Demand Pricing and the Shangri-La Diet

Demand pricing (also called dynamic pricing) is adjusting the price according to demand. More demand, higher price. It is being considered for movie tickets:

If a movie is hot, the price could rise to whatever the market will bear. For example, I’d have paid $20 per ticket to see “Avatar” in 3-D when it first opened; maybe others would have been willing to pay even more. As demand becomes clear through lower ticket sales, prices would drop. So “Avatar” might cost $15 a few weeks after opening, gradually making its way to $10.

The theory behind the Shangri-La Diet says that body fat is adjusted in a similar way. When food is abundant, the set point (which controls how much fat you have) goes up (= you store more fat). When food is scarce, the set point goes down. This is how many storage systems work, of course; they increase the amount stored when the price is low and decrease the amount stored when the price is high.

Perhaps one day weight control will be explained to children by telling them it is like the price of airplane tickets: “When more people want to go to Los Angeles, the price goes up. When more food is available, your set point goes up . . . “

24 Replies to “Demand Pricing and the Shangri-La Diet”

  1. Seth
    As you probably know eating a lot of fat will not cause you to accumulate fat
    (That is my N=1 conclusion anyway). So the price of fat is cheap yet the body doesn’t store it. In my opinion accumulating body fat is the body reaction to a perceived shortage. The signal for food shortage is relatively large consumption of carbs which in ancient times happened only when the preferred food of meet (+ fat) was is short supply.

  2. Seth,

    I think the analogy between demand price and food abundance is not valid. Food abundance is the analogue of supply, and increased supply makes price (“setpoint”) go down, not up. Animals like squirrels and bears tend to get fatter and hoard food in preparation for the scarcity of winter, and shed weight when fresh food is more abundant in summer. The better analogue for demand price is energy expenditure, which tends to make set point go down (at least under certain conditions).

    I think SLD is an effective diet, but I think your food scarcity argument is not the correct explanation. I proposed an alternate explanation based on the role that flavor plays in psychological conditioning of the vagus nerve and the hormonal control of appetite:


    1. Todd, don’t confuse anticipatory changes in set point (e.g., bears) with reactive changes. I assure you that abundant food causes the set point to increase — otherwise we would be suffering from an epidemic of thinness.

  3. Seth, what is the critical flaw in Gary Taubes’s reasoning regarding carbohydrates causing a spike in insulin, which in turn leads to fat accumulation? I know that that theory is inconsistent with your evidence regarding sugar water/Shangri-La, but Taubes marshals some fairly compelling evidence of his own. Not sure what to make of it all.

  4. I doubt that a monotonic sugar or any other diet is a good enough simulation of anything. Low carb diets are casing weight loss (Rferences can be provided on request) and I was merely proposing an evolutionary explanation for this fact.

  5. Seth, I’m not convinced that the obesity epidemic can be blamed on the sheer “availability” of food. Many non-Westerners with abundant food don’t have nearly the obesity we have in the U.S. and Westernized society. There must be other factors at play — most likely the unnaturally high levels of factors like fructose and certain fatty acids, which inflame insulin receptors & glucose transporters and disregulate the hypothalamus. In addition, highly flavored, palatable food in combination with calories induces a magnified insulin response, rapid emptying of glucose and fatty acids from the bloodstream, and consequent overeating in the attempt to normalize blood sugars and fatty acids. Robert Lustig has mapped out the science behind this pretty convincingly.

    Your experience with drinking sugar water does not disprove Taubes insulin theory of obesity. The amount of sugar calories you consumed was probably too small to elicit an insulin response strong enough to elicit the kind of blood sugar crash that leads to overeating. Strong, appetizing flavors magnify the preprandial insulin response that ramps up hunger — your flavorless calories tend to suppress this effect. Try consuming much larger amounts of sugar water or carbohydrates and doing it more frequently, and you’ll find at some point the blood sugar swings will cause increased hunger and overeating. I’ve measured my own blood sugar after consuming small carbohydrate meals vs. larger carbohydrate meals and the difference is instructive.


    1. Todd, Taubes’ theory about insulin causing obesity says nothing about a “blood sugar crash that leads to overeating”. You say I consumed amounts of fructose too small to produce overeating. I don’t remember that proviso anywhere in Taubes (that a mere 300 calories of fructose per day is harmless) — and in any case, keep in mind I lost a huge amount of weight. At the beginning I tried huge amounts of fructose water — much larger than I ended up with. They caused profound loss of appetite, not overeating.

  6. Seth, Taubes account of the connection between insulin, appetite and eating is carefully laid out in Chapter 24 of Good Calories, Bad Calories. Particularly noteworthy in this respect are the experiments of Jacques Le Magnen, including rat studies (on pp. 441-2) demonstrating how more palatable or familiar flavors induce a preprandial first-wave insulin response that “shuts down the mobilization of fat…and stores away glucose…As a result, hunger increases…” My point about fructose is different. Fructose doesn’t induce any insulin response, so it is not involved directly in the above mechanism. However, over time, fructose can alter insulin sensitivity by impairing the GLUT4 receptors, and insulin resistance can make lead to overeating as higher levels of insulin (from larger carb/protein meals) are needed before glucose is readily transported into cells. However this is not a short term effect: in the context of a single meal or drink, Taubes theory would predict that, since you could drink a lot of fructose without triggering insulin, appetite would be reduced since there is no insulin to store away calories. The same is not true for drinks containing glucose or sucrose (half of which is glucose), since large amounts of these will spike insulin and deplete blood sugar.

  7. Yes, it’s a well-known fact. Fructose, unlike glucose, does not stimulate insulin secretion from the pancreas:

    However, over a period of weeks, high levels of fructose in the diet can lead to impaired glucose tolerance, insulin resistance and reduced leptin levels:

    So fructose consumption may ultimately lead to higher than normal insulin levels — when insulinogenic carbs or protein are ingested. This is not due to any direct action of fructose, but rather a gradual metabolic change by which the pancreas begins putting out more insulin to overcome the insulin resistance. Yet consumption of fructose by itself — as in your French sodas – will not cause any appreciable insulin response.

    Try repeating your French soda experiment, but add 300 calories of glucose or 600 calories of sucrose. I think you may find a different effect on appetite.

    1. Thank you for your comments, which have helped me learn about fructose metabolism. Leaving that aside, the French sodas I drank were not “fructose by itself”. They were sweetened with either sucrose or high-fructose corn syrup. Many people have found you can lose weight by drinking unflavored sucrose water. Nor does your theory of why SLD works explain why unflavored oils work and why the sugar water must be unflavored.

  8. Seth,

    One more thought. Adding 600 calories of sugar to a soda may or may not stimulate rebound hunger, depending on the level of insulin resistance of the individual. For carb-addicted, insulin resistant individuals, I expect that would be sufficient. But I expect there is some level of sugar — it may be 1000 calories, it may be higher — that will set up enough of a swing in blood sugar to lead to hypoglycemic hunger. And this will happen with or without added flavor; flavor only enhances the effect. If a pure flavor-calorie association effect explained appetite and set point, it should not matter how many calories you ingest. And yet I think the fact that SLD works best with a limited dose of calories supports the idea that a small dose of calories will not provoke an insulin response, whereas a large enough dose will. The amount needed to trigger hunger will vary depending on the insulin sensitivity of the individual.

  9. Seth, I think my theory does offer an explanation for those facts. It is based upon several observations or hypotheses, leading up to conclusions #6 and #7, which I hope address your objections:
    1. Hunger and weight gain are driven largely by having insulin levels above a certain threshold for more time each day than they are below that threshold. That doesn’t mean just “insulin spikes” — it’s the integrated area above vs. below the threshold over the 24-hour cycle. At the same time, raised insulin levels either pre- or post meal will increase hunger by making glucose and fatty acids less available. (Raised insulin during a meal is no problem).
    2. Having insulin resistance (due to obesity) will make it more difficult to lose weight because insulin levels don’t drop below the critical threshold as quickly, since the pancreas oversecretes even basal insulin, and hunger is more easily stoked.
    3. Glucose (and starches or sugars that metabolize to glucose) and to a lesser extent certain proteins, will raise insulin levels. Small and/or infrequent glucose will not lead to weight, but large or frequent doses will.
    4. Fats and oils (in the absence of insulin or insulinogenic carbs or protein) don’t provoke an insulin response, and thus don’t cause hunger or weight gain.
    Of course, there are often “stealth carbs” in many high fat foods, so you have to be careful with this one and use pure oils and fats.
    5. Flavor, particularly familiar flavors, increase or amplify the insulin response, both before and during the meal, due to classical conditioning. (This has been demonstrated in animals and humans and tied to vagus nerve activation by taste sensors in the nose and palate). This serves a biological purpose in reading the digestive system to “recognize” and prepare for a meal. It is a kind of “feed forward” process control! See the work of Karen Teff documenting this.
    6. Unflavored oils induce neither an inherent insulin response (direct stimulation) or a conditioned response (since there is no flavor), so they provide fatty acid energy to the bloodstream with no storage effect, hence they suppress appetite because the fatty acids remain available.
    7. Sugar water will suppress appetite only if it is provided (a) in small doses and (b) without flavors that condition a significant insulin response. But large doses of sugar water will raise insulin. There is probably an optimum dose where a small dose of unflavored sugar adds available glucose to the blood stream but is too low to elicit a large enough insulin response to quickly store it away. But adding familiar flavor to the sugar will provoke a learned pre-prandial insulin response. This can especially be a problem when the flavor is very reinforcing and the sugar dose is low — whereby the enhanced insulin response leads to depletion of blood glucose and fat, causing hunger.
    8. We perhaps need some critical experiment to decide between your theory of SLD and mine. First, my theory predicts that very large doses of unflavored sugar will not suppress appetite as well as smaller doses, and may actually stimulate hunger later on — there is a point where the SLD benefits “rolls over”. I think your theory would not predict this. Second, my theory predicts that larger doses of unflavored oils will continue to suppress appetite, perhaps even better than smaller doses, so long as they contain zero carbs or protein. Furthermore, my theory predicts that even adding some flavor to oil will also work, because the insulin response from the flavor will be minimal combined with zero insulin response from the oil. I wrote about this theory on your forum, calling these flavored non-insulinogenic fat or oil calories “Platinum Calories”. In short, my theory predicts different qualitative behavior for carbs and protein than for fats and oils. By contrast, I believe your theory lumps them together as “calories” and predicts no qualitative difference in their action.

    I hope I did misrepresent your theory; please correct me if I did. But you, or I or anyone can test out my predictions. There may be individual differences, so it would probably be best to have several different people try these experiments. Let me know if you’d like to do a joint test on this. I’m certainly willing to eat my words if I’m wrong.


    1. Todd, in an earlier comment I pointed out that when I started to drink fructose water, I started with a really large dose — much larger than I ended up with. The really large dose, as I said, produced profound appetite suppression and weight loss, not weight gain. This contradicts your first prediction. You might want to read the work by Israel Ramirez I cite for more examples of what your ideas don’t explain.

  10. I’m only injecting a comment here as it seems the thread here is dead, but Seth, you note that “when I started to drink fructose water, I started with a really large dose — much larger than I ended up with.”

    Assuming you meant sugar water (that contained sucrose or HFCS or glucose and not just pure fructose), this does not contradict Todd’s alternative theory insomuch as you had yet to develop a strong pre-prandial insulin response to the sugar water. The question is: had you continued drinking large doses of sugar water, would the profound AS continue or not? If it did, then this would be a problem for the insulin response theory Todd has developed here.

    On the face of it (and admittedly not knowing all the details of your initial experiment), there is no problem with Todd’s alternative theory to SLD here.

  11. Hi Seth,

    I have recently been diagnosed with insulin resistance. Before this i have been on the diet and have been having the oil regularly for over a month. However i have not lost any weight although my appetite has reduced. Does insulin resistance affect results?


    Seth: Sorry, I don’t know.

Comments are closed.