Talking Points for a debate with Tom Hazlett on TV band white space use

"Innovation, Technology, and Spectrum Policy” mini-conference, GMU School of Law, November 14, 2006

Pierre de Vries

 

Summary

• Not either/or choice between L and UL: need to decide best for this specific case

• UL is best for white space junk bands

• Mixed L/UL is a necessary regulatory hedge, and in fact synergistic

Not either/or

• Don’t face an abstract universal choice of what’s better in all cases, EAFUS/tradable-licensed (“L”) or Commons/unlicensed (“UL”)

• Hazlett’s “marginal allocation challenge” – given x MHz, what do you do?

• Not aware of underlying economic theory that treats L and UL in the same way, and which affords mechanism for choice

Specific case: “digital dividend” repacking of bands leads to

• Open 700-800 MHz band -> auction

• White spaces in remainder -> ? Will argue should be UL, not L

White space is a “junk band”

• Stringent non-interference burden vis a vis TV broadcasts

• Relatively few channels in major metros, most adjacent to TV

• Many channels in rural areas, but few people – high auction transaction costs but low values (cf. AWS auction prices)

UL works well in junk bands – cf. experience in 2.4 GHz

• Spread spectrum, LBT works to share spectrum capacity without need for exclusive assignment to band manager

• Reports of the death of the commons greatly exaggerated

  • Dilemma burden on UL: if unused, then “a waste”; if heavily used, the “tragedy of commons”

• Generates lot of innovation

• Wi-Fi > 60 bn devices, Bluetooth > 1 bn devices, Canopy proprietary standard

• Note: not saying UL leads to more innovation than L; tech ^ regs

• Not just limited to short range, just within property lines: WISPs, offices on different floors, campuses

Benefits of UL in white space

• More economical for rural WISPs, muni mesh – early stage when need coverage, not capacity

• Bigger footprint for hotspots, easier to find/connect

• User-deployed neighborhood meshes

• Sensor meshes (industrial, ag)

• White space offers in-fill range at low power (100’s meters at <100 mW; cf. 10’s m for 2.4 GHz UL, 1000’s m for O(GHz) L)

• Very low electrical power drain at short range and v. low power

Choosing L vs. UL

• No demand curve for UL, so can’t calculate consumer/producer surplus – so can’t compare to suplus calcs for CMRS

• Market size no indication of value – huge-value $0 bn markets, e.g. OSS, wikis

• Other goods have public/private ownership mix: roads, parks, intellectual goods

• “Managed UL” will be underprovided in all-L scenario (collective action dilemma)

Hedge

• Scarcity

  • Can’t predict capacity vs. usage outlook, so can’t predict scarcity

  • If scarce, bias to L; if not scarce, bias to UL (short cut around slow anti-trust proceedings)

• Lock-in

  • Any regulatory regime choice locks in outcomes – no way back if make global choice too early

• Govt Greed

  • Moral hazard: get more revenues if slow down auctions to create artificial scarcity

  • UL outlet for entrepreneurs – pay price of putting up with interference if the price of licenses too high

L/UL synergistic, not just complement

• Combo better, so L+UL not just a hedge argument

• Cf. T-Mo, QCOM: offering Wi-Fi plus cellular packages

• Cf. intellectual property

  • Copyright + public domain

  • E.g. Jasper Fforde, Eyre Affair

  • Fair Use like underlay

• Cf. parks (Nat’l Recr & Park Assoc)

  • Dallas – homes on 14 parks, 22%+ in value

  • 1-3% value of home if < 1,500 ft from park

Problems with spectrum-title/tradable-licenses in white space

• Hard to define property right – secondary to TV, can lose access if new TV is licensed, squirly interference requirements

• Definition of rights (by govt) even just for OOB emissions selects some services over others, just as UL rules do (can’t resolve by Coasian negotiation generally – too many counter-parties)

• Illiquid market in rural areas – high transaction costs of low-yield auctions

• Markets don’t end rent seeking via lobbying – cf. R&D tax credit, commodity subsidies, energy companies, …

• Single national overlay license pipe-dream: politics Þ multiple licenses, lots of small rurals Þ high transaction costs, difficulties with Coasian negotiations with many broadcasters

Show me the money – interests shape analysis

• Cellular eqmt vendors– today mostly infrastructure and edge equipment for L

• Broadcasters – option value of licenses

• IT companies – free infrastructure input with UL

• Audio eqmt companies – most sales are for non-L microphones operating with L protection

• Progressives – suspicious of large spectrum owners, seek diversity

• Marketeers – never saw a property right they didn’t like