Causation is Clear for Short-Term Rentals, Less Clear for Broadband

“Who benefits and who loses?” is a reasonable question for analyses of public policy and economic studies. As with all questions related to public policy and economics, though, correlation is easier to demonstrate than causation, simply because the number of variables is so great. 

In the connectivity business, the issue is most common in analyses of broadband impact on economic growth, household income or job growth. We assume broadband “causes” economic growth, but correlation is not causation.  

We assume quality broadband is associated with job growth, but cannot prove causation. We assume better mobile broadband also “causes” economic growth, so we might also believe 5G will similarly cause growth. We might be wrong. 

It also is plausible that areas with strong economic growth, above-average household income and wages, higher educational levels and wealth create demand for better broadband. In other words, demand for broadband is a result of strong economic growth, rather than its cause. 

Other cases arguably have stronger causation relationships. Consider the argument that use of short-term rental apps causes a reduced supply of housing and higher housing costs. It seems plausible for the simple reason that housing vacancies in any market tend to be fixed, and rental property managers can make choices about how to market their rentals: short term or long term. 

The issue is how much difference a robust short-term rental has. It also seems plausible that the biggest impact should be in “touristy” areas where there is high demand for short-term housing. Areas with modest tourist demand should also have modest demand for short-term housing. 

It also seems plausible that the greatest effect is in “touristy” areas that also are affluent. 

Airbnb leads to a reduced supply of housing as properties shift from serving local residents to serving Airbnb travelers, which hurts local residents by raising housing costs, according to Josh Bivens, Economic Policy Institute director of research. 

That assessment seems mirrored by some other studies. Short-term rentals using apps such as Airbnb contribute to housing shortages and rent increases, according to Felix Mindl and Dr. Oliver Arentz, researchers at University of Cologne in Germany. 

They attributed 14.2 percent of overall rent increases to short-term rentals or 320 euros ($385) per year for new tenants.

“While a large proportion of hosts can be considered home sharers, we find an increasing proportion of providers who have developed a professional business model from short-term rentals,” Mindl said in a statement. “Professional short-term rentals are available to tourists throughout the year, and thus compete directly with long-term tenants, for whom the rooms are then no longer available.”

Researchers also have found that in local neighborhoods with a lower share of owner-occupancy, Airbnb had a higher impact on rising housing prices and rents. In areas with a higher share of owner-occupancy, Airbnb had somewhat less of an impact on property prices and rents.

The study also found that the total supply of housing was not affected by the entry of an Airbnb property in a given neighborhood, and that Airbnb listings tend to increase the supply of short-term rental units, while contributing to a decrease of the supply of long-term rental units.

Aside from the presumed effect on housing, short-term rental apps also shift revenue between lodging suppliers. As with sports stadiums, which arguably mostly shift spending from one form of entertainment to another, short-term rental apps shift revenue from hotels to individual property owners.

“The most obvious benefit stemming from the creation and expansion of Airbnb accrues to property owners who have units to rent,” EPI noted.

There are other issues, though. The housing market is affected by forces other than Airbnb, such as gentrification and economic trends. A one-percent increase in Airbnb listings is causally associated with a 0.018 percent increase in rental rates and a 0.026% increase in house prices, other researchers argue. 

“In aggregate, the growth in home-sharing through Airbnb contributes to about one fifth of the average annual increase in U.S. rents and about one-seventh of the average annual increase in U.S. housing prices,” say researchers Kyle Barron, Edward Kung and Davide Proserpio. 

In contrast, annual zip code demographic changes and general city trends contribute about three fourths of the total rent growth and about three fourths of the total housing price growth.

“These results translate to an annual increase of $9 in monthly rent and $1,800 in house prices for the median zipcode in our data,” they say.

The biggest impact comes if a long-term rental unit is converted to a short-term rental unit on a full-time basis, as that subtracts one living unit from the long-term rental stock. On the other hand, an owner-occupied home that rents a room in that house does not do so.

How Might AI affect Trusted Advisor Businesses?

Information disparity explains why enterprises and smaller businesses buying information technology use consultants and system integrators. IT decisions often are complex, almost always lie outside the buyer’s domain expertise and can be significant choices with long-term consequences. 

Some 64 percent of 500 respondents to an Avant Communications poll of U.S. based enterprise decision makers use third party consultants, for example. 

source: Avant Communications 

The role of the advisors is to reduce buyer risk by providing expert advice, and therefore is an “information” service. At a high level, that raises the question of what happens--eventually--as artificial intelligence advances.

source: Avant Communications 

Predictably, experts in fields ranging from finance to law to  management consulting argue that AI cannot replace human judgment and insight. Others point to growing AI roles and argue AI will replace trusted advisors. To be sure, most expect AI displacement for any jobs with routine tasks.

The issue is that AI will become more capable over time, allowing AI systems to potentially displace many more functions at a higher level, even when those human-supplied functions involve “advice.” 

Consider the advice consultants can provide. Much of that advice concerns the choices buyers should be making to place advertising, buy one technology over another, evaluate the life cycle costs of such choices or use cloud computing in place of premises computing on owned hardware, for example. 

In the connectivity business, many of the buyer choices involve MPLS, SD-WAN, cloud computing, security or unified communications. But the value of AI is, in part, the ability to automate any rule-based systems. And the choices between MPLS and SD-WAN are largely rule-based, in terms of total cost of ownership. 

Increasingly, the total cost of ownership choices between cloud computing and do-it-yourself computing are understood and can be routinized. That does not mean complete replacement of trusted advisors, but arguably moves the value of those advisors “up the complexity stack” to areas where benefits and costs are more intangible. 

To the extent that information disparity creates the value added by third party trusted advisors, AI is going to erode the extent of that information disadvantage, and therefore alter the places where advisors can add value. 

So AI will start to replace the lower-order information mismatches at first, then gradually begin to replace more-complex mismatches over time. McKinsey studies suggest AI will replace parts of tasks, perhaps 30 percent of job tasks by about 2030. That might not fully replace jobs, but will change them. 

But most expect the changes to deepen over time, as AI becomes more capable, and can handle more complex information-related tasks. The shift of U.S. jobs from farming to other pursuits did not happen immediately, either, and was not directly affected by information technology as by mechanical technology. But it happened. 

Some 90 percent of U.S. residents once worked on farms. Today about one percent do so. Digital transformation, defined as the use of digital technologies to create new--or modify existing--business processes, culture, and customer experiences to meet changing business and market requirements, has been going on since the early 1980s, for example. 

Few students today realize that, in 1977, nobody used a personal computer at school. Nobody used the internet or a mobile phone. In 1984 perhaps eight percent of U.S. homes had a computer.  

Businesses did not routinely begin using PCs until after IBM introduced its first IBM PC in 1981. 

The point is that AI, as have other forms of digital transformation, starts slow. But transformation does eventually happen. People do not work, learn or live the same way. Eventually, a wide range of tasks now performed by trusted advisors will be handled by AI-based apps and services. 

“Advice” will be among those tasks. The only issue is how much advice can be provided by algorithmic systems.

Videoconferencing Makes a Breakthrough

If you have followed predictions for use of videoconferencing or telepresence for any length of time, you know that expectations have been high, both for services as well as business instructure. AT&T marketed its Picturephone service as early as the 1970s, with modest uptake. 

By some estimates, AT&T invested $6.6 billion to $6.7 billion in current currency to develop Picturephone, only to shut it down because there was no sustainable business model. 

These days, much videoconferencing--both consumer and business use cases--is an app, not a service. But it is noteworthy that consumer and business use of videoconferencing made a huge leap during the Covid-19 stay-at-home and work-from-home policies, virtually doubling in the spring of 2020.

source: U.K. Office of National Statistics, Benedict Evans  

You would be hard pressed to find any single quarter, in the history of videoconferencing, when usage climbed that much.

Good Illustration of Covid-19 Impact on Underlying Trends

This is a good example of the impact of Covid-19 economic lockdowns on technology adoption in general, which is that the economic lockdowns and work-from-home policies pushed any number of existing trends to a sharply-higher level. 

This data shows e-commerce share of total retail sales, excluding autos, fuel, restaurant activity, got a sharp bump because of the pandemic, presumably shifting the curve upwards, but possibly not altering the long-term growth rate. 

source: Benedict Evans 

This illustrates what is meant when observers say “we saw a year’s worth of growth in a few months.” 

What is not clear is whether the underlying long term rate of growth has been altered, or not.

Some Covid-19 Winners Boosted AI Spending Because Financial Advantage was Seen

In virtually every recession or economic panic, some firms, in some industries, are able to boost sales, take market share and maintain or even boost profit margins. Often known as high-performing firms, there might not be a single clear reason for success.

The firms could be in growing young markets, have leadership or capital access advantages, have made astute technology investments, or simply be in industries that benefit from the particular crises or recessions.

Likewise, some firms, across industries, seem to have continued to invest in artificial intelligence during the Covid-19 pandemic, and those firms seem to have done so because they saw clear advantages to use of AI in ways that helped their business models.

If the Covid-19 pandemic affected enterprise information technology investments, it arguably has slowed such investments at some firms, which have had to shift support to remote workers. On the other hand, some firms who already have found use cases, and invested more heavily prior to the pandemic, seem to have increased their investment level, a  McKinsey survey found.

Respondents from 61 percent of firms who report success with AI also say their firms increased investment in 2020. Patterns across industries show big variations.

 source: McKinsey

Firms in healthcare, pharma, medical products; as well as companies in the automotive industry were most likely to have increased AI investments in 2020. 

 source: McKinsey

Most enterprises likely have not yet found clear financial benefits from AI deployment, though. A survey of more than 3,000 company managers about their AI spend found just 10 percent had gotten significant financial benefits from their investment so far, a report from MIT Sloan Management Review and Boston Consulting Group found. 

A separate survey of U.K. firms found that 40 percent of 750 surveyed U.K. executives plan to invest in artificial intelligence in 2021, a survey by Fountech Solutions finds. 

  

Some 30 percent of respondents say their firms piloted an AI solution for the first time since the onset of the Covid-19 pandemic. New AI specialists will be hired by 41 percent of respondent firms. Also, 48 percent of respondents say their companies will seek AI training for existing staff.

As a rule, some firms managed to grow revenue and profit during recessions and crises, Boston Consulting Group data suggests. While 44 percent of firms might experience shrinking profit margins and sales growth in a recession or crisis, 14 percent have shown growth in both sales and profits. 

Some 28 percent of firms see lower sales but manage to increase profit margins. About 14 percent of firms see higher sales and lower profit margins. 

Firms in health and consumer staples are most likely to see winners in recessions. Companies in energy, information and communications technology and financial industries are least likely to emerge with higher sales and profits in a recession.

source: Boston Consulting Group 

It is possible--even likely--that firms continuing to invest in AI during the Covid-19 pandemic were already finding themselves gaining market share, increasing sales volume and maintaining or increasing profits. 

The McKinsey survey found, for example, that a small number of respondents at some firms attributed 20 percent or more of their firm earnings before interest and taxes (EBIT) to AI. Those companies planned to invest even more in AI during the COVID-19 pandemic. 

That is in keeping with the BCG data suggesting some firms gain market share and boost sales during recessions and crises. If such gains are attributed to AI, it makes sense that firms would maintain or boost such investments.

When Telcos Discount Prices, What are They Protecting, What are They Merchandising?

Some products are attractive for attackers in the fixed networks business for reasons of gross revenue. For T-Mobile, the attraction of home broadband includes average revenue per account that is double that of a mobile account, as well as ability to take market share. T-Mobile has zero share of the home broadband business.

In other cases, defenders are motivated by profit margins, as much as gross revenue. For ISPs, linear video entertainment profit margins have fallen dramatically, while margins for voice and internet access have remained higher.

Without adjusting for differences in currency, internet service providers in the U.S. market might arguably have an advantage over ISPs in some other markets: the average revenue per user is as much as three times the ARPU of the same service in South Korea, or double the revenue in Europe. 

source: ETNO, Analysys Mason 

That might be an illusory advantage.

Adjusting prices using the percentage of gross national income method, U.S. prices in 2014 were among the lowest in the world. Adjusting revenue using the purchasing power parity method, global prices hover around $50 a month. 

There are other adjustments, though. In the U.S. market, 60 percent to 75 percent of internet access plans are bought in a bundle, so there is no way to directly state the internet access price. Price has to be inferred. 

The point is that such bundles always offer lower prices on the product elements purchased. So the “effective price” paid for internet access--purchased in such bundles--is discounted from the posted retail prices for internet access purchased on a stand-alone basis. 

Precisely how big a discount might exist is contingent on the assumptions one makes about each of the constituent elements of the bundle. Some customers buy a triple-play bundle including a service they actually do not want or use (voice), in order to get lower prices on video and internet access services. 

Also, each service provider will make different allocations based on profit margin protection as well as gross revenue priorities. Cable operators arguably used to merchandise voice or internet access to protect linear video revenue. These days, they might lean to protecting internet access revenue and margin and merchandise voice and video to a greater extent. 

Telcos might arguably have protected voice revenue and margin while merchandising either video or internet access. These days, it is not clear what is merchandised. Profit margins arguably still are highest for voice and internet access, so video is the likely candidate for discounts, up to a point. 

Cost of goods is the floor for discounting wiggle room, suggesting the value of streaming versus linear product offers. Even when content costs are equivalent, the other costs of fulfilling a video account are lower when the streaming service is offered, compared to a linear video service. There are avoided truck rolls and customer premises equipment savings, for example.

Business Models Matter More than Access Media

Though the standard prescription for better broadband globally is fiber to the premises, there are some significant differences in a few countries. Looking at where gigabit internet access speeds are now available, In the United States 80 percent of locations are reached by cable operators. About 25 percent of telco FTTH homes supports those speeds, Analysys Mason data indicates. 

In South Korea, about 60 percent of homes can buy gigabit service. About 80 percent of homes served by telcos can do so. In Japan, nearly the same percentage of cable homes can buy gigabit service, while 75 percent of telco homes can do so. 

In Europe, about 40 percent of homes can buy cable gigabit service from a cable operator, compared to about 25 percent of homes able to buy gigabit service from a telco. 

source: ETNO, Analysys Mason 

Two points are noteworthy in this regard. FTTH and HFC refer only to access media. Use of either media does not mean “gigabit per second speeds.” Cable networks also can do so. But most U.S. FTTH networks are not yet supporting gigabit speeds. 

The point is that the traditional telco framing of the FTTH deployment case is about access media, not speeds. If speed, and coverage, are the issues, then hybrid fiber coax often is a major--if not the leading--platform. 

Future proofing also is an issue. Still HFC architects have successfully boosted speeds to gigabit ranges, with a roadmap to 10 Gbps speeds and higher speeds (up to 100 Gbps over the next decade), before the platform possibly reaches a limit. 

As a practical matter, one might ask whether the cable HFC business model ever reaches a point of limits over the next few decades, if speeds can be pushed to 100 Gbps, and made more symmetrical. 

The issue is not simply speed, but what it costs competitors to invest in platforms to do so, what the expected take rates might be, and what the business model therefore delivers, in financial terms, when most consumers rely on mobile service, and mobility drives total revenue and profit. 

The existence of strong cable competition in some markets necessarily limits the financial return any leading telco can expect from new FTTH deployment, and increases the risk of substantial stranded assets which produce zero return. 

In deployments to date, telco FTTH networks have struggled to exceed 40 percent take rates, which means 60 percent of the assets serving consumers are stranded. Conversely, only about 30 percent of cable assets typically are stranded. 

As always, the better mousetrap does not always win, assuming FTTH is deemed the better technology than hybrid fiber coax. The HFC upgrade path seems always to have been more incremental and more graceful (financially), as FTTH is “rip and replace.”

It remains true that for a legacy telco, FTTH remains a “better” technology choice than copper access. Whether it always is the better business decision remains the issue.