Today internet infrastructure encompasses root servers, broadband lines, routers, content delivery net-works, cloud storage and cellular towers. Broadly construed, these physical assets perform two related and essential services for the modern digital economy. Infrastructure acts as an intermediate input for the production of many services by firms and it acts as an intermediate input into the delivery of access and related services to the internet for households.
The improvements to infrastructure receive less heralding than a new and shiny app or device. Fewer financial analysts examine every imaginative aspect of clever business decisions. Nonetheless, improvements arrive apace, both motivated by and enabling advances in a plethora of new devices and platforms.
Today’s column will take a step back and marvel at the economic logic behind these long term changes. Let me issue apologies in advance for skirting past a ton of technical details. The column will try to make a complex economic topic digestible by focusing on “trends” in the recent past.
SHOW ME THE MONEY
Trend 1. Value chains grew. A value chain is a set of activities that together produce an outcome that users purchase and consume. Most networks support a chain around a single output – e.g., the electrical grid produces and delivers electricity. In contrast, internet infrastructure supports several chains, so it is challenging to estimate the value of the markets it supports.
Broadband internet access to homes and businesses is one important value supported by infrastructure. Official GDP statistics in the Service Annual Survey, collected and archived by the US Census, show enormous levels and growth in access fees. In 2017, the last year in which data is available, payments for access in wireline forms contributed over $88.7 billion to GDP, growing more than 30% from 2012 (in nominal terms). Payments for access in wireless forms amounted to over $96.0 billion in 2017, growing more than 57% from 2012.
A related value chain exists for electronic commerce in the same years. It has many parts. Official GDP statistics find online advertising contributed $105.9 billion to GDP in 2017 among Internet Publishing and Broadcasting and Web Search Portals. That has grown 250%. Another part of this chain involves electronic retailing, which the Census puts at over $545 billion for electronic shopping and mail order houses. It grew 65%.
That is not every value chain linked to the internet, of course, but it is enough to illustrate that hundreds of billions of dollars depend on the infrastructure, and this activity has grown tremendously in recent experience. It also illustrates a difference with the speculation of the dot.com era. Today’s investments support valuable services that users regularly buy. Even if every over-optimistic entrepreneur disappeared tomorrow, there would still be a large amount of value affiliated with the internet.
Trend 2. Generativity thrived. Internet infrastructure supports too many products and services to list. Smart phones helped supported a boom in apps, and massive change in platforms. Video and streaming worked their way into every device, altering music and video services, not to mention online advertising. Broadband provides the best experience for streaming movies, and numerous over-the-top services have been proposed and some have been widely adopted by users.
These experiments arrived in fits and starts, and punctuated events with new rollouts and inventive new products. Some of this went mainstream and some stayed with cutting-edge users. There is not enough space to mention even a tiny fraction of them. There is, however, a simple but obvious point to stress: Those experiments and the increase in the value chain could not have happened without associated improvements in the operations of infrastructure. Users experienced more resolution in their video, less waiting time for rendering, and faster reactions in their applications.
Notably, household and businesses paid higher prices for faster access in their wireline and wireless services. That is why the value chains grew in size. Similarly, investments in infrastructure grew where the money flowed, especially where users paid for the improved experience.
NEW VERSUS OLD ARCHITECTURE
Trend 3. Users increasingly received more data than they sent. Networking engineers today talk about ever increasing “asymmetric flows” in data. While the network flows were never perfectly symmetric, the engineers have a point. Blame the increasing popularity of streaming and video. Blame all those smart phones. Blame all the new addictive apps. Flows have steadily, almost inexorably, gotten more asymmetric each year.
These changes were enabled by, and led to the growth of, content data networks (CDNs). CDNs are mirror sites, which put data closer to users in order to reduce delays in delivery. Akamai is the largest third party provider of these services. Large content providers, such as Google and Netflix, also operate their own CDNs to optimize performance to their unique needs. Many cloud providers offer related services, and carriers have entered with some services too.
The incentives of content firms, cloud providers, and carriers align most of the time, but not all of the time. All benefit from growth in the value chains, but each wants a bigger share of some part. This cooperative and competitive stance will continue to be among the most important tensions in the near future.
Trend 4. The new architectural models outcompeted the old. Three competitive outcomes are most salient. First, and visibly, dial-up access long ago lost relevance for most users, who chose faster access through broadband and wireless channels. Less visible, and related, newer forms of access required rights of way, and expensive delivery systems, and management of major capital investments. The new access channels typically exhibited economies of scale in local markets, which resulted in limited options for users in many locations. In spite of the limited options and higher prices, users still preferred broadband over dialup because it was just that much better.
Second, and less visible, the demand for backbone services shrunk in the face of capital deepening around processes that use CDNs and cloud storage. Those also shaped interconnection on the internet. Most interconnection does not take place at quasi-public points of exchange where backbone lines met. Private arrangements for interconnection have become pervasive, and now dominate activity.
The third change is the most remarkable (to me). For all intent and purposes, private firms operate internet infrastructure, while public or non-profit organizations retain responsibility for its interconnected pieces. For example, the World Wide Web Consortium managers upgrades to HTML and related software, the Internet Society oversees upgrades to TCP/IP, ICANN oversees the domain name system, and the IEEE committees that design new upgrades to WiFi. These public organizations design standards, charge little for their services, and, simultaneously, have little away over daily operations for the internet.
SYSTEMIC PROPERTIES
Trend 5. Interconnectedness led to economic interdependence. Interconnectedness has led the economic fortunes of one sector to become linked to another. For example, the rise of the smart phone generated demand for more cellular towers and demand for programmers to develop new applications. More broadly, these links in fortunes are welcome when economic growth in one causes growth in another, but they are less welcome when declines propagate from one to another.
Interconnectedness also creates systemic risks from externalities. Does the internet contain enough redundancy to localize failures during potentially catastrophic natural disasters? Does the geographic spread of interconnection help the internet survive a targeted terrorist attack? Why would we expect the entire system to possess desirable answers to these troubling questions?
Basic economic reasoning suggests the root of the problem. Privatization of investment fosters in-sufficient incentives for protecting the entire system. For example, negative externalities arise because one firm’s (under)investments in security leads to another’s problematic experience. When defenses fail at a poorly managed firm – say, against a virus or a sophisticated hack –many others pay a price for their neglect and/or incompetence.
A major open question today is whether a government-led effort, a private led effort, or non-for-profit organization will emerge to manage these externalities. While there are many examples of initiatives, none have resolved these types of issues.
Trend 6. Governments slouch towards a splintered networks. It was a utopian dream to believe the entire world would want to govern the internet according the same principles. Indeed, the opposite has emerged. Different governments have asserted authority to treat the internet as they see fit in their own home countries. Many have declared no interest in following the US lead.
This has planted the seeds for reducing the degree of interconnectedness, or splintering the net-work. A determined government can already do this, and some have. For example, China has limited the ports at which data can enter and leave the country, and put up a large software firewall between the domestic network and the rest of the globe. As another example, the EU is determined to impose its privacy principles on practices within its borders, and those differ from the mandates in other parts of the globe. Data centers have been built within European borders to support these mandates.
CONCLUSION
The internet does not resemble any other network in the modern economy. Broadcasting and electricity simply do not work the same way, for example.
The primary driver of US infrastructure investment has been the satisfaction of user needs by privately managed firms. Firms largely provide services in relations to the comparative revenues and costs, and invest in infrastructure that supports their projections about revenues and costs.
Despite its optimization to user needs, the present system is rife with collective issues. It is also too soon to declare whether recent expansion in the value chains is a bridge to even more prosperity, or merely the last instantiation of rapid growth.
Copyright held by IEEE. To view the original essay click here.
Digitopoly 