A Vector Autoregressive Model for Electricity Prices Subject to Long Memory and Regime Switching

TY - UNPB

T1 - A Vector Autoregressive Model for Electricity Prices Subject to Long Memory and Regime Switching

AU - Haldrup, Niels

AU - Nielsen, Frank

AU - Nielsen, Morten Ørregaard

PY - 2007

Y1 - 2007

N2 - A regime dependent VAR model is suggested that allows long memory (fractionalintegration) in each of the regime states as well as the possibility of fractional cointegra-tion. The model is relevant in describing the price dynamics of electricity prices where thetransmission of power is subject to occasional congestion periods. For a system of bilat-eral prices non-congestion means that electricity prices are identical whereas congestionmakes prices depart. Hence, the joint price dynamics implies switching between essen-tially a univariate price process under non-congestion and a bivariate price process undercongestion. At the same time it is an empirical regularity that electricity prices tendto show a high degree of fractional integration, and thus that prices may be fractionallycointegrated. An empirical analysis using Nord Pool data shows that even though theprices strongly co-move under non-congestion, the prices are not, in general, fractionalcointegrated in the congestion state.

AB - A regime dependent VAR model is suggested that allows long memory (fractionalintegration) in each of the regime states as well as the possibility of fractional cointegra-tion. The model is relevant in describing the price dynamics of electricity prices where thetransmission of power is subject to occasional congestion periods. For a system of bilat-eral prices non-congestion means that electricity prices are identical whereas congestionmakes prices depart. Hence, the joint price dynamics implies switching between essen-tially a univariate price process under non-congestion and a bivariate price process undercongestion. At the same time it is an empirical regularity that electricity prices tendto show a high degree of fractional integration, and thus that prices may be fractionallycointegrated. An empirical analysis using Nord Pool data shows that even though theprices strongly co-move under non-congestion, the prices are not, in general, fractionalcointegrated in the congestion state.

KW - Cointegration, electricity prices, fractional integration, long memory, Markov switching

M3 - Working paper

BT - A Vector Autoregressive Model for Electricity Prices Subject to Long Memory and Regime Switching

ER -