The U.S. natural gas market is about to do something many experts thought impossible just last year: experience a sharp increase in price. The Organization of Petroleum Exporting Countries (OPEC) has injected considerable volatility into the energy market since many of its members declared that they would voluntarily restrict production in an attempt to rebalance the global market, and along the way, raise crude prices.

The global price of crude has implications for U.S. natural gas production and prices, but the relationship is not well understood. While gas prices exert little impact on the global price of crude, the global price of crude has a significant impact on gas production, thus the U.S. natural gas price.

Depressed crude prices have contributed significantly to a leveling, and more recently, a decline in U.S. gas production. That decline now threatens to become a supply shortage that will lead to a sharp rise in gas prices over the next six months to a year. Ironically, OPEC’s success, while raising oil prices, may place a cap on rising gas prices, or possibly even drive them lower.

The total revenue that a producer receives and that drives return on investment is a function of the price of each contributing commodity. Decisions to drill are predicated on expected revenues from the combination of commodities that the producer projects it will get from each well. In areas such as the Karnes trough in the Eagle Ford Basin, crude volumes are so large relative to the commingled gas volumes that the price of gas is largely irrelevant to the decision to invest capital in the well; rather, it is the price of crude that drives revenues and drilling activity.

In other places, such as the Anadarko Basin, crude production is a lower portion of overall revenues from the well, so the natural gas and/ or natural gas liquids (NGLs) production makes a material difference to the overall return of the well. In these areas, the combination of oil and gas prices is relevant to the investment decision, as depicted below.

In Figure 1, the lines represent the combination of oil and gas prices that provide a 20% minimum acceptable rate of return from the top tercile of wells in various fields within the Permian and Anadarko basins. A steeper slope indicates greater sensitivity to gas prices, as gas comprises a larger proportion of the production stream. The figure illustrates that Permian Basin wells are less sensitive to gas prices than Anadarko Basin wells are, because the relative proportion of gas supply is small relative to oil.

However, because of commingled oil and gas production, it is critical to forecast the combined production stream—oil and gas—at various price combinations in order to accurately depict the economics of supply. In our example, without understanding the gas-oil relationship, an analyst would be likely to underestimate gas production levels if either commodity price changes.

The significance of this point is illustrated by reviewing historical gas production from the past eight years. During the early phase of the shale revolution, drilling activity focused on natural gas wells. That triggered dramatic increases in gas production, from 48 billion cubic feet per day (Bcf/d) to 62 Bcf/d from January 2007 to December 2011, a 28% increase over the period. Not surprisingly, the increase in production resulted in a decline in gas prices to below $2 per million British thermal units (MMBtu).

What was surprising, however, was that despite low gas prices, gas production subsequently increased to a peak of 73.5 Bcf/d in February 2016, as shown in Figure 2.

There are three primary reasons why this occurred. First, there was the advent of the Marcellus Shale, where wells are highly productive and are frequently economic at well below $3/MMBtu gas; second, a significant portion of production was hedged at above-market prices; and third, high NGLs prices, then high oil prices stimulated drilling for oil in the Eagle Ford, Denver-Julesburg (D-J), Permian and Anadarko basins. The result? These areas produced significant amounts of associated gas along with oil.

Let’s consider an extreme example to further demonstrate the dependency of gas production on oil prices. The Permian Basin has some of the lowest break-even production economics in the country. If we assume a $0 gas price and the forward curve for oil for the next five years (ranging from a low of $50 in early 2017 to approximately $56 in 2021, based on the forward curve as of Jan. 20.), there will be sufficient economic wells based on the oil production alone to drive growth in dry gas.

Figure 3 illustrates that even our extreme price example drives 2.2 Bcf/d in growth in Permian dry gas production from 2017-2021. This illustrates perhaps the most significant change in the U.S. natural gas market in the past 20 years: natural gas production is no longer predominantly a function of natural gas prices.

Instead, NGLs can be the principal driver of gas production. The majority of U.S. gas production comes from wells that produce crude or other NGLs, making the decision to drill more dependent on the price of crude than the price of natural gas.

What are the implications of these factors when evaluating the gas market? One needs to dynamically quantify production outcomes based on the interaction of oil and natural gas (and NGLs) prices rather than to simply evaluate production as a function of natural gas prices.

Let’s evaluate how the relationship between crude oil prices and gas supply impacts production in 2017.

Forward curve with base-case gas demand: we are short gas supply

Right now, gas production has been falling everywhere in the U.S. except the Marcellus, where well economics are strong at $3.50/MMBtu gas price, and the Permian, where increased oil production is driving greater production of associated gas volumes. Our models indicate that, under the forward curve as of the end of December 2016, gas production will increase from 71 Bcf/d in 2016 to 72.4 Bcf/d in 2017.

Gas demand, net of Canadian imports, will average 73.9 Bcf/d, thus the market will be approximately 1.5 Bcf/d short.

In this “base case” scenario with assumptions for normal weather, the forward curve fails to clear the market for gas demand. Under this scenario, dry gas inventory will be approximately 3.45 trillion cubic feet (Tcf) as of the end of October 2017, which is well below the five-year average of 3.79 Tcf during 2011-2015. It is not a question of whether gas prices will rise—it is a question of how fast and how much.

Stronger than expected demand: Exacerbates gas storage shortfall

If we anticipate being under-supplied in a base case scenario with normal demand, then clearly any growth in demand will exacerbate the gap. Although gas demand largely depends on weather, there are several additional factors pointing toward an increase in demand.

• Our base case scenario assumes exports to Mexico will average 3.9 Bcf/d as additional capacity comes online, which is an increase of 0.5 Bcf/d compared to 2016. Growth will be limited by Mexican demand, not pipeline capacity, and if Mexico has a warmer than expected summer, the export number could rise above 3.9 Bcf/d.

• LNG exports were approximately 1.5 Bcf/day at the end of 2016, which is modeled in our base case scenario. Exports are expected to increase this summer as Sabine Pass trains 3 and 4 come online, although the realized export volumes will likely ramp over time, similar to the trend observed when the first two trains became operational. Export volumes will also depend on weather and other demand factors in the countries receiving the LNG, as well as the then-current price of U.S. natural gas and global crude prices.

• If the U.S. has a warm summer, power demand will increase by as much as 2 Bcf/d compared to normal weather, and drive up the price of gas. Historically, power generation would switch to coal if gas prices increased above approximately $2.50/MMBtu, but many coal plants have been retired, leaving only approximately 1.5 Bcf/d to 2 Bcf/d of spare coal capacity to bring online. Therefore, the coal-to-gas switching will be less severe than in the past, even if gas prices are higher than $2.50/MMBtu.

OPEC may solve the shortage

OPEC’s actions may prove to mitigate the potential gas supply shortage. If OPEC successfully implements its proposed production cuts, the expected increase in oil prices could generate sufficient associated gas production from oil wells to solve the shortage. Figure 4 illustrates several combinations of oil and gas prices required to produce gas supply sufficient for our base case gas demand.

The forward curve, as of Jan. 13, 2017, for WTI and Henry Hub predicted gas supply (indicated in green) below the average 74 Bcf/d of demand (dotted line) throughout 2017 (net of imports from Canada). If the WTI forward curve materializes, then we need the gas price to rise to $4.25/MMBtu to stimulate production to meet supply by November 2017 (blue bars).

If crude prices increase to $60/bbl, which will only be possible if OPEC’s cuts are successful, this means only a $3.50/MMBtu gas price is required for supply to meet demand.

Accordingly, although OPEC is cutting production to meet oil price targets, those cuts will have an important impact on gas prices as well. U.S. operators concentrated in oil production will benefit if OPEC successfully cuts production and drives up oil prices, while operators concentrated in gas production should be hoping that OPEC will fail.

Editor’s note: This was originally written in January prior to this year’s unusually mild winter which has since weakened demand for natural gas. The author still supports the case there will be a gas supply shortage and prices will increase this year, but to a lesser degree than forecasted in this commentary.

As the vice president of strategic development for Drillinginfo, Tanya Andrien helps identify opportunities to expand into new markets, and she sets and evaluates internal, strategic objectives and metrics to measure DI’s success. Before joining DI, she was a lecturer at the McCombs School of Business at the University of Texas at Austin, and the associate director of the McCombs Energy Center.