ALASKA STATE LEGISLATURE
SENATE RESOURCES STANDING COMMITTEE
January 30, 2017
3:30 p.m.
MEMBERS PRESENT
Senator Cathy Giessel, Chair
Senator John Coghill, Vice Chair
Senator Natasha von Imhof
Senator Bert Stedman
Senator Shelley Hughes
Senator Kevin Meyer
Senator Bill Wielechowski
MEMBERS ABSENT
All members present
COMMITTEE CALENDAR
OIL PRODUCTION FORECAST METHODOLOGY OVERVIEW
- HEARD
SENATE BILL NO. 30
"An Act approving and ratifying the sale of royalty oil by the
State of Alaska to Petro Star Inc.; and providing for an
effective date."
- MOVED SB 30 OUT OF COMMITTEE
PREVIOUS COMMITTEE ACTION
BILL: SB 30
SHORT TITLE: APPROVAL: ROYALTY OIL SALE TO PETRO STAR
SPONSOR(s): RULES BY REQUEST OF THE GOVERNOR
01/20/17 (S) READ THE FIRST TIME - REFERRALS
01/20/17 (S) RES, FIN
01/30/17 (S) RES AT 3:30 PM BUTROVICH 205
WITNESS REGISTER
JIM SHINE, Commercial Manager
Division of Oil and Gas
Department of Natural Resources (DNR)
Alaska Department of Natural Resources (DNR)
POSITION STATEMENT: Provided overview in support of SB 30.
ED KING, Special Assistant to the Commissioner
Alaska Department of Natural Resources (DNR)
POSITION STATEMENT: Commented supportively on SB 30 and helped
provide an overview of the oil production forecast methodology.
DOUG CHAPADOS, President & CEO
Petro Star, Inc.
Anchorage, Alaska
POSITION STATEMENT: Supported SB 30.
DAN STICKEL, Chief Economist
Economic Research Group
Alaska Department of Revenue (DOR)
POSITION STATEMENT: Provided an overview of the oil production
forecast methodology.
PAUL DECKER, Petroleum Geologist and Manager
Resource Evaluation Section
Division of Oil and Gas
Alaska Department of Natural Resources (DNR)
Anchorage, Alaska
POSITION STATEMENT: Provided an overview of the oil production
forecast methodology.
ACTION NARRATIVE
3:30:22 PM
CHAIR CATHY GIESSEL called the Senate Resources Standing
Committee meeting to order at 3:30 p.m. Present at the call to
order were Senators Stedman, von Imhof, Hughes, Meyer,
Wielechowski, and Chair Giessel.
SB 30-APPROVAL: ROYALTY OIL SALE TO PETRO STAR
3:31:13 PM
CHAIR GIESSEL announced SB 30 to be up for consideration. She
said this bill cannot be amended since it is ratifying a
contract for royalty oil between the Alaska Department of
Natural Resources (DNR) and Petro Star. Last year this body
considered and ratified the contract between the Alaska
Department of Natural Resources and Tesoro. She explained that
Alaska's Constitution mandates developing the resources for the
maximum benefit of the people of Alaska and when the state takes
its royalty oil in kind (RIK) it needs to prove that it is
getting more money than it otherwise would taking royalty in the
traditional means of in value (RIV).
JIM SHINE, Commercial Manager, Division of Oil and Gas,
Department of Natural Resources (DNR), Alaska Department of
Natural Resources (DNR), said he would first provide a brief
overview on the contract ratification process and then review
the contract.
ED KING, Special Assistant to the Commissioner, Alaska
Department of Natural Resources (DNR), introduced himself.
MR. SHINE noted that DNR Deputy Commissioner Mark Wiggin was
available on line as well as other members of the Commercial
Team who were involved in the negotiation with Petro Star.
He explained that when the state takes royalty it has a choice
to receive it in kind (RIK) - physical possession - or in value
(RIV). When it elects to receive its royalty in value, the
producers ship, co-mingle, and sell the state's royalty share
with theirs and remits the state's value to it by way of a
check. When the state elects to receive RIK, the state assumes
ownership over the actual oil and the DNR Commissioner disposes
of it through sale procedures described in statute. The state
has regularly sold RIK to in-state refiners dating back to 1979
with Mapco/Williams.
The contract that is before them in SB 30 for ratification has
gone through a thorough public review; the preliminary best
interest finding (BIF) was out for a 30-day public comment and
no comments were received. A revised BIF was presented to the
Royalty Oil and Gas Development Advisory Board (Royalty Board)
in August and it considered the contract, the presentation, and
the BIF and recommended unanimously that the legislature approve
it.
3:35:01 PM
MR. SHINE clarified that there are two contracts, and the one
that is currently in place is a one-year contract commencing on
January 2017 for the sale of royalty oil to Petro Star, which
does not require legislative approval. Following the termination
of that contract starting in January 2018, the four-year
contract in SB 30 would commence. The state will receive a
combined benefit from the two contracts of $29 - $37 million
more than if it had received the same barrels over those five
years in value.
SENATOR COGHILL joined the committee.
CHAIR GIESSEL asked who is on the Royalty Board.
MR. SHINE replied that the Royalty Board consists of eight
members: three commissioners from the Department of Natural
Resources (DNR) who is a non-voting member, the Department of
Revenue (DOR), and the Department of Commerce, Community and
Economic Development (DCCED), as well as five public members:
Bruce Anders (Chair), Dana Pruhs, Kathryn Dodge, Lawrence
Gaffaney, and Steve Selvaggio.
SENATOR WIELECHOWSKI asked if the lower tariffs are factored
into the $29 - 37 million in savings from the increased
production in the pipeline that would result in more revenue to
the state.
MR. SHINE replied the $27 - $37 million figure is spelled out in
the BIF. The real benefit is the difference between the marine
transportation deduction, which is present in the RIV net back
formula versus RIK net back, which is an in-state location
differential, which he would describe later.
SENATOR WIELECHOWSKI asked again if that savings figure takes
into account the lower tariff that would result in slightly
higher taxes to the state.
3:38:12 PM
MR. SHINE answered that both RIK and RIV formulas have a tariff
reduction as part of the net back formula. If the royalty
volumes are coming from Prudhoe Bay, the tariff is from Pump
Station 1 to the Valdez Marine Terminal. Tariffs from any fields
upstream of Prudhoe Bay would include transportation to that
point.
3:38:32 PM
However, before taking RIK the commissioner must find it is in
the state's best interest. The state can dispose of its RIK
through a competitive bid process or a non-competitive,
negotiated sale process. The DNR issued a solicitation of
interest in January 2015 to five refineries - Petro Star,
Tesoro, Flint Hills, BP, and ConocoPhillips - within the state
to determine market interest in purchasing the state's royalty
barrels. The two responses it received were from Tesoro and
Petro Star, and last year the Tesoro negotiated sale contract
was ratified. This year the Petro Star negotiated sale contract
is before them.
Based on the responses, it was determined that there was not
enough competition for a competitive sale, Mr. Shine said, and
while Tesoro agreed immediately to the price terms in the
solicitation, Petro Star was seeking a different pricing
mechanism that is not as advantageous to the state. At that
point the DNR commissioner determined that a competitive bid
sale was not in the state's interest and entered into separate
negotiated sales.
The first contract in effect right now is less than one year in
length to relieve market conditions and allow Petro Star to
secure its feed stock for refineries in Valdez and North Pole in
the near term while also negotiating a long term contract to
provide a secure source of supply for the refineries over the
next four years.
He explained the reason for having both contracts terminate near
the same time in 2021 is because at that point the department
will have a better sense of what royalty volumes are available.
Historically, the state has been able to enter into 10-year RIK
sale contracts, but with declining throughput and uncertainty of
what volumes would be available, it's been determined that five
year contracts are more accurate at this point.
3:41:26 PM
MR. SHINE said AS 38.05.183(e) states that the commissioner must
sell the state's royalty oil to the buyer who offers "maximum
benefits to the citizens of the state," and in making this
determination, the commissioner must consider:
1. The cash value offered,
2. The projected effects of the sale on the economy of the
state,
3. The projected benefits of refining or processing the oil in
state,
4. The ability of the prospective buyer to provide refined
products for distribution and sale in the state with price or
supply benefits to the citizens of the state, and
5. The eight criteria listed in AS 38.06.070(a), as reviewed by
the Royalty Board.
3:42:08 PM
For approval of an RIK sale the DNR must make a Best Interest
Finding (BIF) in support of the sale. In this case, the
preliminary BIF was issued in July 2016 and the final was issued
in September 2016. DNR presented the proposed sale to the
Royalty Board on August 31, 2016, and it unanimously voted in
Resolution 2016-2 that the proposed disposition of ANS royalty
oil to Petro Star meets the requirements of AS 38.06.070. Prior
to finalizing the RIK contract, the legislature must pass a bill
ratifying the contract with Petro Star (HB 70; SB 30).
3:42:58 PM
The Royalty Board's decision criteria was listed on slide 6.
Slide 7 had the actual contract terms:
1-year contract:
from 18,800 bpd to 23,500 bpd for Jan. 2017 -Dec. 2017
4-year contract:
from 16,400 bpd to 20,500 bpd for Jan. 2018 -Dec. 2018
from 13,200 bpd to 16,500 bpd for Jan. 2019 -Dec. 2019
from 10,800 bpd to 13,500 bpd for Jan. 2020 -Dec. 2020
from 8,400 bpd to 10,500 bpd for Jan. 2021 -Dec. 2021
MR. SHINE explained that the range is a minimum nomination per
day and the high number would be the maximum. The numbers
decline over the next five years, whereas the Tesoro contract
last year had a static number of 20,000-25,000 barrels per day.
The Petro Star contract is meant to give them as much royalty
volume as the state can project to not exceed in the next four
years. Typically they don't nominate more than 95 percent of
expected royalty volumes and so the Tesoro contract combined
with the Petro Star contract is about 95 percent of expected
royalty volumes under contract to local refiners within the
state.
The net back formula provides a higher revenue to the state over
RIV that uses a marine transportation deduction, which is
basically the price to ship a barrel of oil from Valdez to its
destination on the West Coast. The in-state location
differential is a deduction in the net back formula meant to
represent the cost of a barrel of oil within the state.
3:45:26 PM
MR. SHINE explained that the flexibility of quantity provides
for a three-month consecutive turnaround clause in which either
refinery may nominate below its minimum range for planned
service interruptions for factory upgrades, de-bottlenecking,
and efficiencies, which is customary in the refining and the
upstream oil and gas industry. The reason the state doesn't want
to keep 5 percent of expected royalty oil is to keep markers on
what the marine transportation deductions are and what the net
back formulas look like to ensure the department is meeting its
statutory mandate to meet or exceed royalty in value when the it
elects to sell the state's royalty in kind. If more royalty is
available than projected, additional volumes will be offered to
both Petro Star and Tesoro on equal terms consistent with the
pricing mechanisms in each contract.
He said that Petro Star has filed a $46 million surety bond with
the state as provided in the one-year contract in the event an
in-state refiner or royalty purchaser defaults (or denominates)
on its obligations to pay the state for royalty already
delivered. Both contracts encourage commercially reasonable
efforts to manufacture refined products within the state and
promote local hire of Alaska residents and contractors. All
communications with Petro Star have indicated that they have no
intent to do anything but refine the products within the state
for local use as jet fuel, home heating fuel, and ultra-low
sulphur diesel, to name a few.
3:48:31 PM
MR. SHINE said the RIK net back formula is: ANS Spot Price -
$1.95 -Tariff Allowance +/-Quality Bank Adjustments -Line Loss.
He explained that the ANS spot price is a monthly average of the
daily average of the two reporting agencies, Platts and Reuters.
The in-state location differential is deducted from that $1.95
as well as the TAPS and upstream tariffs to Pump Station 1. Then
there are Quality Bank adjustments and a small percentage of
line loss. Line loss is a .09 percent, industry-standard
deduction from net back formulas that is meant to represent
small differences in measurement between meters upstream and
downstream as well as any loss in product due to evaporation,
friction, ice build-up, or paraffin as is the case in the
TransAlaska Pipeline System (TAPS).
3:50:00 PM
SENATOR WIELECHOWSKI said if you take 20,000 barrels out of TAPS
he understands that the tariff will go up slightly for everyone
else who put oil in the pipeline. Then that in turn is able to
be deducted from the production taxes that are paid to the State
of Alaska, which ultimately causes a small loss for the state.
He asked if that is factored into this.
3:50:44 PM
ED KING, Special Assistant, Alaska Department of Natural
Resources (DNR) answered that the oil that is being delivered to
Tesoro is presumed to have already been produced and shipped
through TAPS through the Petro Star refinery, so it would
already be calculated into the tariff calculation. If the state
weren't selling oil to Tesoro the assumption is that they would
be purchasing the same volumes of oil from someone else and it
wouldn't have any net effect.
SENATOR WIELECHOWSKI said that didn't answer his question, but
they could discuss it afterwards.
SENATOR COGHILL asked who puts the calculation together and what
value that adjustment means at the end of the line.
MR. SHINE answered that the Quality Bank administrator makes
those determinations, which is meant to measure the difference
in value of the oil streams into the co-mingle point and the
stream of oil coming out in Valdez. So, adjustments are made to
those entities who are contributing a lower quality of oil that
result in a benefit to the upstream producer who has a higher
quality product. Those adjustments can be made as far back as
eight years in this contract when the Quality Bank administrator
or FERC makes a determination.
SENATOR COGHILL said that is interesting, because this contract
has a huge variable.
MR. SHINE said that is the reason for the eight-year tail in the
contract. Even though it is a five-year contract having that
tail move out eight years provides the state with a little bit
more certainty that if adjustments were made that they would
benefit the state, too.
MR. KING noted that the Quality Bank adjustment is also in
existence if they were to take RIV as well as RIK. So this
contract has no actual effect on the value in that regard.
3:53:36 PM
MR. SHINE continued that the contract will yield $29 - $37
million in additional revenue to the state over what it would
have received taking the same barrels in value. The real benefit
between RIV and RIK is really realized in the in-state location
differential and the RIK net back formula as opposed to the
marine transportation deduction, which is present in the RIV net
back formula. The marine transportation deduction for FY17 is
somewhere in the $3.30-$3.40/barrel range and is expected to
move up by about 10 cents per year until 2021 when it will be
about $3.70/barrel. The static $1.95 RIK differential in the
current contract as well as the Tesoro contract is where the
state is getting the most benefit from the sale of royalty over
RIV.
The value realized by the different contracts is:
1-year contract (Jan. -Dec. 2017): from $7.6 to $9.5
million
4-year contract (Jan. 2018 -Dec. 2021): from $22.3 to
$27.9 million
3:55:04 PM
He explained that some of the criteria that the commissioner and
the Royalty Board consider are the impacts to local economies
and local hire; and Petro Star provides benefits to the state in
terms of employment, its in-state refining capabilities, as well
as providing ultra-low sulphur diesel, jet fuel, and asphalt to
local economies. He presented a comparison of the volumes of the
two contracts on slide 10 and said he was available to answer
questions.
CHAIR GIESSEL asked if he had any objections to this contract.
MR. SHINE answered no and no public comments were received; no
adverse comments were presented at the Royalty Board.
3:56:09 PM
DOUG CHAPADOS, President & CEO, Petro Star, Inc., Anchorage,
Alaska, thanked the DNR Division of Oil and Gas for their
efforts and past and present commissioners for the work they had
done in getting this contract in place. These are critically
important contracts for their company, because without oil they
are out of business. With the decline of TAPS throughput they
have found it more and more difficult to source crude oil from
producers on the North Slope. Being a consistent source of crude
in the future, the state has brought Petro Star back and they
look forward to refining this oil and making products for Alaska
consumers.
SENATOR COGHILL commented that the investment Petro Star had
made was immediately beneficial to the North Pole area and he is
very grateful.
3:57:59 PM
SENATOR MEYER asked if it matters to him or the refinery if the
oil is heavy or light, or what the gravity or sulphur rate is.
MR. CHAPADOS said they like to see higher quality crude oil and
Petro Star refineries are designed to process a barrel of crude
in a very simple way. They like to see lots of middle distillate
materials in the crude: kerosene, jet fuel, and diesel fuels.
That allows them to retain more from each barrel they process.
Typically they retain 25-30 percent of each barrel that they
process and the balance is returned back to the pipeline. That
is where this Quality Bank liability is generated, because the
oil they return is considered to be of lower quality than the
balance of the oil that is being shipped through TAPS. Crude
from the Alpine Field is an example of a lighter crude oil that
has a high concentration of jet fuel range material in it.
SENATOR MEYER asked if he gets to choose which oil he gets.
MR. CHAPADOS answered he wished he could, but they are subject
to whatever is being shipped through TAPS, which is a co-mingled
stream from all the fields. Over time the quality of the oil
increases and decreases; it's at a good point now between those
that are coming on line and those that are declining.
4:00:20 PM
SENATOR MEYER said he had heard that pipeline oil is becoming
heavier, so it is encouraging to hear that CD-5 and the Willow
discovery have high gravity rates.
MR. CHAPADOS responded that he believes that the new fields are
of reasonably good quality.
4:01:13 PM
MR. CHAPADOS said Senator Wielechowski questioned whether or not
the sale of this royalty oil to Petro Star would increase the
tariff rates for the remaining barrels that are being shipped
through TAPS, and the very quick and simple answer is no. He
explained that prior to the state royalty oil contracts, Petro
Star was buying oil from another North Slope producer. Those
barrels were being shipped through TAPS just as these barrels
will be. So, at the end of the day, it's really a zero net gain
in terms of how many barrels are being shipped through TAPS and
where it's being shipped to.
4:02:20 PM
CHAIR GIESSEL opened public testimony, and finding none, closed
it.
SENATOR COGHILL moved to report SB 30, labeled 30-GS1873\A, from
committee with individual recommendations and attached fiscal
note(s). There were no objections and it was so ordered.
4:03:05 PM
At ease
^Oil Production Forecast Methodology Overview
Oil Production Forecast Methodology Overview
4:05:32 PM
CHAIR GIESSEL called the meeting back to order and announced an
overview of the oil production forecast methodology. Before them
is the fall forecast and explained that the Revenue Sources Book
is like a paper version of the state's accounts and check book
telling them how much money is coming in to pay for government.
For petroleum the state forecasts the price of oil and the
production each year. These two inputs form a calculation upon
which the budget is based. It is a forecast and some would say a
guess, but it is based on a scientific method, but it's not
infallible. Over the last 10 years this calculation methodology
has changed in an effort to more accurately predict the trends
of the state's oil field activities, and it is in the interest
of the members of this committee and the legislature to
understand this methodology, its terms and the changes made to
it. She invited Mr. Stickel who will explain how it works in the
context of the Revenue Sources Book.
DAN STICKEL, Chief Economist, Economic Research Group, Alaska
Department of Revenue (DOR), said he would talk about how the
production forecast is used, some of the recent history of the
production forecasting process, and what the roles of the
Department of Revenue (DOR) and Department of Natural Resources
(DNR) were in this year's production forecast. He would also
present the forecast overview and then hand it over to DNR for
more of the technical discussion.
4:07:54 PM
MR. STICKEL said the production forecast is one of the key
inputs into the revenue forecast along with oil price and cost
of production - into the production tax and royalty forecast, in
particular, but also to a lesser extent to the property tax and
corporate income tax forecast. This is important because
petroleum revenue provided 72 percent of unrestricted state
revenue in FY16 and is forecast to provide about 70 percent of
unrestricted revenue over the next decade. The production
forecast is also a key source of information the department
provides to policy makers, industry, and the public.
4:09:01 PM
MR. STICKEL said for the last 30 years the DOR has hired an
outside consultant to produce its production forecast, and that
was the case through 2016. However, realizing that DOR forecasts
were being over-optimistic a risk factor was added in the fall
2012 that basically included only a portion of expected
production from new fields. So, a lot of the work the DNR has
done this year built on adding uncertainty to the forecast. In
fall 2016, the over $100,000/year consultant contract had
expired and the department decided not to renew it and to use
in-house expertise instead.
4:10:17 PM
In hopes of providing better information to policy makers, the
department began applying risk factors in 2012 to new oil and
new fields that only incorporated a portion of that expected
production into the revenue forecast. And from fall 2012 to 2015
the actual production came in a lot closer to their forecast.
4:11:27 PM
The chart on slide 7 adds the fall 2016 forecast to the previous
two charts. For the next several years the fall 2016 forecast is
in the range of where the production forecast has been recently
but slightly higher in the out years, which has to do with a
change in the risking methodology (moving from a single risk
factor to evaluating risk on a project-specific basis).
4:12:05 PM
The chart on slide 8 addresses what the roles of DOR and DNR are
in the forecast process. DNR has done everything that the
consultant previously did for the DOR. They review the
production surveys and plans of development that come in from
the companies and create the forecasting model. And they deliver
to DOR a forecast of barrels of oil per day pool-by-pool. One
key difference this year is that that forecast was done
probabilistically, which allows them to look at a range of
possible production levels and outcomes for each field as
opposed to the previous consultant's forecast that just
delivered a one-point estimate for oil production.
The DOR actually publishes the forecast and creates a revenue
forecast out of it. They conduct the surveys and interviews with
the oil companies and find out what their plans are and update
the DOR tax database system that goes back to DNR as one of
their inputs for their forecast. A high level view of the
production forecast is in the fall 2016 Revenue Sources Book.
4:13:21 PM
CHAIR GIESSEL asked if the consultant was one person.
MR. STICKEL answered yes; he was Frank Molly, an engineer out of
Colorado, who had developed proprietary production forecasting
software.
CHAIR GIESSEL asked what kinds of things his software took into
account that differ from DOR's.
4:14:45 PM
MR. STICKEL replied that his forecasting methodology had several
differences that a DNR slide compared side-by-side. He explained
that the first slide was the familiar production mountain chart
showing North Slope peak production in 1988 at about 2 million
barrels a day. Since then with a couple small exceptions, the
first being the start of Alpine in the early 2000s and the
second being last year, production has been on a downward trend.
Last week it averaged about 560,000 barrels per day. Their
forecast anticipates modest 4 percent declines through the next
decade with the majority of oil still coming from the major
fields of Prudhoe Bay, Kuparuk and Coleville River Units.
In putting the production forecast together, Mr. Stickel
explained that the DOR looks at production in four different
categories: currently producing (CP) (including an assumption
for some background drilling work), volumes under development
(UD) (any new fields expected to come on line within the next
year as well as any work in the existing fields above and beyond
the baseline level of development). An example in the current
forecast category is the continued build out of CD-5 in the
Coleville River Unit. Then there is the under evaluation (UE)
category (new fields expected to produce within 2 - 5 years);
some examples are additional developments at Oooguruk and
Kuparuk, Mustang, and Greater Moose's Tooth Unit in the NPR-A).
Anything that doesn't fall into those three categories falls
within the fourth category (oil with an expected start date of
five years or greater or with different types of uncertainty
regarding financing, permitting, economics or resource
definition); some examples are Pikka, Smith Bay, and Willow. The
official forecast is the sum of those categories. However 90
percent is in the currently producing category, a few are under
development (about 5,000 barrels a day in 2018), and about
30,000 barrels a day are under evaluation oil.
4:17:50 PM
In developing the fall 2016 forecast model, Mr. Stickel said DNR
used a "probabilistic analysis," which allows them to examine a
range of possible values. In addition to giving DOR a baseline
official forecast they also give them a high case (P-10), which
means they believe based on the given activity set in the
forecast there is a 10 percent chance that oil production will
come in at that level or higher. They also give them a low-case
(P-90) forecast, which means there's a 90 percent chance that
oil production will come in at that level or higher. These cases
are based on those fields included in the forecast and don't
include the new fields.
4:19:11 PM
Another chart shows what the official, low, and high case look
like for the fall forecast. The official one has oil production
declining to about 331,000 barrels per day by 2026; the P-10 and
P-90 range is plus or minus 40,000 barrels a day.
Slide 15 is the same information in table format and comes from
page 37 of the Revenue Sources Book. In addition to providing
the raw numbers behind that chart, it also provides an estimate
of that production that qualifies for the gross value reduction
(GVR), which is an incentive under the production tax law for
qualifying new oil. He noticed that the GVR-eligible oil goes to
zero by the end of the forecast as it changed from the previous
book, the reason being HB 247 passed in the last legislature and
it implemented phasing-out oil qualifying for that provision.
4:19:58 PM
The chart on slide 16 was a comparison of the fall 2016 forecast
to the previous spring 2016 forecast, the last forecast produced
by the previous consultant, reduces production a little bit for
the next several years. The reasons are reduced drilling in the
company plans of development, as well as reduction in company
spending. Getting into the long-term 2024 and beyond, the
production forecast is actually a little bit higher than the
previous forecast.
4:20:36 PM
On a final note, DOR made a seasonal adjustment to the current
year production forecast from information DNR provided. The
department's revenue models are on a monthly basis. So, they
created a seasonally adjusted forecast by taking the total DNR
production forecasted for FY17 and allocated that out among the
months in the fiscal year based on what seasonality has looked
like for the last several years. When that was done there was no
change to the total numbers in their forecast. In addition, when
they put their revenue forecast together in September and
October they actually had two more months of actual data, which
brought the forecast up a little bit.
CHAIR GIESSEL asked if the consultant would have taken the
Willow and Pikka discoveries into account.
MR. STICKEL replied that they would have left them out, because
the Willow announcement came out after the forecast was
completed. Pikka was left out of the spring forecast because at
the time it didn't meet the level of certainty needed to be
included in the forecast yet. Both will be evaluated for the
next forecast.
4:23:29 PM
SENATOR MEYER said one of the frustrations they must have in
trying to predict future production is the permitting process
that could take many years, especially at the federal level, and
asked if that is why they have a high and a low case scenario.
He gave the ConocoPhillips CD-5 project as an example.
MR. STICKEL said he would let DNR speak to how that uncertainty
is incorporated into the forecast, but permitting has
contributed to some of the over-forecasting in years past with
CD-5 being a great example.
SENATOR MEYER remarked that President Trump has indicated that
he hopes to expedite permits and eliminate some regulations, so
maybe it will get better.
CHAIR GIESSEL thanked them and invited the next presenters to
come forward.
4:25:03 PM
PAUL DECKER, Petroleum Geologist and Manager, Resource
Evaluation Section, Division of Oil and Gas, Alaska Department
of Natural Resources (DNR), said several members of his section
worked alongside Jim Shine and his section to generate the
forecast.
4:25:34 PM
ED KING, Special Assistant to the Commissioner, Alaska
Department of Natural Resources (DNR), said prior to this job he
worked with the commercial group in the Division of Oil and Gas.
Prior to that in 2012/13 he was a petroleum economist with the
DOR; his primary task was to work on the production forecast.
4:26:08 PM
MR. KING said DNR took over the production forecasting
assignment this year and that he kept his DOR relationship very
close. Now production forecasts are done in-house by DNR
independent of all prior forecasts. This is the first time it
has been done exclusively by the department. They didn't take
last year's forecast and make adjustments to it; they actually
took data and generated a new independent forecast, the goal
being to make the most accurate forecast that they could. In
reviewing some of the historical forecasts that had been made
and his experience with the process, he made a couple of
adjustments. They also wanted to make sure that their
methodology is scientifically rigorous as it could be - data
driven, empirical, and defensible. They do not believe the
forecast is conservative and there was no intent to be; it is
realistic and it is based in real data.
4:28:06 PM
MR. KING acknowledged that there is a difference between their
forecast and the last one. There is lower production once the
available information was put in. A lot of the lower production
level is being driven by the producers' reduced plans of
operations (a change in behavior as opposed to methodology).
They also acknowledge a cross-over point that is a by-product of
the change in methodology. Whereas the previous method used
increasing and escalating decline rates to get rid of new
production, he used a probabilistic approach, which spreads the
barrels out instead of removing them from the forecast. So, in
future years one sees less of a penalty on the risking method
that was previously adopted.
CHAIR GIESSEL said in 2016 there were over 500,000 barrels of
production a day and asked why he is starting FY18 so
precipitously below that.
MR. KING answered to begin with, FY18 will begin in this coming
summer, but they do have six months of production for FY17,
which isn't quite on this graph, but on the next one which gets
to her point that current production is actually higher than the
forecast. And while the department is really excited about that,
it doesn't want to be overly optimistic because of the fact that
turnarounds still have to happen in the summertime, maintenance
is going to take production off-line, and the last year or two
have had less investment after the price of oil fell. They do
expect that oil production is going to decrease and that the
flattening will not continue.
He also pointed out that the production forecast is due to the
DOR by October, which means DNR instigated the production
forecasting process in August and the last data they had for it
was from June. The data that was used to develop the 2017
forecast didn't use any data after June of 2016. This was
because data needs to be static in order to do the production
forecast, and if they were continuing to update the forecast, it
would be like starting over every month. June 30, 2016, was the
data cutoff and then the forecast was released. They now have
seven more months of data, and it does look promising.
SENATOR STEDMAN asked what the final date was for actual figures
used in the spring forecast.
MR. KING answered that they are currently in discussions with
the DOR on how the spring forecast is going to look, but
historically the consultant has not instituted another forecast
well-by-well using new data. He replaced the forecast data from
June through March with actual data and that is the process DNR
will probably use this spring. The next fall forecast will take
into account the new data that does exist as well as the new
discoveries that have recently been announced.
He also pointed out that the 10,000-or-so barrels in excess of
the 2017 forecast amount to $20-30 million more to the state.
4:33:10 PM
MR. KING hit the highlights again saying they developed a new
forecast methodology that is an improvement over the change in
methodology that happened in 2012 when the risk concept was
introduced. It uses probabilistic approaches, which honor the
uncertainty, as well, that result in a range of future scenarios
rather than just one. He also pointed out that the 10-year
forecast is about a 4 percent decline and historically the
decline on the North Slope has been about 5.5 percent. It's not
conservative. Rather by employing this probabilistic approach,
the DOR can run different scenarios through its model and
generate different revenue scenarios to get a range of possible
outcomes in terms of revenue rather than just a one-point
estimate.
Finally, the price dependency in the forecast is very important
as an improvement to the process as the distribution around the
forecast can be used to inform DNR's model of how production
will respond to the price environment it is projecting. That has
never been done in the past.
4:35:41 PM
In 2012, Mr. King said, when he was at the DOR, one of his first
tasks in doing production forecasting was looking at the history
of how things had been done in the past. The chart on slide 8
caught his eye as something that needed to be fixed. Said the
production forecasts back in the early 2000s were overly
optimistic as the decline rates are really flat, but year after
year as production forecasts were updated the laws of physics
took precedence over that optimism. That is not the way super
giant oil fields produce; they tend to decline.
4:36:44 PM
Using a shooting analogy on slide 9 he explained the first
column on the left represents all of the forecasts one year into
the future and the "shot group" or the cluster of data is pretty
good, which means there is not a lot of volatility or deviation;
it's also fairly close to target. Moving to the right one sees
years forward or a farther-out target, and when you're shooting
at a farther target it's a lot harder to get consistency. Part
of that is because more variables are involved (price of oil
changes and permitting issues), but it also might be an
opportunity to improve the technique (breathing technique in
shooting, continuing the analogy). Their goal was to both to
make the aim more accurate and to fix their technique.
SENATOR GIESSEL said that was a pretty logical conclusion, but
she still questioned his December forecast that used actual data
up to June 2016 that is pretty far-off already.
MR. KING replied even though it appears to be a fairly sizeable
deviation, in reality it is only about 2 percentage points off,
which is pretty good for forecasting. They used all of the
available data and tried not to use any subjectivity.
MR. DECKER pointed out that their forecast did not attempt to
quantify seasonality effects. The cause of the deviation on
slide 5 is because of the production increases in the winter
months.
4:41:40 PM
CHAIR GIESSEL said she appreciated those variables.
4:42:00 PM
MR. KING explained that a lot of what they are seeing over the
last year and into the beginning of this fiscal year is the
result of investment that happened in 2013/14 in Shark Tooth,
Kuparuk, and CD-5, and Point Thompson came on line. Also,
maintenance issues happened in Prudhoe Bay the prior year that
didn't have to happen this summer, and those fixes actually were
very productive. What is being seen now is the result of
investment that occurred before the price of oil fell by half.
Since then, investment level has not kept on par, and in the
next coming months and years reduced production will be the
result of this lack of investment. It might happen as soon as
this summer. He said their 2017 forecast was based on the full
fiscal year of 2016 without using any actual production months
of fiscal year 2017. Moving forward through the rest of this
fiscal year, the expected annual average of 550,000 barrels
probably won't be met.
4:43:20 PM
CHAIR GIESSEL said the previous methodology included a gathering
of experts and producers who provided them information on
investments being made. She asked how much of that was
confidential and included in this forecast and to what extent
that process was followed this year.
MR. KING replied that DOR was allowed to continue their past
process by which they met with the producers and asked them
specific questions about their investments. They then shared
what they could of that information that wasn't confidential
with DNR. Their forecast used as much publicly available data as
possible including the plans of development provided by the
operators and Alaska Oil and Gas Conservation Commission
(AOGCC)'s data for actual production rates, very similar to the
way DOR's consultant had done in the past (using his own
subjective assessments). After talking to Mr. Molly, the
consultant, DOR would talk to the producers and ask clarifying
questions. That introduced a degree of subjectivity which the
state wanted to remove. So, the operator's information was
definitely considered, but they didn't rely on it.
4:45:20 PM
He said the very glaring reality is that looking into the future
of a development that is meant to come into production five or
more years into the future almost always resulted in some form
of delay. For example, in 1997 Liberty was supposed to come on
line in 2002, but that production still has not occurred, and
that field has entered into and out of the forecast many times
over the last 20 or 30 years. For that reason, because so many
things can change, they elected not to include anything that
wasn't anticipated to come into production within the five-year
window. It is considered imprudent to include that kind of
development into the revenue forecast right now until it's more
certain.
4:46:27 PM
In 2012, the DOR started introducing some risk factors, thus
reducing some production from future developments, and if a
development was forecast to come into production five years from
now at 100,000 barrels, the DNR would hit it with a risk factor
and then only include a portion of those barrels in the
forecast. This was an improvement over prior methods that didn't
use any risking whatsoever.
MR. KING said he uses this technique because it is quick, but it
is not considered the best practice in risk management and the
department wants to move to a Monte Carlo simulation stochastic
approach this year. That method will continue to evolve as they
forecast method gets back-tested.
4:47:42 PM
MR. DECKER continued the presentation (slide 15) and started
with the 2016 method that had three tranches of production:
currently producing, under development, and under evaluation.
The important take-away being that this new forecasting
methodology adjusted the time frame and the criteria for
inclusion or exclusion in the various categories, in particular
in the under development and the under evaluation categories.
4:48:28 PM
He said slide 16 was a table comparing the current methodology
to some of the previous consultants' methodologies. The two most
impactful things are at the top of the chart:
-limiting the inclusion to the first oil window from 10 years to
1 year;
-shortening the under evaluation tranche from 10 years to the 5-
year outlook (anything expected beyond five years out has been
excluded).
There is uncertainty throughout the process and a probabilistic
method was developed so that Monte Carlo simulations based on
the ranges of many of the variables could be used. That is a big
change, because things in the past were deterministic: basically
scenarios or single-point estimates. Oil price dependency was
also incorporated, so as the model is run projects are weighed
against the break-even price versus the DOR's price forecast. If
a project was under water, then it wouldn't be brought into the
forecast.
4:49:57 PM
In general, Mr. Decker said, they have tried to apply
probabilistic risking throughout even to the currently producing
tranche, so the decline-based analysis has a probabilistic range
associated with it rather than a single slope of decline. They
also applied probabilistic pool-by-pool type wells to represent
new production that can be added in as they come in, making sure
that the well's type was appropriate to that pool, and also gave
them a range. Finally, the forecast level uses mostly the
currently producing/decline analysis and that was done on a
pool-by-pool level as opposed to a well-by-well level, another
distinction, although not the most impactful one.
4:51:10 PM
He recapped that the currently producing category constitutes
more than 90 percent of the total forecast. They looked at 34
individual North Slope pools and for Cook Inlet, because those
fields are very mature and only a few produce significant
amounts of oil they were aggregated into a single pool, with the
exception of the Cosmopolitan Field which is still brand new and
has a different development style. So, it has its own
characteristics in the decline curve analysis. This is based on
public Alaska Oil and Gas Conservation Commission (AOGCC) data
that has a two-month lag in availability. Therefore, they chose
the date of cutoff as the end of last fiscal year.
MR. DECKER also pointed out that the decline curve analysis
forecasting, when done at the pool level, inherently includes
and accounts for the background ongoing investments being made
to keep well stock alive and keep projects moving. That actually
shaves some out of the under development category and accounts
for it in the decline of the currently producing category.
He explained that a decline curve analysis looks at historical
data for trends of decline while asking what part best predicts
the future. Because they are making an effort to do everything
possible to be probabilistic, they worked with Schlumberger to
develop a software plug-in for oil field manager software (OFM)
that actually helps quantify various decline rates in terms of
the distribution of possible declines.
4:53:12 PM
An example of a probabilistic decline curve analysis is for the
Tarn Pool (slide 19) in the Kuparuk River Field. If one were to
just look at the historical points, some general trends can be
seen, but things vary greatly from those as well. The software
plug-in for this project helps quantify what would be a
reasonable low-side decline and a high-side decline. This was
done for all the pools.
4:53:49 PM
The time frame for the under development tranche was restricted
to first oil by the end of the current fiscal year, next June
30th. This includes incremental wells added in producing fields
that are in excess of the background level. New fields would be
included that are intended to start up within that time frame,
but there aren't any this year.
MR. DECKER said they also applied a 90 percent chance of
occurrence for each of the under development and under
evaluation wells based on a look-back at plans of development:
if they said they were going to drill 10 wells, typically 9
would be drilled.
The price dependency economic risks are applied to both the
under development and the under evaluation categories. The under
evaluation is for first production expected between June 1,
2017, and the end of the 2021 fiscal year, years 2-5 of the
forecast. Some of the criteria they would apply to make sure the
production is in this category would be having detailed
development plans in place, significant sunk costs or at least
sources of funding in place, maybe inside or outside capital
committed and secure, facilities or facility sharing agreements
developing, and the National Environmental Policy Act (NEPA)
analysis and Environmental Impact Statements (EIS) would be in
progress or completed. The same chances of occurrence in price
dependency are used in the under development category. Examples
in the under evaluation category are Oooguruk Unit, the Nuna
Pool, the Greater Mooses Tooth-1 Development (GMT-1) in NPR-A,
the Mustang, the Kuparuk Field Moraine Development, the 1-H News
in the West Sak at Kuparuk, the Oooguruk Nuiqsut Expansion, and
the Greater Mooses Tooth-2 Development (a separate reservoir in
the Greater Mooses Tooth Unit from GMT-1).
4:56:23 PM
The category excluded from the forecast because it just didn't
meet those criteria were projects that are just a little bit
less defined than those that were squeezed into a five-year
window. Yes, the environmental and permitting challenges are one
of the key variables and that is recognized here. Examples here
would be the Pikka, Ugnu, Placer, Tofkat, the major gas sales
from Point Thomson, Liberty, the Fjord West, and Smith Bay,
Willow, and ANWR - in some cases, just looking at prospects.
4:57:25 PM
The results were on slide 25 in which the North Slope makes up
most of the production mountain. Slide 26 showed statewide
production trends. Basically one could argue from this slide
that if you just look back at the last 10 years and put an
exponential fit to the data, you would get an average decline of
about 5.3 percent and pretty close to what the actual production
was.
4:58:08 PM
MR. DECKER said slide 27 compared that history to the production
forecast and that showed up in a range of dots moving into the
future. The mean decline over 10 years is a 4 percent decline -
the historic decline is around 5.3 percent - so, a little more
optimistic overall than past history since 1988.
4:59:20 PM
He said some "pot of gold" scenarios were looked at: what things
were not well enough known to include them into the forecast
being prepared for revenue generation purposes. A total of eight
projects were looked at including Smith Bay, Pikka, Willow,
Liberty, etc. and excluded from the forecast. If four or five of
the most likely of those projects are brought on, a healthy bump
in production would be seen, but Alaska will never go back up
that production mountain to "the Glory Days of a couple of
decades ago." It's just not likely to happen with the excluded
projects. This will be addressed in a report that will come out
from the Division of Oil and Gas as soon as it can be thoroughly
vetted.
MR. KING added that slide 26 is a classic example of why it's
important not to make overly complex models. In fact, the
efforts that were made by consultants to include everything to
make the model look more like reality actually injured the
ability to forecast and using a simpler approach relying on data
over the last decade would have provided much better results.
5:01:46 PM
CHAIR GIESSEL thanked everyone and finding no further business
to come before the Senate Resources Committee she adjourned the
meeting at 5:01 p.m.