Response To LP Workshop Report



Response to the Load Profile Workshop Preliminary Report Issued on June 11,
1997 by SCE, PG&E and SDG&E
by
Michael Parti
Applied Econometrics, Inc.
email: parti@inetworld.net
voice: 619 - 755 - 1266

The main issues to be covered in this response concern 

1. the amount of load profiling information that would be provided by the IOU's;
2. the question of which load profile would be used for customers who left
IOU's (the IOU's load profile or the ESP's load profile); and, 
3. the timing of the planning and implementation of competition.

These goals will be discussed in order.  After this a short statement about
the Load Profiling interest of Applied Econometrics, Inc. and a
mini-glossary conclude this response.


Issue 1: Load Information To Be Released By The IOU's

The discussion in this section concerns the UDC’s points in section 4 of
their report having to do with the design of load profiles.

There seems to be no stated IOU goal that is served by severely restricting
the amount of load shape information furnished to the interested community.
The IOU's intend to provide only average customer-class-level load
information to the interested community.  To make matters worse, this
minimal load information is to be provided in weather-normalized form with
no information about how the weather normalization was performed or how we
would adjust the load shapes to track actual (as opposed to normal) weather.  

There seemed to be two types of arguments that the IOU's used to justify
their intention to provide only this minimal load information: 1) a cost and
time-constraint argument; and 2) a customer privacy issue.  As regards the
cost/time argument, it would seem that less IOU effort would be required for
the simple transfer of the raw load data on which their weather-normalized
averages would be based.  The load data and any related raw data (weather,
ZIP code, SIC information, etc.) would allow non-IOU analysts to investigate
the underlying regularities and the causes of differences in customer load
shapes.  This type of basic knowledge would be useful for all the players in
the market and would help to provide an analytical basis for system
planning.  On a broader level, if the UDC's cannot provide adequate
segmented load profiles for 1/1/98, I suggest that they allow a competent
body such as the CEC to perform this function.  As regards the data privacy
issue, it would be a simple matter to eliminate the customer names,
addresses and telephone numbers from the files to be transferred.  

In section 2 of their report, the IOU's raised many objections having to do
with the workability of competition (customer churning, etc.).  It is
curious then that they propose to limit our access to the load information
that would increase our ability to solve these problems.  Likewise, it is
difficult to understand why this load information, developed at ratepayer
expense, should not be used to foster competition which would benefit
ratepayers.



Issue 2: The Question of Which Settlement Load Profile Would be Used for
Customers Who Leave UDC's (The UDC's Load Profile or The ESP's Load Profile)

The discussion in this section concerns the points in section 2 of the UDC
report having to do with their contentions that they should have, “financial
indifference to the application of load profiles,” and that “a single load
profile be used for all customers in a rate category regardless of whether
the customer takes service from the UDC or a retailer.”  Though it may not
be obvious, these issues are related.

The UDC's face the problem of maintaining a rate cap of 10% less than the
rate in June, 1996; while securing payment of their stranded asset costs.
These two goals could be in conflict if many of the relatively flat-load
customers move to ESP retailers.  In such a case, the average load of the
UDC might become more concentrated in the on-peak hours.  This, in turn,
could lead to an increase in the average PX price reflecting the actual load
shape for the remaining UDC customers.  If more and more of the flat load
customers leave the UDC, its average PX price for its remaining customers
might rise perilously close to the rate cap making the recovery of the
stranded asset costs difficult or impossible.  A PG&E representative called
this type of scenario a "death spiral".

Clearly, the IOU's have a point here.  It should be stressed, however, that
this scenario is merely a theoretical possibility.  No simple spreadsheet or
any other evidence has been advanced to show whether this effect is large
enough to worry about under a realistic set of assumptions.  In fact, it
would be possible to perform a telling simulation experiment to determine
the probable magnitude of the "death spiral" effect using the existing load
research data along with other information.

Whether or not the danger of the “death spiral” is real, the IOU’s feel that
the possibility of this scenario should be eliminated completely by
requiring the ESP’s to use the average UDC load profile for ISO settlement
purposes.  In this case, the UDC and an ESP retailer using energy provided
by the PX, would pay the same energy charge per kWh even if the ESP's
customers actually had flatter (i.e., less expensive) load profiles than the
remaining customers of the UDC's.  This equality of energy charges will
prevent, of course, any possibility of the "death spiral" effect.  It should
be stressed, however, that this IOU approach also eliminates any effective
price-related competition from ESP's who use the PX.  This means, for
example, that an ESP offering an air-conditioning rebate incentive could not
reap, individually, the PX cost benefits that such a rebate would have for
the system.  Therefore, no part of the actual PX rate differences between
this ESP and the UDC could be used by the ESP to attract customers.   

In effect, then, the IOU's are claiming that the only way to collect their
stranded asset costs is by using a market structure that will most likely
eliminate any retail competition from ESP firms supplied by the PX.  There
is an approach, however, that would foster as much competition as possible,
consistent with the goal of paying off the stranded asset costs.  This
approach would also begin the process of full retail competition in a
gradual fashion that would allow us to iron out any restructuring process
problems before they become large problems.

The approach can be summarized as follows:
1. Set a reasonable reference target date for the completion of the primary
stranded asset cost recovery, say 9/31/2001.  
2. Ask each UDC to prepare a forecasted cost recovery completion date each
month beginning 1/1/98.
3. An independent body (the CPUC?) should review the stranded cost recovery
forecasts to ensure that the assumptions are reasonable (e.g., expected PX
prices should bear some sensible relationship to previous PX prices).
4. Each retailer (UDC or ESP) will compute its load profile as accurately as
possible, using methods approved by an independent evaluation entity (the
CEC?), and present it to the ISO for a system-level settlement procedure.
The ISO will compute a system-level average kWh price for each retailer (and
for each of the retailer’s billing-cycles).
5. A secondary settlement procedure will be performed by the UDC using a
method approved by the CPUC.  This secondary settlement procedure will
recompute the revenue allocation to each retailer so that a) only a fraction
of the ESP-UDC price difference will be credited to the ESP; and b) all of
the PX energy costs attributable to the UDC’s and its member retailers will
be fully allocated to them.
6. In this secondary settlement procedure, the fraction of each ESP-UDC
price difference credited to each ESP will initially be quite small and will
be increased during every month that the forecasted cost recovery completion
date for a month precedes the reference date.  We can speed up the rate at
which the ESP’s are given this fractional adjustment if the estimated payoff
date continues to move backwards in time.

The completion reference target date should be selected as the very latest
date that any reasonable prediction error in cost recovery would be
correctable before the elimination of the stranded asset charge.   

Although it may seem extreme to actually transfer revenue from one retailer
to another, the method proposed here is actually less unfair to the ESP’s
than the approach advocated by the IOU’s.  Using the IOU approach, each ESP
uses the UDC’s load profile for PX settlement purposes.  By contrast, using
the method presented here, each ESP’s energy prices more accurately reflect
its actual customer load profile as the danger of the “death spiral”
scenario diminishes over time.   The advantages of this approach are: 1)
full customer-specific load profiling is phased in gradually so that we can
solve any implementation problems while they are small; 2) the UDC recovery
of stranded assets is given high priority; 3) the advantages of competition
to ratepayers are realized as rapidly as possible, given the cost recovery
requirement; and, 4) it requires the UDC’s to provide supporting analysis to
show that we are not sacrificing the benefits of full competition to solve a
cost recovery problem that doesn’t exist.

As to the estimation of the ESP customer load shapes, the ESP’s would be
responsible for demonstrating that their load shapes were as defensible as
those of the UDC’s before an independent evaluation entity (CPUC?, CEC?,
other?).   Oddly enough, since the IOU’s are proposing a relatively crude
load profiling scheme (static class-level averages uncorrected for actual
weather, not peer-reviewed), the ESP’s normal marketing procedures would
probably give them an advantage in such proceedings. 

In addition, note that the stranded asset costs recovered from the UDC
customers probably will be greater than the IOU's are anticipating. This
could occur because the 10% decrease in price ordered by AB1890 will promote
greater kWh consumption considering standard price responses (negative
short- and long-run price elasticities). Note also that since the ESP prices
will probably be lower than UDC prices, the positive effect on kWh
consumption, and thus the stranded asset cost collections, may be even
greater for ESP customers than if these customers were to stay with the UDC.

Summing Up --

As a practical matter, the LP workshop participants should recognize two
simple truths in this negotiation process:
1. the IOU's will not be denied their stranded asset recovery without
putting up a strong legal and political defense;
2. there will be potentially devastating political consequences for the
legislature, the CPUC, and the IOU's if ratepayers find that the IOU's are
"having their cake and eating it too" by recovering their stranded asset
costs at the same time that they have blocked reasonable steps toward
competition.

Most important, we should not block the benefits of competition for the sake
of an unanalyzed theoretical “death spiral” scenario.  If, indeed, we need
to worry about such a scenario then the UDC's should present evidence that
the threat is serious before proposing a draconian rule that will block the
full benefits of competition.  If we find that there is a danger that the
"death spiral" might exist, we should implement a solution, such as the one
I suggested above (there are many other solutions), that preserves as much
of the benefits of competition as possible. 






Issue 3: The Timing of the Planning and Implementation of Competition.

The discussion in this section concerns the point in section 3 of the UDC
report having to do with their contentions that decisions about more
complete steps toward segmented load profiling be put off until 1998.

I propose that we convene a special-purpose workshop run by the CPUC in
which we would discuss plans for encouraging competition while paying off
the stranded asset costs.  This discussion could take place over the
internet, to reduce the time burden on everyone.  It should include all the
topics that were unresolved in the current proceeding.  Just as the IOU's
feel that there is no justification for putting off the repayment of
stranded asset costs, there is also no justification for putting off
consideration of methods for fostering competition.  

The discussion in this workshop should go on until we reach a conclusion.
If no adequate solution is found by 1/1/98, we could use the utility
approach as an interim solution.  Alternatively, if some ESP’s have
defensible load shapes by 1/1/98, there is no reason that we should not make
some effort to reward their industry.




Load Profiling Interest Of Applied Econometrics, Inc.
 For the record, AEI is a non-IOU company that is in the business of
providing load-related planning and marketing information to natural gas and
electricity firms in the U.S. and abroad.  We have been in business since
1980 and we invented the Conditional Energy Demand technique, one of the
staple techniques for econometric load analysis.  AEI’s goals in the current
proceeding are to aid the transition to competition; and, in the longer run,
to provide load shape and other information to the energy market
participants. If the transition to competition is not expedited, AEI’s
fortunes would be adversely affected along with the ratepayers and the other
non-IOU commercial organizations.  From a strategic point-of-view, we feel
that it would be difficult and probably unproductive to quarrel with the
basic stipulations of California Assembly Bill 1890.  Accordingly, we
suggest that we get on with the business of encouraging the maximum feasible
amount of competition consistent with the payoff of the stranded asset costs.



mini-Glossary: 
CEC – California Energy Commission
CPUC – California Public Utility Commission
customer churning – the movement of customers across retailers (UDC's, ESP's)
ESP --  Energy Service Provider (the future competitors of the IOU's)	
IOU-- Investor Owned Utility (SDG&E, PG&E and SCE will become UDC's)
LP – Load Profile 
PG&E -- Pacific Gas And Electric Company
PX – Power Exchange
SCE – Southern California Edison Company
SDG&E  -- San Diego Gas and Electric Company	
UDC -- Utility Distribution Company



.