Based on our analysis of BLS data, the tech-ecommerce ecosystem added 1.4 million jobs between September 2017 and September 2021, the most recent data available from the BLS. The previous job creation leader, the health care sector, added 500,000 jobs, roughly one-third of the tech-ecommerce total. And the rest of the economy lost 900,000 jobs.

On the state level, the tech-ecommerce ecosystem took the place of health care as the main job producer in most of the country. From our analysis, 40 states gained more jobs from tech-ecommerce than health care and social assistance from 2017Q1 to 2021Q1 (our analysis requires detailed QCEW data from the BLS, which currently goes through the first quarter of 2021).*

The top of the list, not surprisingly, was California, which added 310,000 tech-ecommerce jobs over the four year stretch. That made a big difference. In a June 2021 blog item, we estimated that tech-ecommerce accounted for roughly 42% of the increase in California personal tax revenues from 2015 to 2020.

The next four states, ranked by tech-ecommerce job creation, are Texas, Florida, Washington, and New York. New York, in particular, added 73,000 tech-ecommerce jobs over that 4-year stretch. Meanwhile, the number of jobs in the crucial New York finance and insurance sector was flat or slightly down.

Other states with strong tech-ecommerce job growth included Ohio (61,000); Arizona (58,000); Georgia (56,000);North Carolina (56,000); Illinois (56,000); and New Jersey (55,000). Note how tech-ecommerce jobs are well-distributed around the country.

Amazon is currently building its second headquarters in northern Virginia, with a total of 25,000 workers expected over the next decade. But even before the Amazon build-out, Virginia has experienced a surge in tech-ecommerce jobs. Between 2017Q1 and 20201Q1, tech-ecommerce jobs rose by 38,000, while jobs in the rest of the economy, including health care, fell by 53,000.

Virginia’s tech-ecommerce jobs are also well-compensated, earning an average of $109,700 per person in 2020. That’s compared to an average wage of $65,100 for all Virginia workers, and $63,900 for Virginia manufacturing workers.

Nevada had a 76 percent increase in tech-ecommerce jobs from 2017Q1 to 2021Q, the biggest percentage gain among states. Arizona had a 49% increase in tech-ecommerce jobs, the third highest percentage gain. Arizona tech-ecommerce jobs paid an annual wage (including bonuses) of $83,300 on average in 2020. That’s comparable to the average pay for Arizona manufacturing wages($82,400), and substantially higher than average pay in Arizona health care and social assistance ($57,600). Tech-ecommerce pay in Arizona is 43% higher than average pay for the Arizona economy as a whole.

The raw numbers are not so impressive for smaller states, but tech-ecommerce is still important for a state like Delaware, which gained 2,000 tech-ecommerce jobs between 2017Q1 and 2021Q1, while finance and insurance employment stagnated. Average pay for the tech-ecommerce sector in 2020 was $73,000 per year, compared to $58,000 for health care and social assistance jobs.

New Hampshire gains 5,000 tech-ecommerce jobs, while health care was flat in terms of hiring and the rest of the state economy lost jobs. In Vermont, tech-ecommerce jobs were flat but employment in the rest of the economy, including health care, shrank by 20,000.

One interesting note: Minnesota is one of the few states where health care jobs grew significantly more than tech-ecommerce jobs. Perhaps coincidentally, Minnesota is also the home state of Senator Amy Klobuchar, who is the lead sponsor for a tech antitrust legislation in the Senate.

*Note that the total lost jobs on the state level, outside of tech-ecommerce and healthcare, is much larger because the most recent detailed state level data available is 2021Q1.

The large tech and ecommerce companies have become massive job generating and income creating machines, hiring hundreds of thousands of workers in the United States. This is one of the great hiring surges in history, providing well-paying jobs for an unprecedented number of workers.

But just looking at hiring by the tech giants themselves does not fully answer the question of their impact on the labor market. It could be that, like tall trees, they block the sunlight and keep other tech companies and ecommerce companies stunted.

This “ecosystem dominance” would manifest as weak job and income growth in the tech-ecommerce sector as a whole.  If true, this harm to workers becomes a powerful justification for strong regulatory and antitrust growth against the tech giants. In other words, chopping down the trees would help the rest of the forest grow.

Alternatively, strong job and income growth across all tech and ecommerce industries would show the tech giants–who invested a stunning $65 billion in the United States in 2020—are playing a crucial role in a thriving ecosystem that benefits workers, raises wages and generates tax revenues.  Indeed, from 2015 to 2020—a period that includes the pandemic—the tech-ecommerce ecosystem generated 1.7 million net new jobs and added $289 billion in labor income. By comparison, the whole private sector lost 360,000 jobs. In that case, common sense would call for regulatory prudence.  As the saying goes “if it ain’t broke, don’t fix it.”

California

For this blog post we will focus on the job, income, and tax impact of the tech-ecommerce sector on California, which is the headquarters of three out of the four tech giants. In addition, in the fourth quarter of 2020,  Amazon employed more workers in California (153,000+) than it does in Washington (80,000+).

Our analysis builds on PPI’s April 2021 paper, “Innovative Job Growth in the 21st Century: Has the Tech-Ecommerce Ecosystem Become the New Manufacturing?”. The tech-ecommerce ecosystem includes five tech industries and three ecommerce industries. The tech industries are computer and electronic production manufacturing (NAICS 334); software publishing (NAICS 5112); data processing and hosting (NAICS 518); Internet publishing and search, and other information services (NAICS 519); and computer systems design and programming (NAICS 5415). The three ecommerce industries are electronic shopping and mail order houses (NAICS 4541); local delivery (NAICS 492); and ecommerce fulfillment and warehousing (NAICS 493).

We draw on Bureau of Labor Statistics data from the Quarterly Census of Employment and Wages (QCEW). This dataset reports on all wages, salaries, and bonuses, including ordinary income from exercised stock options. We look at the five-year period from 2015 to 2020, which includes the pandemic year.

Table 1 shows the growth of jobs and labor income in California’s tech-ecommerce sector from 2015 to 2020.  Tech-ecommerce jobs rose by 38% over the five-year stretch in California, compared to 31% in the United States as a whole. Meanwhile, private sector jobs rose by 0.3% in California and fell by 0.3% nationally (not shown on table).

Wages and salaries in California’s tech-ecommerce sector rose by an astounding 76% from 2015-2020, compared to 56% nationally. Meanwhile, private sector wages and salaries rose by 31% in California, and 21% nationally.

Table 2 shows the importance of the tech-ecommerce sector for California’s economy. The tech-ecommerce sector added 350,000 jobs between 2015-2020 in the state, and $100 billion in additional wage and salary income. That means the tech-ecommerce sector accounted for 38% of the entire increase in private sector wages in the state over that period.

Note that Table 1 and Table 2 break out the core tech counties, San Francisco, San Mateo, and Santa Clara, from the rest of the state. Taken together, the two tables show that both the core tech counties and the rest of the state have shown roughly equal rates of income growth from the tech-ecommerce sector.

Table 3 looks specifically at ecommerce and retail jobs in California. Obviously, the pandemic forced a dramatic decline of brick-and-mortar retail jobs in the state. At the same time, the number of ecommerce jobs increased by more than enough to counteract the decline of brick-and-mortar retail. Moreover, the ecommerce jobs were substantially better paid on average.

As a result, when we combine brick-and-mortar retail with ecommerce industries in California, the number of net jobs rose by 28,000. Average annual pay rose by 22 percent.

Finally, we turn to the question of the impact of the tech-ecommerce sector on personal income taxes in California. Tax collections have come in much stronger than expected, with personal income tax collections in the first nine months of the 2020-21 fiscal year running at 17% or $14 billion above forecast. Personal income tax revenues in the 2020-21 fiscal year are now forecast to be 54% about 2015-2016 levels.

How much of that gain is accounted for by the tech-ecommerce sector? There are several issues with making this calculation. The state government reports and forecasts tax revenue data on a fiscal year basis, while our data on the tech-ecommerce sector is on a calendar year basis and stops with 2020. Second, our definition of the tech-ecommerce sector includes a wide variety of industries, with average annual pay that runs from roughly $50,000 to well over $300,000. Third, much of the surge in personal tax revenues is coming from capital gains, which is directly connected with the success of the tech-ecommerce sector but is not reported in the BLS QCEW data.

Nevertheless, we can make a back-of-the-envelope estimate of the personal tax revenue generated by the tech-ecommerce sector. First, let’s start by looking the increases in personal tax revenues coming from wage and salary income (included ordinary income from exercised stock options) over the 2015-2020 period. By our estimate, the increase in tech-ecommerce wages and salaries accounts for roughly 37% of the increase in personal tax revenues from wages and salaries in the 2015-2020 period.

But of course, there has been a surge in capital gains revenues as well. If we attribute half the unanticipated increase in capital gains in 2020 to the tech-ecommerce sector, then tech-ecommerce accounts for roughly 42% of the increase in California personal tax revenues from 2015 to 2020.  This should be viewed as a rough estimate rather than a final number.

On Wednesday, the House Judiciary Committee is going to mark up five tech antitrust bills. Collectively, the bills mark a major departure from the traditional consumer welfare standard that has governed antitrust law over the last few decades. Instead of focusing on consumers, these new laws would single out just five large tech platforms and apply an entirely different set of standards. One bill would effectively ban them from making any future acquisitions, which might have the unintended consequence of reducing startup investment, and therefore reducing competition. Most concerningly, another one of the bills would lead to breakups of all five major tech companies. Vertical integration would effectively be prohibited because, according to the bill’s authors, it presents an irreconcilable conflict of interest.

But what this framing misses is all the consumer benefits that flow from integrated ecosystems. Many digital products are free to access because they are subsidized by ads elsewhere in the ecosystem. A hallmark of a seamless user experience is being able to switch between devices, websites, and apps without needing to re-enter all your information. Crucially, these integrated experiences are also safer for users because fewer players in the market have direct access to user data (which is why most government agencies do not allow federal employees to “jailbreak” their smartphone devices or sideload apps from unapproved app stores). And of course, private label goods on Amazon work just the same as they do in Walmart or CVS — they offer consumers similar quality to name brands at lower prices.

Here’s a more detailed breakdown of the five bills and what they would do to tech platforms (in order from most reasonable to least reasonable):

The Merger Fee Modernization Act sponsored by Representative Neguse

The budgets for the FTC and DOJ to conduct antitrust enforcement have fallen by 18% between 2010 and 2019, after adjusting for inflation. Over the same period of time, the economy has grown by 22%. To properly enforce the antitrust laws on the books, the DOJ and FTC need resources that match the scope of the problems they face. This bill would increase their enforcement resources by almost 30% and change the merger filing fee structure to fall more heavily on larger deals. There is significant bipartisan support for this bill and it is urgently needed.

For the next four bills, you need to understand what a “covered platform” is. All four bills define them the same way (and these new rules would only apply to covered platforms). A covered platform is a “website, online or mobile application operating system, digital assistant, or online service” that meets all three of the following conditions: (1) 50 million U.S.-based monthly active users or 100,000 U.S.-based monthly active business users; (2) greater than $600 billion in net annual sales or market capitalization; and (3) is a “critical trading partner” that can restrict business users access to customers.

As of today, there are only six companies in the U.S. that meet the $600 billion market capitalization threshold. Every commentator assumes Amazon, Apple, Facebook, and Google will qualify as covered platforms, and most agree that Microsoft will be included as well, considering it operates multiple large-scale platforms, such as Windows, Office, Xbox, and LinkedIn. It remains to be seen whether the “net annual sales” metric will be interpreted to cover financial services companies like Visa, JP Morgan Chase, and PayPal, which process a large volume of payments.

What seems clear, however, is that the subcommittee bills target big tech firms instead of probing economic concentration across the U.S. economy.

The Augmenting Compatibility and Competition by Enabling Service Switching (ACCESS) Act sponsored by Representative Scanlon

The ACCESS Act would require platforms to provide third parties with APIs, software that allows access to platform data. The bill leaves the definition of “data” up to the FTC to determine. Data portability done right can lower switching costs and improve competition in an industry. Consider the enormous success of telephone number portability in the telecom industry. Letting consumers own their telephone number lowers the cost of moving to a new provider. And because telephone numbers are a necessary and discrete piece of data that all carriers must use to operate a network, there was no risk of decreasing the incentive to invest in creating this data.

For tech platforms, there might be similar discrete, static, and critical data sets that should be subject to mandatory portability rules. For example, the social graph — the list of all your friends or connections on a social network — is a very important dataset for new startups to have access to. Users are more likely to use a new app if during the onboarding process they are able to share a social graph from another social network and find all their friends on the new platform with a single click.

However, the problem with this bill is that it is not narrowly tailored to discrete and critical data sets like the social graph. It merely says that “a covered platform shall maintain a set of transparent, third-party-accessible interfaces (including application programing interfaces) to enable the secure transfer of data to a user, or with the affirmative consent of a user, to a business user at the direction of a user, in a structured, commonly used, and machine-readable format.” The bill leaves it to the FTC to define what “data” means for the purpose of the bill. It would be very helpful if Congress offered more guidance on what kinds of data it intends to be covered by these rules.

If data is defined too broadly, then there might be unintended consequences for investment incentives. For example, tech companies are all racing to build the next great computing paradigm. Will it be virtual reality? Blockchain technology? Augmented reality? Smart devices? Or something else no one can predict? Regardless of which paradigm wins out, if the future winner is forced to give every one of its competitors access to all of its data, then that would decrease the incentive to invest in the next big platform today. The most tragic part of the scenario is that these will be unseen costs — we won’t know what we lost out on. The future will just be somewhat dimmer because a well intentioned policy backfired due to poor drafting and a rushed process.

Platform Competition and Opportunity Act sponsored by Representative Jeffries

The Platform Competition and Opportunity Act is effectively a ban on all mergers and acquisitions by platform companies. This bill would ban platforms from acquiring companies that:

• “compete with the covered platform … for the sale … of any product or service”;
• “constitute nascent or potential competition to the covered platform … for the sale … of any product or service”;
• “increase the covered platform’s … market position”; or
• “increase the covered platform’s … ability to maintain its market position”

Given how broad this language is, the bill would effectively ban all acquisitions by platform companies. Since more than 90% of startups provide a return for their founders, employees, and investors through an acquisition as opposed to going public, this bill has the potential to backfire and decrease investment in startups. A recent study found that “VC activity intensifies after enactment of country-level takeover friendly legislation and decreases following passage of state antitakeover laws in the U.S.” This bill would qualify as an antitakeover law.

American Innovation and Choice Online Act sponsored by Representative Cicilline

This bill is aimed at remedying the perceived conflict of interest by platforms and businesses that leverage those platforms to reach consumers. In essence, this bill bans self-preferencing by requiring platform owners to refrain from any conduct that gives their own products an advantage over competitors’ products. Section 2 of the bill makes it clear how all encompassing this rule aims to be (emphasis added):

“It shall be unlawful for … a covered platform … to engage in any conduct that … advantages [its] own products, services, or lines of business over those of any other business user, excludes or disadvantages the products, services, or lines of business of another business user relative to the covered platform operator’s own products, services, or lines of business, or discriminates among similarly situated business users.

The bill then provides 10 examples of discriminatory conduct, including tying, anti-steering provisions, retaliation, and restrictions on pricing.

But those specific examples aren’t really necessary when the bill includes a blanket ban on any conduct that “advantages” the platform’s products over those of third parties. While this attempt to fix a conflict of interest may seem intuitive at first glance (think of Elizabeth’s Warren’s baseball analogy), the more you think about the idea, the less it makes sense. For example, consider how this rule would apply to Apple. The iPhone runs on Apple’s proprietary iOS operating system. Apple wouldn’t be allowed to “advantage” its App Store in any way, which means it can’t be pre-loaded on devices and it can’t be the default app store unless users select it. This same logic applies to every layer of the tech stack. Apple makes dozens of popular first-party apps, including FaceTime, iMessage, Mail, and Music. As the bill is currently written, Apple would not be allowed to pre-install those apps on iPhone devices because that would “advantage” them over other video conferencing, messaging, mail, and music apps.

Now consider how this law would apply to Google. If a user typed in “restaurants near me” on Google, the search engine wouldn’t be able to directly offer map results at the top of the page from Google Maps because that would give it an “advantage” over other mapping services. Google would be forced to merely provide links to competitive mapping services rather than give consumers the answer to their question. The same rule would apply to Google Shopping if a user searched for sneakers. Instead of showing the user sneakers, Google would have to show users links to shopping websites that sell sneakers. This would represent a huge loss to consumer convenience that makes these products so popular (91% of Americans have a favorable opinion of Amazon, 90% have a favorable opinion of Google, and 81% have a favorable opinion of Apple).

Most concerningly, this bill would break the safety and security of many features of the iPhone. If Apple has access to a piece of hardware, such as a sensor or communications chip, then it has to give equal and fair access to that same hardware function to all third parties. That sounds like a laudable goal if you want more options when it comes to payments (i.e., access to the NFC chip) or location services (i.e., access to GPS) or the microphone (e.g., the way say Siri is always listening for “Hey, Siri”). But the flip side of more competition in this context is that every bad actor with the intent to defraud consumers or invade their privacy now also has access to sensitive data by law.

Lastly, some argue that the affirmative defense section of this bill would allow some pro-consumer conduct by the platforms to continue (such as continuing to pre-install apps on phones). The platforms can “advantage” their own products so long as they “would not result in harm to the competitive process by restricting or impeding legitimate activity by business users; or was narrowly tailored, could not be achieved through less discriminatory means, was nonpretextual, and was necessary to prevent a violation of, or comply with, Federal or State law; or protect user privacy or other non-public data.” But pre-installation and default settings clearly give a leg up to the products controlled by the platform owner and therefore might “result in harm to the competitive process.” If the intent of the drafters is not to ban this type of conduct, they should clarify this section.

Ending Platform Monopolies Act sponsored by Representative Jayapal

The most extreme and economically destructive of the five bills is The Ending Platform Monopolies Act. It tries to address the same problem as the American Innovation and Choice Online Act — conflicts of interest between platform owners and platform competitors. But instead of requiring platform owners to operate their platforms in a neutral fashion as the non-discrimination bill does, this bill bans vertical integration outright and would lead to the break up of every large tech company across multiple dimensions.

Google would have to spin off YouTube, Android, Chrome, the Play Store, and its apps (Gmail, Google Maps, Drive, etc.) into separate businesses. Of course, that would destroy Google’s current business model where revenue from search and display advertising is used to subsidize an ecosystem of free products for consumers. Post-breakup, the newly independent entities would likely need to start charging subscription fees or create their own advertising business from scratch (and add more ad units to their respective products).

Amazon would be forced to spin off its private label goods business (e.g., Amazon Basics) and Amazon Marketplace because those two lines of business compete with the traditional retailing model where Amazon takes inventory of the product from wholesalers and then resells it at a markup. Amazon would also be forced to spin off its Amazon Prime Video streaming service and Amazon Web Services.

It has not yet been properly appreciated that this bill is aimed at addressing the same alleged conflict of interest issue as the American Choice and Innovation Online Act. If they are passed together, this bill would obviate the other one. As independent technology analyst Ben Thompson pointed out, this could mean that Chairman David Cicilline is attempting to make his bill seem reasonable by comparison even though it also has radical implications for tech ecosystems. Legislators shouldn’t fall for this obvious gambit.

The DOJ and FTC desperately need more resources to adequately enforce the antitrust laws on the books, and a narrowly tailored data portability mandate could enhance digital platform competition. But blanket bans on acquisitions, self-preferencing, and vertical integration would destroy many of the consumer benefits that make the tech giants world leaders in their respective markets. Hobbling America’s tech giants without adequate evidence of consumer harm would be a capitulation to the populists on the far left and far right at a time when we need to be focused on economic recovery.