Biofuels: the Biggest Supply Response to the 2000s Oil Shock

Real annual average oil price according to BP (top graph), and estimates of volumetric production/capacity for various oil alternatives (bottom).  1975-2008/9.  Note that these are total volumetric estimates: about half of tar sands production is sold as bitumen, not synfuel, and the energy density of the biofuel stream is only about 70% or so of that of crude oil.
There are four kinds of liquid fuel alternatives to crude oil in actual commercial production at the present:

• Biofuels - ethanol and biodiesel, primarily from food crops around the world
• Tar Sands - synfuel and bitumen, primarily from Canada
• Gas-To-Liquids (GTL) - from South Africa, Malaysia, and increasingly Qatar
• Coal-To-Liquids (CTL) - primarily from South Africa, but just starting in China

In this piece, I summarize some research I've been doing to look at how each of these sources responded to the oil price increases of 2005-2008.  Two sources, GTL and GTL, haven't shown any particular price sensitivity to date and are at low levels.  Tar sands growth has shown modest price sensitivity but mainly appears to be growing on its own internal dynamics.  Biofuel production growth appears to be extremely oil price sensitive, and increased the fastest and reached the largest volume in response to the mid-to-late 2000s oil shock.  I have argued in the past that there are structural reasons for this: given the comparatively low capital requirements and small plant size of biofuel plants, they can respond much faster to episodes of high oil prices than can the other sources, all of which tend to involve larger, slower-to-build, more capital intensive plants.  This has important implications for food and land prices in future oil price shocks.  Food prices are likely to rise quickly and markedly in response to oil shocks, public policy permitting.

In the rest of the piece, I'll briefly survey the statistics I have for each fuel source, and then draw conclusions about the overall situation.

Coal to Liquids

This is the possibility to use various kinds of chemical transformations to make a petroleum-like liquid fuel from coal. See the Wiki entry on coal liquefaction for more details of the various possibilities. This was done most famously by the Germans during World War II, and has been done for a long time in South Africa; the South Africans needed to get around economic sanctions during the Apartheid era, and that country has a lot of coal and not much oil. Since there are huge amounts of coal underground around the world, CTL is often cited as a potential substitute for oil in future (generally by folks not worried about climate change).

There are at present two plants in the world operating coal liquefaction processes at commercial scale. The first is operated by Sasol in South Africa and has been operating for a long time. The second has just been opened last year by Shenhua in China, and is currently about 1/7 the size of Sasol's operation.

Statistics about coal-to-liquids production are difficult to find, but I did find numbers that get us into the ballpark. In the graph below, the red curve represents actual production from Sasol annual reports, the blue data points represent government statistics for the refinery capacity of Sasol's refinery for the syncrude created from the CTL operation, and the green curve (on the right scale) represents the coal used in the process.

Overall, global CTL production appears to have been roughly flat, until 2009 when a small but not precisely known increment came online in China.

For more methodological details, see Coal to Liquids Production Statistics.

Gas to Liquids

The wiki explains:

Gas to liquids is a refinery process to convert natural gas or other gaseous hydrocarbons into longer-chain hydrocarbons such as gasoline or diesel fuel. Methane-rich gases are converted into liquid fuels either via direct conversion or via syngas as an intermediate, for example using the Fischer Tropsch process.

There are presently three plants in operation around the world: Mossel Bay, South Africa; Bintalu, Malaysia, and Oryx, Qatar. The plant capacity over time is summarized in this next graph:

Note that actual production stats are not available and may well be less than plant capacity. For more methodological details, see Gas to Liquids Production Statistics.

Tar Sands

There are very large amounts of oil that has been degraded to tar mixed with sand up in Alberta, Canada. This can be extracted from the sand with heat (either by surface mining or in-situ). Some of the resulting bitumen is sold directly as tar, while just over half is upgraded to a synthetic liquid fuel (syncrude). Production statistics are available from the Canadian statistics authority as follows:

For more details on how these statistics were obtained and compiled see Tar Sands Production Graph.

Biofuels

Finally, global biofuel production statistics are available in various Worldwatch Institute publications and presentations - see here for the statistics prior to 2000, and here for data from 2000 to 2008. I haven't found global numbers for 2009 yet.

This graph summarizes the global production picture broken down by fuel type:

And this picture looks at how much of the global biofuel total comes from US corn ethanol (according to statistics from the Renewable Fuels Association.

Note that while, in future, there may be production of "second generation" ethanol from cellulosic ethanol from crop wastes or wood, at present almost all biofuel production is either directly from food (eg corn or sugarcane), or is from crops (eg rape) grown on land that would otherwise be used for food. So to a very good approximation, current biofuel production represents the conversion of a fraction of the global food supply to fuel.

Conclusions and Implications

I summarize the volume of production of each of these four kinds of liquid fuels on this next graph (along with oil prices in the top panel).  These levels can be contrasted to total global liquid fuel production of about 86 million barrels/day currently.  Clearly, a fairly small fraction of overall global liquid fuels is coming from anything other than petroleum.

Before we move on, note two caveats about this graph: the tar sands volume below is actually about half tar, not all liquid fuels, and the biofuel energy content is only about 70% of the equivalent volume of petroleum (although, in fairness, ethanol can be burned in engines at higher compression ratios making the engine more fuel efficient).  For simplicity, we ignore all these effects here and just look at total volumes.

Note that by 2008, the biofuels were the largest volume, and also the fastest growing, alternative liquid fuel.  If we corrected for energy density but also excluded the tar sands that is sold directly as bitumen, this would be even more strongly true.

The other obvious features of the data are the price sensitivity: the coal-to-liquids and gas-to-liquids processes are at low levels and have not markedly responded (yet, anyway) to the 2000s oil shock.  The recent increase in the GTL line is due to the onset of the Oryx plant, and planning for that began back in the early 2000s.  GTL production is probably more a product of the desire to exploit stranded natural gas than of high oil prices.

Tar sands production has been increasing fairly steadily for several decades, but is not phenomenally price sensitive.  By contrast, the biofuel production curve can be seen to respond very obviously to oil prices.  It grows sharply in the 1970s and early 1980s when oil prices were high, then only slowly in the late 1980s and 1990s, and then again very sharply in response to the 2000s price shock.

This next graph illustrates the same point by looking at the year-on-prior-year change in production, and plotting it against the real oil price (in $2008) at the time:

The tar sands graph is relatively flat (ie price insensitive) at around the 10% mark.  By contrast, the biofuel graph slopes substantially upward, with zero growth corresponding to oil prices much below $20/barrel, and growth of tens of percent per year when prices are over $60.  The growth rate is volatile, but generally very high.

This is something I argued back in an early 2008 Oil Drum post, Fermenting the Food Supply.  Essentially, converting food to fuel becomes increasingly profitable at high oil prices (and at sufficiently high prices this doesn't require subsidies), and because ethanol plants are relatively small and numerous, it's possible to build more of them very quickly, converting ever more of the food supply to fuel.

In the US, 2009 data is now available for the corn harvest, and combining this with data from the renewable fuels association, I updated one of the graphs from Fermenting the Food Supply.  This shows the total ethanol potential for the entire US corn crop in the background (mid pink), and the capacity of ethanol plants either under construction or in production in the foreground:

The bad news is that we are now in a position to convert over 40% of the US corn crop to ethanol.  The good news is that the process has slowed down, at least for now.  The combination of the 2008 collapse in oil prices, and possibly the beginning of pushback on the ethanol lobby due to the effect on food prices, has caused a reduction in the amount of ethanol capacity buildout.

Still, the ethanol industry is now building from a very high level, and further increases at anything like the growth rates of the mid 2000s would quickly consume most of the remaining corn.

This process is very sensitive to public policy, since the amount of ethanol allowed and/or mandated in gasoline is set by law.  This is currently under active politicking as noted by Ethanol Producer Magazine:

The Iowa Renewable Fuels Association has requested Iowa legislators pass a statewide mandate requiring all motor vehicles to use at least a 10 percent ethanol blend. As the nation’s largest producer of ethanol, with approximately 3.2 billion gallons produced at 39 ethanol plants, the state has never previously issued mandates for ethanol use. According to Monte Shaw, executive director of IRFA, ethanol fuel blends sold in Iowa remain around 75 percent while the rest of the nation stands at 80 percent. “Iowa is lagging behind the rest of the country in ethanol use,” Shaw said at the annual IRFA summit held in January.

The IRFA’s 2010 proposed ethanol mandate is not the first attempt by the organization. In 2006, IRFA unsuccessfully pursued an incentive program designed to boost E10, higher ethanol blends, and biodiesel. Now, according to Shaw, “after four years and no improvement on E10 or biodiesel, it’ time to try again.”

The proposed mandate comes as the ethanol industry awaits a final decision from the U.S. EPA to increase the current allowable blend of ethanol with gasoline from 10 percent to 15 percent. Iowa Gov. Chet Culver believes the E10 mandate “is a step in the right direction.”

It appears to me that the next oil price shock is going to bring these issues to the forefront in a fairly dramatic way.  It appears that the behavior of the overall politico-economic system in the presence of an oil shock is to turn first to biofuels.  However, a biofuel response in future similar to the response to the last oil price shock will cause fairly dramatic increases in food prices.

That's probably not a very good idea.