Skip Navigation
Get updates:

We respect your privacy

Thanks for signing up!

Hal Singer responded to our read on his latest bad think-tank-caliber “analysis.” He asks “Are we still bickering about whether the Open Internet Order had a negative impact on ISP investment?”

Well, yes, at least he’s bickering about it. And we’re replying yet again because Singer is misleading his readers to scare them.

The truth is that no one — not even a “progressive” economist who presents his work these days at such celebrated “progressive” organizations as AEI — can discern an impact that the FCC’s order had on investment.

Yet Singer pretends he can. He says he can calculate the magnitude of that impact, and claims it’s more than enough to swamp the benefits of the Net Neutrality rules.

He also suggests he can isolate the cause for this alleged investment calamity in a part of the law the FCC chose not to enforce. Then he claims this unenforced provision in Title II would impact companies to which that provision likely would not apply even if the FCC had kept it on the books.

In his latest broadside Singer even accuses Free Press of what he’s expert in himself: selective bias. He falsely charges us with making “pivotal mistakes” and errors in our analysis. He glosses over examples that don’t fit his narrative, ignoring the mountains of evidence that don’t support his theory.

Singer also stretches minor pronouncements the FCC made in conjunction with its decision so that he can pretend the entire value of the Net Neutrality rules — which protect the Internet economy for everyone — amounts to no more than $100 million. He sets that number against an amount for the supposed drop in investment that he picks out of a hat and imagines as Title II capital flight.

Singer can condescend all he wants about how economists assess data as he himself ignores data that disprove his claims. He can sneer at Free Press and call us (note the scare quotes) the FCC’s “economists” — in the process ignoring the existence of the FCC’s actual chief economists — while he plays “lawyer” and shows his ignorance of the law.

But none of this does anything to rescue his arguments. 

The facts that Singer doesn’t and can’t dispute

  • The first full accounting period following the FCC’s vote was the second quarter of 2015. Capital investments at the top publicly traded ISPs (those shown in the table included with our earlier post) were higher following the FCC’s vote than they were during the same accounting period the prior year, meaning the second quarter of 2014. For the 18 publicly traded ISPs tracked, capital spending was higher for 10 of the 18 during second-quarter 2015 than it was during second-quarter 2014.
  • Capital investment is up at the large cable ISPs collectively, whether we look at the second-quarter to second-quarter comparison or expand it to a first-half-of-2015 versus first-half-of-2014 comparison.
  • Among those cable companies, network investments (excluding money spent on cable set-top boxes and other such customer premises equipment, or “CPE”) are up by double-digit percentages.
  • For big wireless companies, capital investment is up (again, whether we look at the first-half or second-quarter numbers) at Verizon, T-Mobile and Sprint. That’s true even without Sprint’s new capital purchases stemming from its leased handset program, but those handset purchases are real capital investments that Singer wrongly says shouldn’t count (more on those below). Wireless investment went down only at AT&T Wireless, returning to a lower level after temporarily rising due to AT&T’s “Project VIP” deployments.
  • Capital investments are up at numerous smaller ILECs too, including Frontier and Cincinnati Bell. 

Singer ignores all of these facts. His silence speaks volumes. It’s a strong indication that what matters to him is making the case he’s been charged with making, no matter whether the facts fit.

But this data invalidates his claim that investment has declined across the board and that some of that’s just gotta be from Title II.

The decline is there only if you cherry-pick the companies in the sample, fine-tune the time periods analyzed, and, most importantly, ignore the companies’ own explanations for these changes. Singer excels at all of these tactics.

Those are some of his main sins of omission. Now, on to what he did say.

Comcast doesn’t count for much in Singer’s world 

We said Singer’s initial article in Forbes failed to account for Comcast, because it did. That article selectively highlighted investment numbers for AT&T, Verizon (ignoring its wireless business, where spending increased), CenturyLink, Charter and Cablevision.

We explained last week how Singer hid the explanation for the declines in those instances: These companies made investments in prior years, completed their upgrades, and like rational businesses didn’t just keep spending money for the sake of spending it on things they’d already paid for.

Singer’s presentation to his “progressive” friends at AEI did list Comcast’s marked increase in investment. This increase is fueled in part by Comcast’s decision to undertake the largest, fastest deployment of fiber-to-the-home in history, in a plan that was announced after the FCC’s Open Internet vote.

Yet in Singer’s Forbes piece, all of that Comcast investment warranted nothing but a dismissive mention suggesting it was due to spending on set-top boxes.

As we noted last week, just a piece of Comcast’s increase was CPE-related. A substantial portion of it was due to higher spending in a category Comcast calls “Network Investment.”

Yet Singer dismissed all of it as CPE spending to explain away the Comcast increase totals; he also failed to mention that drops in set-top-box spending accounted for almost all of the declines at Charter and Cablevision. Funny how set-top-box spending counts sometimes and doesn’t count at others. 

Singer’s bad guesses about Title II’s Impact are no better than his (incorrect) observations on changes in investment

“Although I can’t say precisely what portion of the capex decline is attributable to Title II,” Singer writes, “I can say that it is roughly one-third of the predicted effects of mandatory unbundling on ISP capex [ ]. This suggests that ISPs are discounting the likelihood that the Open Internet Order will lead to policies mandating unbundling … In the face of this murky abyss of regulatory uncertainty, the ISPs appear to be hedging their bets against a worst-case scenario.”

This is a very strange thing to say, for several reasons.

1) Again, investment is higher in the second quarter of 2015 than it was in the prior year for that period.

2) Not one single CFO, CEO, or any other executive at these companies has indicated to institutional investors that the FCC’s action is the reason for any investment changes. Each has explained, in great detail, what’s behind their short-term movements in spending, and the FCC isn’t on the list.

3) Not a single CEO or CFO, or any other executive at these companies, or any analysts that rate these companies for investors, has indicated they are “hedging their bets against a worst-case scenario.” Some, such as Cablevision’s CEO, have said the FCC’s move will not have any impact on his company’s investments.

4) As for unbundling, the FCC order explicitly forbore from application of the statute the agency would use to require unbundling. Even if the FCC had kept that statute in place, that unbundling obligation applies only to Incumbent Local Exchange Carriers, which most ISPs are not. Almost all of Singer’s observed capital decline stems from AT&T’s wireless segment, a business not subject to AT&T’s ILEC legal duties, even if the FCC had not forborne.

So what makes more sense as an explanation for the declines that occurred at a few particular ISPs but not across the board after the FCC acted to restore the law Congress wrote in Title II?

Could it be that AT&T’s downturn in capital spending, for example, is due to precisely what the company told its investors? (“[G]etting the Project VIP initiatives completed and when they’re done the additional spend isn’t necessary because the project’s been completed.”)

Or should we speculate instead that AT&T’s downturn is due to some looming threat of unbundling, never mind what the company says and what the FCC order actually did? (And never mind that this part of the law doesn’t touch AT&T’s wireless business, where most of its spending decline occurred.)

Should we attribute the decline to this vague and frankly impossible regulatory consequence, even when Singer claims that CFOs are too stupid or too mendacious to assess the impact of these regulations and react to them accordingly?

For someone whose CV is littered with work for cable and telecom companies, Singer sure has a low opinion of the executives at these companies. Maybe he doesn’t highlight his views on their competence when he’s courting them for assignments.

If AT&T builds it, Singer will … demand that it be built again

As we documented in our first write-up, AT&T initially expected its Project VIP spending bump to last longer than it did. We quoted AT&T executives saying the company had finished ahead of schedule. Singer’s attempted gotcha here is silly, as he suggests that finishing ahead of schedule and under the budget predicted by a two-to-three-year-old press release must be a bad thing.

AT&T’s fiber deployments continue today at a higher level than the company previously suggested they would. The company obviously has no need to keep building what it’s already built elsewhere in its service territory.

People who are serious about gauging companies’ reasons for their investments could do better than speculate about how executives must be acting in fear of regulations those execs don’t even understand. Looking at the investor transcripts and the SEC filings themselves can explain all sorts of things about the data that Singer claims to assess so soberly.

It’s strange for him to suggest that AT&T’s normal capital spending level for 2015 should be $20.5 billion in light of a three-year-old three-year average, which was up temporarily after the T-Mobile merger failed and AT&T actually started investing to improve its network instead of trying to buy out wireless competitors.

Why might AT&T’s investment have declined since then?

Well, the company has since unloaded assets like its former SNET New England wireline business. But more than that, it’s odd that Singer thinks the total amount of capital invested by a single firm (or even the entire industry) always must rise for a policy to be judged beneficial.

AT&T has completed its LTE deployment. It’s upgraded about as much of its wired network as it ever planned, moving up to either VDSL or IP-DSLAM technology, and it did so faster and at lower cost than it initially expected. The company’s shareholders are excited about that and the additional cash flow it creates. 

Singer accuses us of ignoring AT&T’s PR stunt about pausing investments (which happened prior to the FCC’s vote).

But we didn’t ignore it. The supposed pause warrants little attention because once the FCC team reviewing the DIRECTV merger promises asked the company to explain itself, AT&T backed down from its threats. The end result was no pause, even after Randall Stephenson’s saber rattling.

And remember, Singer spends most of his time telling us not to listen to executives’ claims about the impact of rules on investments. But when those claims fit his narrative, he credits them. It’s only when executives’ claims don’t fit Singer’s narratives (yet do match reality) that he dismisses them.

First half-year comparisons versus second-quarter comparisons

Singer dismisses our suggestion that if his theory were correct, he should see declines from second-quarter 2014 to second-quarter 2015, not just first-half 2014 to first-half 2015.

Remember, this is Singer’s theory of the world, not ours. If asked, we’d answer that it’s almost impossible to know the precise magnitude of the impact from the FCC’s actions on industry thinking, or even company-specific investment, over any short-term period.

All we can base our theories on are the statements these companies make to their investors, the deeds of the companies themselves, and the fact that many large and small ISPs have indicated that the FCC’s order had no impact on their plans. Singer is the one who thinks the short-term spending patterns indicate causation.

If he believes this, then he has to reconcile the fact that second-quarter to second-quarter spending is higher, not lower.

Writing off Windstream as just a CLEC is absurd

This is a strange one.

Yes, in parts of its service territory Windstream is a competitive provider, and in other parts it’s an incumbent — one that, in fact, takes universal subsidies administered by the FCC to offer DSL services to Windstream customers. There’s no reason to discount CLEC’s real investment in broadband, as Singer seems to suggest we should.

We also don’t discount cable ISPs’ broadband investments, even though they’re not ILECs either. Nor do we discount wireless companies’ investments, and they’re also not ILECs.

Regardless, Windstream serves over 1 million ISP customers, most with its own facilities, so the company should be curtailing its investments if Singer’s theories about Title II’s impact are correct.

Yet Windstream spending is up, not down. The company keeps accepting FCC subsidies in exchange for a long-term commitment to improve service for its most rural customers. 

Sprint’s increases in capital investments are real, not the result of an accounting change

Most egregious in Singer’s latest broadside is his charge that we “made a significant error in reporting Sprint’s capex,” one that is tantamount to us “fudging on Sprint.”

This is completely untrue. It’s Singer who’s made a significant error on Sprint.

“The problem with Free Press’ number for Sprint,” he writes, “is technical and concerns a change in accounting: In the fourth quarter of 2014, Sprint began capitalizing the cost of the leased devices.”

From this faulty premise, Singer argues that much of Sprint’s 2015 increase in spending is the result not of real capital investment but “technical” accounting changes.

However, in September 2014, Sprint made a major change that is hardly the “accounting change” Singer imagines.

Sprint actually started a new business practice in which the company itself purchases devices and then leases those devices to customers. Sprint owns those devices at the end of the lease. The customer doesn’t. This is very different from the usual wireless-carrier method of having the customer buy the device at the time of purchase (and either pay for it in full with monthly payments or by paying a higher monthly service charge).

In other words, this is Sprint making real, actual and substantial purchases of capital equipment, which its prior business model did not require.

At the end of the lease term, Sprint owns the handset. This is real investment and a real risk of capital too. At the end of the lease, Sprint is stuck with a used device that it must sell to the original customer or on the secondary market.

In starting this new lease program and purchasing these handsets, Sprint is betting that the secondary market for these Internet access devices will be robust — a bet it certainly wouldn’t make if it were worried about how Title II could be a “worst-case” scenario for broadband markets, by the way.

Sprint’s purchase of devices to lease to its customers is also an innovation in service practices, and one that Sprint must have tried to compete and differentiate itself in the market. This is the kind of real capital spending produced by competition that policymakers across the political spectrum say they want to promote.

If Singer’s “catch” were simply an accounting change, Sprint would report results to shareholders in both actual and pro forma methods. It doesn’t do that because this isn’t a pro forma change: It’s capital investment in assets that Sprint owns, and for which it has taken on the additional risk of needing to sell these used devices on the secondary market.

Singer wants to remove Sprint’s new capital spending on devices from the data, all to help him make his flimsy case less flimsy. But if he wants a real apples-to-apples comparison, he has to remove spending on all the devices and set-top boxes purchased and owned by other ISPs.

This shows the peril in Singer’s entire approach to using short-term changes in capital allocation as a metric for regulation-induced causation. Many of these companies had real, material changes in their business practices like this over the prior year, all of which make comparisons difficult without substantial caveats.

AT&T sold its SNET assets. Verizon sold its tower assets and was in the midst of selling off its wired systems in California, Florida and Texas to Frontier. Comcast, Time Warner Cable and Charter were all entangled in a failed merger.

On and on. Simple apples-to-apples comparisons aren’t completely possible here, and any economist interested in truth and not propaganda would know better than to try and claim causation the way Singer does.

But even as he tries to explain away some of Sprint’s capital-investment increase as an accounting oddity (which it was not), he still acknowledges that spending at Sprint is up. Just as it is at T-Mobile and Verizon Wireless.

In fact, Verizon was the first wireless carrier to announce trials for 5G technology next year, something that runs counter to Hal’s “hedging” theory.

The final, phony comparison: guessing the value of open-internet rules

In his last desperate defense of his claims about the net loss the FCC’s order allegedly caused, Singer resorts to his typical sleight of hand.

First, he says that his Forbes piece didn’t mean to suggest that all of the supposed decline in investment was due to Title II. He believes he needs to establish only that the law is responsible for a drop of $101 million in broadband network spending.

Why on earth $101 million? Because Singer can point to a number in the FCC’s “Congressional Review Act Abstract” — a document prepared to comply with the Gingrich-era “Congressional Review Act” (or “CRA”). This statute gives Congress a chance to overturn expert agency rules that lawmakers don’t like. It’s triggered by rules that have “an annual effect on the economy of $100,000,000 or more.”

Now forget for a moment that the very same FCC document suggests an average annual increase of $2.75 billion in broadband investment under the old Net Neutrality rules.

Singer certainly forgot this, or, more likely, just hopes you don’t know it. Forget too the billions of dollars in Internet company venture capital and online video revenues this same FCC document reports, all just as a matter of demonstrating that we’re talking about at least $100 million here.

Focus only on this breathtaking maneuver. Singer:

(a) critiques the FCC for its “casual empiricism” in the CRA document;

(b) ignores the fact that the $100 million threshold is just a statutory minimum; and

(c) hides the fact that the FCC cited billions of dollars of growth to satisfy the threshold question.

He does all this while acting as though this statutory minimum figure is somehow the actual maximum value of the benefits derived from protecting the Open Internet.

In the span of a few short sentences, Singer goes from haughty critique of the FCC’s empirical methods — when the agency is just talking about statutorily set minimums — to complete reliance on that statutory exercise for his “empirical” maximum value of rules protecting commerce and free speech online.

Singer will likely be tempted to selectively respond to this reply since he has a press conference on this topic this Friday that he needs to salvage. If you’re a policymaker or reporter in attendance at any such events, or even just an interested observer, we ask only the following: Think for yourself.

Ask yourself what is the most likely explanation for any short-term changes in investment, both at the company level and at the market level. Is it the FCC’s actions, i.e., something not a single company or credible financial analyst has pointed to as impacting investment?

Or is it the more benign list of reasons documented in our prior post and reiterated here?

More from the Policy Library