed. Revenue per employee? Easily gamed with layoffs and outsourcing to the detriment of the company. (Perhaps we will one day comprehend the destruction wreaked by the metric obsession in modern business management and corporate governance.)

But what about antitrust enforcement? In reading and discussing Amazon and antitrust issues, starting from the Yale Law Review article by FTC Chair Lana Khan, a question arose:  Has Amazon been using low prices to game antitrust enforcement?

Some Antitrust Background

In the early 1900s, Teddy Roosevelt established anti-trust laws to protect society against times when the free market breaks (sometimes it does — not usually). Teddy Roosevelt and Justice Louis Brandeis considered antitrust critical for times when companies amass power that damages society. Of course, it was important that these laws live within fundamental fairness of society and the constitution. Still, they were used effectively to prevent problems which had become obvious around 1900 – questions of power.

During the 1970s, the Chicago School became highly influential and shifted the interpretation of Roosevelt’s antitrust work to focus on one theory:  antitrust damage only matters if it affects customers through higher prices. According to Lana Khan’s article, this shift rejected consideration of power and narrowed interpretations of antitrust in the courts to primarily consider low prices as the indicator of a “healthy free market.”

This gives me pause as using low prices as a proxy for “healthy market” is all messed up. Most concerning, as a marketer I know low price is no proxy for market health — it often indicates dying markets. Healthy markets are also evolving into the future when participants invest in order to deliver additional value to customers. These markets usually have a range of prices, including higher prices, where market health benefits customers with more ways to obtain the value they want.

It’s worth recalling that in the absence of meaning, customers fall back on price. What this means is that if customers don’t understand the value of products, low price is their alternative. So when companies regularly communicate the value their product delivers, it turns out prices will be higher and customers will also be more satisfied.

The Chicago School assumptions also need to be checked. This low price idea comes from their work with classical equilibrium economics. There is some validity within a commodity market. But, for example, online retail is NOT a commodity market. For that type of market, we need to turn to complexity economics where we learn that equilibrium is an exception so we need to be more careful with its assumptions. This is especially true with modern “digital like” startups who have leveraged Brian Arthur’s Increasing Returns theory. Amazon, Uber, Lyft, Dollar Shave, WeWork, and the rest all exist outside of economic equilibrium.

How does this matter with Amazon?

I don’t know for certain that it does. But I think we also need to be clearer about exactly what Amazon has done. If we are, then it might be very important.

Through a detente’ with investors, Amazon was allowed to lose money on retail-like sales for decades as long as Amazon stock price continued to increase (it did). A profitless Amazon returned money to investors.

Losing money early isn’t a problem as antitrust law allows entering a market below market price based on the assumption that the market will drive adjustments when investors demand profit from THAT MARKET. After 25 years, Amazon remains profitless on the bulk of its revenue (retail-like sales online) and investors aren’t demanding profit. That violates all kinds of fundamental free market assumptions. How did that happen?

Amazon used the freedom given by investors to build a “platform.” Where antitrust law assumes competition in the original market forces adjustment for predatory pricing, there are no forces in the original market pressuring Amazon to make profit. That’s because Amazon didn’t underprice to create future profits within their market as in normal economics — they underpriced to build things which they ran off with to vastly different markets. It’s a bit like devastating the land in a small country to extract lithium so you can sell battery powered cars in the US. No economic force will force the company to change its activities in the small country.

In truth, Amazon is only profitable in entirely unrelated markets while within the retail-online market (the bulk of their revenue) they continue predatory pricing and reap no profits. Over half of quarterly profits come from AWS (rentable cloud computing), some profit from selling their own devices, and some from selling digital content. Other than AWS profits, no one — investor or otherwise — can tell where profit comes from because the numbers needed to tell us how they do business are hidden.

It appears that Amazon’s predatory pricing in the online retail market was used to gain experience and technology to build a platform they could sell  profitably in far distant markets. 

None of Amazons profits come in the market where they underprice – the market of retail-like online sales. As a result, the market destruction and chaos they have wreaked is not economically creative destruction because market forces are not at work. Investors are plenty happy with the returns from subsidizing Amazon to destroy parts of the retail market in order to build capabilities they sell outside that market. 

Back to Anti-trust Issues.

I called this a Goodhart’s Law example. How is that? Amazon seems to have gamed anti-trust enforcement as follows:

  • A distant and “top level” look at Amazon suggests their “competition” keeps retail prices low. Of course their prices are low —they’ve been subsidized by investor patience and cash flow gimmicks for 25 years. 
  • A distant look doesn’t distinguish the vast distances between their markets. Amazon is quite careful to keep enough PR focused on retail like items — making it appear they’re serious about profits in that market…profits which somehow never appear.
  • By keeping retail-like online prices low, Amazon avoided serious investigation given the legal interpretation derived from Chicago School theories.

Even Lana Khan’s interesting article in the Yale Law Review fails to note the fact that Amazon is using predatory pricing in one market to build foundation which can only produce profit in a far, far distant market.

Andrew Everett also pointed out that this loophole is clearly present on the FTC website. Note the following quote about predatory pricing and how it assumes within the same market (downloaded August 2021):

This strategy can only be successful if the short-run losses from pricing below cost will be made up for by much higher prices over a longer period of time after competitors leave the market. Although the FTC examines claims of predatory pricing carefully, courts, including the Supreme Court, have been skeptical of such claims.

More Companies have Followed Amazon’s Approach

Learning from Amazon’s success, a bevy of startups seem to be playing the same game. Uber, Lyft, all kinds of food delivery services, Dollar Shave, the WeWork disaster, …..

All of these rely on their investor starting money to underwrite massive losses over long periods of time. While each of these companies argues that they “could be profitable any time they want” there’s no evidence of profit anywhere. Is it possible they will take their platforms well outside the market where they subsidize destruction? Only time will tell.

The game seems to be given away when they get a start with big announcements in sexy and large markets where investors are willing to subsidize predatory pricing based on “volume.” Beyond Amazon, though, we haven’t seen any build platform capabilities which make big profits in entirely unrelated markets. It is early for most of the list, though. Regardless, there is safety in their approach as market forces will never push back hard enough to control the destructive behavior in the original market.

As JP Castlin and JCP Hankins observed about Uber in a recent article, “the company appears lightyears from making money.” Yet Uber continues to offer low prices while offering a rationale that, somehow, being connected digitally makes it all possible. Excepting that the magic pixie dust of digital doesn’t work that way. Their argument sounds very much like Amazon.

The Antitrust Challenge: Highly Adaptable Platforms are Unique to Tech

I lack the insights to clearly state the specifics behind how a tech “platform” like AWS plays into this. That said, it is clear that AWS is a  fundamental technology for hosting databases (and other IP assets) on remote computers with the capability clients need to access to that data. This market has only the tiniest overlap with retail when a retailer hires AWS for their cloud needs.

By definition, a platform capability is “application independent.” So, where an automotive engine technology can only be applied with relative ease in a automotive or truck application, AWS can be easily taken to any market. The independence is what makes platforms highly valuable.

Should we allow companies to use predatory pricing in one application to develop the platform which can only have economic return in another market? Fully answering that is beyond my pay grade. My instinct and ethical sense is that Amazon has pulled a fast one and should be stopped. But I’m also not beyond accepting that, while we don’t see how today, maybe the market will mete out a punishment that is well deserved.

Gaming with Goodhart

Amazon has had a 20-year free ride using investor money and patience to destroy one market while building capabilities and experience in order to abandon that market and walk into an entirely different one. (A friend observed last week that he believes Amazon would jettison the online retail market in a heartbeat with the right motivation.) 

And Uber? Nothing indicates that Uber can ever be profitable either from taxi service or delivery. So the only way Uber can find profit is with a pivot — taking the platform to an entirely new market where they can create huge profits (maybe corporate HR?). Meanwhile, they have destroyed the US taxi market with underpricing and are threatening restaurants with destruction by stealing margins which are already low. (As a marketer, the marketing crime with Uber is that their underpricing wasn’t necessary — their service was so good that they would have succeeded well by charging a premium which would make them profitable today.)

From what I know of Teddy Roosevelt’s vision around antitrust, these all appear to qualify as destructive abuses of power wreaking more damage than mere low prices and competition — and they are executed in ways the market won’t re-adjust.

I certainly don’t see any way in which the economy will correct these power abuses. That would mean that we need the federal government to step in to curb these practices. And I fear to say those words since governmental action always drives unexpected consequences. Yet this IS the role of government for society.

Unfortunately, it’s also my sense from a distance that courts change slowly and their love of the “low price” proxy for “healthy market” is going to be tough to change via case law. But that is certainly not my area of expertise.

At this point, I’m a spectator watching retailers damaged by Amazon’s expectation setting and market destruction struggle to survive in markets where they should thrive.

©2021 Doug Garnett — All Rights Reserved

Through my company, Protonik LLC, I consult with companies as they design and bring to market new and innovative products. I am currently writing a book exploring the value to business from the field of complexity science. Protonik also produces marketing materials for artists including documentaries, websites, and blogs. As an adjunct instructor at Portland State University I teach marketing, consumer behavior, and advertising.

You can read more about these services and my unusual background (math, aerospace, supercomputers, consumer goods & national TV ads) at 

Categories:   Business and Strategy