In my mid-July article on natural gas prices ("Can Natural Gas Avoid the Third Quarter Slump?"), I was looking for a tough autumn for spot gas, and more specifically, the United States Natural Gas fund (UNG), because of seasonal upticks in production and in contango, as well as the potential for declining demand and increasing weekly injections.
These dynamics had, together, led to peaks and gradual declines in UNG share prices - and natural gas spot prices - in mid-to-late summer over the past several years. However, this year, gas has had quite a strong third quarter, with UNG recovering from an August correction and continuing to trend higher in September and early October (see chart below).
Both natural gas spot and futures are now testing technical resistance areas. Near month natural gas futures are testing the upper band of a descending triangle after rallying off of 10-year lows (see chart below).
UNG is currently trading above its 200-day moving average and the consolidation area around $20/share that I identified in my July article (see chart below). Also, in the chart below, notice that UNG has been trending upwards versus a broad basket of commodities (see the grey, relative performance line between UNG and DBC).
So, is this autumn strength sustainable? Will UNG be able to break through the areas of technical resistance and push higher, driven by shifting secular dynamics?
While my call in the summer was early, I continue to believe that the next move in natural gas prices will be down, with spot prices challenging support around $2 per MMCF while UNG pushes down toward its 2012 lows around $15/share. In this article, I focus on dynamics in production, consumption, storage, and contango in the futures curve to make my argument.
Production Adjusting, But Still High
As I argued in my July article, production of natural gas is seasonal, with gross withdrawals-per-day declining into the summer months and then rebounding in the fall (see chart below).
Production in 2012 has followed this basic path, with gross withdrawals / day steadily trending down through July, the most recent month with data (see chart below).
While the path of production/day in 2012 is far closer to the 1990-99 "pre-fracking" production environment than either the 2000-09 or 2010-11 averages, production levels remain high and have consistently registered small year-over-year gains in the weekly production reports, confounding expectations of large decreases in production to adjust to low prices. Indeed, notwithstanding the shut-in production in late August due to the hurricane season, production has been trending up into the fall, registering week-over-week improvements.
Part of the resilience in production comes from the continued intense drilling for oil in many of the shale plays and the production of significant amounts of gas by oil-directed horizontal rigs. Indeed, according to Baker-Hughes data, while gas directed rigs are currently at a 10-year low, total rigs drilling continue to be near record highs (see chart below).
Also, the mix of rigs has continued to shift strongly toward the more-productive horizontal rigs within both oil and gas (see chart below).
In summary, I believe that production will continue to trend up through the end of the year, consistent with cyclical seasonal patterns, and that secular declines in productions will not come until the record number of total drilling rigs working declines to more normal levels.
Potential for Weak Consumption This Winter, Though Higher than 2011-12 Winter
Consumption data through July showed continuing high electric power demand growth (versus the previous year), normalizing residential and commercial demand and weak industrial demand (see chart below). Indeed, electricity demand in July was 146,000 MMCF above July 2011, while residential and commercial demand were a combined 6,200 MMCF below July 2011 and industrial demand was only 19,000 MMCF above 2011.
While electric power demand is still running above 2011 levels in the weekly reports, we are transitioning out of the peak summer cooling (electricity demand) months and into the winter heating months. Heating demand received a big boost in October from colder-than-average temperatures throughout much of the country (see chart below).
Going forward, the U.S. Energy Information Administration recently forecast that heating demand in the 2012-13 winter would be substantially higher than the 2011-12 winter, but that heating demand would still be somewhat less than average due to NOAA's forecasts of warmer than average temperatures. Indeed, the following charts show temperature projections for the Oct/Nov/Dec and Jan/Feb/Mar periods, both of which predict warmer than average temperatures through much of the country.
Higher than average temperatures, even if they are less pronounced than the 2011-12 winter, could lead to a less forceful increase in heating demand. If this is coupled with further declines in industrial demand, it could lead to a relatively weak winter demand picture.
Near Record Gas-in-Storage; Weekly Injections Resilient
Because of a much warmer than expected winter in 2011-12, the 2012 injection season began with far more gas in storage than previous years, leading to fears of a "gas-mageddon."
Over the year, higher-than-expected electric power demand (driven by coal-to-gas switching in power generation), and relatively stable production caused weekly injections into storage to underperform, reducing the glut of storage down to previous maximum levels (see chart below).
Average weekly injections in October, though, have recovered versus history and are currently running above average levels for 1993-2009, though below 2010-11 averages (see chart below).
Given the production and consumption dynamics I discussed above, I would look for weekly injections to continue to hold up pretty well in the rest of October and November. Indeed, the year-over-year change in weekly injections seems to exhibit an interesting cyclicality; the series is currently near the bottom of its normal range and may be set to turn upwards (see chart below).
Contango Widening Out Later Than Usual
The presence of contango in the futures curve - or the situation where longer-dated futures trade at a premium to shorter-dated futures - has been a major driver of UNG performance versus spot. Contango leads to a substantial negative "roll yield" for UNG, which owns the near-month futures and rolls them forward each month, leading to significant underperformance of spot over time (see chart below).
The futures curve was in backwardation - the opposite of contango - as recently as the summer, though contango has widened out considerably in recent weeks, consistent with seasonal patterns (see chart below).
Contango has actually widened out later in 2012 than in previous years, leading UNG to hold up better in August and September (see chart below).
It is clear from the production and consumption sections that the natural gas market is adjusting to low prices through dialing back the record production increases of recent years and increasing consumption in areas like electric power generation. However, while factors such as production shut-ins in August and a colder-than-usual October have supported prices this fall, I believe that a correction of recent natural gas strength is in order.
Natural gas spot prices have increased over 74% in the last six months. In the past, sharp 6-month increases have predicted lower returns over the next year, with a full series correlation of -0.26 (see chart below).
Combined with this reversal effect are the factors mentioned above. Production is resilient and will likely trend up through the end of the year. Consumption has the potential to weaken. And, important for UNG, contango has come later this year and may stay wider than average in October and November.
More medium-term tactical investors can position for a potential correction in natural gas prices through the shares of regulated utilities like American Electric Power (AEP) or PPL Corporation (PPL), which currently trades at a TTM P/E ratio of 9.9x, a Price-to-Book ratio of 1.6x and a Price-to-Sales ratio of 1.2x, all well below their 5-year averages, and which is currently showing improving relative strength compared to the S&P 500 (see chart below).
Shorter-term traders could consider going short UNG with an initial target of $20/share, secondary targets of $17/share and $15/share, and a stop loss at $25/share (see chart below).