In a statement released after the latest consumer price report, President Biden remarked on the “meaningful reduction in headline inflation” but indicated that there was still “more work to do, with price increases still too high and squeezing family budgets.”

In particular, the Biden Administration wants to protect consumers by identifying markets where sellers are taking advantage of the pandemic and supply chain snarls to raise prices. That’s a great plan.

At the same time, it’s also important to recognize and acknowledge those industries where price increases have been moderate and restrained.

In that spirit, we examine the inflation performance of the digital sector of the economy, encompassing tech, ecommerce, broadband, and related industries. These companies have come under fire for a variety of different reasons, some deserved, some not.

In this blog item we will show, based mainly on government data, that digital companies are helping hold down inflation at a time when prices are soaring in many other parts of the economy.  For the Democrats and the Biden Administration, this is a success story they can build on.

The Historical Perspective

We’re used to computers getting cheaper over time, as they become more powerful and versatile. The Internet opened up entirely new dimensions of free websites, with everything from recipes to news to maps and directions. Long distance phone calls have effectively become free. Broadband networks, both wired and wireless,  have become faster, connecting almost every part of the country. Content has become more varied and cheaper, at the same time.

There is no doubt that technology has been a profoundly disinflationary force historically. But what about today? The GDP inflation rate was 4.6% in the year ending with the third quarter of 2021. That’s a big jump from the 1.7% GDP inflation rate in the third quarter of 2019, before the inflation. How much of that acceleration is coming from the tech sector?

The answer is, precisely none.  As part of its calculation of GDP, the Bureau of Economic Analysis (BEA) calculates price changes by industry. It turns out that inflation rates in four key digital industries are not only negative but falling (Table 1).  For example, the inflation rate in the “data processing, internet publishing, and other information services” industry fell from +0.5% in 2019 to -1.1% today.

The same is true for the other three key digital industries. Digital is still following the historical trends of being disinflationary.

Table 1. Digital is Still Disinflationary (change in value-added prices)
	year ending
	2019Q3	2021Q3
Computer and electronic product manufacturing	-0.1%	-1.8%
Broadcasting and telecommunications	-0.9%	-2.6%
Data processing, internet publishing, and other information services	0.5%	-1.1%
Computer systems design and related services	-0.2%	-2.8%
Gross domestic product	1.7%	4.6%
Data: BEA, based on Table TVA104-Q

 

Ecommerce and inflation

Let’s now consider ecommerce prices in particular.  As of the third quarter of 2021, ecommerce accounted for 13% of retail sales according to the Census Bureau. That’s back on the long-term trend line after a temporary pandemic-induced jump.

But still, there’s an important question: Why isn’t ecommerce a bigger share of retail sales, given how much we are all shopping online? One reason might be that online prices tend to rise at a slower rate than brick-and-mortar prices, according to the available evidence. Indeed, government data shows that the long-term trend of ecommerce has been and continues to be disinflationary.

Consider this: The BLS measures changes in gross margins in all major retail industries, where the margin is defined as the selling price of a good minus the acquisition price for the retailer.  Margins include all costs, such as labor, capital, and energy, plus profits, taking into account gains in productivity. A slower rise in margins translates directly into less inflation for consumers, all other things being equal.

Between December 2007 and December 2021—a 14-year stretch that included the financial crisis, the long boom, and the pandemic—margins in the electronic shopping industry rose by 20%, according to BLS data (Figure 1). Over the same stretch, consumer prices rose by 33%. The implication: Ecommerce companies were accepting thinner margins in real terms, and passing those benefits onto consumers.

By comparison, the data for the overall retail industry has shown a much worse inflation performance, measured by margins. Margins for the total retail industry rose by 48% since December 2007, much faster than consumer inflation. As a result, real margins for the retail industry as a whole have risen, putting upward pressure on consumer prices.

Now let’s look at the current situation. Even in today’s inflationary burst, ecommerce stands out as a force holding down margin increases compared to the rest of retail. In the year ending December 2021, overall retail margins rose by 13.1%. Meanwhile margins at general merchandise stores like warehouse outlets rose by 12.6%. Margins at auto dealers and other auto-related retailers rose by a stunning 26%, far outpacing inflation

By comparison, over the past year, ecommerce margins only rose by 7.1%, about the rate of inflation (Figure 2). These results are completely consistent with the economic literature, which mostly concludes that prices for online goods rise slower than the prices for comparable goods sold offline.  A 2018 paper co-authored by Austan Goolsbee (CEA head under President Obama) found that online inflation was more than a full percentage point lower than the corresponding official consumer price index. Sometimes the difference can be much greater. The latest “Digital Price Index” report issued by Adobe shows that the price of furniture and bedding sold online rose by 3% in the year ending November 2021. Meanwhile the official CPI for furniture and bedding, including all brick-and-mortar stores, rose by 12%,

 

Moreover, the slow growth of ecommerce margins came at the same time that ecommerce fulfillment centers were dramatically boosting employment and pay. Over the last year, the average hourly earnings for production and nonsupervisory workers in the warehousing industry rose by 19.4%. That covers the great majority of ecommerce fulfillment centers.

The ability to simultaneously hold down prices for consumers, reduce shopping time for households, and boost pay for workers, represents a rare win-win proposition. What could be better?

 

Smartphones, Telecom, and the Digital Economy

During the pandemic, the daily life of Americans has been supported by wired broadband and wireless networks, by content delivered to the home and to wireless devices such as smartphones. This Digital Economy has been essential for work, school, and social contacts in the midst of these bizarre years.

But equally important, the Digital Economy is also a low-inflation economy. While the price of old economy products like cars, clothing, and gasoline has been soaring, the inflation rate of digital goods and services like smartphones, video and audio services, wireless, and internet access has remained low.

According to our analysis of BLS data, the digital consumer inflation rate was only 1.6% in the year ending December 2021, barely above the 1.4% rate in the year ending December 2019, before the pandemic started (Figure 3). This figure includes computers, smartphones, and other IT commodities; video, audio, and music services; telephone services; and internet services and electronic information providers. We use BLS spending shares to weight the components of the digital inflation rate.

Looking at individual items, the inflation rate for video and audio services, including cable and satellite television service, fell from a 3.1% rate in 2019 to a 2.6% rate in 2021. The inflation rate for telephone services, including wireless, went from 1.6% in 2019 to 0.7% in 2021. Perhaps most striking, the price of smartphones continued their relentless plunge in 2021, dropping in price by 14% after adjusting for quality.

By contrast, there was a huge jump in core consumer inflation, which went from 2.3% in 2019 to 5.5% in 2021. Note that even if some of the components of digital inflation are mismeasured, as some have argued, looking at the change over time should be more accurate if the size of the mismeasurement stays the same.

Tech Inflation

Here we drill down into inflation performance into various components of the tech sector, using data from the BLS Producer Price program.  Figure 4 compares inflation in several tech-related industries with consumer inflation.

According to the BLS, prices for the software publishing industry fell by 0.7% in the year ending December 2021. Prices for data processing and IT support and consulting rose by a measly 0.5%. Computer and electronic product manufacturing prices rose by 2.8%. And the price of internet publishing and search advertising rose 4.1%, considerably slower than the overall consumer inflation rate. That means in real terms the price of internet publishing and search advertising has been getting relatively cheaper.

The App Economy

Finally, we come to the App Economy and the app stores.  Arguably one of the great technological shifts of all time, the introduction of the Apple iPhone in 2007 and then the Apple App Store in 2008 created an entirely new model for delivering services to consumers conveniently and at a low price. It is clear that the App Economy is a profoundly disinflationary force.

The current price statistics do not break out app-relevant price measures, like the price of app downloads or in-app purchases, either from the consumer or app developer perspective. Nevertheless, a careful look at the structure of the pricing structure of the app stores suggests they are contributing to low inflation today.

App store pricing comes in two parts. First, both the Apple App Store and Google Play charge a nominal fee for registering for a developer account. Google Play charges $25 to register, an amount that hasn’t changed in years. Similarly, the Apple App Store charges an annual fee $99 for a basic developer membership, an amount that also hasn’t changed in the U.S. for years (there are a variety of exemptions). As a result, the inflation-adjusted fee has fallen substantially over time.

Most app developers pay no more than this initial fee, or the somewhat higher fee for enterprise developers. As Judge Yvonne Gonzalez Rogers wrote in her September 2021 decision in the court case involving Apple and game developer Epic: “over 80% of all consumer accounts [in the Apple App store] generate virtually no revenue as 80% of all apps on the App Store are free.” These are apps which are free to download, and have no in-app purchases or subscriptions. Many of them, like banking or airline apps, may be quite frequently downloaded and used.  This huge swath of the app stores is disinflationary, with a price that is fixed in dollars over time.

Then there are the small percentage of apps which collect significant consumer revenues on the app stores. Most of these are gaming apps.  For the purposes of assessing their impact on inflation, there are two important factors. One factor is whether the price of the subscription or in-app purchase is rising. The other factor is whether the percentage fee charged by Apple and Google for use of their platform is rising or falling.

We have little visibility into the price evolution of subscription costs and IAP prices. One survey from Sensor Tower suggest that the median price of subscriptions for non-game apps did not change from 2017 to 2020, while the median price of in-app purchases for non-game apps rose by 50%. However, even in the latter case, we have no way of knowing whether consumers are buying the same digital goods or shifting to higher value purchases, which matters for inflation.

We have much better information on the effect on inflation of the fees charged by Apple and Google. The statistical literature makes it clear that if the fee percentages stay the same, they has a neutral impact on inflation. If the fee percentages rise, that is inflationary. If fee percentages fall, that is disinflationary.

In the past year or so, both Google and Apple have voluntarily cut fees for a significant portion of their developer base. Apple, for example, cut the App Store fee from 30% to 15% for all developers who earned less than $1,000,000 in 2019. By one estimate,  that covered 98% of apps with revenue in 2019. Google reduced its fee on subscriptions to 15% (previously it had charged 30% for the first year). These are substantial changes.

With the app store registration or membership fee being held constant in money terms, and revenue-based fee percentages falling, it’s clear that the app stores are contributing to disinflationary pressures.

Conclusion

Both historically and currently, the broad swath of tech, telecom, and ecommerce companies appear to be leaders in the fight against inflation. Data from the government and elsewhere shows no evidence of accelerating price increases in this sector.

 

 

Summary: Because of their mixed urban/rural nature, swing states are a magnet for digital “tech-ecommerce” jobs. Democrats should consider building their 2022 political narrative around support for the digital economy, which brings strong job growth, higher wages, and lower inflation to swing states.

1. This piece is primarily economic, not political. But to focus our analysis, we start with a list of eight potential 2022 Senate swing states: Arizona, Florida, Georgia, Nevada, New Hampshire, North Carolina, Pennsylvania, and Wisconsin. Adding or subtracting a state from this list doesn’t change the analysis appreciably.

2. For the most part, swing state employment is still below pre-pandemic levels, creating a handicap for the party in power. For example, Wisconsin had 260,000 fewer private sector jobs in November 2021 than in November 2019. The only swing state above pre-pandemic employment levels is Arizona.

3. Average hourly wage growth has significantly lagged national inflation in most swing states, That means real wages are falling on average in the swing states, an obstacle to making a positive economic argument. Once again, there is one exception, North Carolina.

4. However, voters in most swing states are benefiting from exceptionally dynamic job creation in the digital “tech-ecommerce” sector. The reason is not hard to understand. Swing states, by their nature, tend to have a mixed urban/rural character. Not surprisingly, ecommerce fulfillment centers are often sited in areas that are within easy driving distance of large populations but where land is relatively cheap, making them a good match for swing states. Similarly, software and internet operations looking to broaden out from the Bay Area and Boston will pick cheaper locations near to good urban amenities.

5. Digital employers gravitated to swing states during the Biden Administration. From the second quarter of 2020 to the second quarter of 2021, the most recent data available, swing states showed a 14% gain in tech-ecommerce jobs  By comparison, the rest of the country only showed an 8% gain in tech-ecommerce jobs over the same period.

6. We’ve bolded the key numbers in the table below, which lays out the job gains by digital industry.  Ecommerce fulfillment and warehousing jobs registered a 22% gain in the swing states, almost double the 12% gain in the rest of the country. The same thing held true for the electronic shopping industry (which mainly consists of those establishments that are in the technology end of electronic shopping but don’t do fulfillment). Also showing strong swing state growth was internet publishing, which also includes search and other “internet-type” companies.

By contrast, job growth in manufacturing and healthcare was weaker in the swing states than the rest of the country.  These states see their future in tech and ecommerce.

Table 1. Swing States Lead In Digital Growth
	Employment change, 20Q2-21Q2
	Swing states	Other states
Computer and electronic manufacturing	-1.6%	-0.4%
Electronic shopping	18.5%	11.8%
Local delivery	12.4%	13.9%
Ecommerce fulfillment and warehousing	22.1%	12.3%
Software publishing	11.8%	8.7%
Data hosting	12.6%	6.1%
Internet publishing and other information services	14.6%	6.1%
Computer software systems	7.1%	3.5%
Tech-ecommerce (total)	14.0%	8.1%
Private	11.2%	10.6%
Manufacturing	5.0%	6.1%
Healthcare and social assistance	5.6%	6.4%

 

7.  What are the positive economic stories in individual swing states?  Digital employers in Arizona, Florida, Nevada, North Carolina, and Wisconsin are adding digital jobs at a faster rate than private sector jobs. Voters in North Carolina, for example, have benefited from 20% growth in tech-ecommerce jobs, almost double private sector job growth.

 

Table 2. Digital Jobs Drive Swing State Growth
	Employment change, 20Q2-21Q2
	Tech-ecommerce	Private sector
Arizona	16.7%	8.6%
Florida	18.2%	10.8%
Georgia	8.6%	10.4%
Nevada	24.6%	21.4%
New Hampshire	8.4%	13.4%
North Carolina	20.3%	11.5%
Pennsylvania	7.8%	12.4%
Wisconsin	10.2%	9.1%
Swing states	14.0%	11.2%
All other states	8.1%	10.6%

 

8. A strong economic narrative around digital growth in the swing states depends on wages as well. On average, tech-ecommerce jobs pay significantly more than the average for the private sector in every swing state. For example, in the second quarter of 2021, tech-ecommerce workers in swing states got paid at an average annual rate of $81,000, 39% more than the private sector average of $58,000. This includes the full range of tech-ecommerce jobs, from software developers to fulfillment center workers.

Table 3. Workers in Digital Jobs Are Paid More on Average
Annual pay, thousands of dollars, based on 21Q2
	Tech-ecommerce	Private sector
Arizona	76	59
Florida	81	57
Georgia	83	60
Nevada	65	57
New Hampshire	111	71
North Carolina	86	57
Pennsylvania	81	62
Wisconsin	77	53
Swing states	81	58
All other states	123	66
All other states except California and Washington	96	62

 

9. The core of the tech-ecommerce job boom in the swing states is the expansion of ecommerce jobs. These jobs pay well for high-school educated workers. Amazon pays its starting distribution workers an average of $18 per hour. That’s roughly comparable to starting manufacturing wages in many parts of the country. Overall, ecommerce industries pay about 30% more than brick-and-mortar retail in swing states, on average. This is at the heart of a powerful political narrative of growth that creates better jobs.

 

 

Table 4. Ecommerce Industries Pay More than Brick-and-Mortar Retail
	21Q2 pay in thousands at annual rates
	Ecommerce industries	Brick-and-mortar retail
Arizona	48	41
Florida	49	39
Georgia	45	36
Nevada	46	39
New Hampshire	57	38
North Carolina	40	35
Pennsylvania	51	34
Wisconsin	45	31
Swing states	47	37
Other states	62	38
Other states ex California and Washington	51	37

 

10. Democrats have a chance to build a powerful economic narrative around strong digital growth in swing states.  They should embrace this opportunity.

 

 

 

 

 

 

Clearly Americans are concerned about inflation. Prices for most food and energy products are soaring, which hits them right in the wallet.  The Biden Administration finds itself on the political defensive, as prices increases are outstripping wage increases for most workers.

But here’s an important piece of good news that is not getting enough attention. Inflation remains low in the digital sector, even as it accelerates across much of the economy.  Start with consumer inflation (as measured by the CPI). Over the past year, prices for digital consumer goods and services tracked by BLS (see graphic) have risen by only 1.9% overall, compared to 4.9% for CPI ex food/energy and 6.8% for total CPI.

Digital consumer goods and services include computers, smartphones and other IT commodities; video, audio, and music services; wired and wireless phone services; and internet services and electronic information providers.

The chart shows that digital consumer inflation has risen only 0.7 percentage points (PP) compared to November 2019, while consumer inflation ex food/energy has risen 2.6 PP. For example, the inflation rate for “internet services and electronic information providers” has only risen by 0.7 PP (from 1.5% to 2.3%).

When we look at producer prices, we see a very similar phenomenon:Producer price inflation in the digital sector is no higher today than it was before the pandemic. Over the past year, producer prices for digital goods and services tracked by BLS (see graphic) have risen by only 1.9%, compared to 7.7% for final demand ex food and energy.

Final demand inflation (ex food and energy) accelerated from 1.2% in the year ending Nov 2019 to 7.7% in the year ending Nov 2021. But digital producer price inflation was roughly constant at 1.9% in both periods (this calculation was done giving equal weights to each component).

 

 

Digital producer prices include computer and electronic manufacturing; electronic and mail order shopping services; software publishers; cable and other subscription programming; wired and wireless telecom; internet access; data processing; and internet publishing and web search advertising.

Some examples: The producer price of internet access services rose by only 0.3% over the past year, up only slightly from the -0.3% rate in November 2019. The producer price charged by software publishers fell by 0.3%, a slightly bigger decline compared to two years ago. The producer price index for telecommunications, both business and residential, is running at a 1.1% rate, slightly down from the 1.2% rate in November 2019.

The digital sector is exerting a dampening effect on both consumer and producer price inflation, reflecting large productivity-enhancing capital investments by digital companies. Meanwhile productivity has lagged in sectors such as food processing, where prices are skyrocketing.

However, the digital sector is not getting enough “credit” in the overall numbers for holding down inflation. For example, there’s no doubt that the digital sector is much more important to Americans today than it was in pre-pandemic 2019. Yet the digital sector gets a smaller weight in the CPI today than it did in 2019, because spending on digital goods and services is a smaller share of the consumer basket. That’s good news, but it has a perverse effect on the calculation of the overall inflation rate.

We may have to move towards a time-weighted versus expenditure-weighted CPI. Certain tasks, like dealing with the DMV or health insurers, are more “costly” in time than money. Conversely, digital services save you time that doesn’t show up in the official stats.

Time-weighted CPI would take into account that digitizing tasks, including shopping, generally reduces time spent by consumers, while adding more regulations generally increases time spent.

A time-weighted approach to CPI would likely show higher inflation for rural and poorer Americans. It would allow us to think about how cutting government bureaucracy and saving the time of Americans actually reduces inflation.

From the political perspective, the Biden Administration can make a strong case that the digital goods and services that are so important to Americans these days are barely rising in price. And America’s digital leadership will continue to hold down price increases once the temporary supply chain issues abate.