ham/0053.1999-12-22.kaminski.ham.txt

Subject: re : working gas price model
vince -
i have a simplified version of brad ' s model in mind .
the " no arbitrage " condition equates trading margins across the country .
costs of transmission rise with congestion on the network . wellhead supply is
almost completely price - elastic , while burner - tip demand is almost
completely price inelastic . storage is rationalized as a perpetual call
option .
the least time - variant parameters are the costs of injecting and withdrawing
gas from storage to the pipeline , followed by the costs of delivering gas
from the wellhead to the pipeline . the intermediate - variant parameters are
the capacity - dependent costs paid to the pipeline ( above shrinkage ) for
transmission . the most time - variant parameters are the trading margins and
the valuations of the storage option .
there are 8 parameters to be estimated at each major node of the betwork .
they are identifiable in either of two straightforward ways : using a short
time series of the last 3 days prices based on the assumed variability
mentioned above , or point - estimates ( " calibrations " ) using only today ' s data
based on a node - based model of competition between pipelines where pipes with
the same region of origination , albeit markedly different terminus , price
versus capacity similarly , " competing " for outflows .
i will write this up for you in scientific word and present it to you at your
earliest convenience .
clayton