A new economic report by Raul Katz for WifiForward appears to inadvertently make a case for switching from Wi-Fi to Ethernet for better performance and a resulting boost to the US economy. The analysis is complicated, requiring us to connect some dots that Katz doesn’t want to connect, but the implications are clear. Let’s start with some background.

The FCC made two major spectrum allocation blunders in recent times: allocating 1,200 MHz to low power unlicensed across 6 GHz and putting low-power CBRS between two high power bands. Both errors are under review by the agency.

Cable industry front groups Spectrum for the Future and WifiForward are in high dudgeon over the reviews because they know the decisions were wrong and can’t stand up to scrutiny. There’s nothing inherently wrong with helping Wi-Fi along or with experimenting on novel spectrum management models, of course, but these cases both have significant flaws.

Both 6 GHz Wi-Fi and CBRS rely on spectrum management agents – SAS for CBRS and AFC for 6 GHz – designed by the Wireless Innovation Forum that respond to queries from registered and authenticated users for information about available channels and their permitted power levels. This is good, a feature that Wi-Fi should have had from the beginning. I have no quarrel with SAS and AFC, but technology only gets us so far with spectrum policy.

The Economics Don’t Add Up

The front groups are minimally interested in what’s good for the private 5G networks enabled by CBRS and the performance gains achieved by Wi-Fi 7 users as such. Rather, they appear to regard these initiatives as means to deprive 5G Fixed Wireless Access (FWA) of spectrum that would enable faster rollout of competitive broadband services.

Cable modem is losing market share to FWA and that’s not good for cable company share prices. WifiForward’s response to the cable modem crisis includes funding for a pair of quasi-economic studies by Raul Katz’s Telecom Advisory Services, Assessing the economic value of Wi-Fi in the United States (2024) and Economic Loss If the Top 700 Megahertz of the 6 GHz Band is Repurposed for Licensed Use (2025.)

Both studies rely on an economic model that finds a posits a precise, linear causal relationship between the download speed and latency of broadband services and GDP per capita [Appendix B of Assessing]:

• A 10% increase in download speed results in a 0.1956% increase in GDP per capita

• A 10% decrease in latency results in a 0.4503% increase in GDP per capita

Katz pulls speed and latency figures out of Speedtest, applies them to Wi-Fi 7, and claims Wi-Fi offload of mobile device traffic automatically gooses GDP regardless of application requirements or any other trivial details. In fact, the biggest consumer of wireless traffic is video streaming, an application whose speed is throttled by the source (Netflix, Amazon, etc.) to conserve server resources.

How Video Streaming Actually Works

Higher resolution means more CPU cycles and storage accesses, both of which cost money. Hence, video streaming services compress content. Here are YouTube’s recommended broadband speeds per format:

Video Resolution	Recommended sustained speed
4K UHD	20 Mbps
HD 1080p	5 Mbps
HD 720p	2.5 Mbps
SD 480p	1.1 Mbps
SD 360p	0.7 Mbps

Hence, increasing the speed of your broadband offering from 50 to 500 Mbps or 5 Gbps has no impact on your streaming experience or the economy. All that matters is that you’re able to meet the minimum requirement, even if watching video in ultra-high resolution actually made you more productive. For this application, connection capacities above 20 Mbps per device go to waste.

Broadband advocates of many stripes have been trying to connect broadband adoption and speed with GDP growth for a very long time, but the intuitive premise remains unmeasurable at the high end of the speed scale even though it feels plausible. Adoption matters and minimum acceptable speed matters, but there is no incremental benefit at the extreme high end of the bandwidth spectrum except for very unusual applications.

Unclear Causality

The relationship between broadband quality and economic output is certainly non-linear.  Something is massively greater than nothing, but having the world’s highest performance doesn’t increase output that wasn’t constrained by poor broadband in the first place.

Moreover, the direction of the causality between income and broadband speed is ambiguous at best. As the World Bank puts it:

Broadband penetration is generally modeled as a function of price and income in the demand equation but is awkwardly linked to the output equation to account for causality. Other models estimate broadband penetration as a function of other factors and use that as the independent variable in an effort to weaken the correlation between income and broadband penetration. Neither is satisfactory and while econometric models are useful tools for estimating relationships, they cannot prove causation. [emphasis added]

But this sentiment doesn’t keep advocates and lobbyists from trying to establish a set of facts favorable to their clients even if their details are squishy.

Reacting to Wi-Fi Slowdown

Setting aside our quibble with the false precision of the Katz GDP estimator, it is a truth universally acknowledged that fast networks are better than slow ones. While Economic Loss If the Top 700 Megahertz of the 6 GHz Band is Repurposed sees nothing but horror coming from a narrower 6 GHz Wi-Fi band, reality tends to be more complex than economic modeling.

Assuming that Wi-Fi will slow down in the future (despite the best efforts of wireless engineers to make it more efficient,) are we justified in assuming that the entire broadband experience for consumers and enterprises will automatically suffer? This could be the case for applications that absolutely depend on Wi-Fi, but such applications are not nearly as prevalent as Wi-Fi connections used where they’re not really needed.

Dean Bubley, a favorite analyst of the “shared” spectrum crowd, offered a sober analysis of the most intense data consuming app for Wi-Fi in a LinkedIn post:

However, being even-handed, I also need to call out a growing habit in Wi-Fi that needs to change: continually referencing its use for 90% of consumer Internet traffic.

Here, the “yes but” is the huge role of Wi-Fi for connecting large-screen TVs in the home for streaming, which I believe accounts for a large % of that impressive figure. An hour of Netflix or YouTube in the living room can mean 1-4GB of data. Of a typical household’s 500GB per month, a significant chunk is just streamed video.

Often, TVs are positioned very close to the home gateway or WiFi router, and to be honest that could be “offloaded” with an Ethernet cable (remember those?). We rightly talk about offloading cellular traffic to Wi-Fi for efficiency – this is a similar concept.

In fact, all or nearly all Wi-Fi traffic is already offloaded to Ethernet at the Access Point. It’s dead simple to connect a TV set, AV receiver, or video streaming box to a nearby gateway router with one of the fabulous new light and flexible SlimRun Ethernet cables. Homes and enterprises with disappointing Wi-Fi performance do so today.

Less Reliance on Wi-Fi Means a Better Economy

The Katz analysis fails to recognize that Wi-Fi is an optional intermediary that is significantly overused today. It can’t reach the Internet until it offloads its traffic onto the local Ethernet for carriage to the cable modem or fiber optical network terminator that connects the premise to an ISP. [Quibble:  A possible exception is all-in-one Internet gateways that combine a modem, an Ethernet switch, and a Wi-Fi access point in a single box. But the internal circuits in these devices route Wi-Fi traffic through the Ethernet switch chip.]  

Because Wi-Fi offloads onto Ethernet, Wi-Fi traffic can never move faster than Ethernet. Choosing Wi-Fi over Ethernet is often a choice made out of convenience rather than necessity. If speed and latency are as important as Katz asserts, we’re better off using Ethernet without the intermediary where we can.

If we take Katz’s claim that high-speed local networks boost GDP at face value, then nothing is more patriotic than an extensive wired network inside every home and office.

Wi-Fi Takes a Big Toll on Quality

For illustrative purposes, I tested a Wi-Fi 7-native Acer Swiftgo 16 laptop and an iPhone 15 Pro Max over Wi-Fi and over Ethernet with both Speedtest and iPerf3. In general, Ethernet is slightly more than twice as fast as Wi-Fi with less latency to boot.

Test	Interface	Download	Upload	Latency
Speedtest
Laptop	Wi-Fi 7	2,562 Mbps	1,643 Mbps	5 ms
	Ethernet	2,906 Mbps	3,240 Mbps	3 ms
iPhone	Wi-Fi 6E	1,420 Mbps	1,230 Mbps	4 ms
	Ethernet	3,120 Mbps	2,590 Mbps	3 ms
iPerf3
Laptop	Wi-Fi 7	1,790 Mbps	1,800 Mbps	ND
	Ethernet	4,530 Mbps	4,530 Mbps	ND
iPhone	Wi-Fi 6E	1,427 Mbps	1,004 Mbps	ND
	Ethernet	3,483 Mbps	1,647 Mbps	ND

The Internet connection under test is 3 Gbps symmetrical FTTH by Quantum Fiber/Century Link. [iPerf uses a local server, hence it doesn’t touch the Internet.] The Ethernet connection uses a plug-and-play UGREEN 5 Gbps USB C adapter.

Cost and Convenience of Ethernet Wiring

Contrary to Katz’s claim that USB Ethernet adapters require users to: “turnoff the Wi-Fi, refresh the Internet page and then access the Internet” they simply need to be connected to Ethernet and plugged into the device. Operating systems are smart enough to automatically choose the wired connection over the wireless one when both are available.

It may be useful to compare these figures with the analysis I did with both desktop and laptop computers over various generations of Wi-Fi from 4 – 7 for a picture of peak Wi-Fi performance with iPerf3. Pure Ethernet performance of desktop machines was ~9.36 Gbps, as fast as the adapters can handle, while Wi-Fi 7 downloads peaked at 3,000 Mbps on desktops and 2,140 Mbps on a laptop with native Wi-Fi 7.

Katz massively over-estimates the cost and difficulty of installing Ethernet cable, declaring the “national average for wiring a 2-room residence with CAT 6 is $660.” That citationless claim must include some hefty labor charges as contractors generally fetch $100 – 200 per room for wiring entire houses with Ethernet cable according to the Reddit Home Networking sub.  

There’s no Avoiding Ethernet in a Typical Home

Setting up a Wi-Fi network for a home with more than two rooms requires more than one Wi-Fi access point (AP); one per two rooms is a good rule of thumb. Each AP requires Ethernet backhaul for peak performance. Such homes can’t readily forgo Ethernet in any case.

Not only is Ethernet essential for Wi-Fi backhaul and useful for video streaming, it’s the most rational way to connect video cameras to a home network. Cameras need power as well as data transfer, which can be supplied by Power over Ethernet (PoE) with an appropriate switch.

PoE also provides power to the ceiling-mounted Wi-Fi access points in a well-designed home network as well as to satellite Ethernet switches that serve multiple devices room-by-room for as little as $30.

Conclusion

Networks are destined to be hybrids of wire and wireless devices for the foreseeable future. Ethernet is the true workhorse of residential and enterprise networks whether they also use Wi-Fi or not. The day will come when wireless is the connection of choice for most information processing devices, but that doesn’t mean we abandon wire altogether. Connecting North America to Europe and Asia is always going to be tricky.

Katz and similar advocates prey on the public’s lack of understanding of key technical concepts such as Wi-Fi, Ethernet, Internet Service, and economics. Some Zoomers call Internet Service “Wi-Fi” because they don’t know what Wi-Fi is made of. Katz himself is reluctant to admit that Wi-Fi is nothing without Ethernet tying the elements of Wi-Fi networks together.

The FCC’s historical concerns about Wi-Fi congestion stem from a lack of understanding of the role Wi-Fi should play in home and enterprise networks, all of which are hybrids. If we stick to using Wi-Fi to connect genuinely mobile devices and we continue to improve Wi-Fi radio performance, there will never be a need for Wi-Fi to use the upper 6 GHz; the development arc for unlicensed spectrum is in the higher frequencies that don’t generate as much co-channel interference as Wi-Fi generates today.   

Properly understood, Economic Loss If the Top 700 Megahertz of the 6 GHz Band is Repurposed for Licensed Use makes a strong case for increased reliance on wired networks simply because they’re faster, cheaper, more reliable, and more responsive than Wi-Fi. That much of the Katz study for WifiForward is valuable.