AD Main Menu

Alaska oil production forecast method invites manipulation

Dermot Cole
OPINION: Instead of making oil production forecasts that are almost certain to be "beat," the Alaska Revenue Department should seek to make the most accurate forecasts possible.
Aaron Jansen illustration

FAIRBANKS -- Oil production predictions in Alaska have turned into a game of numbers as Alaska voters prepare to decide the fate of the oil tax cut in August.

While the production forecast for what might happen in five or 10 years remains uncertain because so much is unknown, there is less disagreement on the here and now.

But in the year since passage of Senate Bill 21, now targeted for referendum by Ballot Measure 1, advocates of the tax cut have been trying to prove that an immediate change in production has occurred despite evidence to the contrary.

A year ago, the state predicted that oil production this fiscal year would average 526,600 barrels a day.

But in December, the state dropped that target to 508,200, choosing, as a Department of Revenue official put it, to use a more conservative number that could be “beat” by the oil industry.

By early April, the Department of Revenue released an updated forecast and said that was exactly what had happened, “based on actual production data.”

In its spring revision, the department raised the forecast by 13,600 barrels a day for this fiscal year, leading Gov. Sean Parnell and other supporters of the tax cut to proclaim the change took place because of SB 21 and that it is the start of a trend toward higher production.

“Just a few weeks ago, the state forecast 13,500 additional barrels of oil moving through the pipeline this year because of the new activity on the Slope that has occurred since oil tax reform was enacted,” Rick Rogers, the  executive director of the Resource Development Council, wrote in a commentary in early May.

The gyrations in the forecast deserve special attention not because the numbers have changed, but because of the purported link to SB 21.

The more logical interpretation of what happened is that the state overreacted in December when it lowered the forecast released a year ago for fiscal year 2014, picking a number that could be “beat.”

The easiest prediction about oil production is the short-term prediction, but there was already evidence by November that the reduction published the next month was too steep.

And there is an equally strong argument to be made that the updated forecast the department released this spring -- with the increase of 13,600 barrels per day compared to the December report -- was also too low.

It is probably true that the oil companies are doing whatever they can in advance of the August vote to show higher production numbers than predicted by the Department of Revenue, as that would help the chances of defeating the repeal effort.

They would like nothing better than to advertise in July that North Slope production in the fiscal year ending June 30 surpassed state predictions.

And unless there are major maintenance problems on the North Slope or some other unexpected issue, they will be able to make that statement, with a margin of about 6,000 barrels a day.

But using the state predictions as a dependable gauge is doubtful because the administration said it wanted a prediction that could be surpassed by the oil industry.

In a report prepared for the trans-Alaska pipeline property tax case, Anchorage petroleum engineer  Bill Van Dyke said the state’s 2014 production estimates were “far from accurate” and too low, “possibly out of design.”

He backed up that assertion by quoting Deputy Revenue Commissioner Bruce Tangeman, who told legislators in April: “My personal goal was to put a production forecast line out there and beat it for a change.”

Underestimating production guarantees production numbers that exceed the forecast. This can create an illusion of a change in activity and a trend when the only certainty is that there has been a change in the forecast.

Tangeman testified before the Senate Resources Committee that it was good news that the oil companies exceeded the state forecast.

“Again, we put a forecast out and we beat it, so that’s very, very encouraging,” he said.

In February, during an appearance before the Resource Development Council, Tangeman talked about the goal of issuing a forecast that could be exceeded by production.

“I think we might be actually on our way to doing that,” he said.

With the revision in April, the state is predicting a decline rate of 1.8 percent between fiscal years 2013 and 2014.

 “No matter who you talk to in this building, more production was the goal. So I’m enthused by the 1.8 percent,” Tangeman told lawmakers in April.

“We got there. That’s the whole point is throughput, throughput, throughput and taking advantage of the resource we have. So yes, we’re under 2 percent now, that’s great news for all of us.”

But Van Dyke’s report said it was clear in advance that the forecasts published in December and April were too low.

Oil production usually drops in July and August when major planned maintenance activities take place. It is highest in the winter when the weather is coldest.

Last summer production did drop early in the fiscal year, with many days in July and August with production far below 500,000 barrels per day, followed by increases in the fall.

Van Dyke said that by Nov. 26, 2013, nine days before the December forecast was released, daily oil production had reached the average of 508,200 barrels and continued to climb, according to the normal seasonal pattern.

He said that by Jan. 10, average production had reached 521,800 barrels per day and continued to climb in the weeks and months that followed.

Three months later, the department predicted that fiscal year 2014 production would reach 521,800 barrels, another target that could be “beat.”

Production is expected to continue to drop somewhat in June with warmer weather and maintenance activities, but will likely be several thousand barrels above the daily average predicted in April.

In other words, look for the oil company production level to beat the state estimates again.

Van Dyke said this performance “should not instill confidence in the short-term forecast methodology” and that it raises a question about the accuracy of the long-term forecast, which is subject to far more uncertainty.

Two changes should come out of this.

First, when the state releases its reports predicting oil production in the middle of a year or with three months left in a year, it should include a statement about whether the year-to-date averages are higher or lower than those in the report.

Second, the state should aim for a short-term production forecast that is accurate, not one that can be “beat,” which invites manipulation.