Electricity%20pricing%20and%20market%20power:%20Evidence%20from%20the%20Hungarian%20balancing%20energy%20market

1
Electricity pricing and market power Evidence
from the Hungarian balancing energy market

• László Paizs
• Institute of Economics
• Hungarian Academy of Sciences
• Mannheim 4rd Energy Conference, May 7-8, 2015

2
Motivation

• Bids for decremental energy show that generators
  offer decremental energy with a significant
  markdown below marginal cost
• average bid 5.7 EUR/MWh (in 2012)
• High market concentration
• average number of bidders in the daily auctions
  for decremental energy 2.3 (in 2012)
• Reasons for high concentration
• balancing market is national in scope (foreign
  generators cannot supply balancing power)
• not all domestic generators can or willing to
  provide balancing power
• some gas-fired generators are switched off due
  to weak market conditions

3
Empirical research on market power

• Previous research has focused on the electricity
  wholesale market, much less attention has been
  paid to the balancing market
• Niu (2005)
• studies ERCOT (Texas balancing energy market)
• predicts equilibrium prices in a linear SFE using
  estimated cost data
• shows that price data fit the theoretical model
  quite well for upward balancing
• Sioshani and Oren (2006)
• study ERCOT
• use similar methodology as in Niu, but also
  individual behavior analyzed
• show that SFE model produces a good prediction of
  bidding behavior of large power plants only
• Heim and Götz (2013)
• examine the drastic price increase in the German
  power reserve market
• show that the two largest generators in pivotal
  position abused their market power

4
Market design for the Hungarian balancing market

• Upward and downward reserve power/balancing
  energy are procured as separate products
• Procurement consists of two phases
• 1st capacity-selection
• 2nd capacity-ordering
• Reserve capacity auction
• held once a year, simultaneous auction of 365
  service periods (each of one day-duration) ?
  significant variation in the mix of suppliers
  across days
• the TSO awards market maker contracts to
  successful bidders, who are selected on the basis
  of their reserve capacity price offers
• data released cover accepted bids for each day
  (identity of bidders, amount of reserve capacity
  allocated to each bidder, accepted reserve
  capacity fee)
• Balancing energy auction
• held daily, 24 separate sealed-bid auctions, one
  for each hour
• participants
• market makers (obligatory)
• non-contracted parties (voluntary) no capacity
  fee paid
• activation order on the basis of energy price
  bids
• remunerated if called, corresponding to their
  energy price bids
• data released cover the simple hourly average of
  all submitted energy price bid, hourly max and
  min energy price bids

5
Providers of downward regulation

• Providers of downward regulation in 2012
  (contracted parties)
• Dominance of gas-fired plants are explained by
• the lack of hydroelectric power plants
• the unwillingness of nuclear and lignite-fired
  based load plants to provide downward reserve

6
Market environment for gas-fired generators in
2012

• General market conditions
• low electricity prices
• average hourly spot price 51.5 EUR/MWh
• in peak-load hours 61.3 EUR/MWh
• high natural gas prices
• average price of natural gas for power plants 42
  EUR/MWh
• Consequences
• Operation of gas-fired units was not profitable
  in most hours.
• Gas-fired plants operated only in periods in
  which they had a contract with the TSO to provide
  balancing power, allowing them to recover their
  loss in the capacity fee.
• In the balancing energy auctions, the TSO
  received bids only from market makers, as other
  (gas-fired) suppliers were not on line.

7
Concentration in the decremental energy market

• Available data cover only bids from contracted
  parties
• But, for reasons mentioned before, others (most
  probably) did not submit offers in most hours
• Daily auction markets for decremental energy is
  characterized by high concentration
• average number of bidders 2.3
• 1 bidder on 74 days
• 2 bidders on 160 days

8
Supply-demand balance in the decremental energy
market

• The expected amount of negative balancing energy
  tends to be higher in the heating season
• The TSO tends to reserve more down-regulating
  capacity in the winter months
• No clear seasonal pattern in supply-demand
  balance
• the expected supply-demand balance seems to be
  constant throughout the year

9
Theoretical framework

• Discriminatory multi-unit auction model (Fabra et
  al., 2006)
• Setup
• two single-capacity firms with asymmetric
  capacities and costs
• each supplier makes one bid, constrained by the
  market reserve price (set by the auctioneer)
• demand is perfectly inelastic and stochastic
• Main results
• demand thresholds that distinguish between
  low-demand realization and high demand
  realizations
• unique pure Nash-equilibrium under low-demand
  realizations
• both firms bid the marginal cost of the less
  efficient supplier
• unique mixed Nash-equilibrium under high demand
  realizations
• with identical costs, the low-capacity supplier
  bids more aggressively
• price competition is more intense when there are
  more bidders, capacities are more symmetrically
  distributed and the reserve price is lower

10
The choice of the theoretical framework

• The discriminatory auction model considered in
  Fabra et al. (2006) captures the most important
  features of the auction mechanism applied in the
  Hungarian balancing energy market
• suppliers are allowed to submit one price per
  unit
• the price in the offer must not be lower than the
  price-floor specified in the market maker
  contract
• The market structure modeled in Fabra et al.
  (2006) is very representative of the Hungarian
  market for decremental energy
• high market concentration
• only two firms were bidding for decremental
  energy on 160 days in 2012
• supply of negative balancing energy is perfectly
  inelastic and stochastic

11
Comparative static results for the duopoly model
in Fabra et al. (2006) the effect of capacity
asymmetry on expected average bid price

• Capacity asymmetry (adjusted for cost
  differences) reduces the intensity of competition

Note Teta demand level
12
Comparative static results for the duopoly model
in Fabra et al. (2006) the effect of capacity
asymmetry on price dispersion

• Capacity asymmetry has a non-monotonic effect on
  price variance
• Increasing capacity asymmetry leads to an initial
  increasing and an eventual drop of the price
  variance

Note Eabs(b1-b2) Emax(b1b2)
-Emin(b1b2)
Note Teta demand level
13
Hypotheses for the empirical analysis

• The number of bidders has a positive impact on
  the average decremental energy price (calculated
  as a simple average of all submitted hourly price
  bids).
• A more equal distribution of capacities among
  bidders has a positive impact on the average
  decremental energy price.
• An increase in capacity asymmetry initially
  increases and then decreases the price dispersion
  (measured by the difference between the highest
  and lowest price bids).

14
Regression model of the average price of
decremental energy
average bid over all bidders and over all hours
between 0AM and 6AM in day t
15
Dependent variable daily average of decremental
energy price bids (HUF/kWh)
heteroskedastic-consistent standard errors are in
parentheses for OLS models number of bidders is
used as weights for WLS models significant at
10, significant at 5, significant at 1
16
Regression model of price dispersion
17
Dependent variable the difference between the
maximum and minimum decremental energy price bids
(HUF/kWh)
Imply that price dispersion is increasing at
HHIasym levels below 0.17 and decreasing at
HHIasym levels of 0.17 and higher
heteroskedastic-consistent standard errors are in
parentheses for OLS models number of bidders is
used as weights for WLS models significant at
10, significant at 5, significant at 1
18
Thank you for your attention!paizs.laszlo_at_krtk.mt
a.hu