spectrum auctions_june 2014
DESCRIPTION
Scott McKenzie's presentation to the European Spectrum Management Conference in Brussels, June 2014, covering spectrum auction formats, issues, case studies, policy objectives and rules.TRANSCRIPT
Spectrum Auctions Session 4, Breakout 2
Presentation to European Spectrum Management Conference Scott McKenzie 17th June 2014
Since 2001, Coleago has offered a wide range of advisory services to the telecom industry
Strategy and Business Planning
Spectrum and Licences
Due Diligence Improving Network
Performance
Business Transformation
and Cost Reduction
Marketing and Customer
Management Regulation and
Interconnect
Digital Content and Media
Training
2 © Copyright Coleago 2014
We are experts in spectrum issues
3
UK 2008-2013 (CCA): Valuation and bid strategy support for 2.6GHz and 800MHz auction. Lobbying support on spectrum caps and floors.
Switzerland 2012 (CCA): Lobbying, valuation and bid strategy for existing spectrum and 2.6GHz as well as 800MHz
Ireland 2011 (CCA): Lobbying and bid strategy for 900MHz, and 800MHz CCA auction
India 2010 (SMRA): Bid strategy and live auction support for regional BWA
New Zealand 2011/2012: Lobbying auction design and LTE Strategy
Australia 2010-2013: Lobbying, AIP, auction design, valuation and bid strategy
Belgium 2011 (SMRA) : Auction strategy and live auction support
Russia 2012: LTE Strategy
Poland 2009: Spectrum valuation
World Bank 2011: Review of international best practice in spectrum allocation
Netherlands 2012 (CCA) : Auction strategy and live auction support
© Copyright Coleago 2014
Spectrum Auction Formats
4 © Copyright Coleago 2014
The process of awarding mobile spectrum has become more sophisticated over time
Lottery! – Dentists, actors, hairdressers etc. entered the
lottery for US wireless licences in the early 80s – Chuck Hagel (current US Secretary of Defence)
was one of the many who made a fortune this way – This led to legislation in 1993 mandating the use of
auctions for future FCC licenses
Beauty contests – Widely used outside the US for initial GSM licences – Can be difficult to administer and bureaucratic – Open to dispute and vulnerable to corruption
Auctions – Transparent process (lower subjectivity) – Policy objective: maximise economic efficiency – Theoretically whoever values spectrum the most
will produce the greatest social good
~1980s
~1990s
~2000s
5 © Copyright Coleago 2014
6
Sealed bid
First or second price
l Single bidding round l Bid on packages rather than individual lots l Can be first or second price i.e. pay what you bid or
the opportunity cost of denying other bidders respectively
l Used in France, Hungary, Norway and Portugal
SMRA
Simultaneous Multi-Round Ascending auction
l Bid on specific blocks of interest (between minimum and maximum set by auctioneer for each block)
l ‘Standing high bids’ for each lot in each round l Auction ends when there is no excess demand l ‘First price’: pay what you bid
l Pioneered by US FCC in 1994 and now the most widely used format
CCA
Combinatorial Clock Auction
l Bid on packages of generic lots rather than on individual lots
l Primary (clock) round followed by supplementary sealed bid round
l Pay second price: opportunity cost of denying other bidders
l Separate assignment round for positioning in the band
l First used in 2008 in the UK and now used in Australia, Canada, Denmark, Ireland, Switzerland, Slovenia and the Netherlands
Three main formats for auctioning multiple spectrum lots
© Copyright Coleago 2014
All auction formats have issues
© Copyright Coleago 2014 7
Auction Format Comments
First price single round sealed bid auction
No opportunity for price discovery and can lead to the winner’s curse
Second price single round sealed bid auction
Winner pays the price of the second highest bidder and therefore reduces risk of the winner’s curse. No opportunity for price discovery which can lead to price discrepancies
Simultaneous Multi-Round Ascending (SMRA) auction
Simple ascending auction gives opportunity for price discovery but aggregation risk* is present as there is no package bidding
SMRA auction with augmented switching
More complex form of SMRA where bidders can in certain circumstances alter bids to reduce aggregation risk*
Combinatorial Clock Auction (CCA) based on second price rule
In theory, economically efficient and avoids aggregation risk* but extremely complex and not very transparent with clock and supplementary rounds. Can lead to major price discrepancies.
* Aggregation risk is the risk that a bidder might win an unprofitable subset of its desired package if there are complements between the lots
There have been some controversial CCA auctions in Europe
8
UK 800 / 2600 MHz multi-band auction 2013. The most spectrum constrained operator was left with a small package
Switzerland 800 / 900 / 1800 / 2100 / 2600 MHz multi band auction 2012. The number one operator secured a much larger package for a much lower price than the second operator
Netherlands 800 / 900 / 1800 /2100 / 2600MHz multi-band auction 2012. The auction raised almost 8x the amount expected and the prices were extremely high for the incumbent operators
© Copyright Coleago 2014
Austria 800 / 900 / 1800 MHz multi-band auction 2013. The implied prices paid in the auction are very high. In addition the largest operator managed to acquire a disproportionate share of the available spectrum especially sub 1 GHz.
Case study: Swiss multiband auction (February 2012)
9
*
l First multiband CCA, involving both existing and new spectrum l Swisscom paid CHF122m less for a superior package l Auction design provided potential for predatory bidding
Spectrum acquired (MHz) Price paid
800 900 1800 2100 2600 2600 TDD Total CHF
$/MHz/POP
Swisscom 2x10 2x15 2x30 2x30 2x20 1x45 255 360m 0.20
Sunrise 2x10 2x15 2x20 2x10 2x25 0 160 482m 0.43
Orange 2x10 2x5 2x25 2x20 2x20 0 160 155m 0.14
Total 2x30 2x35 2x75 2x60 2x65 1x45 575 996m 0.25
© Copyright Coleago 2014
Case study: Netherlands multiband auction (December 2012)
10
*
l Multiband CCA, involving both existing and new spectrum l 2x10MHz of 800MHz reserved for a new entrant leading to sub 1 GHz scarcity l KPN paid €29m less for a superior package l The auction raised 8x the expected revenue and has been criticised for its cost
Spectrum acquired (MHz) Price paid
800 900 1800 2100 1900 TDD
2600 TDD Total €
$/MHz/POP
KPN 2x10 2x10 2x20 2x5 1x0 1x30 120 1,352m 0.91
Vodafone 2x10 2x10 2x20 2x5 1x0 1x0 90 1,381m 1.23
T-Mobile 2x0 2x15 2x30 2x0 1x14.6 1x25 129.6 911m 0.57
Tele2 2x10 2x0 2x0 2x0 1x0 1x0 20 161m 0.65
Total 2x30 2x35 2x70 2x10 1x14.6 1x55 359.6 3,804 0.85
© Copyright Coleago 2014
Case study: UK multiband auction (February 2013)
11
*
l Multiband CCA for new spectrum l Final prices circa 60% of final clock prices, and 45% of winning bid values
– Defeats ‘price discovery’ element in clock phase – Difficult to gauge price exposure; major problem when managing budget limits
l Telefonica O2, the most spectrum-constrained operator secured just 2x10MHz
Spectrum acquired (MHz) Price paid
800 2600 2600 TDD Total £ $/MHz/POP
EE 2x5 2x35 1x0 80 589m 0.18
Telefonica O2 2x10 2x0 1x0 20 550m 0.67
Vodafone 2x10 2x20 1x25 85 791m 0.23
3 2x5 2x0 1x0 10 225m 0.55
BT 2x0 2x15 1x20 50 186m 0.09
Total 2x30 2x70 1x45 245 2,341m 0.23
© Copyright Coleago 2014
Case study: Austria multiband auction (October 2013)
12
*
l Multiband CCA, involving both existing and new spectrum l There was an absence of effective caps on the sub 1GHz spectrum l A1 Telekom was able to capture 54% of the valuable sub 1 GHz spectrum l The auction has been criticised as being very expensive and limiting competition in
the market
© Copyright Coleago 2014
Spectrum acquired (MHz) Price paid
800 900 1800 Total € $/MHz/POP
A1 Telekom 2x20 2x15 2x35 140 1,030m 1.23
T-Mobile 2x10 2x15 2x20 90 654m 1.21
3 2x0 2x5 2x20 50 330m 1.10
Total 2x30 2x35 2x75 280 2,014m 1.20
While CCA offers attractive features, it also poses significant problems for participants
13
l A bidder may win an inferior package yet pay a higher price
than its rivals (even when following an optimal bid strategy)
Embarrassing outcomes
l Allocation and price exposure are uncertain, and there is no
opportunity to react to unexpected outcomes
Low transparency
l Forecasting the impact of bids is difficult. In theory, a strong bidder could be knocked out by a group of weaker bidders, despite being a “last man standing” in the clock phase
Lack of control over outcome
l It is difficult to communicate the risks to key stakeholders and protect the bid team from post-auction criticism in the event of unfavourable outcomes
Stakeholder management
l Managing complex bid decisions during a live auction is difficult, with limited time-intervals between rounds; this calls for robust processes and auction support tools.
Live auction execution
© Copyright Coleago 2014
14
Fixed budget constraints introduce further dilemmas in a CCA – trade-offs between risk and opportunity
“Red line” package Conservative target
Realistic target
Ambitious target
Max package
Bid values
“Red line” package
Conservative target Realistic target Ambitious target Max package
Avai
labl
e B
udge
t
Bid values
l Chance of winning bigger package l Drive prices paid by rivals (minimises risk
of adverse price differentials) l But increases risk of knock out
Focus budget on larger packages? Hence squeeze bid values at the bottom
Focus budget on smaller packages? And squeeze marginal bid values at the top
l Lower risk of complete knock out l No chance of winning bigger package l Significantly reduced impact on prices paid
by rivals
© Copyright Coleago 2014
Summary – key characteristics of multi round auction formats
Design Principle SMRA CCA
Supports simultaneous award of spectrum in multi-bands ✔✔✔ ✔✔✔
Aggregation risk ✘ ✔✔✔
Transparency of bidders and bids ✔✔✔ ✔
Certainty over lots awarded ✔✔✔ ✘
Certainty over prices paid ✔✔✔ ✘
Avoids adverse price asymmetries ✔✔✔ ✘
Simplicity and ease of presentation and transparency of results ✔✔ ✘
© Copyright Coleago 2014 15
Regulator policy objectives should determine allocation process, auction design and licence conditions
Policy objectives differ between countries l Maximise immediate revenue generation
from a spectrum auction l Make spectrum available to mobile
operators as cheaply as possible l Increase mobile broadband access in
rural and remote areas l Increase competition l Introduce new, faster speed services l Ensure development of a connected
society to deliver long-term economic benefits by making best use of spectrum
l Promote economic efficiency
16 © Copyright Coleago 2014
Detailed auction rules may matter more than formats
Rules need to be set to prevent gaming while ensuring that all spectrum is sold efficiently l Spectrum packaging l Spectrum caps l Spectrum set-aside l Activity rules l Provision of information l Bid increments l Spectrum trading l Reserve prices
© Copyright Coleago 2014 17
18
Thank you
Scott McKenzie
Tel: +44 7825 294 576 [email protected]
Director Coleago Consulting Ltd
© Copyright Coleago 2014