Norwegian Offshore Wind Auctions - Strategic recommendations for Phase 1 of Sørlige Nordsjø II

Rapport 2022/44

Andreas Stranden Hoel-Holt, Haakon Riekeles, Åsmund Sunde Valseth, Haakon Vennemo, Ana Amazo Blanco og Bastian Lotz

Norwegian Offshore Wind Auctions - Strategic recommendations for Phase 1 of Sørlige Nordsjø II

Kategori	Rapporter
Underkategori(er)	Strategi og prosessrådgivningKraft og energiReguleringer og konkurranseøkonomiSkatter og offentlig økonomi
År	2022
Rapportnummer	44
Forfatter(e)	Andreas Stranden Hoel-Holt Haakon Riekeles Åsmund Sunde Valseth Haakon Vennemo Ana Amazo Blanco Bastian Lotz
Last ned	file_download (934.5 kB)
Les i nettleser	find_in_page

Report 2022/44	 | For 	the 	Ministry of 	Petroleum and Energy	 	
 	
 	 	
Norwegian Offshore Wind Auctions	 	
Strategic	 Recommendations for Phase 1 of Sørlige Nordsjø II	 	
Ana 	Amazo Blanco, 	Andreas 	Hoel	-Holt, 	Bastian Lotz, 	Haakon Riekeles, Åsmund Sunde	 	
Valseth	 and Haakon Vennemo

Norwegian Offshore Wind Auctions	 	
 
Vista Analyse	 | 2022/44	 	2 	
 
D	ocument details	 	
Title	 	Norwegian Offshore Wind Auctions	 – Strategic 	Recommendations for Phase 1 of Sørlige 	
Nordsjø II	 	
Report Number	 	2022/44	 	
Author	s 	Ana Amazo Blanco, Andreas Hoel	-Holt, Bastian Lotz, Haakon Riekeles, Åsmund Sunde Valseth 	
and Haakon Vennemo	 	
ISBN	 	978	-82	-8126	-602	-5 	
Project Number	 	22	-HVE	-12	 	
Project 	Manager	 	Haakon Vennemo	 	
Commissioned by 	 	Ministry of 	Petroleum and Energy	 	
Date of completion 	 	November 18	th, 2022	 	
Availability 	 	Public	 	
Keywords	 	Strategi og prosessrådgivning, 	ekspertuttalelse, 	kraft og energi, 	regulering, 	offentlig økonomi	, 	
auksjoner	 	
 
 
 
 
About us	 	
Vista Analysis	 is a social science consultancy with its main emphasis on economic research, policy analysis and advice, 	
and evaluations. We carry out projects to the highest professional standards, with independence and integrity. Our 
key thematic areas include climate 	change, energy, transport, urban planning and welfare issues.	 	
Our employees have high academic credentials and broad experience within consulting. When needed we utilise an 
extensive network of companies and resource persons nationally and internationally.	 The company is fully employee	-	
owned.	 	
Guidehouse	 is 	a leading global provider of consulting services to the public sector and commercial markets, with 	
broad capabilities in management, technology, and risk consulting. The company has more than 12,000 profe	ssionals 	
in over 50 locations globally. With over 700 consultants, our global Energy, Sustainability, and Infrastructure segment 
is the largest in the industry. Guidehouse Europe has offices in Germany, the United Kingdom, Netherlands, France 
and Lithuania	. We collaborate with and serve as trusted advisers to the public sector, utilities and energy companies, 	
large corporations, investors and NGOs to help them thrive in the rapidly changing energy, resources, and infrastruc-
ture environment. Our clients incl	ude various European governments and the European Commission. Guidehouse 	
Germany GmbH	 is a fully owned subsidiary of Guidehouse Inc	.

Norwegian Offshore Wind Auctions	 	
 
Vista Analyse	 | 2022/44	 	3 	
 
Preface	 	
This report marks the initial deliverable from the 	advisory project on designing 	the auction	 for 	wind power 	capacity	 	
at 	Sørlige Nordsjø II	. Sørlige Nordsjø II	 is the first 	auction for wind power capacity in Norway, and Vista Analyse carries 	
out the advisory in 	a collaboration with Guidehouse 	and 	Procurex	. 	
The 	report	 is a high level study on strategic	 choices to be made 	in the design of the auction. 	Authors are with Vista 	
Analyse and Guidehouse	. Marty Barc	lay from Procurex has provided peer review on chapter 	3. Cassandra Velten has 	
been the main contact point at the Ministry	. W	e thank her	 and the Ministry team	 for 	a good cooperation with	 con-	
structive discussions. 	 	
 
November 18	th, 2022.	 	
 
Haakon Vennemo	 	
Partner	 	
Vista Analyse AS

Norwegian Offshore Wind Auctions	 	
 
Vista Analyse	 | 2022/44	 	4 	
 
Contents	 	
Executive Summary	 ................................	................................	................................	................................	......................	 6 	
1 	Introduction	 ................................	................................	................................	................................	........................	 10	 	
2 	Choice of	 support scheme	 ................................	................................	................................	................................	.. 11	 	
2.1	 	Which schemes are relevant?	 	11	 	
2.2	 	Assessment of relevant schemes	 	18	 	
2.3	 	Summary and recommendations	 	29	 	
3 	Auction formats	 ................................	................................	................................	................................	..................	 35	 	
3.1	 	Which auction formats are relevant?	 	35	 	
3.2	 	Assessment of the most relevant auction formats	 	39	 	
3.3	 	Reservation price	 	46	 	
3.4	 	Recommendation	 	46	 	
4 	Further considerations of relevance for the auction	 ................................	................................	..........................	 49	 	
4.1	 	Requirements to pre	-qualification criteria from the perspective of the auction	 	49	 	
4.2	 	Pre	-development responsibilities between state and bidders	 	50	 	
4.3	 	Grid connection and payment	 	50	 	
4.4	 	Realization deadlines	 and penalties	 	52	 	
5 	References	 ................................	................................	................................	................................	..........................	 54	 	 
Figur	e 	
Figure 2.1	 	NVE estimates of LCOE for Utsira Nord and Sørlige Nordsjø II, including offshore grid	 .................	 17	 	
Figure 2.2	 	Illustration of the effect of different caps on payment	 ................................	................................	... 33	 	 
Tab	le 	
Table 2.1	 	Some possible support schemes	 ................................	................................	................................	.....	 12	 	
Table 2.2	 	Distortions in 	cases with low/negative value of production	 ................................	...........................	 20	 	
Table 2.3	 	Distortions in the timing of production	 ................................	................................	..........................	 21	 	
Table 2.4	 	Distortions in supply of ancillary services, without storage	 ................................	............................	 23	 	
Table 2.5	 	The consequences of price risk for the three support schemes	 ................................	.....................	 25	 	
Table 2.6	 	Pros and cons of different support schemes	 ................................	................................	..................	 30	 	
Table 3.1	 	An explanation of auction theory terms in the context of Sørlige Nordsjø II	 ................................	.. 36	 	
Table 3.2	 	Common auction formats and their general rules	 ................................	................................	..........	 37	 	
Table 3.3	 	Overview of relations between auction types	 ................................	................................	................	 38	 	
Table 3.4	 	Common and private factors	 ................................	................................	................................	..........	 41	 	
Table 3.5	 	Pros and cons of different auction types, relative to first	-price sealed	-bid	 ................................	....	 47	 	 
Text Frame	 	
Text frame 3.1	 	Summary of winner’s curse	 ................................	................................	................................	............	 42	 	
Text frame 3.2	 	Summary of efficiency	 ................................	................................	................................	....................	 43	 	
Text frame 3.3	 	Summary of attitudes to risk	................................	................................	................................	...........	 43	 	
Text frame 3.4	 	Summary of familiarity	................................	................................	................................	....................	 44	 	
Text frame 3.5	 	Summary of collusion	 ................................	................................	................................	.....................	 45

Norwegian Offshore Wind Auctions	 	
 
Vista Analyse	 | 2022/44	 	5 	
 
Text frame 3.6	 	Summary of entry	 ................................	................................	................................	...........................	 46

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	6 	
 
Executive 	Summary	 	
We provide advice on the design of the 	auction for phase 1 at Sørlige Nordsjø II. 	A two	-sided CfD and 	
upfront 	investment 	support	 are feasible 	schemes with pros and cons, but upfront	 investment 	support	 	
has not been much used in past	 auctions	. If a CfD is chosen we recom	mend a long reference period. 	We 	
recommend that 	the auction 	should be ‘Anglo	-Dutch’, that is, a period of open bidding followed by a 	
final 	sealed	 bid 	from the top two b	idders. 	 	
Auctions of 	offshore wind energy 	are	 coming to Norway	 	
Auctions 	to allocate acreage for renewable energy production 	are	 fairly common across Europe and 	
North America	. In some auction	s bidders indicate the highest price they are willing to pay	. In 	others they 	
indi	cate the lowest support they are willing to receive	, and in some it is not determined beforehand. 	 	
Norway has decided that 	auction is 	the main model for all	ocating acreage 	for offshore 	renewable energy 	
production. The initial auction will c	omprise Sørlige N	ordsjø II	. Sørlige Nordsjø II is developed in two 	
phases, each with a capacity of 1500 MW	, and there will be one auction for each phase.	 	
Presently the Ministry is preparing for the auction of 1500 MW 	at Sørlige Nordsjø	 II phase 1. 	The Min	istry 	
has set “an efficient allocation of the licenses to be auctioned, with the purpose of realizing the highest 
sum/lowest support” as the objective for the auction. 	Vista Analyse, Guidehouse and Procurex 	have 	
been engaged t	o help set up the auction. 	Thi	s report is the 	first deliverable 	of our engagement	. 	
In th	e report we primarily discuss 	which support scheme to use, and which auction 	format	 to use. By 	
support scheme we mean 	whether the	 object of the auction 	should be investment support, 	the 	elec-	
tricity price or something else	. By 	auction 	format	 we mean whether the auction should be of the 	sealed	-	
bid type, or 	ascending type	, or some other type. 	 	
The 	auction at Sørlige Nordsjø II is to be preceded by a pre	-quali	fication stage. 	In the final chapter of 	
the report we provide comments on pre	-qual	ification and other regulatory elements (grid connection, 	
penalty framework, timing) from 	the perspective of the auction.	 	
Both upfront	 investment 	support	 and 	a two	-sided Cont	ract for Differ-	
ence are viable support schemes	 	
While there are several 	possible support schemes	, we find the 	two	-sided Contract for Difference 	(CfD) 	
and 	upfront 	investment su	pport 	to be 	the two most promising candidates for Sørlige Nordsjø II phase 	
1. An 	auction for upfront investment support	 simply ask	s bidders to state the lowest level of support 	
they 	are willing to receive for the acreage, or, if the auction crosses zero, the highest 	amount they are 	
willing to pay for the acreage. Support or payment is 	carried out 	prior to 	operation	 and	 possibly prior to 	
investment	. 	
An auction for a two	-sided CfD 	asks participants to state the lowest price they accept for the	ir produc-	
tion. The state will guarantee that price	 and takes on 	most of 	the 	market price risk. Th	e winning bid may 	
imply a CfD price that is higher, or lower, than the expected market price.	 A high CfD indicates support.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	7 	
 
Upfront investment support has the advantage 	of not distorting	 the design an	d operation of the 	wind 	
energy plant. The 	CfD, by contr	ast, 	partially or wholly (depending on 	various adaptations)	 shelters the 	
operator fr	om 	information about market conditions 	inherent in the 	price 	signal	. A CfD with a short ref-	
erence period 	means that the operator 	effectively faces a fixed price. A CfD with a long reference period 	
ensures that the operator faces short term price signals, whilst 	removing risk from 	long term fluctua-	
tions in the electricity price.	 	
The most 	important advantage of the CfD contract is tha	t it reduces uncertainty about future income. 	
This 	reduces the risk that the winning bidder 	finds that it has overestimated the future income from the 	
project, and 	defaults	, abandons or delays	 the project.	 	
If the developers have a higher cost of risk than 	the Norwegian government, a 	CfD 	scheme	 has	 the 	
further advantage that it 	transfers	 market price risk	 from the developer to the government	. We note 	
on this point that 	a CfD probably will reduce the cost of financing for the winning bidder, but this is by 	
itself 	not a sure sign that the bidder faces a higher cost of 	risk	. A change from an uncertain to a certain 	
income stream should always 	reduce the cost of financing, no matter who experiences it.	 	
The upfront 	investment subsidy is 	rarely 	used	 in renewable energy auctions, while the 	two	-sided CfD is 	
common. The reasons for the relative popularity of CfD seem to be that the financing cost of the bidder 
goes down,	 and the risk of defaulting on the project is lower. There also seems to be a 	liquid	ity concern, 	
especially whenever the auction 	turns into a payment mode. If the winning bid is a payment	, the inves-	
tor of the 	upfront scheme 	will have to 	fund	 both the	 physical installation and 	provide payment to the 	
state	 at the same time	. 	
If a CfD is cho	sen	, we recommend 	an annual reference price.	 An annual reference price 	insu	lates the 	
operator from deviations to the annual price trend, which arguably matters the most for long	-term 	
profitability. But it exposes the operator to 	seasonal price volatility, 	which is important for 	operational 	
choices	. In	 effect an annual reference price significantly reduces the distort	ionary impact of a CfD and 	
removes most the objection to the scheme from the objective of market integration. 	 	
We also provide advice on 	the maximum support implicit in the CfD	-contract, the length of the contract, 	
and price indexing. We advise against a maximum payment from the 	winning bidder	, and we recom-	
mend that the cap on payments from the state to the develop	er is set sufficiently high 	as 	to not be 	
binding for the winning bids	. 	
We recommend an Anglo	-Dutch auction	 	
In ideal circumstances all auction formats give the same outcome. In practice, however, which format 
to use is important for the success of the aucti	on	. 	
Open bidding for the CfD price	 is considered to yield a 	lower price (	or lower investment support	) than 	
other auction formats	, and are called efficient	. In an open	-bid 	system 	the auctioneer declares a (high) 	
reservation or ceiling price and participants	 bid 	lower prices in successi	on 	until one bidder obtains the 	
winning bid. 	(Confusingly this format is called the ascending type). 	Another advantage of th	e format is 	
that participants may learn from bids of others what the true value of the object 	might be.	 Le	arning 	
from others 	will 	reduce the risk that the winner defaults on his obligations and the project is not carried 	
out. 	This is particularly 	valuable when the object is unfamiliar and there is significant 	market 	uncertainty, 	
which is likely the	 case in the auction for Phase 1 of Sørlige Nordsjø II.

Norwegian 	Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	8 	
 
An alternative to the open bids of the ascending type auction is the 	first	-price sealed bid auction 	
whereby bidders put in one bid	 for support	, the lowest 	bid	 wins, and the winner receives the suppor	t 	
level that she asked for. 	This 	format is the most common renewable energy auction 	format	, to the ex-	
tent that 	it is not always considered an auction, but a tender (perhaps particularly 	so 	when combined 	
with other 	criteria). 	 	
The sealed bid auction has some advantages over the open bid 	ascending type. In particular, it 	makes 	
sure that the winner 	receives 	the 	minimum support 	he himself is willing to 	receive 	(as opposed to that 	
of the second highest bidder in an open 	format (ascending type)). 	This is particularly important 	since 	
Sørlige Nordsjø II is a “flagship type” auction. 	In an ordinary first	-price sealed bid auction bidders are 	
expected to 	“shade” their bids, i.e. bid lower than their true valuation	 in order to g	ain some profit for 	
themselves	. In a flagship type auction bidders are 	expected to be risk	-averse 	and will not risk shading 	
their bids. In other words we expect the winner 	in such an auction 	to bid his true valuation	. 	
The auction format we recommend combi	nes attractive features of both the open	- and final bid formats. 	
We recommend	 open bidding that is 	stopped when two bidders remain. These two bidders are then 	
allowed a final sealed bid	, which determines the outcome. The efficiency of this format 	(i.e. sup	port 	
level) 	is theoretically 	almost as good as 	fully open bidding (ascending type	). In fact, 	efficiency	 can be 	
better	 if the final bidders are risk averse and determined to win the auction. 	Further, t	he risk of collusion 	
is lower	 when final bids are sealed	, and it has been shown that the format	 attracts more participants 	
than other formats. 	 	
Open bidding followed by a sealed bid is 	a fairly common format in commercial auctions around the 	
world. In the economics literature is known as an ‘	Anglo	-Dutch	’ format. Practitioners sometimes call it 	
a ‘final blind’ auction.	 	
The reservation or ceiling price is an important feature of the auction format. We recommend bas	ing	 	
the 	reservation price on the current LCOE (levelized cost of electricity) estimate for of	fshore wind at 	
Sørlige Nordsjø from NVE, with 	a reasonable margin. 	 	
The ultimate 	sign of 	success is that the wind	-farm 	produces electricity	 	
A successful 	outcome of the 	auction is not guaranteed until the wind	-farm begins operation. 	To put it 	
differently th	e auction would 	ultimately 	be a failure if the 	wind	-farm 	is not built. 	With a slight misrepre-	
sentation of theory we denote th	e risk of non	-operation	 as the winner’s curse. 	Minimizing 	the 	risk of 	
the 	winner’s curse 	should be 	emphasised 	both when 	choosing the support scheme and the auction 	
format	, and	 we 	have phrased our recom	mendations accordingly. 	 	
In a strategy to reduce the risk of winner’s curse the 	Ministry can 	also help by 	removing uncertainty 	
about common 	costs	 and regulatory elements	. T	he Ministry can reduce the risk of winner’s curse in the 	
following ways:	 	
• 	Remove regulatory uncertainty	: The Ministry should aim to clarify all 	profitability	-relevant regula-	
tions before the auction begins. 	Important	 example	s are	 the 	possibility for future t	axation, or the 	
outcome of the application for concession.	 	
• 	Reduce uncertainty about grid costs	: The Ministry should 	reduce uncertainty about the radial con-	
nection point and grid connection costs.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	9 	
 
In addition, it is helpful that the Ministry funds and 	shares information about geotechnical assessments	. 	
Well	-designed pre	-qualification criteria, and penalties for defaults are 	other tools.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	10	 	
 	
1	 	Introduction	 	
The report covers three main themes:	 	
• 	Pros and cons of different support schemes	, that is, what should be	 the good to be auctioned and 	
the object of bids	. We 	recommend that the Ministry 	either chooses upfront investment support or 	
a two	-sided CfD. If a CfD is chosen the reference period should be one year. 	 	
• 	Pros and cons of different auction formats	. We recom	mend 	that the Ministry cho	oses 	open bid-	
ding 	followed by a final sealed bid among the two 	top	 bidders	, the so	-called Anglo	-Dutch format.	 	
The reference price should 	equal	 the 	most recent NVE estimate of the levelized cost of electricity 	
at Sørlige 	Nordsjø II	, plus a reasonable margin	. 	
• 	Additional regulatory questions of relevance for the auction, including 	some implications of 	pre	-	
selection for the auction and 	a discussion of penalties and timelines	. We recommend that 	uncer-	
tainty about future regula	tor	y processes and costs be removed as much as possible, and 	that the 	
number of pre	-qualified bidders should be at least eight if possible.	 	
This is a 	high	-level strategic document. When 	the support scheme and 	auction format is chosen by the 	
Ministry 	we int	end to provide advice on 	the 	details of the chosen scheme and format.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	11	 	
 	
2	 	Choice of support scheme	 	
An important issue in auction design, and indeed in the regulatory 	embedding of the auction, is	 the	 type 	
and form of support	 payment	 (support scheme)	, and in 	an auction the	 auctioned good	. Should it be a 	
support payment	 per 	unit of energy produced	 (operating support)	, or per	 capacity installed (	investment 	
support	) or 	something else? 	We call this the choice of support scheme. In this chapter w	e discuss pros 	
and cons of 	different 	schemes	. We draw on theory and on practical experiences from previous auctions	 	
in the renewable energy field. 	 	
Although we 	mainly 	frame the discussion in terms of 	the structure of the suppo	rt from the state to the 	
opera	tor	, the 	discussion is just as relevant for 	an outcome with 	payment	s from the 	operator to the 	
state, and it is an important consideration that both 	outcomes should be possible and yield good results.	 	
There are many 	types of support schemes, and many possi	ble design variations for each broad type of 	
scheme. 	Our first conclusion is that two schemes are most promising: upfront 	investment 	su	bsid	ies	 and 	
two	-sided CfD	s. 	
Our second conclusion is that upfront 	investment 	subsid	ies	 per capacity installed	 and 	CfDs 	have diff	er-	
ent advantages and disad	vantages, and which support scheme is preferable depends on 	what criteria 	is 	
given most weight as well as specific be	liefs regarding how re	levant	 the problems with each support 	
scheme 	are	. If a CfD is chosen	, we 	recommend the use of a long (annual) reference period to 	ensure 	
that the operator faces short term market signals.	 	
2.1	 	Which schemes are relevant? 	 	
We briefly present 	a range of possible 	support 	schemes	 in 	Table 	2.1, before 	narrowing	 it down to the 	
most relevant schemes 	for	 the more detailed discussion in	 Section	 2.2	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	12	 	
 	
Table 	2.1 	Some possible support schemes	1 	
Scheme	 	Description	 	
Upfront 	investment 	subsidy	 	
NOK	 per 	MW installed	 	
 
The object of bids is the investment 
subsidy and the winning bid is the low-
est subsidy.	 	
The developer receives a	n upfront	 subsidy 	per capacity installed	 in NOK	, an invest-	
ment subsidy	. 	
Fixed premium	 	
NOK per MWh	 produced	 	
 
The object of bids is the subsidy per 
unit produced and the winning bid is 
the lowest subsidy in NOK per unit pro-
duced.	 	
The developer receives a subsidy in NOK per unit produced	 on top of market re	ve-	
nues	. R	evenue 	per unit time 	is given by:	 	
(�+������)������ 	
Here, p is the market price	, s is the subsidy and x is production, all per unit of time. 	 	
 
 	
One	-sided CfD 	 	
Payment from state to bidder when 
market price is below 	CfD	- strike 	price	; 	
no payback from bidder to 	government 	
in case market price is a	bove strike 	
price	 	
NOK per MWh	 produced	 	
 
The object of bids is the strike price 
and the winning bid is the lowest strike 
price.	 	
The developer is guaranteed a minimum price, as the state pays the difference be-
tween a pre	-defined strike price and the market price. With strike price 	�min	 per 	
unit of time 	the state pays the difference 	�−�min	 if �<	�min	. 	
 	
Two	-sided CfD	 	
Payment from state to bidder when 
market price is below CfD	- strike 	price	, 	
and from bidder to 	state when it is 	
above	 	
NOK per MW	h produced	 	
 
The object of bids is the strike price 
and the winning bid is the lowest strike 
price.	 	
The state pays/receives the difference 	between strike price and reference market 	
price 	whenever the market price is below/above the strike price. 	With strike price 	
� per unit of time 	the 	state (1) pays the difference 	�−� if �<	� and (2) receives 	
the difference 	�−� if �>	�. 	
A two	-side	d C	fD is a subsidy if the strike price is high	er than	 the 	expected market 	
price. 	 	
 
Strike	 price: 	If the reference market price is 	determined 	hourly	, the CfD is essen-	
tially a fixed price contract	 since prices in NO2 are set on an hourly basis	. If the ref-	
erence market price is the 	yearly 	or monthly 	market price, or some other longer 	
period, the developer receives/pays the difference between the strike price and a 
reference price which is not necess	arily equal to the actual hourly market price	 	
(i.e.	, the CfD premium is adjusted annually / monthly in that case).	 Hence, longer 	
reference periods 	provide	 short	-term market integration incentives, as	 short	-term	 	
market price 	signals remain in place	. 	
 
 
 
 
 	
 	
1 Payment schemes are symmetric to the listed 	support schemes	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	13	 	
 
Experiences from other countries	 	
Contract	s for difference (CfD), 	either 	one	-sided or two	-sided 	(also known as double	-sided) 	are a	 com-	
mon format of remuneration in Europe	, and for offshore wind support schemes in particular	. For exam-	
ple, the UK has been using hourly 	two	-sided sliding premiums	 in its offshore auctions since 2015. France 	
has been allocating	 monthly 	two	-sided CfDs in its offshore wind auctions since 2017 	(e.g., for 	Dunkirk 	
and Normandy sites and floating offshore	 OWPs in Mediterranean Sea & Brittany). 	 	
Denmark 	and the United Kingdom 	ha	ve	 used two	-sided CfD	s in multiple 	successful 	auctions. 	The 	third 	
round of auctions in the UK in 2019 saw bid prices of around 	£ 40 per MWh. 	In Denmark successive 	
rounds of auctions 	have generally seen lower prices, with 	prices 	of € 49.9 per MWh for Kriegers Flak 	in 	
2015	 (Jansen, et al., 2022)	. In the 	latest 	Danish 	auction	 of this type	, the	 Thor tender, 	two	-sided CfDs 	
based on an annual 	reference market price	 have been implemented	. Moreover, payment streams that 	
can be made from the winning bidder to the state wer	e capped at 2.8 billion DKK.	 Thor OWP was 	
awarded in November 2021 to RWE at a strike price of 	0.01 ore/kWh. Since five bidders bid the same 	
price, the winner of the tender was decided by lot. A CfD with a strike price of (almost) zero implies that 
a developer must pass on all revenues generated from energy production to the Danish state	 for the 	
dur	ation of the 	support scheme	. The reason why this was a rational decision by bidders is because of 	
the	 cap on the payment streams. Hence, once the first 2.8 billion DKK are paid to the Danish state, the 	
CfD is fulfilled, and RWE can keep all revenues from a	ny further production at the market price. The cap 	
is thus similar to a negative bid to pay 2.8 billion DKK to the Danish state to build the OWP.	 Generally, 	
Denmark ha	s seen a 	high rate of realization of winning bids for offshore wind	, whilst low penalties	 for 	
non	-completion in the UK has given rise to some 	option bidding 	(Welisch & Poudineh, 2019)	. 	
The	 Netherlands	 and Germany	 have been using 	one	-sided CfDs	 in the past	. In 	Germany three one	-sided 	
CfD auctions in 2021 	had winning bids of € 	per MWh. 	Reasons for these low bids could be that the 	
owner have a long time span 	to make a final investment decision	 and can therefore wait for lower costs 	
and higher prices, that grid c	osts are socialized	 and that	 winning	 bids were required to 	retain the option 	
of delivering planned projects in the future	 (Jansen, et al., 2022)	. The specific design of these auctions 	
can therefore be said to encourage a 	certain level of option bidding.	 	
The Netherlands has recently switched to 	qualitative criteria 	(also sometimes referred to as 	a beauty 	
contest	) selection of projects	 without allocating any form of subsidy. Moreover, 	Germany will switch to 	
a new offshore wi	nd support scheme that differentiates between centrally and not centrally predevel-	
oped sites as of January 2023.	 For 	not 	centrally pre	-developed sites, 	an auction	 for the allocation of	 one	-	
sided CfD	s continues to be 	foreseen	. In case of multiple zero bids, 	an 	additional dynamic bidding proce-	
dure on 	a concession 	payment from the developer to the state	 that will be	 split over 	20 	years	 is orga-	
nized	. This means that 	the auctions contains to possible	 bidding component	s, with the s	econd being 	
introduced if the CfD	 price is zero	. This is therefore not a pure CfD	-auction.	 For centrally pre	-developed 	
sites, s	election of a bidder is based on a negative bid on the amount of a payment to the state plus 	
additional qualitative 	criteria (but no support payment is allocated)	. 	
Fixed premiums have been used in 	the Danish open	-door procedure	, under which developers can 	sub-	
mit 	unsolicited application	s for	 a license for preliminary investigations	 (i.e., 	no auctions). 	Upfront	 invest-	
men	t support	, to our knowledge, has not been 	allocated	 in European offshore wind auctions.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	14	 	
 	
2.1.1	 	Removing the least promising support schemes	 from consideration	 	
The first step in our analysis is 	to narrow the options by removing the least promising support schemes 	
from further consideration. 	This allows us to go in	to more detail in analy	sing 	the more promising op-	
tion	s. Our method for narrowing down is a pair	-wise comparison o	f two similar support 	schemes. When 	
one 	has clear drawbacks	 according to important criteria	 for this auction	 compared to a broadly similar 	
support scheme, it 	will 	be dropped for further analysis.	 	
One	-sided vs two	-sided CfD	 	
The first pair	-wise comparison is between one	-sided and	 two	-sided CfDs. 	They are similar in that they 	
change the price fa	ced by the project owner	 from a variable market price to the CfD	-price	, and 	shift risk 	
from the bidders/project owners to the state. 	The primary difference is that a two	-sided CfD 	allows 	for 	
payments from the project owner to the state as well as from the state to the project owner	 if the 	
project is expected to be profitable without support or if the market price becomes higher than ex-
pected.	 In a two	-sided CfD, the expected bid is 	close to the developers	 total	 LCOE	.2 With a one	-sided 	
CfD	 payments can only go one way	, and with a profitable 	project the expected bid is zero	. An overarch-	
ing goal of the 	auction is to achieve the “lowest support/highest payment”. 	 	
If there is 	a need for 	support to realise the project, there wil	l be no difference in the 	expected	 net 	level 	
of support 	given with 	the two schemes. The bidding price with one	-sided CfDs would be somewhat 	
lower to account for 	the upside the project owner can be expected to 	receive in periods of high prices	, 	
but the expected net level of support would be the same.	3 The 	realised	 level of support/payment can 	
however differ between the schemes. If market prices end up higher 	than expected at the time of bid-	
ding, a two	-sided CfD	 ensures that the state captures this upside	, giving a lower realised support. If the 	
market prices end up lower than expected, the somewhat higher 	CfD	-price 	in a two	-sided CfD	-auction 	
means that the realised level of support might be higher than for one	-sided CfD.	 	
If there is no need for support 	from the state to realise the project, then the auction result in 	a one	-	
sided	 CfD will be the minimum possible bid	, generally a CfD	-price of zero	, whilst with a two	-sided CfD	 	
one can expected bids at or close to th	e bidders estimated LCOE	. Zero bids for one	-sided CfD	 are some-	
times presented as a success	, but this situation	 has two undesirable 	effects. First, the state forfeits pos-	
sible payments from the project owner, 	undermining the goal of “highest payment”. Secondly, 	there 	
can be several zero bids, making it difficult to allocate the project to the most efficient potential opera-
tor. 	This can	 be dealt with by 	specifying some 	alternative bidding rule once a 	CfD	-price of zero is met	, 	
as is being planned for some future auctions in Germany	, but this needs to be planned in advanced and 	
complicates the auction.	 We do not recommend an auction with t	wo possible bidding components for 	
the first auction carried out in Norway. 	For SNII	 it is not 	certain	 whether 	there is need for support to 	
realise the project	, even if cost estimates by NVE indicate that	 some support is necessary	. Based on 	
these considera	tions 	we conclude	 that one	-sided CfD are inferior to two	-sided CfD 	given the goals of 	
this auction 	and can be dropped as an 	option in this preliminary stage.	 	
Experiences from other countries	: An important trend in offshore wind auctions in Europe has been 	
lower required levels of support over time, as the technology has matured. There has in the last few 
years been carried out several auctions with bids at, or close to zero. In 2017 and 2018 au	ctions in 	
 	
2 More complex designs, such as a 	cap on paymnents and longer reference periods will change the expected bid.	 	
3 This 	conclusion rests on some assumptions regarding risk and bidding behaviour	, that we will not go into detail here.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	15	 	
 
Germany gave bids of € for more than half the awarded capacity in one	-sided CfD auctions. In 2021 	
three German auctions all gave winning bids of zero in one	-sided CfD auctions. This means that no direct 	
price support was necessary	 and m	arket pa	rticipants signalled that they would be able to construct and 	
operate the offshore wind parks in questions with market revenues alone	. In several auctions there 	
were multiple bids at the minimum price of zero, and the winner had to be awarded by lot. A sim	ilar 	
situation occurred in the recent Danish Thor tender. Here, five of the six bidders bid the minimum price 
of 0.01 øre/kWh, and the winner was chosen by lot. The Thor tender was a two	-sided CfD auction with 	
a maximum pay	back	 from the concession owner to	 the state set at 2.8 billion DKK	 (see above)	. The 	
multiple zero bids show that several companies were willing to pay the state this maximum payment in 
order to receive the right to develop the project.	 	
However, 	it should be noted that 	zero bids are not ta	ntamount to the development of offshore wind 	
being free of public support. In the German auctions, the cost of grid connection is socialised. Grid con-
nection can account for as much as 30 percent of total investment cost and is especially significant 
where	 the distance to the shore is high. In Sørlige Nordsjø II grid connection is bound to be a significant 	
issue. The distribution of grid costs is therefore another important policy choice for the auction. Fur-
thermore, CfDs imply that risk is transferred from	 the producers to the government.	 In addition, 	caps 	
on paybacks 	in two	-sided CfDs 	allow bidders 	to recuperate investment costs 	with the additional ex-	
pected market revenues 	beyond the	 payback amount. 	 	
Fixed premium 	vs ad valorem fixed premium	 	
A second 	pair	-wise comparison is betw	een a fixed premium and a	 premium that is a percentage of 	mar-	
ket price	 (ad valorem premium	). The ad valorem 	premium implies higher transfers from the state when 	
prices 	are high, which	 is unattractive from a political point of vi	ew.	 From	 a market disturbance point of 	
view 	an ad valorem 	scheme 	may have	 less negative effects on 	operation as it does not alter relative 	
prices, but 	we 	do not consider this a major advantage and focus on the	 fixed premium	. 	
Three schemes for further consi	deration	 	
We assess these 	three	 schemes as the most relevant:	 	
• 	Upfront 	investment 	subsid	y 	
• 	Two	-sided CfD	 with 	a short	 (hourly)	 reference period	 	
• 	Two	-sided CfD with a long (mo	nthly or yearly)	 	
In addition we also discuss a fixed premium	 scheme. 	This can in some 	respects be seen as an interme-	
diate 	option between upfront investment subsidy and two	-sided CfD, with many similarities with 	the 	
latter when it has a long reference period.	 	
A fundamental difference between the 	upfront scheme and the two others is that the 	upfront scheme 	
incentivizes	 investment	 through a subsidy at the investment stage, while the others 	incentivize	 invest-	
ment thro	ugh a subsidy 	during	 the operations phase. 	As	 will be clear from the subsequent discussion, 	
an advantage of the upfront scheme is that it does not affect the design or operation of the 	wind farm	. 	
A disadvantage	 is that it 	may 	increase the risk of 	project default.	 	
Being the least complex	 scheme, the upfront scheme 	will be	 part of the more detailed discussion	 in 	
section 	2.2	. As two	-sided CfDs are widely used and have some interesting properties, th	ese 	will also 	be

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	16	 	
 
discussed more in detail. 	For two	-sided CfDs 	the properties change in important ways	 with 	some design 	
variation, particularly the cho	ice of a long or short reference period (discussed in more detail in section 	
2.3.1	). A short reference period	 means that 	CfD	-payments to or from the 	developer are calculated 	based 	
on the 	difference between 	the CfD price and 	the 	market price at any given time (typically hourly). This 	
means that the project owner is 	completely shielded from price fluctuations. A longer reference period 	
of a month or even a year, means that the 	payments are based on the 	difference between average 	
ma	rket price over the re	ference period and the CfD	-price. This exposes 	the project owner to 	more short	-	
term variation in market prices. 	Hence, 	less 	market risk being shifted from the 	project owner to the 	
state, but 	improved incentives	 to 	adapt to market pric	es. This is an important trade	-off.	 	
Finally	, we have included 	the fixed premium scheme	. This scheme retains some properties of the up-	
front 	subsidy, in particular the exposure, at the margin, to volatile prices; while also retaining some 	
properties of the CfD, in particular be	ing a subsidy during the operation phase. 	By extension, a two	-	
sided CfD with an annual reference period is equal	 in its economic properties to a 	fixed premium that is 	
revised annually.	 	
2.1.2	 	The scheme should allow for 	ou	tcomes with payment	 	
We recommend that the auction allow	s for outcomes where the winner 	pays	 the government for the 	
right to apply for a license. 	Allowing for payment ensures that the government receives a compensation 	
for letting a developer 	utilize	 valuable natural resources, 	should	 offshore wind 	be 	profitable 	at the auc-	
tion stage	 without support.	 	
This recommendation is 	in accordance with one of t	he primary objectives of the auction, which is t	o 	
achieve low support or 	high income	. Furthermore, 	based on cost and price estimates from the Norwe-	
gian Water Resources and Energy Directorate (“NVE”)	, we 	find that 	there is a possibility that 	offshore 	
wind a	t SNII is profitable without government support.	 The NVE estimates are presented	 in more detail	 	
below.	 	
All of the examined schemes can allow for 	payments to the state from the developer with the proper 	
design. 	For a two	-sided CfD, any bid price below 	the expected market price entails an expected payment 	
to the state. For an upfront subsidy	, negative bids must be allowed in order for this 	possibility. One can 	
envisage mechanisms where a	 modified	 one	-sided CfD also 	allows for payments to the state if t	he	 type 	
of bidding changes if a price of zero is reached. 	Such a hybrid	-scheme would be complex and is not 	
examined further here.	 	
Profitability of 	offshore wind at 	SNII	 	
In 202	1 NVE	 estimated	 the 	profitability of fixed offshore	 wind at SNII	, in relation to 	a government white 	
paper	 on energy policy.	4 NVE 	estimated 	that the	 levelized cost of energy (LCOE) 	would be	 78 øre/kW	h5 	
if such a 	plant 	was to be built 	today	, inclu	ding 	offshore 	grid costs	 and using a 	6 % real discount rate	. 	
Taking into account expected cost reductions, NVE estimates 	an	 LCOE 	between 48 and 68 øre/kWh	 in 	
2030	. The 	LCOE essentially 	is the 	average (real) electricity 	price 	a project must 	have throughout its life-	
time	 in order to break	-even.	 	
 	
4 Meld. 	St. 36 (2020	–2021) Energi til arbeid	 	
5 10 NOK = 	€1. 	Equals €78/MWh.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	17	 	
 
The NVE estimates 	can b	e seen from 	Figure 	2.1, which also includes the estimates for floating offshore 	
wind at Utsira Nord.	 	
Figure 	2.1 	NVE estimates of LCOE for Utsira Nord and Sørlige Nord	sjø II	, including offshore grid	 
 	
Source:	 	White paper on energy 	policy (Meld. St. 36 (2020	–2021) Energi til arbeid)	 	
In NVE’s most recent 	long	-term power market analysis	, from 2021, the 	electricity 	price in Norway’s 	
southernmost bidding zone NO2 was 	estimated 	to be in the range 	of 51	-54 øre/kWh	 in the 	analysed	 	
period from 2025 to 20	40. 	 	
Whether the	 LCOE and the expected market price	s are close or far apart is a matter of judgment	. On 	
the one hand the	 estimated cost exceeds the price by around 50 percent	. On the other hand	, the	 price 	
and costs	 are not an order of magnitude apart. It 	cannot be ruled out 	that 	fixed offshore wind at SNII	 	
will be	 profitable	 without support.	 Although the current sky	-high 	electricity prices 	throughout Europe 	
and southern Norway 	are expected to fall	, we believe an	 upd	ated long	-term price estimate 	could	 po-	
tentially	 be above	 the estimate from 2021. 	For instance,	 NVE assumed a	n EU ETS price of 	44	 EUR/ton 	
CO2 in 2025, increasing to 55 EUR/ton in 2040, while the price 	has fluctuated between 60 and 90 	
EUR/ton through	out	 most of 2022	 ind	icating that the long	-term estimates of ETS prices in the forecast 	
might be too low	. Higher 	emission prices 	lead to 	higher electricity prices so long as fossil 	fuel	-based	 	
production is the marginal 	producer.	 Statnett’s most recent Short	-term market analysis for 2022 to 2027 	
revise price expectations upwards and indicates a price in Southern Norway of 50	-70 € per MWh in 	2027 	
(Statnett, 2022)	. Whilst 2027 might be before SNII can be expected to come online, this nonetheless 	
indicates that the expected market prices	 in the medium term	 are close to the LCOE	 estimated by NVE	. 	
Moreover,	 costs 	estimated 	without carrying out detailed planning of the project 	are uncertain and 	NVEs 	
estimates of costs may 	therefore 	be overstated	. There have been several 	auctions for offshore wind in 	
Europe where the result has been 	no support 	required to reali	se the projects. These have typically been 	
in shallower water and closer to shore than 	SNII, and in some case	s the grid connection costs have 	been 	
borne by the state	. These auction results nonetheless show that offshore wind is approaching 	profita-	
bility in Europe. In D	enmark 	the 	Danish Energy Agency (DEA)	 has instituted 	an «open	 door scheme	” for 	
new offshore wind, and has received interest for subsidy	-free construction of several projects at similar 	
or greater distance from the shore	 to that	 of SNII 	(Energistyrelsen, 2022)	. An argument going in the 	
opposite direction is that the	 current high prices of raw materials and inflation could 	lead to a lasting 	
increase in the cost of new projects, whilst financing cost ma	y increase due to higher interest rates.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	18	 	
 
Furthermore, there could be learning effects which further increase the value of this project to devel-
opers. Developing th	e project could also have a symbolic value to some potential bidders, as it is the 	
first larg	e-scale offshore wind project in Norway. For these reasons as well, it is conceivable that the 	
winning bid will take the form of a payment.	 	
2.2	 	Assessment of relevant schemes	 	
In this section we assess the properties of the relevant schemes 	with regards to the following issues	: 	
• 	Market integration and possible distortions 	 	
• 	Market p	rice risk 	 	
• 	Risk	 of winning bidder abandoning the project	 	
• 	Liquidity	 	
• 	Fiscal policy framework	 	
• 	Uncertain	ty about future t	axation	 	
• 	Administrative costs	 	
• 	Possibility of 	overcompensation (ex	-post)	 	
We will argue in section 	2.3	 that market integration/avoiding distortion	s incentives, and 	price risk	, are 	
the most relevant and pressing concerns	, whilst the other issues listed are of secondary importance	. 	
2.2.1	 	Market inte	gration and 	possible distortions 	- Effects on	 design	 and operation	 	
In this section we look at several different ways in which the support scheme can give distorted incentives	 	
to the producer	. Overall	, the conclusion is that an upfront subsidy has the best characteristics on this 	
issue, and that a CfD with a short reference period has the worst 	characteristics	. Longer reference peri-	
ods, as well as other adaptions of a CfD, reduce 	and in some cases remov	e the distorting effects on 	
producer incentives.	 	
The 	CfD	 and the fixed premium have the property that the price received by the developer/operator of 	
the plant 	generally differs from the market price. 	The market price is known to send a signal of willing-	
ness to pay 	for electricit	y. Hence, in 	a CfD	 and the fixed	 premium the price signal is distorted. 	The dis-	
tortion might affect how 	the wind farm is designed and 	operat	ed	. 	
Even though the scope for distortions is limited 	in intermittent power generation 	where production is 	
largely given by the wind resources, we ha	ve identified potential challenges related to the 	following 	
situations/issues:	 	
• 	Low/negative value of production	 	
• 	Decisions affecting the t	iming of production	 	
• 	Grid integration and 	supply of reserves (	ancillary services	) 	
Note that 	the list is of	 potential	 distortions of developer behaviour. 	 We 	discuss in each section 	whether 	
the	 distortions 	are likely	 to be an issue	 and the 	severity of their consequences. We consider 	issues re-	
lating to low/negative value of production to be likely to 	arise in pra	ctice, but 	the consequences can be 	
largely mitigated through 	design choices. Issues relating to timing of production are modera	tely likely

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	19	 	
 
to arise	. Problems of g	rid integration	 becoming 	a significant issue is something	 we consider to be more 	
hypothetical	 in the Norwegian	 context with ample flexible production	 in NO2	. 	
The upfront 	subsidy does not distort the price signal	. A	 subsidy on capacity may 	induce	 the 	operator to 	
emphasise capacity/effect over	 operation	/energy	 but	 in the case of SNII capacity	 of the grid connection 	
is capped at 1,4 GW and the maximum allowed production capacity	 for the project is 1,5 GW	 (taking 	
into account loss in the grid connection and 	production not being at capacity at all times)	. 	
Low/negati	ve value of production	 	
During certain	 times	, the 	gross 	value of 	electricity 	production 	(MWh)	 from the wind farm can be low, 	
zero or negative. This could be the case with high intermittent electricity production, demand saturation 
and/or limited grid 	capacity.	 In such situations it could be efficient for the	 wind farm to limit or stop 	
production.	 The 	CfD	 and fixe	d premium 	will 	completely (	CfD	 with short reference period	) or partially 	
(premium	 or CfD with a long reference period	) insulate the wind farm 	from the 	market price and stim-	
ulate the farm to continue operation	 in these cases	. 	
Requirements in the revised EU 	Climate, Energy and Environmental 	State 	Aid	 Guidelines	 (CCEAG) 	that	 	
the 	CfD	 payment 	must	 be discontinued if the market price is 	negative	, effectively removes this issue	.6 	
With CfDs with long reference periods	, problems can still arise when the 	short	-term 	market price is low	, 	
if the 	average market price for the reference period is above the CfD	-price	. When the 	average market 	
price 	for the reference period is above the CfD	-price	, the operator must pay	 per kWh	 the state the 	
difference 	between these two prices. If the short term (hourly) market price goes below th	e sum the 	
operator 	must	 pay per kWh, then 	the operator has an incentive to 	shut off production. 	This is not 	de-	
sirable as production with a low but positive market price still has value. This 	issue can be dealt with by 	
suspending CfD	-payments to the state	 for all pro	duct	ion 	at times when the market	 price is below the 	
level where this issue arises.	 	
Properties of the	 various schemes are discussed in 	Table 	2.2, with	 c as 	marginal cost,	 � as the market 	
price, 	������ as the fixed premium subsidy, 	� as the CfD strike price and 	���������	 as the CfD reference	 price in 	
the case where it differs from the mar	ket price.	 	
 	
6 Point 123, CCEAG: «beneficiaries should not be incentivised [...] must not receive aid for production in any periods in which	 	
the market value of that production is negative». 	https://eur	-lex.europa.eu/legal	-con-	
tent/EN/TXT/PDF/?uri=CELEX:52022XC0218(03)&from=EN

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	20	 	
 	
Table 	2.2 	Distortions 	in cases with	 low/negative 	value of production	 	
Scheme	 	Distortion	 	
Upfront subsidy/payment	 	None	 	
Fixed premium	 	Wind farm stops 	production	 too late: 	The producer price is 	
�+������. The m	arket price 	� must then be below 	������−������, possi-	
bly below zero, for the wind farm to stop.	 The scheme 	
causes the wind farm to continue production in cases where 
production has negative value.	7 	
Two	-sided CfD (Hourly)	 	Wind farm does not stop	 production	: The producer price is 	�, 	
i.e. fixed. The wind farm thus is unresponsive to the market 
price	. The scheme causes the wind farm to continue produc-	
tion in cases where 	�<	������, price is lower than cost, 	where 	
production has negative value.	 	
Two	-sided CfD (Yearly or another period 	
longer than hourly)	 	 	
Wind farm stops	 production	 too late: 	The producer price is 	
�+(�−���������	)>	�, where 	(�−���������	) is the annually 	
based	 CfD subsidy. 	The CfD essentially works in the same 	
way as the fixed premium subsidy, with 	�−���������	=	������. The 	
scheme causes the wind farm to continue production in 
cases wher	e production has negative value.	 	
Timing of production	 	
Although the time of production is largely given by the wind resources, 	there are some degrees of free-	
dom	 both in the design and operations phase	, which means that the price faced by the 	operator	 is 	
relevant for	: 	
• 	Maintenance:	 It is likely that	 there is	 some	 flexibility in when	 planned 	maintenance 	is carried out	.8 	
It would then be beneficial to perform maintenance at times where the 	value	 of production is low-	
est	, not when the 	volume	 of production is lowest	. 	
• 	Batteries and other storage solutions:	 Storage solutions might make it possible to 	shift the time at 	
which power is delivered to the grid, from low to high price hours. 	We have not 	investigated	 	
whether this is a 	current 	possibilit	y for SNII.	 It may be	come	 more likely 	as 	the 30 year	s concession 	
time 	progresses	. 	
• 	Choice of 	configuration in design phase:	 there might be 	trade	-offs in the design phase	.9 Maximising 	
production volume 	could lead to different choices 	than maximising production 	value	 if choosing a 	
design that allows for more pro	duction when 	prices are expected to be high	 entails lower produc-	
tion at other times.	 	
These issues are relevant only for support schemes where the 	operator faces a 	different price at the 	
margin than the market price. With a CfD with a 	hourly reference period, the operator faces 	only 	a 	
constant	 CfD	-price	 and will 	therefore seek to maximise produced volume and minimise costs, 	with no 	
regard to the market price. 	By contrast, an operator that receives 	an upfront subsidy	 or 	an annual 	
 	
 
8 The 	ease and cost of doing maint	enance is also highly seasonally dependent	, with maintenance in harsh winter months 	more 	
complicated and expensive. This is in all likelihood a more 	important consideration than adaptation to short term price 	
fluctuat	ions	 	
9 The trade	-off could be between d	ifferent 	turbine designs with different 	power curve (	e.g. a different cut	-off or cut	-in wind 	
speed), or in 	the lay	-out of the 	wind farm	 to prioritise output for different wind directions and 	wind 	strengths.

Norwegian Offshore W	ind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	21	 	
 
average CfD	-price will be exposed to hourly market prices at the margin and there will	 be no distortion 	
to the timing of production.	 	
 	
Table 	2.3 	Distortions in	 the timing of production	 	
Scheme	 	Distortion	 	
Upfront subsidy/payment	 	None. 	 	
 
When costs are constant the unit gain from switching production between peri-
ods is 	(�ℎ−�������), i.e. the difference between	 the high price	 (�ℎ) and the low 	
price	 (�������). 	
Fixed premium	 	None.	 	
 
The loss from 	low production when the price is low:	 	
−((�������+������)������−������) 	
The gain from high production when the price is high	: 	
((�ℎ+������)������−������) 	
 
Here, 	������ is cost, x is production and s is the subsidy. 	�ℎ and 	������� as above. 	The 	unit 	
gain from 	switching production between periods	 	
(�ℎ+������)−������−(�������+������)+������=	(�ℎ−�������) 	
 
(�ℎ−�������) is the same expression as in the case of upfront subsidy, in other 	
words, no distortions. 	 	 	
Two	-sided CfD (Hourly)	 	Wrong	 incentive to	 change production from low	-price to high	-price hours. 	 	
 
The loss from low production when the price is low:	 	
−((�1+������)������−������) 	
The gain from high production when the price is high:	 	
((�2+������)������−������) 	
 
Here, q	1 and q2 are two hourly 	CfD	-price	s and the other symbols as above. 	 	
The unit gain from switching between periods	 	
 	
(�1−������)−(�1−������)=	(�1−�2) 	
 
(�1−�2) is in general different from 	(�ℎ−�������), although it is 	possible 	that it 	
will have the same tendency (the bidding will 	take costs during 	periods of high 	
market prices into account	). 	
Two	-sided CfD (Yearly or 	
another period longer than 
hourly)	 	
None. 	 	
 
The CfD essentially works in the same way as the fixed premium subsidy, with 
�−���������	=	������. 	
I.e., the 	absolute price difference	 is unaffected	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	22	 	
 
Incentives to 	integrate 	efficiently	 to the grid and provide a	ncillary services	 	
The increase of intermittent 	renewable power sources makes managing 	electricity grids 	more 	challeng-	
ing and	 can 	give rise to a	 need for expensive	 system	 investments. 	The 	low market prices 	in periods with 	
high production of intermittent renewable energy	 and higher prices 	when intermittent production is 	
low	, gives a	n incentive for energy producers to 	invest in 	production or storage capacity that reduces 	
these problems. CfDs with short reference periods shield 	producers 	from this market price. This has 	
bee	n raised as an issue in the UK	 and 	has been put forward as an argument for 	discontinuing the use of 	
CfDs 	(Catapult Energy Systems, 2021)	. Norway has 	large amounts of 	flexible renewable energy in the 	
form of hydropower and has 	an electric	ity market where most consumers	 and producers	 face 	short	-	
term price	 signals	, giving ample incentives for 	adaptations. 	We therefore consider this to be a 	less	 im-	
portant issue	, than in ot	her countries that have noneth	eless opted to use CfDs.	 	
Anc	illary services 	(frequency stabilization reserves) 	are services 	which	 producers of electricity provide 	
beyond simple generation in order to balance the grid in the 	short term. 	Ancillary services fr	om offshore 	
wind power could become a	 viable alternative in the future, in some circumstances. We have not as-	
sessed the 	potential value of such reserves. However, we note that the TSO Statnett in recent years 	
have been 	investigating	 other sources 	than hydropower with reservoir for contributing reserves	, with 	
the development	 of fast frequency responses (FFR	) to	 help stabiliz	e the grid in situations 	with little in-	
ertia in the power system	, for instance	 due to high intermittent power production.	 	
It is thus relevant to assess whether the support/payment scheme affects the incentives of offshore 
wind farms to offer ancillary serv	ices. 	 	
Without energy storage, provision of a	ncillary services (frequency stabilization reserves) impl	ies	 re-	
duced	 “standard” 	production	: 	
• 	Downwards adjustment is by itself a reduction	 	
• 	Upwards adjustment require	s available capacity, i.e. that 	ordinary	, baseline	 production 	is reduced 	
beforehand	 	
Distortions with different schemes	, without energy storage,	 are discussed in 	Table 	2.4.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	23	 	
 	
Table 	2.4 	Distortions in 	supply of ancillary services, without storage	 	
Scheme	 	Distortion	 	
Upfront subsidy/payment	 	None	 	
Fixed premium	 	To provide ancillary services the operator gives up unit pro-
duction at 	cost	 (�+������). This 	cost	 is higher 	than the market 	
price (	�) 	
�+������>	� 	
Will 	supply	 too little	 ancillary services.	 	 	
Two	-sided CfD (Hourly)	 	Cost of 	lower production is	 higher, and 	too little	 ancillary ser-	
vices supplied, if the strike price is above the market price:	 	
�>	� 	
Cost of producing less will be lower, and 	too much	 ancillary 	
services supplied, if the strike price is below the market 
price:	 	
 	
�<	� 	
If reserves from	 offshore wind power 	primarily are	 interest-	
ing in the case 	of much wind, little hydropower production 	
and low prices, the former case (	too little	) is the most likely.	 	
Two	-sided CfD (Yearly or another period 	
longer than hourly)	 	
The CfD essentially works in the same way as the fixed pre-
mium subsidy, with 	�−���������	=	������. 	
Thus, offshore wind 	will supply 	too little	 ancillary services.	 	
 
 	
With storag	e, such as batteries, 	the	 discussion of distortions is much the same as in the discussion of 	
timing of production. However	, whether the supply of ancillary services is too low or too high depends 	
on whether the actual case necessitates shifting from low to high price hours, or the	 other way around.	 	
We consider the issue of disincentives for the provision of ancillary services to 	be of 	less importance in 	
Norway compared to other countries	, as hydropower production 	largely fulfils this role in the Norwegian 	
grid today	, but the importance of this issue is increasing 	(Statnett, 2016)	. 	
Offshore	 buyers of	 electricity	 	
There is a possibility that the wind park could in the future 	have the option of selling the 	ele	ctricity	 that 	
is produced to other destinations than the Norwegian grid. This could for example be to 	offshore uses, 	
such as oil platforms or 	perhaps 	offshore 	production of green hydrogen or synthetic fuels.	 With upfront 	
subsidies, the operator of the wind park wil	l be willing to enter into such contracts as long as the price 	
offered is higher than the expected price in the grid	 onshore	, taking into account the possible transmis-	
sion loss 	when	 transferring power via the DC	-cable to the grid.	 Offshore uses of power, i	f they materi-	
alise, would be expected to have a high 	willingness to pay, as it is costly to supply energy to 	such uses. 	
With a CfD,	 normally 	only sales to the grid give rise to CfD payments. Therefore	 the operator would only 	
be willing to enter into contra	ct if the price paid by the offshore user of po	wer exceeds the CfD	-price	. 	
With a CfD	-price that is above the expected market price, this produces a distortion	 as the most efficient 	
outcome is if the 	operator 	is willing to sell 	power offshore as long as the	 price	 offered	 is higher than the 	
market price	. This inefficiency only arises if offshore users	’ willingness to pay is between the market 	
price and the CfD	-price.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 202	2/44	 	24	 	
 
With a CfD	-price 	below	 the market price	, the operator would have an incentive to sell the power at a 	
price that is 	above	 the 	CfD	-price	. If they sell the production at a price that	 is below	 the market price	, 	
this 	reduc	es the total value of produced power and depriving the state of 	income from th	e CfD. Prohib-	
iting sale of 	electricity 	to offshore uses 	if the CfD	-price is below the market price	, would solve this prob-	
lem, but at the same time 	stop	 economically beneficial sale of the power to 	offshore uses at prices 	
above the market price.	 To solve t	his issue	 there would need to be an arrangement 	whereby the 	state 	
is paid the difference between the market price and the CfD	-price	, whilst allowing the operator to divert 	
the power to the offshore use	. 	
A better option 	to deal with all issues re	garding offshore uses 	would 	however be to 	include	 power sold 	
to offshore uses in the calculation of CfD payments	. In this case the 	payment 	to or from 	the operator 	
sh	ould 	be the difference between the 	onshore market price and the CfD price for all 	production	, giving 	
the 	operator an incentive to sell to offshore uses only when the offshore price is higher than the onshore 	
market price.	 	
2.2.2	 	Market price risk	 and financing cost	 	
In this section we look at how the different support schemes shift risk, and wh	ether this is a relevant 	
consideration. The conclusion is that a CfD shifts market risk from 	the developer to the state. 	This will in 	
turn reduce the 	developers	’ financing costs.	 If the state has a lower cost of carrying (systematic) risk, this 	
is desirabl	e, but if the cost is the same, this issue is not relevant for the choice of support scheme.	 	
There is 	risk to	 the future income 	from offshore wind power production.	 The 	relevant risk 	to an owner 	
and the state alike	 is the risk that cannot be 	dealt with through diversification	, called systematic risk.	 	
One major source of 	risk	 is the future power price. 	A feature of the two	-sided CfD is that it	 transfers 	
market price 	risk from 	the developer to the state. 	With 	the hour	ly market price as the reference price, 	
a two	-sided CfD is essentially a	n hourly based	 fixed price	 contract. The entire 	market price	 risk	 (up to an 	
eventual cap 	on payments)	 is transferred to the state, which covers 	(receives) the difference with low 	
(high	) market prices.	 With a longer reference period, some market price risk is 	transferred from the 	
developer to the state, whilst some market risk remains with the developer.	 If there is positive system-	
atic risk associated with the income from 	offshore wind production, a risk reduction is valuable to the 	
developer. To put it simply, there is positive systematic risk if income is higher than expected in good 
times and lower than expected in bad times. A high discount rate is then warranted, since	 the value of 	
higher income in good times does not outweigh the cost of lower income in bad times.	 	
There is still a possibility that the grid connection to SNII will be 	required to be ready to become part of 	
a future meshed offshore grid, or 	to supply offs	hore users. 	A meshed offshore grid would most likely 	
have its own set of prices in an offshore bidding zone. It is 	very difficult for bidders to 	predict what the 	
market prices will be in a	 hypothetical	 future 	offshore bidding zone	, or when 	a meshed grid mi	ght be 	
ready. 	This is an additional source of price uncertainty	. 	
In principle the 	CfD and fixed premi	um may also amplify volume uncertainty and risk	 as the expected 	
support is a function of 	the volume produced	, but we consider this a lesser problem in prac	tice	. 	
A third 	source of 	risk 	is uncertainty regarding building cost. This uncertainty is not affected by 	the types 	
of support scheme discussed here and 	is therefore not relevant	 for our discussion	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	25	 	
 
The question 	then becomes	 whether 	risk is less, 	equally or more costly to the state, than to the devel-	
opers.	 The	 cost of risk has implications for the appropriate	 discount rate	. If the developer 	has a higher 	
cost of risk,	 the value of the project for the developer has to be 	computed at a higher discount	 rate.	 And 	
vice versa.	 	
Table 	2.5 lays out the consequences of 	price 	risk for 	the 	three	 scheme	s under discussion. 	If risk is less 	
costly for the state	, the risk alleviation of CfD pulls in the direction of that scheme. 	If ris	k is 	equally costly 	
for the state	, the allocation of risk does not affect the choice of scheme. 	We find it unlikely that risk is 	
more costly for the state than for the developer, but if so	, risk pulls in the direction of the upfront 	
scheme.	 	
Table 	2.5 	The cons	equences of 	price 	risk for the three support schemes	 	
Cost of risk	 	Consequence for choice of scheme	 	
Less cost	ly for the state	 	Pulls toward CfD	 	
Equally costly for the state	 	Allocation of risk 	does not affect choice of 	
scheme	 	
More costly for the state	 (unlikely)	 	Pulls towards upfront scheme	 	
We	 stated	 above that the developers 	should 	only care about systematic risk.	 We believe this is a rea-	
sonable assumption, given that the potential bidders as (	consortia	 of) 	large corporations. They or their 	
owners have access 	to 	capital markets and can diversify away unsystematic risk.	 The state, as well, cares 	
only for systematic 	risk. 	 	
There is no doubt that transferring risk from the developer to the state will reduce the formers financing 
costs and therefore the cost of realising the project. Stable income from CfDs will contribute to making 
the project more 	“bankable”. This however does not in itself mean that it is desirable to transfer this 	
risk. Carrying the risk has a hidden cost for the state, and the question is whether this cost is lower than 
the reduction in the developers	’ financing costs.	 	
There is h	owever a question of whether systematic risk as seen from a company per	spective is equal to 	
systematic risk as seen from a state perspective. 	Some arguments in this regard are:	 	
• 	Owners of 	the winning operator may own 	stakes in many firms, but even as a grou	p they are 	gen-	
erally less diversified than the state	 	
• 	Managers of 	the winning operator care about risks that from an owner perspective are unsystem-	
atic	 since owners have difficulty separating 	risk events from bad management	 	
• 	The level of uncertainty regardin	g future grid connections 	is lower	 for the state, as it has 	a central 	
regulatory role that can influence the outcomes	. Benefits or 	disadvantages 	for SNII can be inter-	
nalised in 	the decisions of the state.	 	
• 	Owners of the winning operator are free to 	diversify as broadly as the state	 	
• 	The Norwegian state holds 	30 per cent of the considerable 	petroleum fund in bonds	 with low risk, 	
suggesting that the state 	is averse to the average market risk of 	physical investments (as repre-	
sented by the	 stock market	). A higher expected return 	could be achieved 	by 	increasing the stock 	
share and assuming risk	, but it has been chosen not to do so. 	 	
If the cost of	 systematic risk is lower from a state perspective 	in this sector the same would be the case 	
in many more sector	s. 	We don’t see the state 	providing loan	s and guarantees on a grand scale	. The

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	26	 	
 
financial sector is mostly 	commercial.	 On the other hand	, the state in Norway arguably operates as a 	
guarantee fund 	in times of crisis (	e.g.	 bank crisis of 1990 and financial crisis of 2008	-9). 	
2.2.3	 	Risk of 	winning bidder abandoning 	or delaying 	the project	 	
In this section we 	analyze	 whether the support scheme affects the possibility of the operator abandoning 	
the project after the auction due to o	ver	-optimistic bidding. We find that CfDs lessen this risk.	 	
There is a further 	consideration that merits attention	 regarding risk	: Might the operator 	pull out of its 	
obligation to 	build 	the plant if 	expected 	operating	 and ma	rket	 conditions 	become less 	favourable than 	
expected	? Reneging on the obligation is a breach of contract and 	can	 be met with penalties	. Neverthe-	
less	, there might be situations where it is less costly for the winning bidder to	 pay	 those penalties than 	
to 	complete the project o	r even	 continue operations	. From the perspective of the operator, the option 	
of reneging on the contract is a way 	of 	reducing downside risk. From the perspective of the state	, it is a 	
factor that increases risk.	 The probability of the 	developer	 reneging on t	he contract 	increases if there is 	
a “winner’s curse”	, where the winning bidder is the company that has most overestimated the 	value of 	
the project. This phenomenon is discussed in more detail 	in section 	Error! 	Reference source not found.	. 	
If the 	winner’s curse arises from overly optimistic assessment of future market prices, then 	a CfD	-con-	
tract will 	remove	 this	 risk, while 	an	 upfront subsidy at the investment stage 	does not	. Other possible 	
sources of 	wrong expectations	, such as overly optimistic assumptions on costs, are not affected by 	the 	
choice of support scheme. 	From a practical point of view the risk of discontinuing 	the p	roject	 might be 	
more important to 	account for when choosing support scheme than 	systematic risk	 (the preceding sec-	
tion).	 	
A related issue to the risk of the winning bidder abandoning the project, is that the winning bidder 
submits a bid 	with the intention o	f only 	realising it 	when market prices 	eventually 	make	 the project 	
profitable. 	Even if there are penalties 	associated with delaying the project, these penalties can be 	fac-	
tored in as the cost of holding the real option. 	Bidders 	applying 	real option theory	 will value the project 	
higher	 (Martinez	-Cesena & Mutale, 2012)	, and therefore have more competitive bids. 	This is undesira-	
ble 	for the state	 if the goal is a timely realisation of the project. 	Bidding under one’s true cost 	to get a 	
real option 	often takes place in the expectation of higher carbon prices leading to higher electricity 	
prices	, and this problem is 	removed by the use of 	a CfD	 as 	support system	 (Welisch & Poudineh, 2019)	. 	
2.2.4	 	Liquidity	 	
In this section we look at the effect of 	support scheme on liquidity for the developer. We conclude that 	
this is an issue of little importance,	 as the bidders are expected to be 	companies and consortia with 	
significant financial resources and 	broad	 access to capital markets	. 	
The choice of support 	scheme affects the liquidity of the 	project developer. 	The developer is likely to 	
face some financing cost associated with 	increased liquidity needs.	 	
With an upfront subsidy, the developer receives payment from the state	 prior to or close in time to 	
incurring the major investment costs of building the wind par	k. Support given 	in the form of a CfD 	means 	
that the payments from the state occur in the operational phase	, when the developer has lower costs 	
as well as access to income from the sale of the electricity produced. 	If the project requires net support

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	27	 	
 
from 	the 	state to the developer, upfront support is therefore more advantageous from a liquidity per-	
spective for the bidders.	 	
If however the 	wind park is profitable without state support, and the result of the auction is 	that the 	
winning bidder will pay the sta	te, the situation is reversed. An upfront payment will coincide with 	the 	
investment costs, whilst payments 	in the form of CfD will coincide with income generated from the 	
production of electricity.	 	
In any event, we consider 	the effect of the choice of support scheme on the liquidity position of devel-	
opers to be of little importance. 	All bidders are expected to be companies and consortia 	with significant 	
financial resources and 	broad	 access to capital markets. 	Indeed	, the pre	-qualification requirements are 	
supposed to ensure that bidders have sufficient 	financial capacity to carry out the project. 	Therefore	, 	
the bidders are unlikely to face any major liquidity constraints, and the 	effect of the choice of support 	
scheme on their	 liquidity is not a material concern.	 	
2.2.5	 	Fiscal policy framework	 	
The Norwegian state might 	prefer 	payments that are predictable	, transparent and 	which 	fit within the 	
ordinary budget process. An upfront subsidy is 	the support scheme that is most consistent wit	h those 	
goals. 	Using CfDs as the support scheme however means that the annual payments are 	less predictable, 	
that the total cost at the time of the auction is unknown	10, and that 	the state takes on	 a hidden cost 	
through the assumption of risk. 	The 	budg	et process in 	the Norwegian central government has 	methods 	
of dealing with 	these issues	, but they are not perfect. 	It is likely that budget rules in any event will require 	
setting a maximum sum that ca	n be paid in support. 	Whether this is a relevant concern or not 	should 	
be deter	mined by the 	relevant 	ministries and the government.	 It is likely that it is desirable for the state 	
to cap the total payments that can go to the 	operator.	 	
2.2.6	 	Uncertainty 	about fut	ure 	fees and 	taxation	 of offshore wind	 	
In this section we 	analyze	 whether the support scheme can mitigate uncertainty for the bidders regarding 	
future fees and taxation. We conclude that 	a CfD 	might lessen this uncertainty somewhat.	 	
Norway has 	resource rent taxation on hydropower	 and oil	, and 	there have been discussions	 on the 	
poss	ibility of	 introduc	ing 	similar taxation on 	wind power, in	cluding potentially on offshore wind	 at some 	
time in the future	. The Ministry of Finance has stated that it will look into how 	a future resource rent	 	
tax 	on offshore wind	 can be designed.	11 Furthermore, the 	current government has 	recently proposed a 	
temporary 	tax on 	profits form 	hydropower and land	-based wind power on 	revenue when the power 	
price exceeds 	0,7 NOK/kWh. 	The unexpected introduction of this tax	 illustrate	s for 	bidders the uncer-	
tainty of the 	tax environment in Norway for the 	electricity 	sector.	 It is therefore likely that bidders will 	
take into account the possibility of 	extra	 taxation being introduced for offshore wind in the future. 	Both 	
a resource rent	-based tax and 	other	 tax	es would 	affect the profitability of SN	II. 	It is	 rather	 difficult for 	
bidders to 	estimate the probability of such taxes being introduced	 and what form they could take in the 	
future. 	 	
 	
10 While the	 precise	 fina	l level will be unknown, 	introducing	 a cap on 	total support	 can 	give certainty that the payments will 	
not exceed 	the level of th	is cap	. We discuss this design op	tion below in 	2.3.1	. 	
11 https://www.regjeringen.no/no/aktuelt/kraftfull	-satsing	-pa-havvind/id2912297/

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	28	 	
 
Uncertainty about taxation should be reduced to th	e greatest possible 	extent	 before the auction. At 	the 	
very least, it should be established whether the support resulting from the auction will be taxed or not.	 	
Operator profitability will also be affected by fees, in cas	e of 	a concession fee	. Uncertainty a	bout fees is 	
similar to uncertainty about tax. 	 	
This type of regulatory uncertainty 	negatively affects	 the auction result	 and	 may	 cause	 the bidders to 	
submit less favourable bids. Furthermore, there is a risk that 	the highest bids are placed not by the best 	
producer, but by an inferior producer with a more optimistic estimate of 	the likelihood of future tax 	
changes.	 	
Whilst 	future taxes 	can in theory be applied regardless of 	the support scheme that is chosen	, there are 	
support schemes that make 	future adverse tax changes more or less likely. 	Adverse 	tax changes are 	
most likely in situations where the profits from the offshore wind 	become un	expectedly high. 	If t	he 	
chosen support scheme is an upfront subsidy, it is possible that 	the 	winning bidder ex ante requires a 	
subsidy to be willing to 	carry out	 the project, but that the power prices 	end up higher than expected, 	
giving profits. I	n this scenario i	t can be tempting for 	the 	state to claw back the value of the 	subsidy 	and 	
those profits. 	Bidders 	may take 	the probability of this happening	 into account	 in the bids submitted	, i.e. 	
increase the required subsidy	. This increases the 	expected	 costs for the state	. 	
With 	a two	-sided CfD as the support scheme, the state already captures all	 (hourly CfD)	, or most 	(annu-	
ally) 	of the upside	 from periods of 	high prices	. Th	is re	moves much of the 	potential justification for in-	
troducing 	new taxes on 	offshore wind in the futur	e, giving the bidders less uncertainty	 in the auction.	 	
On the other hand, 	a tax would not have negative effects on 	auctions happening	 after the introduction 	
of the tax.	 There are 	several arguments for introducing a resource rent t	ax on offshore wind power, 	
which it is outside of our scope to investigate. A tax may be introduced even if it is not really necessary 
with respect to profits from SNII. One option is then to keep the SNII investments outside of the tax. 
However, this coul	d be administratively 	demanding. Thus, it is a possible outcome that a tax which co-	
vers SNII is introduced even if a two	-sided CfD is chosen for SNII	, e	specially in light of SNII being small 	
relative to the Norwegian offshore wind ambitions.	 	
2.2.7	 	Administrative	 costs	 and familiarity 	for market participants	 	
An up	-front subsidy is easier to administer for the state than a 	fixed premium, which is easier again than 	
CfDs. 	Simpler and less costly administration 	is a relevant consideration, but 	in our view 	no	t very im-	
portant 	given that even the CfDs are not unreasonably costly to administer.	 	
That the support scheme is one that is used in other countries 	makes it more likely that market partici-	
pants are familiar with it and can properly analyse the consequences. 	CfDs are widely used as a support 	
scheme	. Short reference periods are somewhat simpler for the market participants to analyse, but long 	
reference periods have also been used in several previous auctions. 	Upfront	 investment 	support 	is more	 	
rare	ly used	 but	 is on the other hand a very simple concept to analyse. 	In any event 	the potential bidders 	
are sophisticated actors	 that have the capacity to 	analyse and understand 	the different 	support 	
schemes. We therefore do not consider this to be an important considera	tion.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	29	 	
 	
2.2.8	 	Possibility of ex	 post overcompensation	 	
Ex ante 	the risk	-adjusted 	expected level of support for the different support schemes 	can be assumed 	
to be the same	.12 	
However	, the situation ex post 	will 	differ depending on whether the 	actual 	market 	prices end up higher 	
or lower than expected at the time of bidding. 	Specifically, if the market prices end up being higher than 	
expected, 	an upfront investment support will lead to higher net payments than expe	cted, whereas 	a 	
CfD will lead to 	the same level of payment as expected. 	Conversely, with lower market prices	 than ex-	
pected	 after the project has been built, 	upfront investment support will lead to the 	project owner having 	
received less 	revenue from the project	 than 	desired	 at the time of bidding	. 	
If the state 	has a strong preference for	 avoid	ing	 the 	possibility that	 the conclusion 	ex post 	is that the 	
project has been 	overcompensat	ed	, then 	a CfD 	ensures 	this, whilst	 upfront 	support	 does not	. However, 	
a more rational approach would 	be to 	make decisions based on 	the 	present expected value in which 	
case scenarios with ex post overcompensation should not 	carry higher weight than other scenarios.	 	
With such an approach this ceases to be a	 relevant	 issue.	 	
2.3	 	Summary and r	ecommendation	s 	
Table 	2.6 summarises the 	considerations described in the previous sections. 	 	
 	
12 This s	implifying assumption 	is based on inter alia, no 	option bidding or difference in 	the cost of holding risk	, issues which are 	
discussed elsewhere in this report.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	30	 	
 
Table 	2.6 	Pros and cons of different 	support schemes	 	
Issue	 	Import	ance 	and rele-	
vance 	of issue	 	
Upfront 	invest-	
ment 	subsidy/	 	
payment	 	
CfD	 short 	
reference 
period	 	
CfD with long ref-
erence periods 
and other possible 	
adaptations	 	
Risk of winning bidder 
abandoning	 or delaying	 	
project (a variant of W	in-	
ner’s curse	/option bid-	
ding	) 	
High to medium, can also 	
be reduced through 
choice of auction format	 	
and 	requirements for 	
bidder	 	
- 	++ 	+ 	
Market integra-
tion/avoiding distortions 
in incentives	 	
High, but can be miti-
gated effectively	 	
++ 	-- 	- 	
Efficient distribution of 
market price risk	 	
High if state has a lower 
cost of risk, irrelevant if 	
not	 	
(-) 	(+++) 	(++) 	
Liquidity effect for oper-
ator	 	
Not important	 	+/- 	-/+ 	-/+ 	
Alignment with fiscal 
policy framework	 	
Medium	 	++ 	-- 	-- 	
Taxation uncertainty for 
bidders	 	
Medium	 	- 	+ 	+ 	
Administrative costs	 	Low	 	+ 	-- 	- 	
Avoiding 	ex-post over-	
compensation	 	
Low	 	-- 	+++	 	++ 	
Familiarity for market 
participants (use in other 
countries and ease of 
understanding)	 	
Low	 	+ 	++ 	+ 	
Source:	 	Vista Analyse	. Note: The ideal 	outcome	 would be given 	+++	. A strongly negative outcome would be given 	---. Signs 	
in parenthesis means 	that the	 relevance of the	 issue is conditional on	 a priori 	assumptions 	that are outside the 	
scope of this report	. 	
Three	 issues	 stand out a	s being potentially of high importance. These are 	(1) 	reducing the risk of 	overly 	
ambitious bidding leading to the project being abandoned	 or delayed	, (2) 	ensuring 	good market inte-	
gration 	by 	avoiding distorted incentives, and (3) 	allocating risk in an effici	ent manner. 	 	
The first of these favour	s the 	use of CfD	 as the support mechanism, although this issue can also be 	
addressed by 	the choice of auction format (se chapter 	3). The importance of m	arket integration and 	
avoiding distortions depends on what adaptations developers can make to respond to variable prices to 
maximise value of production. Some possible ad	aptations, such as the possibility of offshore wind 	
providing ancillary services or implementing energy storage, are mostly hypothetical. Other possible 
adaptations could be more important, but their magnitude is unknown. 	This issue favours the use of an 	
upfront subsidy	 over a CfD. However, the issues of market integration with a CfD can be mitigated ef-	
fectively through several measures. The most important is choosing a CfD with a long reference period. 
Th	is ensures that the operator faces short	-term market	 prices, thus removing most distortions. 	A re-	
maining distortion with 	a CfD with a long reference period	 is the situation when 	temporary 	low market 	
prices	 in a situation of payment from the operator to the state	 make	 production unprofitable. This can 	
be dea	lt with 	by suspending 	payments from the operator to the state when these conditions arise.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	31	 	
 
Because a long reference period 	greatly reduces the problem of distorted incentives, we recommend a 	
CfD with long reference period over a CfD with short reference pe	riod. 	 	
Efficient allo	cation of m	arket price risk is a consideration	 that is	 only relevant if the state has a lower 	
cost of carrying market price risk. This is a 	controversial 	question, with no universally agreed answer. 	It 	
is outside of the scope of this report to 	settle this debate. 	Our recommendation is that the Ministry 	
decide what they think is the correct answer regarding this question and weigh the considerations of 
market price risk accordingly.	 	
All in all we concl	ude that both 	a CfD with a long reference period and upfront subsidies are 	support 	
schemes with many positive attributes, that are well suited to 	achieving the goals of the 	Sørlige Norsjø 	
II auction. The 	advantages and disadvantages of each differ, and 	cho	ice between the 	two therefore 	
depend on the state’s precise weighing of the different issues	. 	
If a CfD is chosen, there are several design components that must be 	addressed. These are discussed 	in 	
the next section.	 	
2.3.1	 	Design 	recommendations	 for CfD support s	chemes	 	
Length of the reference period	 	
The 	most significant issue with the use of a CfD is the distortion of market prices and incentives for the 	
producer. 	It is however possible to design a support scheme with a CfD that 	allows for short	-term 	mar-	
ket integration	 by keeping short	-term (sub	-annual) market price signals in place	. The precise length of 	
the reference period 	is a 	trade	-off between 	incentives and what market fluctuations it is desirable to 	
shield the bidders from.	 	
There is a considerable seasonal variation in electricity prices in Norway, with higher prices on average 
during the winter and early spring, when heating needs are higher and hydroelectric production is lower.	 	
An annual	 reference period gives the develop	er incentives to take these variations into account in the 	
design and operation of the wind park	, which is desirable	. A long reference period 	means that	 the op-	
erator takes some market risk and bidders must make independent assessments of the intra	-year var	i-	
ability of power prices. However, the intra	-year variability of prices is a lot smaller and more predictable 	
than the overall uncertainty of the future level of market prices.	 Exposing the project operator to this 	
variability therefor	e does not obliviate 	the advantages of a CfD over upfront subsidies.	 It is	 for example	 	
unlikely that 	different and overly optimistic assumptions of the intra	-year variability of prices 	amongst 	
bidders 	alone is enough to	 give rise to 	a winner’s curse. 	We therefore 	consider this to be a good trade	-	
off.	 	
We recommend 	a reference period	 of one year	 if CfD is chose	n as the support scheme. 	With such a long 	
reference period, the developer is paid the difference between the 	average market price for the year	, 	
and the 	CfD	-pri	ce	. With a long reference period 	the operator might choose to enter into a PPA	. This 	
should not be seen as a problem	 or something that needs regulation.	 It can be economically efficient 	
for the 	parties of the PPA to 	shift	 market price risk.	 	
There are 	several	 additional questions to resolve, such as	 wh	ether the reference price is determined in 	
the previous period or in the given period (with payments ex post), whether the price is a simple average 
or a mean weighted by volume, and whether the periods of	 zero prices with no CfD	-payment should be

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	32	 	
 
excluded from the calculation of the average price. These are issues that we propose be raised in the 
public consultation.	 	
Periods of low market prices	 	
To deal with the problem of the incentives for the producer w	hen the 	market 	power price is near zero	, 	
we recommend th	at the	 volume produce	d when the market price is below a given threshold 	close to 	
zero 	is excluded from the calculation of 	a CfD	-payment	.13 For example, in the Danish Thor tender,	 in 	
years where the of	fshore wind park operator has to pay a CfD premium to the government, 	the 	premium 	
payment is stopped 	in hours where the market price is smaller than the premium to be paid, so that 	
incentives to generate are maintained. 	Moreover, no support should be paid 	in any negative price pe-	
riod.	 This should be done on an hourly basis	. 	
In years when	 the	 average	 marke	t price 	for the annual 	reference period is above the CfD price	, the 	
project owner has to pay the difference between these two prices 	to the state for each 	unit of 	energy 	
produced. If the market price	 net of production costs	 in a given hour is at or below t	he difference be-	
tween the CfD price and the 	annual reference price, the producer 	has an incentive to shut off produc-	
tion	. This is not desirable 	because even if the 	market prices	 are low, any p	ositive price 	net of product	ion 	
costs 	means that	 the 	production has 	value. 	We therefore recommend that 	the CfD 	payment 	be 	reduced 	
or suspended 	for the hours	 when 	the market price is at or below the difference between the CfD price 	
and the reference price.	 	
Incentives to provide ancillary services or power to	 offshore uses	 	
The issue of incentives to provide 	ancil	lary 	services or power to offshore uses	 can be partially dealt with 	
through only offering CfDs for part of the production. 	Already	, the maximum capacity of the wind park 	
is set at 1500 MW, while the Norwegian grid can only accommodate connections of up to 1400 MW. 
The 	maximum 	production capacity is set higher because wind park	s do not	 produce at full capacity	 most 	
of the time. This difference 	incentivises 	the project operator to find uses for 	the excess power produced 	
when the park is producing at full capacity	, such as deliver	ing power to offshore uses	. A possibility 	could 	
be 	to apply 	the 	CfD	 to 	only a part of the total production	. This would give the 	operator incentives to find 	
uses for the remainder of the pro	duction. 	We do not at this time recommend this 	solution but	 propose	 	
that this possibility is raised in the public consultation.	 	
Cap	s on total payment and duration of the support period	 	
CfDs introduce uncertainty for the state regarding 	the 	exact volume of required support	 payment	s to 	
the bidder	. This is typically 	dealt with by setting a cap on the total payment	. In the Thor 	wind auction in 	
Denmark, the state set such a cap at 6.5 billion DKK. 	Figure 	2.2 illustrates	 how a cap o	perates. 	After the 	
payments to the 	developer has	 reached the level of the cap	 (marked by the d	otted vertical line)	, no 	
more payments are made	. Such a cap 	raises several practical issues	 and can have a significant impact 	
on the auction result. 	The best way of controlling the total level of payment from the state is by setting 	
a reservation price in the auction. 	We recognise that a cap may 	nonetheless be 	necessary within the 	
Norwegian 	fisc	al framework	. The level of and design of such a cap must be chosen with great care.	 A 	
first 	consideration is that the cap needs to be set 	high 	enough for 	the p	roject to be expected to be 	
 	
13 As noted above	 this is becoming a requirement 	in EU 	state aid regulations

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	33	 	
 
profitable	 by 	at least one, but preferably multiple bidders	. If the c	ap is set lower than this, the auction 	
could fail	. Since the state does not know in advance the 	assumptions of the 	potential bidders	, the cap 	
should be set so that the project is profitable 	even with 	quite 	pessimistic assumptions on costs and 	
market prices	. 	
 
Figure 	2.2 	Illustration of the effect of 	different caps on payment	 	
 	
Source:	 	Vista Analyse	 	
A cap should ideally 	also 	not be set any lower than the difference between the 	reservation price 	and 	
the expected market price with reasonably pessimistic assumptions	, multiplied by the duration of the 	
support scheme. 	 	
If the cap is set at such a relatively high level, it is unlike	ly that it will be reached for any bids 	for CfD	-	
prices 	below the reservation price	 and will therefore 	have a limited effect on bi	dding behaviour. 	If the 	
cap is set so low that it is likely to 	be binding for the leading bidders	, this 	makes it difficult to compare 	
bids. 	A bid with a lower CfD	-price then simply means that the 	cap is reached somewhat later	, whilst the 	
total 	level of support is the same. 	While the present value 	of the total sum being paid 	generally	 is lower 	
if the 	cap is r	eached	 later	, one can imagine situations where this is not the case. For example if one 	
bidder expects a period of low prices early in the support period, followed by a long period of higher 
prices	 with the cap being reached 	only after this period, the exp	ected present value of these payments 	
can be higher than 	the expected present value for a bidder that expects a 	less variable market price.	 In 	
such a scenario 	it is the 	different bidders 	expected timing of a fixed sum of payments that d	etermines 	
the auction result, rather than who 	has the best project.	 This is a highly undesirable result.	 	
In some auctions countries have also set a cap on the level of payment from the operator to the state. 
In the Thor wind auction in Denmark, the state s	et this cap at 2.8 billion DKK. We advise against setting 	
such a cap. Such a cap can give problems in resolving the auction if there are several companies that 
place a higher value on the project than the cap. 	As noted in previous chapters, this	 was the ca	se for 	
the Thor auction in Denmark, where there were five bidders offering the minimum price of 0.01 øre per 
KWh (which is equivalent to paying the cap over the first 2	-3 years), and the auction had to be resolved 	
by the drawing of lots. This means that it	 is up to chance whether the bidder with the highest valuation 	
wins, and the state “leaves money on the table".

Norwegian Of	fshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	34	 	
 
If there is 	a cap 	on the payment from the state, but no pre	-set cap from payments in the other direction	, 	
a question arises relating to 	whether 	the cap should apply to net or gross payments.	 We	 recommend 	
that 	any	 cap is defined in terms of gross yearly payments, not net payment. This means that if the state 	
over a number of years with prices below the CfD	-price pays the operator a sum equal to the	 cap, and 	
this period is followed by a period with high 	electricity	 prices where the operator pays the state, this 	
does not mean that the state again becomes obliged to pay the operator when prices are below the 
CfD	-price.	 If not the state 	could end up pay	ing the sum of the cap multiple times, 	after the operator has 	
received the CfD	-price for a long period. This can 	then be considered to be an unnecessary overcom-	
pensation ex post. 	The disadv	antage 	with this is that 	even with the same average price	, the cap is 	
reached more rapidly if there is 	more volatility in the 	average annual prices.	 	
Another design element for CfDs is 	the 	duration of the support period. 	A common length for the support 	
period is between 15 	(UK)	and 20 years	 (Denmark, Germany and t	he Netherlands). The shorter the sup-	
port period, the more 	weight 	the bidders 	must	 place on	 the 	market prices after the expiry of 	the period	 	
(Welisch & Poudineh, 2019)	. This	 somewhat	 reduces 	the advantages of a CfD in terms of not risking 	
option bidding and winner’s curse, without any commensurate benefits	. All else equal a longer support 	
period should not affect the present value of 	net expected support	. A 	longer period typically entails 	
lowe	r bids. 	The support period should not be longer than 	the depreciation period of the investment 	
according to 	EU state aid guidelines, but this should not	 hinder a 20	-year period as used in Denmark	. 	
When the support period should start 	is also an issue that 	requires consideration. Should it start when 	
production starts, or when the last turbine 	is ready to deliver power to the grid? 	The latter approach 	is 	
the one chosen in Denmark. 	If the support period starts from the very first production, this 	means that 	
the project owner might be weary of connecting part of the 	planned capacity to the grid as soon as is 	
ready, as this will “waste” part of the support period 	on 	a period with only partial production. Turbine	s 	
that are ready to produce, but are not connected, is 	a loss to society. On the other hand, the Danish 	
approach can lead to a significant period without CfD	 at the beginning of the life of the wind park. The 	
ideal solution is that 	the support period start	s for 	the capacity that is connected to the grid, and 	capacity 	
that is connected later has a support period that starts and ends later. 	It needs to be checked that this 	
is not unduly complicated to administer.	 	
If the Ministry 	considers there to be a need t	o limit the	 size of	 total 	possible 	payment	 and therefore set 	
a low cap	, it is better 	if the support period is also shortened so that the lower cap is unlikely to be 	
reached.	 However	, there is in any event a limit to how low the cap can be set without riski	ng a failed 	
auction.	 	
A final	 issue is whether to index the prices in the CfD	, and 	if so	, according to what index. 	The	 practice in 	
Denmark	 is not 	to index	. This would reduce the real value of the CfD over time	, which is particular	ly 	
problematic if 	inflation is high	. In the UK 	the CfD price is 	adjusted annually according to	 an inflation 	
factor	 based on the Consumer Price Index	 (Low Carbon Contracts Company, 2017)	. We propose 	that 	
the CfD 	be indexed to inflation	, and that the choice of index be 	a question in the public consultation	. 	
Penalties in 	the	 case of 	the project being cancelled or delayed by the winning bidder	 is discussed in 	
section 	4.4	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	35	 	
 	
3	 	Auction 	formats	 	
In this chapter 	we discuss which format should be chosen for	 the 	auction, in order achieve the 	overall 	
objective	s of the 	government. 	The	 objectives are (1) 	to efficiently allocate the licenses	 and (2) to	 achieve	 	
the highest	 sum/low	est	 support.	 As the 	two parts of this objective 	are not in all cases fully compatible, 	
we 	have chosen to	 treat 	them as two separate 	entities here	. 	
The objectives correspond with 	established properties in auction theory: 	Efficienc	y in an auction	 context	 	
means that the	 license is won by the 	bidder with the highest valuation	. Auctions 	that realize the 	highest 	
possible sum/lowest support	 for the seller	 are known as 	revenue	-maximizing 	or optimal auctions	. 	
The outcome of this auction can be either a 	subsidy from the government to a developer, or a payment 	
from the developer to the government. 	Furthermore, what the bids look like will differ depending on 	
whether an upfront or a CfD scheme is chosen. To	 avoid c	onfusion	 about the terms used	, we will briefly 	
explain the three most relevant auction types here in the introduction	: 	
• 	First	-price sealed	-bid auction	: Each bidder submits 	a “best and final”	 bid.	 In the case of an 	upfront 	
scheme, the bid will be 	the size of the subsidy (or payment). With a 	CfD 	the bid will be for the 	
strike price	. The highest bidder 	(lowest s	ubsidy or lowest strike price)	 wins and	 pays its own bid.	 A 	
first	-price sealed	-bid auction is sometimes referred to a	s “Dutch” auction.	 	
• 	Ascending auction	: The term “	ascending	” refers to how bids in a standard auction 	would 	start 	low 	
and increase until only one bidder remains. However, in this case	 it gets a little more complicated	. 	
With an 	upfront scheme	, bid	s start at a high subsidy	 and decrease, possibly turning into payment	s 	
and	 then	 increas	ing	 as in a standard auction	. With a 	CfD scheme	, bid	s start at a high	 strike price	 	
and decrease	 all the way	, until only one bidder remains. 	An ascending auction is someti	mes re-	
ferred to as an “English” or open	-bid auction.	 	
• 	Anglo	-Dutch auction	: This is a hybrid auction. There is first an ascending stage. The 	upfront 	subsidy 	
or 	CfD 	strike price starts high and gradually decreases until only	 two	 bidders remain. Then these 	
two bidders 	submit sealed final bids as in a first	-price sealed	-bid auction	, and the winner pays its 	
own bid.	 	
We first give an overview of standard auction formats and 	explain why we consider the formats men-	
tioned above to be the most relevant. Then we discus	s how 	these formats 	address	 important concerns 	
in auction design.	 Experiences from other countries are 	presented throughout the subchapters. 	We also 	
discuss the topic of a reservation price	 (section 	3.3	). Finally, we recommend 	an auction type to use 	for	 	
phase 1 of	 Sørlige Nordsjø II	 (“SNII”)	. 	
3.1	 	Which 	auction 	formats	 are relevant?	 	
In this 	section	 we present standard 	auction formats and discuss 	which formats are most relevant to 	
investigate more in detail for SNII	. 	
As the governments auction a single license for SNII phase 1, w	e focus on single	-object 	auctions and 	
largely ignore the literature on multi	-object auctions. 	Note, however, that the bidders still may think 	
strategically and consider future license allocations which may affect bids in the current auction. We 
discuss such issues 	below (	3.2.5	).

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	36	 	
 
In this 	section	 we will to a large degr	ee use 	standard terms from auction theory. 	As explained in the 	
introduction these	 terms can sometimes be confusing in 	our	 context, because the auction outcome can 	
be either a subsidy or a paymen	t, and because of the different support schemes. In 	Table 	3.1 we 	explain 	
what is meant by standard auction theory 	terms	 in the context of this auction.	 	
Table 	3.1 	An explanation of auction theory terms in the context of Sørlige Nordsjø II	 	 	
Upfront investment subsidy/payment 	 	Two	-sided CfD 	 	
Concepts	 	 	 	
What is a bid, and 
who wins?	 	
A bid is for 	the 	total amount of upfront 	
subsidy (or payment). 	 	
A bid is for 	the 	price at which electricity is 	
sold	 (the “strike price”)	. 	
Low 	bid	 	High subsidy	 	High strike price	 	
High 	bid	 	Low	 subsidy or high payment	 	Low strike price	 	
Winning bid	 	Winning bid has 	lowest required sub-	
sidy	 or the highest payment.	 	
Winning bid sells electricity 	at 	the lowest 	
price.	 	
Auction formats	 	 	 	
First	-price sealed	-	
bid	 	
Bidder submits “best and final” bid, 
which will be a subsidy or a payment.	 	
Bidder submits “best and final” 	bid, which 	
will be a strike price.	 	
Ascending type 
(English)	 	
Initial bid is a very high subsidy. Subse-
quent bids decrease 	the 	subsidy.	 Even-	
tually bids may turn into payments, and 
increase.	 	
Initial bid is a at a high 	strike 	price. Bids re-	
duce the 	strike 	price. 	 	
Anglo	-Dutch	 	In the ascending stage, the upfront sub-
sidy starts high and gradually decreases 
until two bidders remain. T	hen these 	
two bidders submit sealed final bids as 
in a first	-price sealed	-bid auction.	 	
In the ascending stage, the strike 	price starts 	
high and gradually decreases until two bid-
ders remain. T	hen these two bidders submit 	
sealed final bids 	for the strike price 	as in a 	
first	-price sealed	-bid auction.	 	
Source:	 	Vista Analyse	 	
3.1.1	 	Overview o	f auction	 formats	 	
The two main auction formats	 are 	open	-bid and sealed	-bid auctions	. In open	-bid auctions, each bid is 	
announced to all participants 	to the auction. In sealed	-bid auctions, 	by	 contrast, each bid is known only 	
to the 	seller	. Open	-bid auctions 	(especially	 the ascending auction type	) are game	-theoretically 	often 	
classified as dynamic games in which the participants make sequential moves. 	Sealed	-bid auctions are 	
considered 	static games, in which the participants only move once.	 	
Both auction 	formats	 can 	be d	ivided into two classes	 depending on how the 	final	 price 	is set	: first	 price 	
and second	 price	. In first	-price auctions, the winner pays the winning bid. In second	-price auctions, the 	
winner pays the second	-highest bid. 	In 	Table 	3.2 below, w	e present the resulti	ng four 	types of auction	s 	
and their rules.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	37	 	
 
Table 	3.2 	Common	 auction 	formats	 and 	their 	general 	rules	 	 	
First price	 	Second price	 	
Sealed bid	 	First price	, sealed bid	 	
The seller gets sealed ‘best and final’ bids 
from each bidder. The highest bidder wins 
the auction and pays the 	winning bid.	 	
Second price, sealed bid	 	
The seller gets sealed ‘best and final’ bids 
from each bidder. The highest bidder wins 
the auction and pay	s the second	-highest bid.	 	
Open bid	 	Descending	 	
The seller sets a very high initial price. Then 
the seller lowers the price until a bidder pub-
licly accepts the price. The bidder wins the 
object and pays the bid. Descending auctions 
are sometimes 	called “Dutch” auctions.	 	
Ascending	 	
The seller sets some initial price	. The bids 	
can be open	-outcry (often called “English” 	
auctions), continuously increasing with time 
(“Japanese”) or some other technique in 
which the bids are ascending and common 
informat	ion to all bidders. The bidder with 	
the ultimate highest bid wins the object and 
pays the bid.	 	
Source:	 	Vista Analyse	 	
It is possible to combine these four “pure” auction types into hybrid auctions, meant to capture “the 
best of both worlds”. One 	highly 	relevant 	format	 is the 	Anglo	-Dutch	14 auction 	(Klemperer, 1998)	. The 	
Anglo	-Dutch	 auction has two rounds. The first round is ascending until only two bidders remain. Then 	
these two participate in a first	-price sealed	-bid auction, with a reservation price equal to the bid that 	
initiated round two.	 	
When 	bids 	are	 subsid	ies	 (neg	ative payments) 	an ascending 	auc	tion 	involves 	a decrease	 from 	a high 	
subsidy	 leve	l or high electricity price to a lower level.	 	
First	-price and second price	-auctions 	differ fundamentally in what the optimal bidding behaviour is. This 	
will be explained next.	 	
The 	optimal bidding behavior in	 first	-price auctions	 is to bid somewhat below one’s own valuation. This 	
behavior is often referred to as “bid	-shading”. The reason for 	bid shading	 is that in a	 first	-price auction	, 	
bidders get zero profits if they bid their exact valuation	, because they pay their own bid. Therefo	re, they 	
look to gain profits by reducing their bids somewhat compared to their valuation. This 	implies that a 	
bidder could lose the auction even if it has	 the highest valuation.	 First	-price sealed bid auctions are 	
strategically equivalent to descending au	ctions. There is no fundamental difference between choosing 	
the price you would like to bid in a first	-price sealed	-bid auction, and deciding at which price you will 	
“cry out” in a descending auction (Dutch auction). The optimal bid strategies are the same	 in these two 	
auctions. Because of this, we often refer to both jointly as first	-price auctions.	 	
The optimal bidding behavior in 	second	-price auctions	 is to bid one’s own valuation. 	Bidders	 gain noth-	
ing from bidding higher 	or lower. 	Bidding their own valua	tion is a 	dominant strategy	15. This is straight-	
forward in second	-price sealed	-bid auctions, bud also applies to	 ascending auctions	: Bidders 	should 	
drop out when the standing bid reaches their valuation. 	If they drop	 out sooner	, they	 risk 	losing the 	
auction when they could have won. 	The price at which the second	-last bidder drops out of 	an ascending 	
auction is the 	ultimate	 standing bid	. Thus, t	he winner in practice 	pays 	a price equal to 	or 	marginally 	
above	 the 	valuation of the 	second	-highest 	bidder.	 	
Table 	3.3 gives an overview of the relations between the different auction types.	 	
 	
14 “Anglo” refers to 	the English	-type first round. “Dutch” refers to the first	-price	-type second round.	 	
15 A dominant strategy means that 	the strategy is optimal regardless of what others do.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	38	 	
 
Table 	3.3 	Overview 	of relations between 	auction types	 	
Auction	 	Alternativ	e 	
name	 	
Optimal bidding be-
havior	 	
Outcome similar to	 	Common uses	 	
Ascending 	 	English, Jap-
anese	 	
Exit when standing bid 
reaches your valuation	 	
Second	-price 	sealed	-bid	 	Norwegian real es-
tate market, art, 
collectibles	 	
Descending 	 	Dutch	 	Bid	-shading	 	First	-price 	sealed	-bid	 	Dutch flower mar-
ket, tobacco	 	
First	-price	 	
sealed	-bid	 	
Blind	 	Bid	-shading	 	Descending auctions	 	Private and public 
procurement	 	
Second	-price	 	
sealed	-bid	 	
Vickrey	 	Bid your valuation ex-
actly	 	
Ascending auctions	 	Not much used	 	
Anglo	-Dutch 	 	 	Bid	-shading, but less 	
severe	 	
Begins as ascending, ends as 
first	-price	 	
E.g, 	demand 	re-	
sponse (	flexibility	) 	
New York City re-
gion.	 	
Source:	 	Vista Analyse	 	
Under certain conditions	 the four 	standard 	auction types are revenue equivalent	, i.e.,	 the seller’s reve-	
nue is independent of auction type	.16 That means that, if the 	assumptions of the theorem hold, it does 	
not matter which auction type 	the seller uses because the revenue will be the same for all standard 	
auction 	types. 	This result is	 called	 the “	revenue equivalence theorem	” and is	 important in auction 	theory. 	 	
However, 	there will	 often	 be deviations from 	the necessary 	assumptions	, which allow for a ranking of 	
auction types with re	spect to expected revenue	. 	
3.1.2	 	Common formats in 	renewable energy auctions	 	
The most common auction type used in RE auctions is the sealed	-bid auction, whereby the bidder re-	
ceived the priced offered (“first price”). According to the AURES auction database	17, out of 820 auction 	
rounds captured for 20 EU 	countries for various RE technologies, including offshore wind, between 2011 	
and 2021, only 28 auction rounds were implemented as open	-bid auctions	 (ascending or descending 	
auction)	. The seven countries that 	had	 implemented offshore wind auctions until 202	1 have imple-	
mented them as first	-price, sealed bid auctions. 	 	
While the RE industry is more familiar with sealed	-bid auctions than with open	-bid auctions, open	-bid 	
auctions can also work well. Countries like Portugal, Greece, and Brazil have implemented op	en	-bid 	
auctions. In Brazil, multi	-object RE auctions 	have been	 hybrid. The first stage operates as a	n ascending	 	
price auction. The auction is initiated with a high price (high subsidy) that is expected to create excess 
supply and bidders state the quantity	 (MWh) they would supply at this price. While there is still excess 	
supply, the auctioneer decreases the price (subsidy) until demand is met, in addition to a certain margin. 
This margin would be used in the next stage to keep competition among the bidders	 who pass the first 	
stage. The second stage is a first	-price, sealed	-bid auction, where bidders cannot bid higher than the 	
 	
16 See, e.g.,	 Klemperer	 (1999)	 for an overview of	 the literature and the necessary assumptions for the revenue equivalence 	
theorem to hold.	 	
17 http://aures2project.eu/auction	-database/

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	39	 	
 
price (subsidy) disclosed in the first stage. The second stage is held to meet the actual demand and 
assure that there is no collusio	n between a small numbers of bidders	18. 	
The Netherlands	 RE multi	-object auctions, known as the SDE+,	 defined an 	annual	 subsidy	 budget	 and 	
operated as a sealed	-bid auction with ascending ceiling prices	. The auction consisted of two rounds each 	
year, each r	ound consisting of three phases with ascending ceiling prices. 	Each	 phase is open for one 	
week and 	the 	budget is auctioned on a first	-come	, first	-served basis	. On the day the budget is exhausted, 	
all applicants from	 that day are ranked based on their bids 	and the lowest are accepted first. If there are 	
multiple 	bids	 at the same price, a lottery will take place to decide which projects are awarded	19. 	
An example of Anglo	-Dutch is 	the 	“Demand Response” auction for Consolidated Edison, which is the 	
electric uti	lity serving the New York City region. This auction was done as a part of a “Demand Response” 	
project. Demand Response is needed by electric utilities to reduce the load on the electric grid duration 
peak demand periods, such as very hot summer days. Th	e goal in Demand Response is to pay large 	
electricity users to reduce their electric usage “on	-demand”, i.e., within hours of the request by Con Ed. 	
in exchange for their agreement to reduce their electric demand, Con Ed pays them. The auction was 
conduc	ted among pre	-qualified bidders who demonstrated HOW they would reduce their usage when 	
called on. The auction was done on a price	-per	-kWh basis, with the maximum load reduction procured 	
for the lowest cost per 	kWh. Many awardees were given Demand Response contracts as a result, far 	
exceeding the expectations of Con Ed.	 	
3.1.3	 	The	 most relevant auction formats	 	
We consider these auction formats to 	be most relevant for SNII:	 	
• 	Ascending	 	
• 	First	-price sealed	-bid	 	
• 	Anglo	-Dutch	 	
The reason we do not further	 investigate the 	descending	 auction format, is that it is similar to a first	-	
price sealed	-bid auction in all relevant aspects. Thus, the subsequent discussi	on of the latter also applies 	
to descending auctions.	 	
We also 	do not further investigate the second	-price sealed	-bid auction	. This format is equal to an as-	
cending auction in one important aspect, they are both second	-price. However, they also differ in som	e 	
aspects	. An ascending auction has better properties with respect to winner’s curse, although 	a second	-	
price sealed	-bid is still better than a first	-price sealed	-bid. On the other hand, a second	-price sealed	-bid 	
auction has better properties than an ascending auction with respect to collusion. However, as we will 
discuss more in detail 	below, we do	 not believe collusion to be a major concern for SNII phase II 	– at 	
least not in a way that can be easily remedied by the auction format.	 In sum	mary	 we find the ascending 	
auction to be the most promising of the second	-price auctions.	 	
3.2	 	Assessment of 	the most	 relevant 	auction formats	 	
In 	this	 chapter, we assess the 	most 	relevant auction 	formats	 with respects to the following concerns:	 	
 	
18 http://aures2project.eu/wp	-content/uploads/2021/07/design_elements_october2015.pdf	 	
19 http://aures2pr	oject.eu/wp	-content/uploads/2019/12/AURES_II_case_study_Netherlands.pdf

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	40	 	
 
• 	Winner	’s curse	 	
• 	Efficiency	 	
• 	Risk	-averse bidders	 	
• 	Familiarity	 	
• 	Collusion	 	
• 	Entry	 	
3.2.1	 	Winner’s curse	 	
The winner’s curse is a 	phenomenon where the bidder with the most optimistic expected valuation of 	
the project wins the auction. Winning the auction provides new information that was not known before 
winning, namely that all the other bidders have lower expected valuations than y	ou. As the optimistic 	
winner you may be at risk of building and/or operating at a loss	. 	
With regards to this auction, there are two different ways in which winner’s curse could be a challenge:	 	
• 	The classic winner’s curse problem from auction theory, in whi	ch the possibility of winner’s curse 	
leads all bidders to bid lower, resulting in lower auction revenue.	 	
• 	The problem of the winner not carrying through the project if the actual value is lower than ex-
pected.	 	
The 	classic 	winner’s curse 	from auction theory 	could really be described as a “seller’s curse”. If all bid-	
ders correctly anticipate that they may be too optimistic, they will shade	 (lower)	 their bids to avoid this 	
risk. This leads to lower expected auction revenue. However, in situations where all bidde	rs 	don’t	 be-	
have optimally and fail to consider this effect or fail to correctly calculate how much to shade their bids, 
the outcome may be an actual winner’s curse. This outcome is also 	bad for	 the 	government	, which	 is at 	
risk that the winners leave their 	obligation to build and operate the project because of lower	-than	-	
anticipated project profits.	 	
In the remainder of this section	, we 	describe in some more detail what 	the winner’s curse is	 and 	whether 	
there is reason to expect	 winner’s curse 	challenges 	in S	ørlige Nordsjø II. 	Then we describe measures to 	
reduce the challenges 	including 	the use of a	n ascending auction 	type.	 	
Information structure	 	
Whether or not the winner’s curse is an issue depends on the information structure. There is a higher 
risk of winner’s curse in auctions with higher degrees of 	common value	. Common value means that the 	
true value	 of the acreage	 (ex-post	) is the same for ev	ery bidder, for instance because their costs are 	
approximately the same. Before the auction (	ex-ante)	, bidders form beliefs about the true value, and 	
update their beliefs after learning new information. Learning that another bidder has a high expected 
valu	ation makes it more likely that the true value indeed is high.	20 Thus, learning that all other bidders 	
have lower expected valuations than you	, upon winning the auction,	 will inform you that the true value 	
likely is lower that you expected.	 	
In (independent	) private value	 auctions, however, there is no such problem. In such auctions, the value 	
of the object solely depends on bidder’s own private valuations irrespective of what others think. 	For 	
 	
20 This means that the values are “affiliated”, which breaks the revenue equivalence theorem 	(Milgrom & Weber, 1982)	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	41	 	
 
instance, your 	costs of building and operating the plant 	may be 	lower than that of others	, and 	your 	
winning bid signals your low cost. Learning other’s expected valuations will in this case not affect your 
own expected valuation.	 	
There are both common and private factors in Sørlige Nordsjø II. Such factors are relevant	 on both the 	
cost	-side and the income	-side 	for	 firms because the bids will depend on the 	profit	 bidders expect to 	
accrue from the project. We present an overview over potential common and private factors in	 Table 	
3.4, where some factors have both common and private elements.	 	
Table 	3.4 	Common and private factors	 	
Factor type	 	Income side	 	Cost side	 	
Common	 	• 	The price of electricity. 	 	
• 	The sales volume (e.g., electricity mar-
ket conditions and wind conditions)	 	
• 	Correlation between price and volume	 	
• 	Common learning effects	 	
• 	Grid connections costs	 	
• 	Seabed conditions	 	
• 	Other regulatory uncertainty	 	
• 	Factor costs	 	
Private	 	• 	Portfolio diversification 	 	
• 	Corporate 	social 	responsibility	 	
• 	Private learning effects	 	 	
• 	Firm	-specific technology and cost	-struc-	
ture	 	
• 	Bargaining power in factor market	 	
Source:	 	Vista Analyse	 	
Other measures to mitigate winner’s curse	 	
Before we turn to the properties of different auction formats with respect to winner’s curse, we will 
briefly discuss 	other measures to mitigate the risk of winner’s curse.	 	
The Ministry can reduce the risk of winner’s curse by reducing or removing uncertai	nty about common 	
factors	. It should	 reveal all relevant information that may reduce common value uncertainty 	(Milgrom 	
& Weber, 1982)	. 	
On the cost side, the Ministry can reduce the risk of winner’s curse in the 	following ways	 (to be further 	
discussed in chapter 	4 below	: 	
• 	Remove regulatory uncertainty	: The Ministry should aim to clarify all 	profitability	-relevant regula-	
tions before the auction begins. 	Important	 example	s are	 the 	possibility for future taxation	, or 	the 	
outcome of the application for concession.	 	
• 	Reduce uncertainty about grid costs	: The Mi	nistry should 	reduce uncertainty about the radial con-	
nection point and grid connection costs.	 	
In addition, it is helpful that the Ministry funds and shares information about geotechnical assessments. 	 	
As commented in chapter 	2 the chosen type of support scheme will directly affect uncertainty on the 	
income side. The important difference is between a CfD on 	the 	one side, and fixed premium and upfront 	
subsidy on the other. Using a	n hourly based	 CfD will remove price uncertainty for	 bidders because they 	
will always sell at a fixed price, no matter the actual price. An upfront subsidy or a fixed premium scheme 
will 	by	 contrast not reduce uncertainty about price and these schemes will therefore be more exposed 	
to the winner’s curse.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	42	 	
 
However, with regards to volume uncertainty it is the other way around. If wind resources are worse 
than expected, the support from operations phase schemes 	such as a CfD	 will be lower. If wind re-	
sources are better, the support from these schemes will be hi	gher. Thus, these schemes amplify volume 	
uncertainty. With regards to volume uncertainty, an upfront subsidy has better properties.	 	
Properties of different auction formats with respect to winner’s curse	 	
Ascending auction types have good properties when the	re is a risk of winner’s curse 	(Milgrom & Weber, 	
1982)	. The sequential bidding structure of such auctions help bidders learn about other bidders’ valua-	
tion by observing at which price others drop out of the auction. Learning th	at many bidders drop out 	
early will make remaining bidders vary of being aggressive and cause them to (correctly) lower their 
valuation. Or, if many bidders remain when the standing bid gets closer to their own expected valuation, 
they will (also correctly	) increase their expected valuation and thus make 	improved	 bids resulting in 	
higher expected auction revenue. 	 	
This mechanism may be familiar to many from real estate auctions. One might be unsure about the 
actual value of a house. Seeing that there are ma	ny other bidders might lead one to bid higher than one 	
would if it was not an ascending auction.	 	
We expect common value factors to be more important than private value factors for Sørlige Nordsjø 
II. 	This pulls towards an ascending auction.	 	
Text frame 	3.1 	Summary of winner’s curse	 	
• 	Winner’s curse is an important concern and may 	lead to lower auction revenue or 	in the worst 	
case non	-completion of the project.	 	
• 	The Ministry should aim to reduce regulatory 	and cost uncertainty as much as possible to reduce 	
risk of winner’s curse.	 	
• 	A CfD scheme will remove price uncertainty and thus reduce risk of winner’s curse compared to 
the other to support schemes.	 	
• 	A CfD 	scheme will 	amplify volume uncertainty, 	but 	we con	sider that volume uncertainty less im-	
portant 	than 	price uncertainty.	 	
• 	Ascending auction types help bidders gain information during the auction and will reduce risk of 
winner’s curse.	 	
3.2.2	 	Efficiency	 	
Efficiency means that	 the bidder with the highest valuation wins	 the auction. Efficiency and revenue are 	
positively correlated, because a higher general willingness to pay for the object leads to higher auction 
revenue. Sometimes, however, optimality comes at the cost of effic	iency.	 	
Ascending auctions	, which are (similar to) second	-price auctions	 are robust to inefficiency because a 	
stronger bidder can always outbid weaker competitors. First	-price auctions, however, have some risk of 	
inefficiency. The reason is the strategic un	certainty in first	-price auctions, where bidders must shade 	
(lower) 	their bids to gain profit. Miscalculation in this process may lead to a situation where a strong 	
bidder shades their bid too much, resulting in a weaker bidder winning the auction.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	43	 	
 
Text f	rame 	3.2 	Summary of 	efficiency	 	
• 	Ascending auctions ensure an efficient allocation.	 	
• 	First	-price auctions may cause inefficient outcomes, because of strategic uncertainty.	 	
3.2.3	 	Risk	-averse bidders	 	
The bidder’s attitudes to risk can affect the auction outcome. Bidders can either be risk averse, risk 
neutral or risk seeking. Risk	-neutral bidders are indifferent between an uncertain but expected profit of 	
������ and getting 	������ for certain. Risk	-averse bidd	ers will prefer to get 	������ for certain over getting 	������ in expecta-	
tion. They need a risk premium 	������ to be indifferent between having an expected profit 	������+	������ and getting 	
������ with no uncertainty. Risk seekers, in contrast, prefer the possibility of 	obtaining	 a higher	-than	-ex-	
pected profit over getting something for certain; they have negative risk premia. We do not consider 
risk	-seeking attitudes further.	 	
First	-price auctions will provide higher expected revenue for the seller than second	-price auctions, if 	
bidd	ers are risk	-averse 	(Klemperer, 1999; Vasserman & Watt, 2021)	. Only first	-price auctions (descend-	
ing or sealed	-bid) are affected by the risk attitudes of bidders. In these auctions, the optimal strategy 	
for 	bidders is to bid below their own valuation	, but above the expected valuation of the second	-highest 	
bidder. A risk	-averse bidder is averse to losing the auction unnecessarily and would prefer to bid a little 	
higher to increase the probability of winning at	 a cost to the value of winning. This is not the case in 	
second	-price auctions, where risk attitudes do not matter for optimal bidding strategy. In such auctions, 	
the dominant strategy is to bid your valuation and you risk nothing in doing so. 	 	
We put some	 weight on this issue, given that this is the first offshore wind auction in Norway and 	that 	
we expect several of the bidders to not want to lose this auction.	 	
Text frame 	3.3 	Summary of attitudes to risk	 	
• 	First	-price auctions will provide higher expected auction revenue if buyers are risk averse.	 	
3.2.4	 	Familiarity	 	
As discussed in 	chapter 	3.1.3	 first	-price 	sealed	-bid	 auctions have been common for renewable energy. 	
Furthermore, such auctions are common in public procurement more generally, but then often 	as a part 	
of multi	-criteria auctions. Its simplicity 	and widespread use is an argument in favor of the first	-price 	
sealed	-bid auction.	 	
However, 	although less used 	for renewable energy, we do not believe that an ascending auction would 	
be particularly unfamiliar. It is a well	-known 	format in auction theory and is also in widespread use, e.g.	, 	
in housing markets.	 	
The Angl	o-Dutch auction is likely to be less familiar. However, it has been used for renewable energy 	
purposes 	before in Brazil	 and 	the state of 	New York	, and is well described 	in articles by 	recognized auc-	
tion expert 	Paul Klemperer. Furthermore, our software part	ner Procurex has extensive experience with 	
conducti	ng	 Anglo	-Dutch auctions, in their experience often referred to as “Final Blind” auctions.	 Finally, 	
it is really just a hybrid of two well	-known auction formats.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	44	 	
 
Text frame 	3.4 	Summary of 	familiarity	 	
• 	Althou	gh a	 first	-price	 sealed	-bid auction	 is the most familiar format in the renewable energy con-	
text, the other auction types 	should be relatively well	-known, are well	-defined in auction theory 	
and als	o see widespread use.	 	
3.2.5	 	Collusion	 	
We have assessed whether collusion is an issue which 	should affect the choice of auction type for SNII 	
phase 1. 	Collusion	 in an auction context 	means that bidders 	co	ordinate	 to 	obtain an	 auction 	outcome 	
more 	favourable	 to 	them	. Collusion can either happen 	explicitly	, by 	the 	forming 	of 	cartel	 agreements	 	
and commonly prohibited by law	, tacitly	, by	 signalling	 intentions only through behavior, 	or 	a combina-	
tion of the two.	 Explicit communication is helpful for 	coordination, but not necessary	 for collusion 	to 	
happen 	(Fonesca & Normann, 2012)	. There 	ha	ve	 been several examples of 	government 	auctions with 	
lower	-than	-expected revenues	, likely 	because of	 tacit	 collusion 	(Klemperer, 2002a; Klemperer, 2002b)	. 	
We do not view collusion as a major issue for SNII phase 1, 	mainly 	because it is a single	-uni	t auction. 	For 	
firms to collude, 	there 	must	 be some way of sharing the gains f	rom collusion. This is much harder in 	
single	-unit auctions than in multi	-unit auctions. B	idders cannot easily “divide the pie” between them, 	
because there is only a single lot to be allocated.	 We 	consider the possibility of side	-payments between 	
the winner	 and other colluders as very small.	 	
The upcoming SNII phase 2 auction could 	make the two auctions combined resemble a multi	-unit auc-	
tion, but we	 are still not 	particularly	 worried about collusion.	 Due to the sequentiality, 	it w	ould be more 	
difficult to sustain collusion. The winner of the first 	phase would have 	little 	incentive to honour an agree-	
ment and bid low in phase 2, unless there is a credible threat	 in terms of future auctions in wind energy	. 	
Implications for auction design 	if within	-auction 	collusion w	as a concern	 	
Ascending auctions are vulnerable to collusion	. Because the auction format is open, deviations can be 	
discovered and retaliated against during the auction and not only afterwards. 	Fu	rthermore, bids can be 	
used for signalling.	 	
Collusion is more difficult to 	sustain 	in a first	-pric	e auction. 	In 	an ascending or other 	second	-price auc-	
tion there is no incentive among the colluders to deviate from the agreement since the bidder with the 
highest valuation would win anyway. In first	-price auctions, however, collusion relies on everyone bid-	
ding very low while the designated winner bids marginally more. But then everyone would have an 
incentive to deviate by bidding marginally more than the a	greed sum, winning the auction themselves. 	
Knowing this, everyone increases their bids even further, and so on. This mechanism in first	-price auc-	
tions makes collusion agreements unstable and non	-credible.	 	
The possibility for collusion across auctions	 	
As we	 have touched upon above with regards to phase 2, 	firms can collude 	across	 auction	s. Across	-	
auction collusion can either 	over time	 or 	across region	s at any given time	, or a combination of the two	. 	
Collusion across regions essentially turns individual single	-object auction	s into 	a larger multi	-object auc-	
tion	. Coll	usion 	over time is	 possible in repeated, dynamic interactions where competitors realize that a 	
self	-enforcing agreement can be sustainable	 since (i) retaliation is more costly	 than	 the short	-term gain 	
of deviation and (ii) retaliation is credible 	(Ivaldi, Jullien, Rey, Seabright, & Tirole, 2003)	. As offshore	-

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	45	 	
 
wind auctions happen more frequently, bidders may devel	op bid	-signalling	 strategies, effective punish-	
ments, and entry	-deterrence strategies that may make future auction rounds more vulnerable to collu-	
sion 	(Klemperer, 2002b)	. 	
It is difficult to design auctions that 	protect against across	-auction collusion.	 In general, though,	 using 	
first	-price auctions will make collusion more difficult	, as explained in the previous paragraph	. If an as-	
cending auction is used,	 however,	 it should be designed 	to make signal	ling 	through bids more difficult	, 	
by for example only allowing fixed	-increment bids	 or using an open	-exit design	21. This would prevent 	
so	-called jump bidding, where 	a bidder bids higher than the minimal increment	. Such bids are	 irrational, 	
because it	 in essence	 means bidding against yourself 	at the risk of	 paying a higher than necessary price	. 	
It can be used, however, to signal aggressive intentions	 which make tacit collusion possible	 (Avery, 	
1998)	. 	
We do not put weight on the 	possibility for across	-auction collusion.	 Not because it is not a potential 	
issue, 	but 	because it is difficult to see how auction design could signi	ficantly reduce the risk of such 	
collusion.	 	
Text frame 	3.5 	Summary	 of collusion	 	
• 	Collusion is not a major issue for SNII phase 1	 	
• 	First	-price auctions are preferred if collusion is a concern	 	
• 	Ascending auctions should be open	-exit or fixed	-increment to make collusion more difficult	 	
• 	Across	-auction collusion is more	 difficult to prevent than within	-auction collusion	 	
3.2.6	 	Entry	 	
Having 	many	 bidders	 enter the auction reduces the risk of low	 revenues, 	inefficiency,	 and collusion 	
(Klemperer, 2002b)	. Entry is costly to all bidders, 	because the time and resources spent preparing to 	
participate in the auction have	 alternative uses and thus	 opportunity costs.	 Weaker	 bidders may choose 	
to opt	-out 	of the auction 	altogether	, expecting to lose against a perceived stronger 	competitor	. In 	the 	
worst	-case scenario, all weak 	bidders stay out, leaving only 	the	 strong	est	 bidder to 	participate	. This 	
bidder	 would then	 pay 	a very low price for the object. Knowing this, stronger bidders 	may try	 to 	deter	 	
entry	. 	
Firms	 have incentives to establish a r	eputation for being aggressive in auctions	, causing others to stay 	
out of later auctions if 	they know that the perceived aggressive competitor will participate	 (Klemperer, 	
2002b)	. This mechanism may affect bids positively in Phase 1, because firms 	can try 	to establish a 	rep-	
utation	 or get a	 toehold in the Norwegian offshore wind market. However, it may 	negatively affect	 Phase 	
2 revenue 	if the winner of Phase 1 successfully deters Pha	se 2 entry.	 The net effect on auction outcome 	
is ambiguous.	 	
The choice of auction type can affect entry decisions	 (Klemperer, 2002b)	. In 	ascending	 auctions, 	bidders 	
can always outbid each other	 in the end	 even if 	they were 	outbid 	early	. Therefore, b	idders expect that 	
the strongest 	or most aggressive 	bidder always wins	 and may stay out because the low perceived chance 	
of winning. 	The 	auction 	outcome is not as clear in 	first	-price auctions	, because of strategic uncer	tainty 	
 	
21 Open	 exit means that the standing bid automatically 	increases,	 and bidders announce their irrevocable exit 	from the auction.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	46	 	
 
about how much to optimally shade bids. 	This uncertainty gives weaker bidders a chance of 	winning 	
against	 stronger competitors, which all else equal increase	s entr	y relative to ascending auctions	. 	
We do not consider 	low entry an important issue in SNII phase 1, as our impression is that there 	is great 	
interest for participating in this auction. However, we believe entry could be an issue in phase 2.	 	
Text frame 	3.6 	Sum	mary of entry	 	
• 	First	-price sealed	-bid auctions have better properties with respect to entry, as it yields weaker bid-	
ders a chance of winning.	 	
• 	We do not consider low entry an important issue in SNII phase 1.	 	
3.3	 	Reservation price	 	
Reservation prices, also 	known as c	eiling prices	, define the maximum 	subsidy	 leve	l and 	limit the risk of 	
high support costs for governments/consumers in case of uncertain or limited competition in the auc-
tion or collusive 	behaviour	 between bidders	. As a general point, maximum prem	iums should be seen as 	
a safeguard against the risk of inadvertently awarding projects to very high	-priced bidders, not as the 	
mechanism by which RE prices are driven downwards 	– that should be achieved through competition	. 	
The 	reservation price	 could be 	defined	 using an LCOE calculation	 or based on a previous feed	-in tariff 	
(FIT). Since there are no FITs in Norway and 	NVE 	has	 estimate	d the	 LCOE for Sørlige Nordsjø II, 	the 	
reservation	 price could be set based on an LCOE calculation. 	The 	reservation	 price	 should be set at a 	
level that allows sufficient room for competitive price discovery and should thus not be set too low	, 	
implying a very low subsidy which could deter auction participation	. 	
Disclosing the 	reservation price	 to bidders in advance has the advantage that it prevents otherwise 	
qualifying	 bids	 from being rejected simply because bidders did not know the 	reservation	 price. The dis-	
closure of the	 reservation	 price also gives bidders more planning security, increasing	 the acceptance of 	
the auction. A	 disadvantage of disclosing ceiling prices in 	sealed	-bid auctions	 where competition	 is low	 	
is that it can weaken the	 competitive price discovery of the auction if bidders orient their bids toward 	
the ceiling price	. 	
Accordi	ng to the AURES auction database, 	all 	EU countries 	that implemented RE auctions	 between 2011 	
and 2021	, including	 for offshore wind, defined a reservation price	. The reservation price was disclosed 	
ahead of the auction in all countries. Only Spain in 2021 for multi	-object solar PV and onshore wind, and 	
Denmark for a multi	-object solar PV auction in 2018, have no	t disclosed the reservation price before the 	
auction	. Considering the advantages and disadvantages	 of and country experience with	 reservation 	
prices	, the reservation price should be disclosed to bidders ahead of the auction	. 	
3.4	 	Recommendation	 	
Table 	3.5 summarizes our assessment of the different auction types. 	The ascending and Anglo	-Dutch 	
types are compared to the first	-price sealed bid 	auction	, which is set as the reference. The pluses and 	
minus	es thus indicate strengths and	 weaknesses relative to	 first	-price	 sealed bid	. Pluses and minuses 	
cannot be counted across rows.

Norwegian Offshore Win	d Auctions	 	
 	
Vista Analyse	 | 2022/44	 	47	 	
 
Table 	3.5 	Pros and cons of different auction types	, relative to first	-price sealed	-bid	 	
Issue	 	Importance	 	First	-price	 sealed	-bid	 	
Sealed final bids 	for 	
subsidy/payment (up-
front) or strike price 	
(CfD).	 	
Ascending	 	
Upfront:	 High to low 	
subsidy, eventually 
low to high payment	. 	
CfD:	 High to low strike 	
price.	 	 	
Anglo	-Dutch	 	
As ascending until two 
bidders remain, then 
sealed final bids	. 	
Winner’s curse	 	High	 	Reference	 	+++ 	++ 	
Efficiency	 	High	 	… 	+++ 	++ 	
Risk	-averse bidders	 	Medium	 	… 	--- 	- 	
Familiarity	 	Medium	 	… 	- 	-- 	
Collusion	 	Low	 	… 	-- 	- 	
Entry	 	Low	 	… 	-- 	- 	
Source:	 	Vista Analyse	 	
We assess 	winner’s curse as the most 	important 	concern	 with regards to the choice of auction 	type	 for 	
SNII. This is partly due to the classic winner’s curse problem in auction theory, but mainly due to the risk 
of the project not being carried out.	 This 	pulls toward	s an ascending auction, secondarily towards an 	
Anglo	-Dutch auction.	 Although there is a risk of winner’s curse in the second stage of an Anglo	-Dutch 	
auction, we believe this risk to be much smaller than in a first	-price auction.	 	
We consider 	efficiency, t	hat the lot is allocated 	to the bidder with the highest valuation	, as the second 	
most import concern. 	This also pulls towards an ascending auction, secondarily towards an Anglo	-Dutch 	
auction	. Again, we believe the difference between these t	wo types to be small. While an ascending 	
auction ensures that the bidder with highest valuation wins the auction, an 	Anglo	-Dutch auction ensures 	
that the winner is one of the two bidders with highest valuation. 	 	
We	 also put some weight on risk	-aversion	, due t	o the flagship nature of this first Norwegian offshore 	
wind auction.	 In a first	-price sealed	-bid auction, bidders 	will weigh the possible profit from 	a low bid 	
(high subsidy) against the probability of not winning. A risk	-averse bidder will 	want to bid hig	her (lower 	
subsidy) to increase the probability of winning.	 The bidder does not want to risk losing. This can lead to 	
a high winning bid (low subsidy), which increases government revenue (reduces government expenses).	 	
This is beneficial. 	The ascending auct	ion does not allow for this, as the winning bid is likely to be just 	
above the second highest bid.	 However, the Anglo	-Dutch allows for this to some degree, because 	there 	
is a strategic consideration in the final first	-price sealed	-bid round.	22 	
Finally, fam	iliarity is also a concern we put some weight on	. Most offshore wind power auction have 	
been sealed	-bid first	-price auctions. However, 	ascending auctions are used in many different contexts 	
and are relatively well known. The greatest challenge with ascendi	ng auction might be which terms 	to 	
use, as the 	ascending bids 	are really 	decreasing subsidies	 in this context, as long as the standing bid still 	
implies a subsidy. 	In the case of a CfD, the 	ascending bids	 are 	decreasing strike prices	. We believe this is 	
a challenge that can be overcome. 	A solution could be to use the term 	open	-bid	 auction, which is less 	
precise but also less confusing.	 
 	
22 Note that with respect to this concern, the term risk is used in another way than in the discussion of risk allocation in cha	pter 	
2.2.2	 on support schemes.	 This is about the strategic 	considerations of bidders. 	Uncertainty about	 profitability is dealt with 	
in the discussion of the winner’s curse.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	48	 	
 
Although Anglo	-Dutch auctions are 	much less used, we do not discard them due to lack of familiarity	. 	
As mentioned in the section on experiences from other countries, 	such auctions are used for renewable 	
energy in Brazil	 and for demand flexibility in the state of New York	. Furthermore, our partner Procurex	 	
has extensive experience with such auctions, 	whi	ch in their experience are often referred to as “final 	
blind” auction	s. 	
Collusion and entry are generally 	important concerns in auction design, but we do not believe them to 	
be crucial in the auction for SNII phase 	1. As there is only one lot, it is diffic	ult for collusion to be a major 	
issue. Cross	-auction collusion, across 	phases or regions, could be a concern, but it is hard to see how it 	
can be mitigated through the choice of auction type. 	Due to the considerable interest in participating in 	
the auction	, we do not regard entry as a challenge. However, note that with respect to both these con-	
siderations, an Anglo	-Dutch auction has better properties than a pure ascending auction.	 	
Based on our assessment, 	this is our ranking of the auction types	: 	
1. Anglo	-Dutch	 	
2. Ascending	 	
3. First	-price sealed	-bid	 	
We recommend 	the Anglo	-Dutch auction because it retains most of the good properties of the ascend-	
ing auction with respect to the concerns we find 	the 	most important, while also having some of the 	
good properties 	of the first	-price sealed	-bid auction with respect to the less important concerns. The 	
main argument against is that it is a relatively novel auction type in this context. However, 	as it is not 	
overly complex and Procurex has extensive experience with this	 type of auction, we consider this a 	
manageable challenge.	 	
This is	 our	 recommendation regardless of 	which	 support scheme	 is chosen	. However, if an upfront 	
scheme is chosen, the case for an 	Anglo	-Dutch or ascending auction is even stronger. This is due to t	he 	
good properties of these auctions with respect to the 	winner’s 	curse.	 	
Furthermore, we recommend that a reservation price is set 	as a safeguard. In the case of an upfront 	
scheme, the reservation price will be the highest possible subsidy. In the case of 	a CfD scheme, the 	
reservation price will be the highest possible strike price.	 While the reservation should be set taking into 	
consideration LCOE estimates for SNII, it should not be set so 	low	 that it 	causes potential bidders to not 	
enter the auction.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	49	 	
 	
4	 	Fu	rther considerations of rele-	
vance for the auction	 	
This chapter collects 	regulatory 	issues and problems that 	have implications for auction design and/or 	
the success of the auction in terms 	of 	achieving 	its 	objective: 	“An efficient allocation of the licenses to 	
be auctioned, with the purpose of realizing the highest sum/lowest suppor	t”. 	The 	chapter has the fol-	
lowi	ng subsections:	 	
• 	Requirements to pre	-qualification criteria from the perspective of the auction	 	
• 	Pre-development responsibilities between state and bidders	 	
• 	Grid connection and payment	 	
• 	Realization deadlines and penalties	 	
4.1	 	Requirements to pre	-qualification crit	eria from the perspective of the 	
auction	 	
Pre	-qualification requirements are 	used	 to increase the probability of timely project realization and to 	
prevent bidders from participating in 	the auction	 with no 	robust probability	 of realizing	 the project. 	Pre	-	
qualification criteria can be financial or technical/environmental in design, and t	hey can relate to the 	
bidder in general or to the Sørlige Nordsjø II project specifically. 	Pre	-qualification requirements for the 	
auction 	of Sørlige Nordsjø II	 will be developed by the 	Ministry and are 	not 	part of our assignment. How-	
ever, 	the design and im	plementation of 	pre	-qualification criteria 	may have a bearing on the 	success	 of 	
the auction	. This section deals with 	the interaction of pre	-qualification criteria and auction 	success. 	 	
A viable auction requires a number of 	qualified bidders. 	It is difficul	t to think of a 	reason why 	a limited	 	
number of bidders is optimal in the sense that 	an additional	 qualified 	bidder will reduce the 	expected 	
success of the auction. The reason is of course that the 	additional	 bidder may be the one that goes on 	
to win the auction. 	 	
If the number of qualified bidders for some reason should be restricted it is important that the	 number 	
is sufficiently lar	ge. 	From the perspective of the auction 	we recommend that 	the number of bidders 	
should be at least eight	 if there is a sufficient number of qualified candidates.	 	
From the perspective of the auction 	a fixed	 number of participants	 determined by the Ministry 	will 	
probably	 create an “arms race” between contestant	s who seek to be included as bidders. This will in-	
duce them to 	invest heavily in producing the required documentation	 at the societal cost of providing 	
a number of duplicates. More importantly	, perhaps	: it will 	induce them to make promises 	for their pro-	
ject design in environmental, technical and other dimension	s which	 will add costs to the project should 	
it be implemented. 	This arms race may lead to over	-fulfilment	 of qualification criteria at the cost 	of 	
higher support	 to the winning bidder	. 	
Another possibility is that 	too few contestants will bother to participate in pre	-qualification given the 	
up	-front costs	. This may lead to fewer than 	eight bidders in the auction, and 	– depending on who drops 	
out 	– perhaps not the ideal composition of bidders.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/	44	 	50	 	
 
The alternative to a competition for a fixed number of slots in the auction is t	o set fixed hurdles that 	
contestants should pass, and give a slot to everybody that passes. A compromise would be to set fixed 
hu	rdles 	in some dimensions (financial strength?) and allow a competition in other dimension (environ-	
mental safeguards?).	 	
Many NSEC countries have 	in pre	-qualification 	opted for the sole use of financial pre	-qualification crite-	
ria (	e.g	., Denmark, Germany, and the Netherlands) 	and/	or bidder qualification requirements (	e.g	., 	
France	, Denmark	), given that sites are pre	-developed to a large degree by authorities. The UK makes 	
use of various technical pre	-qualification requirements (	e.g	., pla	nning authorizations and lease agree-	
ments for the site), given its decentralized system of site selection. 	 	
4.2	 	Pre	-development responsibilities between state and bidders	 	
Site selection and the cost of geotechnical studies are issues that generally are of conc	ern in wind energy 	
auctions	. In this case	 the site is selected and 	seabed mapping	 at the site will 	be carried out prior to the 	
auction on 	behalf of 	the Ministry	, and later paid for by the auction winner	. 	
Norwegian practice on this point	 corresponds to Euro	pean best practice as it is being practiced in Bel-	
gium, Denmark (except nearshore & open	-door procedure), German and the Netherlands.	 	
The Offshore wind energy act and associated regulation put forward requirements for pre	-development 	
applications and permi	ts that the winner of the 	SNII 	auction needs to undertake within predefined time 	
limits. From the point of view of the auction the	 uncertain approval of applications 	are	 uncertainties	 	
that it would be advantageous to reduce as much as possible.	 	
The uncertainties are somewhat reduced if the pre	-qualification criteria are aligned with the legal re-	
quirements. For instance, a bidder may commit in the pre	-qualification stage to pursue a pro	gram for 	
the study of bird migration and adapt the location of his 	turbines 	to what 	the 	program reveals.	 If the 	
evaluation is aligned with the legal process and in effect is a 	pre	-approval of (this part of) an EIA some 	
regulatory 	uncertainty 	is removed 	at 	the bidding stage. 	The main uncertainty in this case, however, is 	
what the program reveals and how the location is affected. 	 	
In Europe 	some variation exists regarding the responsibilities attributed to the state body or the devel-	
oper. In general, a minimu	m level of preliminary and basic site investigation performed by a public au-	
thority seems advantageous as it contributes to de	-risking 	developers,	 and it allows for a more con-	
trolled rollout of offshore wind capacity to meet deployment targets. 	It is possi	ble to go further	, and 	let 	
a state body provide a one	-stop	-shop where the successful developer receives all the required permits 	
and licences to start developments, as is the case in the Netherlands. 	In other NSEC countries develop-	
ers are still required 	to fill in 	pre	-development applications	, as is the case in Norway.	 	
4.3	 	Grid connection and payment	 	
The selected 	grid connection regime	 or grid delivery model for offshore wind farms defines stakeholder 	
roles and responsibilities during the different project phas	es. The grid delivery model determines who 	
finances the grid connection and where the interface lies between a developer and the governing TSO, 
transmission asset owner (TAO), or offshore transmission owner (OFTO). The adopted regime balances 
cost to consu	mers and government control over planning and realization timelines with de	-risking of 	
developers in the various stages of the project.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	51	 	
 
There are two main grid delivery models for offshore wind farms, developer	-built/decentralised and 	
TSO	-built/centralised	. Under the 	developer	-built	 or decentralized grid delivery model (also applicable in 	
Norway), the offshore wind farm and offshore wind transmission assets fall under the responsibility of 
the developer; this represents a deep grid connection regime. 	 	
In NS	EC countries and the UK, various models are adopted that combine responsibilities of the two main 	
models. The UK is currently the only market with a full developer	-built offshore grid delivery approach. 	
In Denmark, the developer 	of the 	Thor 	offshore wind farm is responsible for the offshore substation and 	
export cable. The grid delivery models in other NSEC countries with an offshore market have generally 
evolved from direct connections established and operated by commercial parties towards a	 TSO	-built 	
model where the TSO or TAO has a legal obligation or a government mandate to design, build, and 
operate the offshore grid. Other NSEC countries have no significant amount of offshore wind farm ca-
pacity in place yet and have a more decentralised 	model (increased grid delivery responsibilities for the 	
wind farm developer), for example, Ireland’s current system.	 	
For the upcoming 	auction for Sørlige Nordsjø II,	 it has 	not	 been 	de	cided	 whether	 the project developer 	
or Statnett 	will develop the radial 	connection to shore. 	In any case, it is likely that the radial will be 	
classified as customer specific	, which implies that the 	developer pays the full cost	. However, this is yet 	
to be concluded.	 	
Statnett would be responsible for potential onshore grid rei	nforcements. Statnett is currently in the 	
process of identifying connection points in the onshore grid that have sufficient capacity. 	While nowa-	
days many countries have moved to a TSO	-led development (due to higher efficiency/better grip on 	
offshore wind d	evelopments), many NSEC countries had developer	-led developments in the early 	
stages. Overall, Norway would not be contradicting the general trend with the choice to go first devel-
oper	-led (with the option to later switch to TSO	-led). 	 	
According to Norwegi	an law, the developer should pay 100% of the costs of onshore grid infrastructure 	
built specifically to serve the offshore wind farm. These costs should be determined and announced 
before the auction. A	t the very least 	bidders should know in advance of the	 auction the exact location 	
of the substation, the voltage level, and the capacity. This information will be crucial for their bid. 	 	
However, even	 if the location is made known to the bidder, there are still risks for the project developer 	
related to the l	andfall (delayed permits, uncooperative local authorities). To 	reduce	 these uncertainties, 	
the government could start 	preparing 	the permitting process for the 	radial connection	 cable	 as soon as	 	
the onshore substation has been identified by Statne	tt, so that such issues 	to some degree 	can be 	re-	
solved before the auction	. This could shorten the realization time needed by the project developer. 	 	
If investments in the meshed grid are necessary, the developer will have to pay 50% of its proportional 
share	 of these costs. The proportional share depends on the number and size of others benefitting from 	
the investment. Our impression is that the need for such investment is small for SNII phase 1, but may 
be larger for phase II. We recommend (1) that need for 	and cost of such investments is determined 	
before the auction and (2) that the Ministry considers letting the government pay these costs. If the 
government pays these costs, they will likely be more than recuperated in the auction. This will reduce 
uncerta	inty for the bidders and reduce the probability of non	-completion.	 Alternatively, the government 	
could set a cost cap above which it covers the costs.	 If the cap is not set too high, this will reduce much 	
of the uncertainty.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	52	 	
 	
4.4	 	Realization deadlines and penal	ties 	 	
The 	realization period	 specifies the time during 	which	 the	 auction winner needs to apply for a conces-	
sion, i.e., the exclusive right to build the offshore wind farm within a specific area in Sørlige Nor	dsjø II, 	
and the time until which the 	project ne	eds to be commissioned, i.e. be in operation. If the realization 	
periods 	is exceeded, i.e., a project fails to be completed in time	 or the auction winner exceeds the time 	
to conduct the necessary project specific assessments in an area and apply for a conc	ession to develop 	
and produce renewable energy in that area	, the right to the acreage will be withdrawn and additional 	
penalties may be imposed. Realization periods need to reflect realistic project delivery periods while 
avoiding lengthy 	delays	 that would	 encourage speculative 	behaviour	 and thus increase the risks that 	
projects are not realized. 	 	
The Regulation to the Offshore wind energy law (“havenergilovforskrifta”) states that the winner of the 
auction	 (in this case	) “initially” shall submit a 	proposal for a project specific research program.	23 The 	
program will include an environmental impact assessment. The Ministry approves the program. 	 	
Within two years from that point the 	winner must submit his application for concession, appending the 	
EIA and other documentation. 	According to present rules, the concession owner should submit his “de-	
tailed plan” two	 years after approval of concession	. However, the government is assessing 	wh	ether the 	
concession application and the detailed plan can be processed simultaneously to shorten the process. 
Finally	, the plant should start operation within three years following approval of the detailed plan. All in 	
all we are looking at a period of ab	out 	eight	 years depending on approval processes from a winner is 	
selected to production should start	 (i.e. the realization period is about eight years)	. However, the period 	
might be shorter if the 	concession application and the detailed plan can be process	ed simultaneously. 	
The Ministry may grant multiple two	-year extensions to the deadlines, but this is at 	their 	discretion. 	
Otherwise	, oversitting a deadline may lead to withdrawal of the right to get to the stage of operation.	 	
Of course	, these are maximum t	ime limits. 	 	
These regulatory deadlines compare to the ideal 	duration of the realization period	, which 	depends on 	
several elements, including 	 	
• 	Realistic timeframe to conduct necessary assessments	 and commission the project	: The pen-	
alty	-free period to conduct site	-specific assessments and commission the project should be in-	
formed by international best practice. Otherwise, adverse implications for risk levels incurred 
by bidders and as a result potentially higher bid price in	 the auction may be the result, while at 	
same time increasing chances of project non	-realization. 	 	
• 	Infrastructure development and permitting:	 Realization periods should be aligned and coordi-	
nated with the required infrastructure development lead timelines.	 The realization period 	
should consider both the lead times for the offshore infrastructure as well as the lead times for 
the required onshore infrastructure. 	See our discussion in 	4.3	. 	
• 	Supply chain:	 Supply chains are currently under great pressure as internationally the offshore 	
wind roll	-out increases in pace. This could result in production slots being fully booked to the 	
point that a short realization period may	 not be achievable by the developers.	 	
• 	First of a kind:	 Since this will be the first offshore wind tender in Norway, some discretionary 	
time may be considered in case of unforeseen lags in permitting or slow coordination of the 
different actors. 	 	
 	
23 §3: «Dersom eit føretak som oppfyller krava i havenergilova § 3	-5 ønsker å søke om konsesjon til eit energianlegg, skal det 	
fyrst sendast inn ei melding til departementet, med framlegg til eit prosjektspesifikt utgreiingsprogram.»

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	53	 	
 
For the 	realization period, different systems are in use across NSEC countries. In the UK, contracts are 	
awarded for delivery in a particular year. In the Netherlands, 	for the most recent tenders the 	realisation 	
period 	was 4	 years	. However, 	in 	the 	Netherland	s, the government has performed the environmental 	
impact assessment prior to the award. 	In Germany, the realisation period starts from grid completion 	
onwards 	– projects then have 18 months to realize. 	It seems that the current Norwegian regulation 	
gives con	siderably more leeway than this. 	 	
Bid bonds	 may be asked from bidders before participating in the auction to ensure the successful bid-	
der’s commitment to 	enter 	a contract after being awarded. If a successful bidder does not sign the 	
support contract, the s	upport counterparty/auctioneer will retain the bid bond, otherwise, the bidder 	
receives the bond back.	 	
Furthermore, 	financial 	guarantees	 can be demanded, which together with	 penalties	 in the	 case of non	-	
delivery or delay beyond the contractually agreed realization period	, can contribute	 to timely delivery 	
of projects. Beyond the already foreseen withdrawal of the right to build the park in case the offshore 
wind park is not commissioned withi	n the set realization period, 	submitted financial guarantee	s at the 	
time of entering the auction 	may be confiscated partly or in full	, depending on the extent of the delay 	
or non	-commissioning (so	-called 	completion bond	). In line with international best pr	actice, imposing 	
financial guarantees in the range of up to 10% of estimated project costs may be suitable to ensure 
project delivery. However, this only serves as a rough indication and should be revisited as the auction 
design is defined in more detail.	 	
When determining the extent of penalties, a right balance needs to be struck between maximizing re-
alization rates and avoiding excessive risks for project developers. If penalties are set too strict and fi-
nancial guarantees are too difficult to obtain, 	it may lead to low participation, and bid prices might in-	
crease. If penalties are too low, 	it might lead to	 a longer realization period or even non	-realization.	 There 	
is a risk of the bidders treating 	the auction 	as only the first stage of a 	two	-stage investm	ent decision, 	
where 	winning the auction secures a real option that can be exercises 	if the conditions prove favourable	 	
(Welisch & Poudineh, 2019)	. The limited 	cost of entering the auction and low p	enalties, can be seen 	by 	
the bidders 	as the price of 	acquiring 	the real option.	 Beside	s the full or partial confiscation of the finan-	
cial guarantee, penalties can take several forms	, such as 	lowering of support level	 or the 	shortening of 	
contract validity	 period	 by the time of the delay. Penalties may also be escalated over time to account 	
for the extent of delays or deviation from contractual obligations.	 Termination of the contract would be 	
a penalty 	by itself 	if there is profit in the contract. 	 	
As note	d in the final bullet 	point	 above, since this is the 	first offshore wind auction in Norway	, the 	risk 	
of delay may be high compared to other markets in NSEC. Comparatively high penalties may add to th	is 	
risk	 from a developer point of view. 	At the same time,	 ensuring 	timely realization	 by ensuring only seri-	
ous bids are considered is key for creating confidence in the Norwegian market. Implementing suffi-
ciently high penalties also decreases the winner’s curse risk	 and e	nsure that bidders submit adequate 	
bids t	hat are likely to allow for successful project completion. 	In this sense well	-designed penalties may 	
reduce	 the 	need for strict	 pre	-qualification 	criteria	.

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	54	 	
 	
5	 	Refer	ences	 	
Avery, C. (1998). Strategic Jump Bidding in English Auctions. 	Review of Economic Studies	. Retrieved from 	
https://oconnell.fas.harvard.edu/files/cavery/files/avery_1998	-04.pdf	 	
Energistyrelsen. (2022, August 30). Oversigt over først til mølle. 	Retrieved from	 	
https://ens.dk/sites/ens.dk/files/Vindenergi/oversigt_over_foerst_til_moelle_i_aaben_doer_o
rdningen_30082022_0.pdf	 	
Fonesca, M. A., & Normann, H.	-T. (2012). 	Explicit vs. Tacit Collusion 	– The Impact of Communication in 	
Oligopoly Experiments.	 Discussion pap	er no 65. Düsseldorf Institute for Competition Economics. 	
Retrieved from https://d	-nb.info/1026225752/34	 	
Ivaldi, M., Jullien, B., Rey, P., Seabright, P., & Tirole, J. (2003). 	The Economics of Tacit Collusion 	- Final 	
Report for DG Competition, European Comm	ission.	 Toulouse: IDEI. Retrieved from 	
https://publications.ut	-capitole.fr/id/eprint/1200/1/tacit_collusion.pdf	 	
Jansen, M., Beiter, P., Reipin, I., Müsgens, F., Guajardo	-Fajardo, V., Staffell, I., . . . Kitzing, L. (2022, May 	
24). Policy choices and outcom	es for offshore wind auctions globally. 	Energy Policy	(167).	 	
Klemperer, P. (1998). Auctions with almost common values: The `Wallet Game' and its applications. 	
European Economic Review	. Retrieved from https://doi.org/10.1016/S0014	-2921(97)00123	-2 	
Klemperer, 	P. (1999). Auction Theory: A Guide to the Literature. 	Journal of Economic Surveys	. Retrieved 	
from 
https://www.cs.princeton.edu/courses/archive/spr09/cos444/papers/klemperer_guide.pdf	 	
Klemperer, P. (2002a). 	Using and Abusing Economic Theory.	 Marshall Lecture to European Economic 	
Association. Retrieved from http://www.nuff.ox.ac.uk/users/klemperer/usingandabusing.pdf	 	
Klemperer, P. (2002b). What Really Matters in Auction Design. 	Journal of Economic Perspectives	. 	
Retrieved from http://darp.lse.a	c.uk/PapersDB/Klemperer_(JEcPersp02).pdf	 	
Maskin, E. (2001). 	Auctions and efficiency.	 Institute for Advanced Study and Princeton University. 	
Retrieved from https://scholar.harvard.edu/files/maskin/files/auctions_and_efficiency.pdf	 	
McAfee, R., & McMillan, J. (1987). Auctions and Bidding. 	Journal of Economic Literature	. Retrieved from 	
https:	//vita.mcafee.cc/PDF/JEL.pdf	 	
Milgrom, P. R., & Weber, R. J. (1982). A Theory of Auctions and Competitive Bidding. 	Econometrica	. 	
Retrieved from 
https://www.cs.princeton.edu/courses/archive/spr09/cos444/papers/milgrom_weber82.pdf	 	
Vasserman, S., & Watt, M. (2	021). 	Risk Aversion and Auction Design: Theoretical and Empirical Evidence. 	
International Journal of Industrial Organization	. Retrieved from 	
https://shoshanavasserman.com/files/2021/05/RiskAversionSurvey.pdf

Norwegian Offshore Wind Auctions	 	
 	
Vista Analyse	 | 2022/44	 	55	 	
 
 	
 
 
 	
Vista Analyse AS	 	
Meltzers	 gate 4	 	
0257 Oslo	 	
 
post@vista	-analyse.no	 	
vista	-analyse.no