Natural Gas Amateur Hour: The Supply Shortage That Never Was

Many expected soaring natural gas prices by winter. They were wrong. Many blame the failed price rally on a warm November. They are also wrong.

Production is Only Part of Supply

The idea behind the bull gas thesis was sound: less oil production, less associated gas, lower gas supply. Associated gas from tight oil production accounted for 24.6 bcf/ in 2020 (Figure 1). That’s 25% of total U.S. gas output so a decline in tight oil production would have a significant effect on gas supply.

Moreover, gas production was declining before the Covid-19 economic closures. Output had fallen -4 bcf/d before the Covid-19 economic closures. It has now fallen -8.7 bcf/d from the December 2019 peak of 97 bcf/d to 88.3 bcf/d in December (Figure 2).

The same thing happened after the last oil-price collapse. Oil production dropped more than 1 mmb/d after April 2015 and gas production fell -4.5 bcf/d. It is notable that spot gas price increased from its March 2016 of $1.73 low in response to lower supply but it only averaged about $3/mmBtu for the next two years despite low production. Even during cold winter periods, prices only exceeded $4 for one month in late 2018.

Markets knew that this was not a supply crisis but sent a somewhat higher price signal to producers to ensure adequate supply for the next winter heating season.

There’s more to supply than production and markets also knew that ample spare capacity was available from Canada. The U.S. has imported an average of about 9 bcf/d from Canada over the last twenty years. That’s a hefty chunk of supply.

Annual imports from Canada have fallen from 12 bcf/d in 2007 to 6.7 bcf/d in 2020 (Figure 3). When production dropped from 2015 to 2017, Canadian imports increased from about 7 bcf/d in 2014 to 8.4 bcf/d in 2017.

Net gas imports from Canada in 2020 averaged 4.5 bcf/d and have been 6 bcf/d so far in 2021 (Figure 4).

Then there’s storage. If production and imports are supply’s checking account, storage is its savings account. When we have a lot of money in our savings accounts and an unexpected expense comes along, it is unfortunate but not a big deal. When we have little in our savings account, it is a crisis. So it is with gas storage. When storage is high, a cold snap results in higher gas prices but not much higher. When storage is low, however, cold weather can push prices much higher.

Storage was at very high levels throughout 2020 compared to previous years (Figure 5). It was at record levels since early August and finished the year at 3.3 tcf.

Consumption is Only Part of Demand

A key component of the bull gas thesis is that decreasing production and increasing consumption will result in higher gas prices. Disappointing gas prices this winter are incorrectly blamed on warm weather.

The problem is that consumption has been falling—not rising–and it’s not because of weather. Figure 6 shows consumption in red and normalized spot gas price in blue. U.S. consumption was fairly flat from 2000 through 2009 largely because of high gas prices. Price averaged more than $7/mmBtu in the five years before the Financial Collapse in 2008.

Then, shale gas and tight oil associated gas over-supplied the domestic market resulting in a decade of low prices. In 2009, prices fell and averaged about $4 from 2009 through 2011. Since 2011, price has averaged less than $3/mmBtu.

Between 2010 and early 2019, almost 550 coal-fired power units (102 GW of generating capacity) were retired resulting in a steady increase in natural gas consumption. That has largely ended and now, gas-fired plants face increased competition from wind and solar.

The result is that consumption is falling.

Figure 6 also shows degree days—a measure of the energy used for space heating and cooling—in orange. Figure 6 suggests no obvious correlation between weather and either price or consumption. This does not hold on a higher-frequency scale but it suggests that the conventional wisdom about weather and price merits examination. I will discuss this important subject later in the post.

Just as production is only part of supply, consumption is only part of demand. The U.S. now exports a considerable volume of pipeline gas and LNG (liquefied natural gas). It became a net exporter of gas in late 2016, and November 2020 net exports were 11 bcf/d (Figure 7). That means that although November consumption was 82 bcf/d, demand was 93 bcf/d. Exports have added 13% to consumption.

Exports are the gas producers’ solution to their lack of discipline. If there is insufficient domestic demand (consumption), create demand through exports. There’s nothing wrong with that if it works, but it doesn’t benefit domestic consumers or industries that rely on low-cost gas. So far, it looks like a lose-lose since it hasn’t done much for prices.

Natural Gas Amateur Hour

In June 2020, I wrote:

Lower tight oil production should mean lower associated gas production so U.S. natural gas prices should rise, right? That is the conventional wisdom. Too bad it is mostly wishful thinking. Markets are more complex than that. Conventional wisdom is, after all, called conventional for a reason.

That wasn’t a lucky guess. It was based on comparative inventory (C.I.). Those who ignore C.I. do so at their peril.

Comparative inventory is an approach to interpreting how the market prices a commodity. It is based on comparing working supply for a particular period with the same period in previous years. Markets seem to have a kind of collective memory and an expectation for what the components of supply—domestic production, net imports and storage—should be.

Storage or inventory is like supply’s saving account. When inventory/savings is increasing, there’s nothing to worry about. When it’s decreasing to pay expenses that cannot be otherwise met, there may be a problem with liquidity.

The C.I. value is automatically adjusted or normalized for seasonal variation by month or week. The moving 5-year average, therefore, becomes a proxy for market expectation. When C.I. is in deficit and demand exceeds supply, markets typically increase price— or yield—forcing consumption to decrease. Higher price sends a signal to producers to drill more wells. Conversely, when C.I. is in surplus, lower prices stimulate consumption while causing producers to drill fewer wells.

Figure 8 shows comparative inventory from January 2012 through the present. Positive comparative inventory is shown in green and negative C.I. in yellow. Spot gas price is shown in red.

Higher winter gas prices correlate with negative C.I. and lower winter prices with positive C.I. Comparative inventory was positive during this and last winter and gas prices were relatively low. Markets pay attention to year-over-year data and have acknowledged falling supply with higher prices than during the previous winter. Markets also pay attention to inventories and knew that there was more than ample gas in storage to maintain supply even in a very cold winter this year.

The important lesson from comparative inventory is that C.I. is primary and weather is secondary. Weather acts as an accelerant when C.I. is negative. That goes against conventional wisdom about gas markets but that is why it is called conventional.

C.I. has increased +106 bcf (+55%) since early November and is +294 bcf more than the 5-year average (Figure 9). That is not supposed to happen during the winter heating season. Some of that was because of an extraordinarily warm November but December was colder than normal. Increasing C.I. can only mean one thing namely, that consumption is lower than normal for reasons other than weather.

I expect production to decrease in earnest beginning in May or June and it is likely that supply may become much tighter by November. At the same time, Canada has at least another 6 or 7 bcf/d that it would love to sell to the U.S. beyond the 6 bcf/d we have bought in December and January.

The idea that gas price might soar this winter because of lower production reflected a profound misunderstanding of how natural gas markets work. It was natural gas amateur hour.