Auction or merit based comparative evaluation approach
1. Country “A” plans to provide end to end Internet connectivity to its residents at affordable cost. Though, it is as small as New York State, but the estimated population is almost 150 million. Country “A” is concerned that the Internet penetration rate is as low as 0.4 percent till date. The country is connected to only one submarine Internet backbone, installed at a cost about 94 million U.S. dollars back in 2006, but almost 90% of its capacity has remained un-utilized, due mostly to its initial high tariff rates, lack of proper nationwide telecommunications backbone infrastructure and its pricing model which failed to promise the ability to reach it’s prospective subscribers in under-privileged areas. As usual, the state owned incumbent phone/Internet company could not reach those unconnected as expected and they long ago amortized the cost of their fewer copper network, as they hardly added new connections over the years, and spent barely anything on expanding it. The e-commerce could not start, as you have guessed it right - where is the connectivity?
2. Let’s get to some of the basic facts which will determine the Internet business cases viable for this country. The country is very flat, I would say, it is the radio propagation heaven. The population is heavily dense, which means you get to meet more than 2700 people in a square mile. This will bring the thought in your mind that investment on infrastructure will have its return over night. That’s why the local subsidiary of foreign cell phone companies is earning their fortune here. The single radio BTS don’t even need to reach out to cover cell planning area, because it gets filled up way before half of its expected coverage.
3. The per capita GDP is close to US$ 470 (as estimated in 2007) which will depict the baseline of affordability of the common people. I would say, the lowest Internet connectivity tariff may start at 4 USD, because that *really* can be termed as affordable for this country. I understand Maintaining QoS of Internet at this price can be challenging, but the existing broadband policy sets 128 kpbs as a minimum for time being. Here most of the existing (mostly unlicensed) ISP’s don’t care about the contention ratio for the home users, because with the existing higher bandwidth pricing, it is nearly impossible to provide good contention ration like 1:20 or 1:30 with 8-9 USD. The taxation on CPE (modems) is also arbitrarily high to get the cost down.
4. Say, right now, the policy makers of this country really want to connect all the unconnected like providing universal access services obligation. Let’s set the initial connection speed to 64 kpbs for some time to get the initial cost down. While there is very less infrastructure on copper or fiber (FTTH and ADSL relies heavily on infrastructure) in this country, Broadband Wireless Access (BWA) technologies can be life saver for reaching everyone. The population density, the vibrant private sector with state support for ICT should be able to deliver when it comes to connecting those unconnected. Here technologies like WiMAX/Wi-Fi and HSPA can be good contenders for connecting everyone wirelessly. For most countries, these BWA services are complementary to those offered by ADSL/cable and FTTH, but here the scenario is different now. For reaching everyone, it (BWA) has to be the primary stop gap solution before the real infrastructure comes up. We have to give it a time. With regard to spectrum costs and charges, telecommunication experts sometimes opine that charges paid for radio spectrum exceeding the costs for administrative spectrum management - either caused by auctions or set in beauty contest’s spectrum fees - should be avoided as far as possible and in any case should flow back into the telecommunications sector for the benefit of the information society of a country. The attribution processes should be judged on merits how the country is benefited, and not to consider this as an opportunity for governments to withdraw additional money from this sector in order to reduce their budget deficit. For future expansions, experts generally support the principle of “technology neutrality” as proposed by other regulatory authorities for the optimum delivery of BWA services.
5. Now, let’s say, we have devised BWA services to be the best methodology to bridge the digital divide quickly. Before opening up these BWA operations to private sector, the country needs to figure out what is best way to reach its goal. That is why the authorization of Telecommunications Services to the companies and right to use radio spectrum and licensing rural and universal access services should be dealt differently. Other countries which has ADSL and FTTH up and running might have the luxury to earn extra cash by auctioning, but we have be little careful about that. Let me put something from ICT regulatory toolkit document ….
“There are significant differences in the authorization practices in force in different countries. At one end of the spectrum are wide-open authorization regimes, where no form of governmental approval is required to start a telecommunications service business or to operate network facilities. At the other end are individual licensing regimes with lengthy licence documents customized to the circumstances of a specific service provider. In between are many forms of general authorization or “class licences” that authorize and provide generally applicable regulatory conditions for classes of telecommunications service providers.”
6. We hear loads of good thing about spectrum auctioning but it does not work when it comes to using it for universal access services. But, it is true that as per the Telecommunication Regulatory Handbook, in least-cost subsidy auctions, a selection is made based on which qualified applicant requires the lowest subsidy to provide a non-economic service. The services authorized using a least-cost subsidy auction is generally subsidized as part of a country’s universal access program. This country “A” is yet to work on that kind of program right at this moment. Others would say, prices for mobile communications services in the United States continue to drop despite the billions of dollars paid for licenses in FCC auctions. Those who claim that bid prices will be passed through to consumers have yet to provide a sound economic argument or to show any empirical evidence that this has occurred anywhere in the world. It’s confusing, it is!
7. On the contrary, beauty contest seems to the most adequate method to allocate the spectrum licenses because the regulatory authority is in a position to design the comparable criteria around the country’s need. It (merit based comparative evaluation approach) has its downside for not being that transparent like auction. Most of the comparative evaluations require the applicants to make the best use of the limited resources associated with the license to serve the public. That saves the hidden cost passed on to the subscribers. It also relies in part on quantitative measures, such as the time it might take to cover the whole area (coverage obligation), the number of years of operational experience. Others rely on more qualitative (and thus subjective) criteria, such as the management skill. But, it brings out the applicants technical competence, experience, and cost efficiency. Some regulators might even put specific criteria like the lowest tariff for Internet connectivity to the home users. This requires the applicants to review their business case to offer the lowest tariff based on its ROI. This exercise makes the prospective service provider to open up its business case to the evaluator. The evaluation committee might have to get down calculating the cost per Hz per population, which seems a great idea to minimize the cost as this will be passed on to the subscribers. But, its success largely depends on building a smart merit based comparative evaluation metrics (criteria) with some pre-set list of selection criteria with the distribution of weight-age. Those can be listed under some broad head like, Business Plan, Telecommunications Sector Experience and Financial Strength and Stability and Network roll-out obligation.
8. Here’s how I thought of building one from some of the regulatory consultation papers …
Merit Based Evaluation on Defined Performance Metrics
1. Past Performance (should the country consider licensing from existing operators).
a) Service Roll-out and Coverage
aa) Coverage Provided (a. Number of transmitters, b. Number of base station sites and their demographic distribution, c. Percentage of Land mass and Population receiving coverage from Applicant’s network, d. throughput)
bb) Spectrum Efficiency (a. Earlngs per MHz of spectrum allocated, b. Quality of Service based of allocated spectrum, spectrum reuse schemes)
b) Delivered Telephony Penetration
aa) Number of subscribers served, Subscriber Growth, Subscribers per MHz of spectrum.
bb) Urban Penetration (Contribution to penetration over 5 Year period, Spectral efficiency in terms of cell planning in dense area)
cc) Rural Penetration (Contribution to penetration over 5 year period, aggressive pricing scheme to reach the under privileged area)
aa) Financial Standing (Audited financial statements for the past 3 years to assess operational track record)
bb) Investment Record (Annual investments on network and non network capital assets over the last 5 years)
d) Regulatory Compliance
aa) Compliance Confirmation (Minor and major violations if any, show causes also to be considered)
e) Technology Development
aa) New Services (Number of new services introduced over last 5 years)
bb) New Technologies (Pioneering technologies introduced over the last 5 years)
cc) Local R&D (Locally developed products and the contents)
dd) Technology (development contribution to ICT sector partnerships established to promote/assist local ICT sector)
f) Quality of Service
aa) Network Performance, Fault Incidence and Repair, Customer service complain handling, throughput as promised
bb) Ability to provide broadband service with the expected contention ratio
9. Now, when inviting new players it to this BWA regime, we need to evaluate their business case to evaluate their capability to roll-out throughout the country.
a. The Business Plan
1) Pricing of services, the introduction of services and distribution plans.
2) Local employment, skills transfer and training.
3) The involvement of local businesses either as shareholders, partners or as suppliers of goods and services leading to reinvestment in the Country “A”s economy, and minimizing the exportation of economic benefits.
4) Geographical coverage and coverage in terms of population, of network and services, and rate of roll out. (Number of transmitters and roll out plan, number of base station sites and their demographic distribution along with implementation time-line, i.e. service launch time-line of pilot service, full commercial service, network deployment plan etc.)
5) Promotion of applications and content services development,
6) Promotion of Human Resource Development such as job creation, knowledge development etc.
7) Range and quality in the BWA services provided by the Applicant.
8) Evidence of primary and secondary market research supporting forecasts and financial plans.
9) Financial aspects related to the planned BWA network, particularly evidence of the necessary funding and commitment to funding the business.
10) Experience and track record under comparable conditions, and proven ability and capability to deliver commitments made in competitive markets.
11) The overall soundness and quality of the business plan.
b. Miscellaneous Obligations
a) Country “A” would love to have proposals from the applicant to provide free basic 256 kbps wireless broadband services to all the public schools for the first year. This enables all the schools to be connected to Internet. It makes a CSR (corporate social responsibility) case for them.
b) Valued users to be provided the enhanced mobility services including voice-over-IP, video conferencing, online gaming, location-based services etc.
c) Right to share its own infrastructure and should not invest on newer infrastructure when competitors have that already.
10. Some experts says that new operator should not be barred from providing some services like voice when home broadband phone can never take away Cell Phone’s position, at least for next 10 years here in this country “A”. Voice is still a prime service here in this country. This can be done in order to allow new operators to join the market simply without constraints until a SMP state is reached. In this later case, some constrains might be imposed to guarantee that the free competition is in conformity with the free competition regulatory framework. Also, required modification in regulations to make spectrum available and harmonize use of spectrum across international boundaries can only pave its way to connect the unconnected.
11. Please provide me some feedback on making a great Merit Based Evaluation form on defined performance metrics. What more should I include to make it somehow full proof to get the widest coverage of Internet with the lowest tariff?
[Consultation papers from NTRA, OFCOM, TRCSL, Maravedis, TRA (Bahrain), CSK, OECD, OFTA, Industry Canada, MCA, ECC, GSA, HTLL, IDA, ETNO, IEEE 802 LMSC, MPT and NTA were really helpful to make these metrics]
Saturday, May 31, 2008
Re: BWA Licensing Issues: Auction or Beauty Contest [telecom]
>1. Country "A" plans to provide end to end Internet connectivity to
>its residents at affordable cost. Although it is as small as New York
>State, the estimated population is almost 150 million. Country "A" is
>concerned that the Internet penetration rate is as low as 0.4 percent
>up to this date. The country is connected to only one submarine cable
>for its Internet backbone, installed at a cost of about 94 million
>U.S. dollars back in 2006, but almost 90% of its capacity has remained
>un-utilized, due mostly to its initial high tariff rates, lack of
>proper nationwide telecommunications backbone infrastructure and its
>pricing model which failed to promise the ability to reach it's
>prospective subscribers in under-privileged areas. As usual, the state
>owned incumbent phone/ Internet company could not reach those
>unconnected as expected and they long ago amortized the cost of their
>small copper network, as they seldom added new connections over the
>years, and spent barely anything on expanding it. The e-commerce could
>not start, as you have guessed it right - where is the connectivity?
You're already showing a fatal flaw. The incumbent was given control of a monopoly infrastruture, including an undersea cable, and they have been allowed to manage it for profit maximization, not national interest. The unit price of the undersea cable is a function of utilization; at 10% utilized, its unit cost is high, but high rates which maintain such scarcity are more profitable for the owner. So the incumbent's wings need to be clipped and that cable has to be treated as a public highway facility, not as a private franchise.
Now on to the main questions...
>2. Let's get to some of the basic facts which will determine the
>Internet business cases viable for this country. The country is very
>flat, I would say, it is the radio propagation heaven. The population
>is very dense, which means you get to meet more than 2700 people in a
>square mile. This will bring the thought in your mind that investment
>on infrastructure will have its return over-night. That's why the
>local subsidiary of foreign cell phone companies is earning their
>fortune here. The single radio BTS don't even need to reach out to
>cover cell planning area, because it gets filled up way before half of
>its expected coverage.
For those who haven't figured it out yet, Raqueeb is talking about Bangladesh. It is indeed flat and densely populated. I do note that radio propagation at higher frequencies (especially >5 GHz) is harmed by heavy rain, which happens there. But for the broadband access piece, the density is high enough to allow towers to be close enough to the users so that it isn't critical. In rural areas it might be, especially in the beginning.
What's not given here is how Bangladesh expects the BWA operator to provide backhaul from the tower to the rest of the world. In the US, for instance, it is often hard to get a point-to-point microwave antenna on a cell site, so >90% of cell sites are fed by incumbent-telco wireline. This is extremely expensive (such rates are mostly unregulated and there is rarely competition) and is posing a problem for wireless broadband, which needs much more capacity. If however the Bangladeshi operator is given access to adequate spectrum for point-to-point backhauls, separate from its access spectrum, then it can build a workable backhaul network. The value of the licensee is thus very dependent on its independence from a predatory incumbent.
>3. The per capita GDP is close to US $470 (as estimated in 2007) which
>will depict the baseline of affordability of the common people. I
>would say, the lowest Internet connectivity tariff may start at 4 USD,
>because that *really* can be termed as affordable for this country. I
>understand Maintaining QoS of Internet at this price can be
>challenging, but the existing broadband policy sets 128 kpbs as a
>minimum for the time being. Here most of the existing (mostly
>unlicensed) ISP's don't care about the contention ratio for the home
>users, because with the existing higher bandwidth pricing, it is
>nearly impossible to provide good contention ration like 1:20 or 1:30
>with 8-9 USD. The taxation on CPE (modems) is also arbitrarily high to
>get the cost down.
I don't see how high taxation can hold down prices! It should raise prices.
Those retail prices for ISP service are indeed low, especially given the high cost of reaching the rest of the world's Internet. One has to be realistic about expectations. I would caution that it is reasonable to cap usage, such that a given rate's subscription is either surcharged or throttled-down if it exceeds some usage limit; on the other hand it is not reasonable to use deep packet inspection to make value judgements based on protocol or content. Some bounds on abuse are reasonable, though in the US, it has stirred controversy. However, this is best handled by a market structure, not regulatory edict...
>...With regard to
>spectrum costs and charges, telecommunication experts sometimes opine
>that charges paid for radio spectrum exceeding the costs for
>administrative spectrum management - either caused by auctions or set
>in beauty contest's spectrum fees - should be avoided as far as
>possible and in any case should flow back into the telecommunications
>sector for the benefit of the information society of a country. The
>attribution processes should be judged on the merits of how the
>country is benefited, and not to consider this as an opportunity for
>governments to withdraw additional money from this sector in order to
>reduce their budget deficit. For future expansions, experts generally
>support the principle of "technology neutrality" as proposed by other
>regulatory authorities for the optimum delivery of BWA services.
>6. We hear loads of good things about spectrum auctioning, but it does
>not work when it comes to using it for universal access services. But,
>it _is_ true that (as per the Telecommunication Regulatory Handbook),
>in least-cost subsidy auctions, a selection is made based on which
>qualified applicant requires the _lowest_ subsidy to provide an
>(otherwise) non-economic service. The services authorized using a
>least-cost subsidy auction is generally subsidized as part of a
>country's universal access program. Country "A" is not ready to
>work on that kind of program right at this moment.
Here's where traditional First World answers don't really make sense for a developing state. Why are there auctions? Because there are more entities who want spectrum than there is available spectrum. So if you don't auction it publicly, then the un-auctioned license will become a corporate asset, and the corporation's value will be
realized at auction, either by selling the licensed entity (which happened all the time in the US when spectrum lotteries were used) or by selling shares in the company (share values reflect asset values).
And why do subsidies exist? Because in some places, costs exceed revenues, so the business has low inherent value, and nobody would invest without subsidies. This is of course the opposite end of the continuum; if there were an auction, it would have to go into negative values to get a taker (I'll build that system if you pay me; I'm not paying for the license).
But why should there be scarcity anywhere? It is an artifact of spectrum allocation policies. How busy is the microwave spectrum in Bangladesh? I know India has problems because the military is sitting on most of the spectrum, to hell with the civilian sector, but that is a failure of governmental will.
How much spectrum do you need? It is cheaper to have one tower and 100 MHz than ten towers with 10 MHz. I have worked with consulting teams that were helping bidders on spectrum auctions, and seen, and done, some of the quantitative analysis used in setting a rational maximum bid. When the 3G auctions were taking place in Europe at the end of the last decade, the key value proposition turned out not to be broadband access but a lower cost of voice, simply because additional spectrum lowered cell-splitting costs. Of course the bidders went nuts anyway and bid far too much. (This was a chapter in my book The Great Telecom Meltdown.) Also, the mobile voice business was so lucrative that it made wireless Internet access look like a bad use of costly spectrum.
So the whole question is framed wrong. It should not be "auction off a license, or have a beauty contest". In both cases you're creating monopolies, and then scratching your head about how to regulate them. When you frame the question that way, you've already arrived at a wrong answer.
>Others would say,
>"prices for mobile communications services in the United States
>continue to drop despite the billions of dollars paid for licenses in
>FCC auctions. Those who claim that bid prices will be passed through
>to consumers have yet to provide a sound economic argument or to show
>any empirical evidence that this has occurred anywhere in the
>world." It's confusing, isn't it?
Those "others" are wrong. The prices are passed through. Usage has skyrocketed in the US so the per-minute cost is still low. That's simply proof that our mobile-party-pays system is far superior to Europe's calling-party-pays system. We average around 1000 minutes/month per mobile phone, but it's predominantly low bit rate voice (efficient use of spectrum and backhaul). Wireless (licensed) data services are still on shaky ground here.
>7. On the contrary, a "beauty contest" seems to the most practical
>method to allocate the spectrum licenses, because the regulatory
>authority is in a position to design the comparable criteria around
>the country's need. It (merit based comparative evaluation approach)
>has its downside, since it is not as transparent as an auction.
What you're doing is accepting monopoly and simply regulating it, or worse, not regulating it and hoping they live up to expectations. Monopolies are the problem.
Here's a suggested alternative. First, come up with a realistic amount of spectrum, so there is not serious scarcity from day one. This probably means that in urban areas, heavy usage (business sites, for instance) can use wireline at reasonable rates. Divide this spectrum into multiple license blocks, and cap how much one bidder can hold. And divide it geograhically so that there are really valuable urban licenses, and many smaller licenses with less value. The point is to isolate subsidy-needing areas from surplus-producing areas.
Now hold a *reverse* auction. That is, put *all* of the licenses on the table at once, urban and rural. With the help of a consultant, determine the maximum conceivable value of any license. Offer all licenses at that price per license. As soon as somebody wants one, they get it. Now lower the price a little. So if you have a license price that starts at say 1000 taka crore (I think about US$140M; I'm just pulling a number out of thin air for this), and nobody bids, then you lower it to 900 taka crore, etc. As the price declines a few percent per round, more and more are sold off. You end up with at least two or three (per policy) licensees in the cities, so there's competition. Now as the price declines towards zero, there may be some licenses left, so you stop lowering it in percentage and move to fixed values. If nobody bide 1M taka ($14k USD) for a license for a rural area, then try 0. Then offer to *pay* somebody
1M taka. Then 2M, etc. Those are the subsidies, and they come from the money paid in for the urban licenses. You just need one to start in the rural areas, but they need strict regulation to see that they actually spend the money correctly and don't steal it. What you've done is find the taker who'll need the least subsidy, using the same process.
Also, I would not restrict the licensees as to carrying voice, since having application-specific regulation brings in a whole different set of problems. But it might be a good idea to maintain a separation of carriage and content, so that the radio system owners must provide access to wholesale ISPs. That way the radio systems can be financed on a wholesale domestic price/bit basis, without thinking about application value, while ISPs compete freely. That is the only real way to have workable "neutrality", and it is economically more efficient than trying to regulate Internet service per se.
Fred Goldstein k1io fgoldstein "at" ionary.com
Sunday, 11 November 2007 10:01:43 - 0800 GMT