Zhen Zhu, Donald Murry, Mike Knapp* (Zhen.Zhu@chguernsey.com)
Since the beginning of the recessionary period, i.e., December 2007, industrial natural gas consumption generally tracked GDP, reaching a low point in the Spring of 2009. It then recovered, although it still has not reached the pre-recession levels. However, our econometric analysis of the seasonally-adjusted industrial natural gas consumption revealed that a longer-term trend, most likely resulting from more efficient capital equipment and use of industrial natural gas, apparently was more significant influence on industrial consumption than the recession. This indicates that industrial gas demand—all other things equal--will not recover at the same rate that it declined and probably not to pre-recession levels. Moreover, the modeling indicated that the absolute and relative declines in natural gas prices have significantly lessened the recessionary impact on natural gas consumption. This means that the expected and actual future natural gas prices probably will determine the rate and ultimate level of recovery of industrial natural gas demand. If natural gas prices increase with economic recovery, this can only dampen the rate of industrial gas consumption.
Industrial Natural Gas Consumption
As the only component of the overall U.S. natural gas consumption that has declined over the years, industrial natural gas consumption decreased from a share of total natural gas consumption of 34% in 2001 to approximately 28% in recent years. Figure 1 shows this decline in U.S. industrial natural gas consumption. Not surprisingly, the recession triggered a more rapid decrease in the industrial gas demand. The seasonally adjusted monthly gas consumption declined from a total of 586 Bcf in January 2008 to 490 Bcf in May 2009, or a 16.6% decline. The recessionary industrial gas consumption reached a bottom, as did national GDP, in the second quarter of 2009, and then recovered somewhat as did the GDP. In July 2010, the seasonally adjusted total industrial gas demand was at 550 Bcf for that month; this was 9.9% higher than the July 2009 level. Although the observation that industrial gas demand tracked the GDP rather closely is interesting, our econometric disaggregation of this demand reveals influences that are probably more revealing regarding the prospects for the post-recession recovery of industrial natural gas demand.
Figure 1: Deseasonalized U.S. Natural Gas Industrial Consumption
(Original Source: Energy Information Administration, Department of Energy; Deseasonalization done by the authors)
Natural Gas Prices and Industrial Gas Demand
Changing natural gas prices may affect gas demand in two ways. First, in the near-term, changing natural gas prices may induce changes in production levels. This may be especially important in industrial processes where natural gas is a feedstock. Second, in the longer term, prices affect the selection of gas as a fuel or feedstock and may influence the selection of capital equipment. Absolute natural gas price levels will affect both these shorter- and longer-term decisions. Relative natural gas price levels may trigger fuel switching in the shorter- and as well as longer-term commitments, such as selection of capital equipment.
Figure 2 shows the short-term relationship between natural gas prices and industrial gas consumption. Obviously, the gas price spikes were not accompanied by sharp declines in the industrial gas consumption. This suggests that industrial gas demand is relatively resistant to temporary price shocks.
Figure 2: Gas Prices and Industrial Gas Consumption
Longer-term, the relative price of natural gas, may be more important than the absolute level of natural gas prices. Figure 3 shows a large variation in the oil/gas price ratio over the last ten years. Overall these data show a decided upward trend in the oil/gas price ratio over this period.
Figure 3: Oil/Gas Price Ratio
Energy Efficiency, Alternative Fuels and Government Policies
Technical progress and shifting to more efficient capital equipment and government policies and regulations encouraging energy efficiency and alternative fuels did and will reduce the rate of industrial natural gas consumption over time. One probably can expect that technical change and capital equipment change-outs will continue, and may continue to be rather gradual. We, however, think that economic recovery, coupled with low interest rates, potentially could increase the rate of capital equipment change-outs, at least for a period. We believe that the impact of government policies and regulations on industrial demand could be much more abrupt.
We modeled the industrial natural gas consumption over the past 10 years in order to isolate the recessionary effects and to infer the likely influences on the rate of recovery of the post-recession industrial natural gas demand. Specifically, we analyzed the relationship between industrial natural gas demand and an industrial production variable, natural gas prices, oil prices, and energy efficiency.
The industrial production variable was a gas-weighted index that reflected production trends of industries using natural gas more intensively in production processes. That is, the industries comprising the gas-weighted industrial production index include food processing, paper and pulps, chemical, petro products, stone and clays, metal, resin, agricultural chemicals, and manufacturing industries. Because we could not measure energy efficiency and the introduction of more energy-efficient capital equipment directly, we used time trend as a proxy for these gradual changes in industrial natural gas consumption. In order to control for weather influences, we introduced both Heating Degree Days (HDD) and Cooling Degree Days (CDD) variables, but we found that HDD was the only statistically significant weather variable influencing the level of industrial gas consumption.
To model the monthly industrial natural gas consumption, we used a simple linear regression, and we found that only five variables accounted for 98.5% of the variations in industrial natural gas consumption over this ten-year period. We have illustrated the statistically significant coefficients of the respective variables in Figure 4. According to our estimate, every unit change in the industrial production index increases gas demand by 3.41 Bcf per month; however, the trend, or possibly the energy efficiency effect, is, on average, a 2.33 Bcf per month reduction in gas demand. In addition, every $1 gas price increase has a demand destruction effect of 4.70 Bcf per month. Aside from the pure price effect, for every one unit increase in the oil/gas price ratio, fuel switching amounts to 7.29 Bcf/month more industrial demand for gas. To put the fuel switching estimate in perspective, if the oil price increases from $74 to $78 while gas price remains at $4, this would result in an increase of industrial gas demand of 7.29 Bcf per month. Finally, we estimated that heating demand also increased demand for gas from the industrial section in the amount of 0.19 Bcf/month for every degree-day increase in heating need. This weather effect was relatively trivial.
Not surprisingly, our modeling revealed a delay in the effects on industrial gas consumption for both the direct price changes and fuel-switching. Econometrically, for the ten-year period, we identified an average two-month lag between the change in natural gas prices and the impact on industrial natural gas consumption. We identified an average delay of about four months between the change in the oil/gas price ratio and the impact on industrial production. These results are also consistent with the observation that the short-term industrial gas consumption is resistant to sharp natural gas price changes.
Figure 4. Effects of Variables on Industrial Demand (MMBtu/Month)
Industrial Natural Gas Consumption during the Recession
The econometric results permitted us to identify some of the major influences on industrial natural gas demand since the beginning of the recession. For the first seven months of 2008, the average monthly gas consumption was 574 Bcf, and two years later, the monthly consumption had declined to 553 Bcf, or a net decline of 21 Bcf per month. (See Table 1) This was a 3.67% decrease. By our estimation, however, energy efficiency (trend) probably represented a loss of 56 Bcf per month compared to two years ago; the slow economy contributed a 22 Bcf per month loss. The warmer winter weather in the later year contributed another 1.7 per month Bcf reduction compared to two years earlier.
Significantly, during this period, this potential total demand loss of approximately 80 Bcf per month was offset by the effects of the absolute and relative declining gas prices. The absolute gas price decline (the pure price effect) stimulated demand growth by 20 Bcf per month. The fuel switching (relative price effect) had even a larger impact of 40 Bcf. As shown in Figure 5, the net effect of all these changes is a decrease of 20.7 Bcf, approximately the same as the actual decline of 21 Bcf per month between 2008 and 2010.
Table 1. Changing Industrial Gas Demand
Figure 5: Explaining the Net Change in Industrial Consumption between 2008 and 2010
Prospective Near-Term Industrial Gas Demand
Our modeling indicates that non-industrial growth variables may have a large influence on the recovery of industrial natural gas demand following the recession. Figure 6 shows the gas-weighted industrial production index including a one-year forecast. December 2007 marks the beginning of the decline in the index, which hit bottom in March 2009. Industrial gas consumption turned upward at the same time exhibiting a direct relationship between the two. However, just as the industrial recovery has been slow, so has the recovery of industrial gas consumption. In the 17 months during the recession (Dec 2007 through May 2009), seasonally adjusted gas consumption dropped from 586 Bcf to 489 Bcf per month, an average of 5.7 Bcf per month. In the 14 months since the bottom of the recession (from June 2009 to July 2010), seasonally adjusted industrial gas consumption increased from 498 to 550 Bcf, an average 3.7 Bcf month increase. There is a decided asymmetric response of the gas consumption between the declining and recovery economic stages.
Looking forward, the EIA has projected a two percent increase in the industry production index for the period from August 2010 to December 2011. Our modeling indicates that this increase in industrial activity, in and of itself, should lead to approximately a 6.5 Bcf/month increase in natural gas demand, or only half the growth rate of the industrial production index. Compared to July 2010, this is only an approximate one percent increase in the industrial gas demand, even if everything else remains the same. Moreover, according to our modeling, this 6.5 Bcf per month increase resulting from this increase in economic activity will not even be sufficient to offset the cumulative effect of the decline because of energy efficiency, i.e., the longer-term trend that we measured over the ten-year period. Furthermore, if the EIA forecasted growth in industrial activity does not materialize, or if the U.S. economy experiences a double dip, the level of industrial gas demand growth will be slower or, of course, even decline again. For example, if the industrial demand index dips again to the low level set in March 2009, then the 12% decline in economic activity, all other things equal, would result in a 7% decline in gas consumption.
Recognizing that the decline in absolute and relative natural gas prices mitigated the decline in industrial gas consumption during the decline in economic activity, if the economy recovery stimulates an increase in natural gas prices, this will, of course, restrain the growth in industrial gas demand. For example, an absolute increase in natural gas price of $1 will probably stimulate a 4.5 Bcf/month decline in industrial gas demand.
The impact of economic recovery on potential fuel switching in the industrial gas demand is harder to estimate since this will depend on the relative oil/gas price changes. The prices of both are likely to be affected by economic recovery. The relatively high oil/gas price ratio at this time implies that gas prices may increase relative to the oil prices upon recovery. This would inhibit the increase in industrial gas consumption. On the other hand, oil is an international commodity and oil prices are likely to be influenced by many factors beyond U.S. economic recovery. Oil prices may accelerate more rapidly than natural gas prices during an economic recovery. 
Figure 6. Gas Weighted Industrial Production Index
(Source: EIA of Department of Energy)
Industrial Gas Demand Outlook
Although we limited our modeling to the historical relationships of industrial gas demand and these other variables over a ten-year period, we believe that it provides insight into the potential shifts in industrial natural gas demand in the near-term. For example, the analysis demonstrates that a main force for industrial gas demand reduction is the energy efficiency over time, and this will not respond to economic recovery positively. In fact, we think that economic recovery with accompanying low interest rates will stimulate replacement of less efficient capital equipment, and this could, for at least a period, accelerate this trend. The relative weights of the model coefficients suggest that the absolute and relative price increases may offset the impact on industrial gas demand associated with economic recovery.
As to factors that may stimulate industrial gas demand, we believe that a very important influence on near-term is a variable not captured by our historical modeling. That is, we think that the influences of government policies concerning alternative fuels and fuel efficiency are surely underrepresented in our analysis. Therefore, our coefficients underestimate these effects. Given the historical relationship, we believe that these effects are likely to be significant and even could determine the overall direction of the industrial natural gas consumption during economic recovery.
* Don Murry is an economist at C.H. Guernsey and Company and Professor Emeritus at University of Oklahoma. Mike Knapp is an economist at C.H. Guernsey and Company and Adjunct Professor of Economics at University of Central Oklahoma. Zhen Zhu is a consulting economist at C.H. Guernsey and Company and Professor of Economics at University of Central Oklahoma.
 The heat rate parity of about 5.8 between oil and gas apparently is not an important factor influencing the level of total industrial gas consumption. The ratio has been consistently higher than 10 in the last couple of years.
 Aggregating the industrial consumption data from these diverse industries to forecast industrial gas consumption and demand surely masks some industry trends.
 We obtained the industrial gas consumption, gas-weighted industrial production and weather data from EIA. Gas price is the Henry Hub price and the oil price is the spot West Texas Intermediate price. Both these prices have been obtained from Gas Daily.
 U. S. oil prices may be more responsive to worldwide economic activity than are U. S. industrial natural gas prices. If other nations experience economic growth ahead of the U. S., this could increase the relative oil/price ratio. This is likely to stimulate an increase in industrial gas demand.