Turns out prices matter in oil production

[energiewende]

The problem is: what is the impact on the economy when the derivative of the oil price, with respect to time, becomes very large? I submit that it would be arrogant and overconfident to assume that we can continue to model this in isolation.

I don't think it would, because if people anticipate large price changes in the near future they will buy/sell which tends to levelise the price. Current prices are an expression of estimated future supply and demand as well as current supply and demand ("speculation", if the price is personally inconvenient to you).

For instance, when food prices rise to the point where many people can't earn enough to eat, something's going to change.

That sort of thing won't happen. The price of food can increase far beyond any reasonable level due to fuel shortages and still remain affordable.

[Simon_Jester]

You have not thought through your objection to EROEI < 1.

If the oil is going to be used in a car for example it only needs to beat the EROEI of whatever alternative fuel is available. For a concrete example consider that a current electric car has an EROEI of 0.3 at best, as batteries are that inefficient, yet they are not automatically worthless.

Even then, if the EROEI is still worse, if you need it to be used for some niche application where there is just one characteristic that it wins on (eg energy density) that makes it vastly preferable in certain applications (flight) then there is still a good reason to continue production/extraction.

I worded that poorly; that kind of thing was supposed to be covered under "some other purpose than 'get energy out of it.'"

For aviation, petrochemical fuel isn't just being used to "get energy," it's being used to get energy from a high-density medium. If the EROEI on oil goes below one, that might still matter. But you'd never consider burning that same oil in a boiler to heat your house; it'd be more efficient to burn coal or gas, or to use electricity.

The problem is: what is the impact on the economy when the derivative of the oil price, with respect to time, becomes very large? I submit that it would be arrogant and overconfident to assume that we can continue to model this in isolation.

I don't think it would, because if people anticipate large price changes in the near future they will buy/sell which tends to levelise the price. Current prices are an expression of estimated future supply and demand as well as current supply and demand ("speculation", if the price is personally inconvenient to you).

This is what in physics I'd call a 'damping effect,' which acts to neutralize rapid changes in the quantity we're looking at. The question is, are there limits on what the damping can do? Can the speculator market in oil futures keep prices changing slowly and smoothly in a predictable "oil is becoming rapidly harder to get and harder to get in quantity" scenario?

For instance, when food prices rise to the point where many people can't earn enough to eat, something's going to change.

That sort of thing won't happen. The price of food can increase far beyond any reasonable level due to fuel shortages and still remain affordable.

You are completely missing my point. Check your reading comprehension.

Food shortages are an example of how a really large price shift in an essential commodity can cause nonlinear effects. Disastrous effects that are NOT covered under standard supply/demand simulations. But they happen anyway, because at some point the practical consequences of "commodity not available to X% of the population" causes something to happen that breaks the rules of a simulation.

Oil shortages may or may not cause food shortages. It doesn't matter, that's not the point. The point is that oil shortages might cause something with unexpected secondary effects, or trigger some unexpected positive feedback loop. Or cause some kind of equivalent of the permanent deformation of metal you see in an overstressed spring.

There are a LOT of candidates for how that might happen. And it is that which creates the fears about peak oil. That at some point, the supply will fail to stretch enough to allow slow, gentle changes in price. Instead we get rapid changes in price (much as food changes rapidly in price when shortages occur and can't be fixed in a hurry). And the rapid changes in price could have other effects besides "demand decreases haha."

A rapid rise in price would have no serious side-effects if we were talking about the price of chocolate or iPods; it certainly has such effects if we talk about food. Is oil more like the luxuries, or more like the necessities?

______________________

One of my greatest complaints about economics as it is often presented by the laissez-faire is that they try to make economics out to be as precise a science as physics. Which isn't a bad thing, except that a physicist is willing to admit that there are limiting cases where his model of the situation breaks down. Real economists are or should be; the kind we get on political forums talking about FREEDOMARKETS so often aren't.

[energiewende]

But not a probable example. Can you give an example of this ever happening, other than perhaps during warfare?

Peak oil frightens people because they are ignorant of economics. They see things in vague qualitiative terms: we use a lot of oil right now, we are consuming it faster than it is being produced, therefore society is one day going to implode. They don't see a gradual increase in price of oil and decrease in price of alternatives leading first to reduced consumption of oil, and secondly to alternatives gradually supplanting oil.

---

Markets sometimes break down, the problem is that they work far better than people who aren't trained in economics think they do. Worse, history shows they work much better even than people who are trained in economics think they should, see the disastrous socialist experiments of the 20th century. There should be a huge presumption in favour of markets in all areas. In some of these, huge, overwhelming evidence may justify something else instead. But only if it defeats the presumption first.

It's like reading the evolution vs creationism site hosted on here. You see all these thousands of little nitpicks, most or all of which are wrong, or at least don't mean what the authors think they do. Often a reader will not have an immediate answer to all of these criticisms, nor necessarily understand precisely why they are irrelevant. But does that mean they should immediately cast aside their belief in evolution as the most credible theory of the development of life? No, because given the evidence so far, the presumption in favour of it should be extremely strong.

[montypython]

Markets sometimes break down, the problem is that they work far better than people who aren't trained in economics think they do. Worse, history shows they work much better even than people who are trained in economics think they should, see the disastrous socialist experiments of the 20th century. There should be a huge presumption in favour of markets in all areas. In some of these, huge, overwhelming evidence may justify something else instead. But only if it defeats the presumption first.

It's like reading the evolution vs creationism site hosted on here. You see all these thousands of little nitpicks, most or all of which are wrong, or at least don't mean what the authors think they do. Often a reader will not have an immediate answer to all of these criticisms, nor necessarily understand precisely why they are irrelevant. But does that mean they should immediately cast aside their belief in evolution as the most credible theory of the development of life? No, because given the evidence so far, the presumption in favour of it should be extremely strong.

I'd say that this type of argument should make people view markets more critically because since markets are dependent on the mass of human behaviours for their function (unlike the non-conscious actions behind evolutionary trends), conflating the two as the same type of general presumption would be erroneous. The proper scientific study of economic interactions is long overdue for that matter.

[Surlethe]

The proper scientific study of economic interactions is long overdue for that matter.

what do you think economics is?

[Surlethe]

This is what in physics I'd call a 'damping effect,' which acts to neutralize rapid changes in the quantity we're looking at. The question is, are there limits on what the damping can do? Can the speculator market in oil futures keep prices changing slowly and smoothly in a predictable "oil is becoming rapidly harder to get and harder to get in quantity" scenario?

The key point that I was trying to make, at least, is that when prices are introduced into the model, there are multiple damping effects that all act to smooth oil consumption over time and soften the landing. You're correct that there may be second- or third-order effects that point toward chaos, but it's not so clear that they would be strong enough to dominate price-induced damping effects.

[K. A. Pital]

The proper scientific study of economic interactions is long overdue for that matter.

what do you think economics is?

Experimental economics is a very new area, actually. Before that economics often presented non-theories, that is, "research programs" and theoretical constructs that are not and cannot be considered scientific since they are not falsifiable.

[Surlethe]

I wouldn't draw a distinction between experimental science and other types of science; astronomy and paleontology are no less scientific because they are not experimental. I agree that some sorts of economics, like abstruse general equilibrium modeling, are probably unfalsifiable, but it's certainly not correct to apply that to the entire body of pre-1980s economic theory.

[K. A. Pital]

I wouldn't draw a distinction between experimental science and other types of science; astronomy and paleontology are no less scientific because they are not experimental. I agree that some sorts of economics, like abstruse general equilibrium modeling, are probably unfalsifiable, but it's certainly not correct to apply that to the entire body of pre-1980s economic theory.

Paleontology is strictly the observation of past events; for that we have history, and economics is not reducible to history precisely because it has an experimental component and its theories claim to have predictive power.

Applied general equilibrium, faulty as it is, is actually a scientific research program (although some argue it is degenerative, or even the entire orthodox-neoclassical field is degenerative), since it is falsifiable, and not only in parts, but in the program core as well.

The scientific method can be applied to past and present and future events; sometimes you try to explain past events, in this case the theory should have explanatory power. If you also seek to explain current and future events, the theory must necessarily have predictive power. Both research programs would be falsifiable; an explanation of past events in paleontology might be corrected or overthrown by a discovery (and the less this is possible, the less scientific a claim about the past is).

[mr friendly guy]

The proper scientific study of economic interactions is long overdue for that matter.

what do you think economics is?

Ok, maybe I am just ignorant, but do economists use principles of empiricism in their theories? Or have falsifiable predictions? For the second example I know Michael Petis predictions of China doom are unfalsifiable which automatically raises red flags for me. For example if Beijing doesn't experience the doom he predicts, or they refuse to do what he says they should do but shit doesn't happen, he will just hand wave it away as Beijing did some funny stimulus to delay the inevitable, so he is still correct when he says China shouldn't do this or that.

For the first part, if observations don't match your predictions, your theory is simply wrong. I have a suspicion economists don't use this empirical method this given the various opposing arguments on economic management, from opposing sides. One side (or both) must be wrong, and it should be relatively easy to see which observation closely matches which theory's prediction. Yet if both sides continue to argue, therefore one side (or both) must be ignoring observation in favour of theory.

Edit - based on my limited knowledge, the only economist I know of who actually proposes empiricism is Australian Steven Keen, who (claims) he also has a physics degree so he understands the principles of empiricism. Keen predicts economic doom by the way.

Edit 2 - if economists don't use principles of empiricism or falsifiability, then its hard to call their discipline "scientific" no?

[Steel]

You have not thought through your objection to EROEI < 1.

If the oil is going to be used in a car for example it only needs to beat the EROEI of whatever alternative fuel is available. For a concrete example consider that a current electric car has an EROEI of 0.3 at best, as batteries are that inefficient, yet they are not automatically worthless.

Even then, if the EROEI is still worse, if you need it to be used for some niche application where there is just one characteristic that it wins on (eg energy density) that makes it vastly preferable in certain applications (flight) then there is still a good reason to continue production/extraction.

I worded that poorly; that kind of thing was supposed to be covered under "some other purpose than 'get energy out of it.'"

For aviation, petrochemical fuel isn't just being used to "get energy," it's being used to get energy from a high-density medium. If the EROEI on oil goes below one, that might still matter. But you'd never consider burning that same oil in a boiler to heat your house; it'd be more efficient to burn coal or gas, or to use electricity.

I still disagree. A material with an EROEI of less that one can be used for ANY purpose with the sole exception as an input to produce more of that material.

[Simon_Jester]

Re: Steel:
As long as the other purpose is one where the overriding value of the material makes it worthwhile, yes; otherwise, no.

Since we're all SF fans here, the ultimate example would be... yeah. Making solar power from a Dyson sphere, to run ludicrously inefficient factories that make antimatter, as fuel for a spacecraft. The EROEI on the antimatter is terrible, probably some number with a "micro" or "nano" or "pico" hanging off of it. But if there is truly no better way, then such is life.

Thing is, in that case you're paying the staggering penalty of EROEI for something other than "energy" as such. You're paying for portable energy.

[Steel]

Yes, however I feel this is a nitpick which is fundamentally not helpful. It turns out that as efficiency is always less than one doing anything with any resource is less efficient than using it in its most raw form, and so there is no process whereby your objection provides any information as to what the best way to perform a given task actually is.

[Surlethe]

Stas, I'll get to your post sometime tomorrow probably.

***

The proper scientific study of economic interactions is long overdue for that matter.

what do you think economics is?

Ok, maybe I am just ignorant, but do economists use principles of empiricism in their theories? Or have falsifiable predictions? For the second example I know Michael Petis predictions of China doom are unfalsifiable which automatically raises red flags for me. For example if Beijing doesn't experience the doom he predicts, or they refuse to do what he says they should do but shit doesn't happen, he will just hand wave it away as Beijing did some funny stimulus to delay the inevitable, so he is still correct when he says China shouldn't do this or that.

... (below) ...

Edit - based on my limited knowledge, the only economist I know of who actually proposes empiricism is Australian Steven Keen, who (claims) he also has a physics degree so he understands the principles of empiricism. Keen predicts economic doom by the way.

Edit 2 - if economists don't use principles of empiricism or falsifiability, then its hard to call their discipline "scientific" no?

Claims of impending doom are often (usually?) unfalsifiable, just like claims of bubbles about to burst. Nonetheless, economics does have a tradition of empiricism; this was emphasized in the mid-20th century by Samuelson. Friedman wrote a good exposition of the principles of positive economics (he's basically reiterating Popperian falsifiability).

For the first part, if observations don't match your predictions, your theory is simply wrong. I have a suspicion economists don't use this empirical method this given the various opposing arguments on economic management, from opposing sides. One side (or both) must be wrong, and it should be relatively easy to see which observation closely matches which theory's prediction. Yet if both sides continue to argue, therefore one side (or both) must be ignoring observation in favour of theory.

You're not quite right here. The question is not whether observations match your predictions, but how well do observations match your predictions in the realm where your predictions are supposed to apply. Newtonian mechanics is correct, up to high tolerance, to situations where v << c and d >> hc. In the same way, we should view economics as a collection of overlapping models which are empirically testable (and tested) in their realms of application. Some are not always and everywhere correct, they're just approximations, but that's okay because to be correct, a theory just has to work to within reasonable tolerances.

Another point to bear in mind is that economics is way harder than physics or other hard sciences, because it's so hard to conduct controlled experiments and because the use of uncontrolled agents makes discriminatory predictions difficult. Example: Milton Friedman's thermostat.

[K. A. Pital]

Another point to bear in mind is that economics is way harder than physics or other hard sciences, because it's so hard to conduct controlled experiments and because the use of uncontrolled agents makes discriminatory predictions difficult.

Correlation does not equal causation (contrary to what the guy in the article says, most researchers are aware of this fact precisely in this form; there's no need to say that this is a Milton thermostat or something). However, economics is universally falling into this correlation-causation trap (worse yet for the field, the ones who aren't falling into the trap are useless people, since they don't even strive to explain and thus voluntarily seclude themselves from the general process of knowledge acquisition). But there are agent-based simulations and a vast variety of game experiments that try to perfect our understanding of economics, and these things are becoming more popular and more readily availaible with the rise of computing power.

[Surlethe]

Paleontology is strictly the observation of past events; for that we have history, and economics is not reducible to history precisely because it has an experimental component and its theories claim to have predictive power.

Maybe we are using "predictive" in a different sense, but I would argue that paleontology and astronomy and other purely observational sciences are also predictive sciences: theories must predict data.

Applied general equilibrium, faulty as it is, is actually a scientific research program (although some argue it is degenerative, or even the entire orthodox-neoclassical field is degenerative), since it is falsifiable, and not only in parts, but in the program core as well.

Agreed. The key is falsifiable by data, whether predicted or observed in the past. Now, I am somewhat concerned that because often economic data are noisy (high error), a sophisticated theory may be in principle falsifiable but in practice unfalsifiable. What I mean is, if the chain of operations to produce prediction from input is long and the input data are noisy, the error in the prediction will be so wide as to include all observations.

The scientific method can be applied to past and present and future events; sometimes you try to explain past events, in this case the theory should have explanatory power. If you also seek to explain current and future events, the theory must necessarily have predictive power. Both research programs would be falsifiable; an explanation of past events in paleontology might be corrected or overthrown by a discovery (and the less this is possible, the less scientific a claim about the past is).

I wouldn't distinguish between past, present, and future events. Rather, a theory must predict data, whether those data are observed past events or controlled observations. The only way in which time enters this scheme is as new data are added to test the theory. So I regard paleontology as much a science as physics; just, paleontology is much harder, because the data are noisier and it is hard to add new data or check ceteris paribus hypotheses.

Correlation does not equal causation (contrary to what the guy in the article says, most researchers are aware of this fact precisely in this form; there's no need to say that this is a Milton thermostat or something). However, economics is universally falling into this correlation-causation trap (worse yet for the field, the ones who aren't falling into the trap are useless people, since they don't even strive to explain and thus voluntarily seclude themselves from the general process of knowledge acquisition). But there are agent-based simulations and a vast variety of game experiments that try to perfect our understanding of economics, and these things are becoming more popular and more readily availaible with the rise of computing power.

The point is not that researchers have to distinguish between correlation and causation, but rather that if a system is operated by an agent, it is an empirical difficulty to distinguish between the operation of the system and the inclination of the agent. In the example he gives, the speed of a car is ceteris paribus proportional to the depression of the accelerator, but it is empirically difficult to deduce this because the relationship between accelerator and speed is correlated to the operator's forecast error. Everybody's aware of the basic correlation-causation problem (unless they're in a political debate and it suits them to not be aware of it), but I'm inclined to trust Rowe that not so many researchers are aware of this agency problem. I certainly hadn't heard of it when I studied econometrics.

[Thanas]

The trick most economists use is the homo oeconomicus - a human being who is characterized as being free of prejudices and likes. In short, a robot. This is then used to falsify economic models.


And then there is a huge surprise when business fails despite the model being sound because human beings are not robots.

[Admiral Valdemar]

There is also the problem of economists taking natural resources for granted. Most classical schools will assume that energy is abound, and if it gets too expensive, then an alternative will be found. There is no alternative to fossil fuels, so when they run out, nay, when they get to the point of diminishing returns of a certain size, then something else will have to be used that will not be as ideal.

There are people who take the Club of Rome route and factor in energy, sinks, and resource limits. They are not, sadly, the mainstream types, since we are still abusing our resources, we are still polluting in the extreme and we still rely on fossil fuels.

The fact that we're using fracking now is testament to how desperate the fossil fuel industry is. Take a look at Bakken's rig count and depletion rates, then seriously say we're in some new second renaissance of oil. We're hearing the same thing over here regarding the "Dallas of the north-west" that is the Lancashire shale deposits.

[energiewende]

The criticisms of economics in this thread are not very fair. Homo economicus is just a mathematical assumption to make the equations tractable, and not a wholly unrealistic one by any means. It is just the same as in physics first solving a problem in a vacuum or with convenient symmetries, etc. This is done because it is easier and, while one cannot learn everything from these, nor can one learn nothing. Economics isn't high school physics, but remember the physics taught in high schools is carefully selected for a reason. Applying those simple equations to realistic situations quickly leads into messy numerical approximations, and mechanics and E&M are simpler and easier to measure systems than the economy.

There is also the problem of economists taking natural resources for granted. Most classical schools will assume that energy is abound, and if it gets too expensive, then an alternative will be found. There is no alternative to fossil fuels, so when they run out, nay, when they get to the point of diminishing returns of a certain size, then something else will have to be used that will not be as ideal.

There are people who take the Club of Rome route and factor in energy, sinks, and resource limits. They are not, sadly, the mainstream types, since we are still abusing our resources, we are still polluting in the extreme and we still rely on fossil fuels.

Synthetic fuels are certainly possible, and the Club of Rome's predictions were mostly wrong (not that it caused their gaggle of politicians, history professors, and celebrity hangers-on to consider changing their minds). Accounting for pollution externalities is as mainstream as comparative advantage.

[Admiral Valdemar]

There are people who take the Club of Rome route and factor in energy, sinks, and resource limits. They are not, sadly, the mainstream types, since we are still abusing our resources, we are still polluting in the extreme and we still rely on fossil fuels.

Synthetic fuels are certainly possible, and the Club of Rome's predictions were mostly wrong (not that it caused their gaggle of politicians, history professors, and celebrity hangers-on to consider changing their minds). Accounting for pollution externalities is as mainstream as comparative advantage.[/quote]

They're not predictions, they're part of what's called scenario planning. They were a series of scenarios looking at what would happen given a set number of variables being tweaked e.g. efficiency gains from energy usage. There have been various updates to these forecasts over the revised books, of which I have one, given the model assumptions and the technology in the model, have changed over the decades. They're only wrong in the same sense Malthus was "wrong".

In any case, I don't see energy shortages as being the problem. It's climate change, which lost the limelight when oil prices started spiking and the likes of shale and oil sands came to the fray in a desperate effort to prop up oil production (at the expense of our environment).

Also, pollution externalities being accounted for doesn't explain why CO2 emissions are not only still rising, they're accelerating. So this is patently false for the most important pollutant globally (ignoring the crap that still gets flushed into the rivers and then oceans).

[energiewende]

They were economic models of the sort everyone else was just criticising. Models that produced wrong outcomes. Incidentally, none of the founders of the Club of Rome nor authors of Limits to Growth had degrees in economics or worked as economists.

Malthus was wrong. Horribly, horribly wrong.

No one in the economics profession disagrees with you about pollution externality (though some disagree that it's especially important including, though he didn't say it like this in the Executive Summary, Howard Stern).

[Admiral Valdemar]

They were economic models of the sort everyone else was just criticising. Models that produced wrong outcomes. Incidentally, none of the founders of the Club of Rome nor authors of Limits to Growth had degrees in economics or worked as economists.

Actually, the business-as-usual model trends closely to data we have of the last thirty years. There's a CSIRO study that looks into this.

I also don't know what the authors credentials, or lack thereof, in economics really has any relevance to something that is biophysical.

Malthus was wrong. Horribly, horribly wrong.

No one in the economics profession disagrees with you about pollution externality (though some disagree that it's especially important including, though he didn't say it like this in the Executive Summary, Howard Stern).

Malthus was only wrong in the time he presented because of the completely unforeseen discovery of fossil fuels and their heralding of the industrial revolution in full swing. The planet could not have kept growing or feeding such numbers of people without the bounty we dug up in the form of coal, oil and gas. In that respect, Malthus was wrong. But the idea that overshoot occurs and resources should be closely monitored is what he brought to the table.

[Thanas]

Malthus was wrong. Horribly, horribly wrong.

Why?

The criticisms of economics in this thread are not very fair. Homo economicus is just a mathematical assumption to make the equations tractable, and not a wholly unrealistic one by any means. It is just the same as in physics first solving a problem in a vacuum or with convenient symmetries, etc. This is done because it is easier and, while one cannot learn everything from these, nor can one learn nothing.

Sure but my point is it is not a science if it deals with very little proof.

[mr friendly guy]

Claims of impending doom are often (usually?) unfalsifiable, just like claims of bubbles about to burst. Nonetheless, economics does have a tradition of empiricism; this was emphasized in the mid-20th century by Samuelson. Friedman wrote a good exposition of the principles of positive economics (he's basically reiterating Popperian falsifiability).[/url]

So what you are saying is that economics use empiricism and some falsifiability when applying to its models, but a lot of economists break these rules straight away by making unfalsifiable predictions. So the problem isn't with the discipline, its with its practitioners. I am not trying to be smarty pants, just trying to understand what you are saying.

You're not quite right here. The question is not whether observations match your predictions, but how well do observations match your predictions in the realm where your predictions are supposed to apply. Newtonian mechanics is correct, up to high tolerance, to situations where v << c and d >> hc. In the same way, we should view economics as a collection of overlapping models which are empirically testable (and tested) in their realms of application. Some are not always and everywhere correct, they're just approximations, but that's okay because to be correct, a theory just has to work to within reasonable tolerances.

Another point to bear in mind is that economics is way harder than physics or other hard sciences, because it's so hard to conduct controlled experiments and because the use of uncontrolled agents makes discriminatory predictions difficult. Example: Milton Friedman's thermostat.

So shouldn't it be obvious which situation is the one where your model is supposed to apply, and then not even bother applying the model. That doesn't seem to be the case when I hear economists argue, because both sides wouldn't argue if for example one side's model isn't supposed to apply in that particular situation.

[K. A. Pital]

Maybe we are using "predictive" in a different sense, but I would argue that paleontology and astronomy and other purely observational sciences are also predictive sciences: theories must predict data.

That is a bit different, however. I was using predictive in the sense of predicting the outcome of events, not merely predicting historical data (much less so when the data is actually complete, as is the case for our current economic activity records, for examle). Astronomy can predict the movement of bodies in space and even the very evolution of the galaxies, etc. and the universe. It is not just predicting the collected data, like historical sciences, but can actually predict the evolution of complex systems. Geology is the same.

I wouldn't distinguish between past, present, and future events. Rather, a theory must predict data, whether those data are observed past events or controlled observations. The only way in which time enters this scheme is as new data are added to test the theory. So I regard paleontology as much a science as physics; just, paleontology is much harder, because the data are noisier and it is hard to add new data or check ceteris paribus hypotheses.

Yes, both are science. However, paleoantology and history in general can have unfalsifiable research programs, which are unscientific, precisely due to the low amount or quality of data available. On the other hand, economics is studying the reality of our current activities, which are damn well recorded compared to even a century ago.