A Critique of Mises's Praxeology (Part 1?)

Economics Economics Questions
81

Because that isn’t the choice they are offered. They are offered a chance to get paid without work vs. a choice to suffer in the short term for a long term healthier economy.

I seriously doubt that people don’t vote for Ron Paul because of time preference. People don’t vote for him because they think his ideas are wrong. Many people support environmental regulations even though such regulations would hurt them in the short term. These people with “lower time preferences” generally oppose Paul. If the choice is really a matter of time preference, then why aren’t more Austrian economists against Paul? Surely, one doesn’t have to have a lower time preference to be an Austrian.

I never said this. Look carefully again at what I said. You have the greater than symbol backwards.

Then what’s this: “The only reason to delay till A+1 is if the satisfaction at that time could in fact be greater. … Something other than just the end of consuming at time A must present in the end of consuming at A+1, and that this new factor makes the End of consuming at A+1 a different and more highly valued end than A.”

It sure seems like you are saying: [A+1: Eat sandwich] > [A: Eat sandwich]

If “all things are equal”, I’ll choose eating at A to choosing “not eating now” and eating at A+1. If I choose instead to eat at A+1, then all things were not equal!.

No, I don’t see how that makes sense. What variables exactly would be equal? What if I said that if “all things equal” I’ll post on Mises at A, and if they are not equal, I’ll post on Mises at A+1. In that case, if I eat a sandwich at A, things will be equal, but since I then can’t post on Mises until A+1, that means that things are also not equal. How can all things be equal and also not equal? Do you see my confusion?

What if I cancel my doctor’s appointment and reschedule it to a later date? How does that bring my goal closer in time?

How does the fact that no one has an infinite time horizon mean that we always prefer something sooner to later? There is a strict conceptual difference between infinity and any finite number, no matter how large. Lending someone something for “infinity” would amount to the same thing as giving them a gift. Do gift givers demonstrate an infinitely negative time preference? If so, then Salerno is wrong about no one having an infinite time preference. If not, then an “infinite time horizon” has nothing to say about the phenomena of lending and interest. Lending simply means deferring the possession of an object, not deferring satisfaction.

So what does this mean? How could this say anything about interest? If I lend money with the expectation of receiving it back in 10 years, then clearly that 10 years isn’t the point at which any goal is valueless, since receiving the money is a goal. If I lend money out expecting to get it back after 20 years, then that isn’t the point at which any goal is valueless either. If I lend it out for 30 years, then…well you get the point.

Logically contradictory? How so? If I prefer something tomorrow when I could have it today, how would that not constitute a negative time preference? What’s contradictory about that?

Zero time preference for what exactly? This just echoes the incoherence of Mises’s and Rothbard’s explanations of scarcity.

82

Crap this is gonna get tedious. I do seriously believe that people look at the program that he puts forth and decide that the short term belt tightening is scarier than being told everythings fine. To be honest if the Republican Governers just balance the budget, and a Republican House, Senate and President simply balanced the budget, that would be a huge step.

you ignore the part in between the two highlighted sections. Something other than the end of consuming at time A is present in the end of consuming at A+1. It’s not simply [A+1: Eat sandwich] > [A : Eat sandwich] it’s that we know that if you defer consumption, there was something ELSE present at A+1 which was not present at A, the comparison then is [A+1: Satisfaction of Eating sandwich] > [A : Satisfaction of Eating Sandwich]. We KNOW if you defer it that something else is present. People wait to eat, to pray, and put a napkin on their lap. They might wait for a friend. But if there is no other factor, they don’t wait. If the satisfaction at A+1 is no different than the satisfaction at A then they eat at A. Time preference keeps saying if :

IF [Satisfaction of eating at A+1] = [Satisfaction of eating at A], then the preference is higher for [Eat at A] > [Eat at A+1], Ugh… I don’t know how many ways to say this… Anyone else take a crack at it? If someone eats at a later time, there is something about eating at A+1 that is MORE satisfying than eating at A. By definition…

No, I’m not confused. I see you attempting to find variables to add, to come up with an example that doesn’t agree. These are logical conclusions from the fact of action, you’re reading way to much into it. Every time you add some external causal relation to push one’s preference higher at a later date, you have confirmed the theory. This is the problem with what you are doing, you’re attempting to create an experiment that falsifies the relation.

This is where I’d like to do some simulation of extremely simplistic goal oriented agents. I’m a programmer, I’ve played around in my spare time with goal oriented systems, and scheduling. If you accidently screw up the value function goal oriented systems, they can stop acting. They can’t do anything. To push off an action requires in those systems, that they 1) wait for a condition that is a precondition to perform the action. 2) or you preempt that queued action, by pushing a more urgent demand on to the stack. But you use an event loop to force the agent to go through a little cycle. The end of every cycle is that it starts, continues, or changes the current operation. That’s how you get directed behavior in an autonomous agent simulation.

I didn’t say that if all things are equal you will post at A, but if they are not you will post at A+1. I can deduce from your action simply that you did prefer to post now, that if you deferred posting I can deduce that you preferred not to post now, but to do that at A+1. I don’t have to know what was different, but I know that something was different. What you aren’ t explaining is how one could post at A+1 if the satisfaction at A and A+1 are the same. That’s the question you aren’t getting. If you give me a specific reason, then you’ve said that the satisfaction is NOT the same. That something changed the equation. It’s apodictic, and so any example you come up with necessarily introduced “all things are not equal” when you come up with a thought experiment in which you find a reason to post later. It required a reason, it must require a reason not to post now. That reason is by definition a difference in the anticipated satisfaction at that later time.

It might be really useful to do some programming, in particular look at simple game AI. It’s pretty funny how they can fail and how they work when they do work, but the elements of praxeology are there in order to bring the appearance of goal-oriented action into existence. And when you “combat” the AI as a gamer, you have to interpret their behavior as goal oriented, and in fact, you presume time preference in the way you evaluate the other players actions. If you play chess you might find yourself doing the same thing against a human opponent. In fact, mistakes in the AI of a computer simulated entity or opponent, often make you aware very quickly that they aren’t human and if you look very closely, it will be caused by simple ridiculous mistakes that humans don’t make. Often it’s in fundamental glitches of the Programming that implements the categories of human action. The time preference one is so basic, that 0 time preference = no action, and negative time prefernce is the equivalent of a glitch in the value function. Meaning that the computer is accidentally inverting the value and returning items as having less value than the gamer (or game designer) expects.

You might consider thinking about the programming challenges of creating an computer opponent in a strategy game, like Starcraft, Warcraft, Command and Conquer, one of my favorites Civilization. Value functions are key, time preference is built into the value function. Preferring brings something close to the top of the event stack. The only reason one delays acting is because there is no action to take which improves things.

There is some reason, rational or not, in the way you value that visit AND EVERYTHING ELSE that pushed other priorities higher on the preference stack. It’s apodictic and a priori. The a posteriori fact that you delayed us tells us when interpreted with the a priori logic, that there was a difference in the expected satisfaction from the visit and all other considereations that pushed that preference lower on the stack. That the making the appointment happen later was more satisfying, or we should more properly say, that made it appear to be more satisflying to delay the appointment and do something else instead. Having watched loved ones struggle with major illness recently, and in fact die, the reasoning behind delaying an appointment is not always rational, sometimes the low-level fear that something might be wrong is preferred to the more fearsome knowing somethings wrong.

I’m going to hope and pray that at this point you can figure this stuff out in your own head. I think I’ve given you the two pieces of information you need. One you’re digging too hard to find more in the statement than is there, it’s apodictic. Two, understanding how to make a goal oriented system happen, requires an evaluation method, and an action stack and an event loop. Together they exhibit time preference. Time preference is what makes an action come out off the action stack and get performed in the event loop.

Get the time preference issue in your head, before you move on to interest. You’ve concluded interest is bad and so you’re attacking the time preference issue. If you want to attack interest find another method, time preference isn’t it. But, I’ll say it anyway, technically you didn’t say anything. You just said “but some periods of provision are really long.” That’s not a counter-argument. He simply said they must be finite.

I’m actually rethinking a bit of this logic in light of the thinking and articulation I did above. I did some rethinking of work I’d done historically and never related to praxeology. If you assume an action must happen, then the value function is what pops an action out the top. The action that appears is the action performed. So I wonder if it doesn’t mean that instead of popping from the top, you are popping from the bottom? A zero time preference would probably result in no action being popped.

Here’s an interesting article about connecting more advanced agents to existing game engines. If you happen to read it, pay attention to the assumptions about the existence and the propogation of time in achieving action. There’s no “time preference” explicitly stated. But the various execution threads, event loops, and message queue processing mechanisms all produce the very thing we’re talking about. Actions happen, and they happen when the conditions are met, and the goal is suffiiently valued to happen (above all alternatives). This IS time preference. If it’s ready to happen, and you want it more than anything else you could pursue instead, it happens.

Umm you went a different way, Please elaborate Mises’s and Rothbard’s explanations of scarcity.

83

you ignore the part in between the two highlighted sections. Something other than the end of consuming at time A is present in the end of consuming at A+1. It’s not simply [A+1: Eat sandwich] > [A : Eat sandwich] it’s that we know that if you defer consumption, there was something ELSE present at A+1 which was not present at A, the comparison then is [A+1: Satisfaction of Eating sandwich] > [A : Satisfaction of Eating Sandwich]. We KNOW if you defer it that something else is present. People wait to eat, to pray, and put a napkin on their lap. They might wait for a friend. But if there is no other factor, they don’t wait. If the satisfaction at A+1 is no different than the satisfaction at A then they eat at A.

Yeah, sorry, I still disagree. You seem to imply that something else is present only if we defer eating the sandwich. My point is that something else is also present when we choose not to defer eating the sandwich. This is due to the fact that we have to do something else in place of eating the sandwich at A+1.

Time preference keeps saying if :

IF [Satisfaction of eating at A+1] = [Satisfaction of eating at A], then the preference is higher for [Eat at A] > [Eat at A+1], Ugh… I don’t know how many ways to say this… Anyone else take a crack at it? If someone eats at a later time, there is something about eating at A+1 that is MORE satisfying than eating at A. By definition…

If the satisfaction gained from both times is equal, then by definition neither is preferred to the other. You essentially just wrote that if things are equal then they are not equal.

Here are some valid conclusions we could draw (I’m adding a “+” between variables–I think that makes more sense):

if

[A: Eat sandwich] + [A+1: Do something else] > [A: Do something else] + [A+1: Eat sandwich]

and if

[A: Eat sandwich] = [A+1: Eat sandwich]

then

[A+1: Do something else] > [A: Do something else]

or, reversing the sign, 2)

if

[A: Eat sandwich] + [A+1: Do something else] < [A: Do something else] + [A+1: Eat sandwich]

and if

[A: Eat sandwich] = [A+1: Eat sandwich]

then

[A+1: Do something else] < [A: Do something else]

or, assuming one had to eat a sandwich at one of the two times, 3)

if

[A: Eat sandwich] = [A+1: Eat sandwich]

and if

[A+1: Do something else] > [A: Do something else]

then

[A: Eat sandwich] + [A+1: Do something else] > [A: Do something else] + [A+1: Eat sandwich]

None of these proofs tell us that sooner is preferred to later as some sort of general rule.

No, I’m not confused. I see you attempting to find variables to add, to come up with an example that doesn’t agree. These are logical conclusions from the fact of action, you’re reading way to much into it. Every time you add some external causal relation to push one’s preference higher at a later date, you have confirmed the theory. This is the problem with what you are doing, you’re attempting to create an experiment that falsifies the relation.

My point isn’t that adding external causal relations will push one’s preference to a later date, but that these external causal relations are also at work in pushing the preference to an earlier date. There is no neutral way to conceive of this that would result in a sooner time preference. If you eliminated all of these external causes there would simply be no preference at all, no action.

The time preference one is so basic, that 0 time preference = no action, and negative time prefernce is the equivalent of a glitch in the value function. Meaning that the computer is accidentally inverting the value and returning items as having less value than the gamer (or game designer) expects.

But anything that the AI is not seeking to do at the moment represents a negative time preference. The existence of a positive time preference for one thing means that there is necessarily a negative time preference for another.

There is some reason, rational or not, in the way you value that visit AND EVERYTHING ELSE that pushed other priorities higher on the preference stack. It’s apodictic and a priori. The a posteriori fact that you delayed us tells us when interpreted with the a priori logic, that there was a difference in the expected satisfaction from the visit and all other considereations that pushed that preference lower on the stack. That the making the appointment happen later was more satisfying, or we should more properly say, that made it appear to be more satisflying to delay the appointment and do something else instead. Having watched loved ones struggle with major illness recently, and in fact die, the reasoning behind delaying an appointment is not always rational, sometimes the low-level fear that something might be wrong is preferred to the more fearsome knowing somethings wrong.

The point is that when we see one activity pushed backwards and another brought forwards, there is nothing there to suggest that the thing brought forward is the determining factor and not the thing pushed backwards.

Get the time preference issue in your head, before you move on to interest. You’ve concluded interest is bad and so you’re attacking the time preference issue. If you want to attack interest find another method, time preference isn’t it. But, I’ll say it anyway, technically you didn’t say anything. You just said “but some periods of provision are really long.” That’s not a counter-argument. He simply said they must be finite.

I agree with a certain version of time preference: that people prefer to do different things at different times. I understand how that works, which I’ve demonstrated in my formulas and syllogisms. Now I’m showing how it alone doesn’t say anything about interest. Mises clearly ties the concept in with interest. It’s the only reason why he comes up with his particular conception of time preference.

How about dropping the accusations of ideological motives and sticking to the descriptive claims at hand? I thought you had a problem with Marxists dismissing their opponents for such reasons.

Here’s an interesting article about connecting more advanced agents to existing game engines. If you happen to read it, pay attention to the assumptions about the existence and the propogation of time in achieving action. There’s no “time preference” explicitly stated. But the various execution threads, event loops, and message queue processing mechanisms all produce the very thing we’re talking about. Actions happen, and they happen when the conditions are met, and the goal is suffiiently valued to happen (above all alternatives). This IS time preference. If it’s ready to happen, and you want it more than anything else you could pursue instead, it happens.

I don’t fully understand the references you’ve made to computer science, but what you allude to sounds interesting. I’ll try to read that article at some point.

Umm you went a different way, Please elaborate Mises’s and Rothbard’s explanations of scarcity.

I mentioned them earlier, the end of my first entry and the ensuing discussion. But to return to Salerno’s example, I think of zero time preference* as no preference or no action. But Salerno’s example refers to such concepts as satisfaction, consumption, the act of cloning, which all seem to imply the existence of action and preference. If they don’t have to clone themselves an infinite number of times, then they must prefer to clone themselves an infinite number of times. So then they must not have a zero time preference for cloning themselves.

*Or perhaps another way to conceive of it would be to prefer two things at the same time. It kind of seems like Salerno is trying to demonstrate this in his example. But if that is what he is trying to communicate, why not give a simpler example, like the preference for eating dinner while watching TV? Well, probably because that would contradict his claim that zero time preference doesn’t occur in our world.

84

Right, entrepreneurs always benefit society by spending their money…

Edit: This comment was aimed at a post by a spambot, which has since been deleted.

85

First I apologize for my tone, you are right, it was inappropriate. We are discussing the merits on the arguments themselves. The extra rhetorical banter by me was rude

I’m going to reply to a part of this instead of the whole thing. Because I think we’re getting to the root of the disconnect in it. If you don’t agree we can come back and look at more parts later. This formulaic representation is I think helping us more than anything else we could do. I’ve actually argued in other places that it would be useful to do this more often. With that said…

You are correct if the way you’re constructing this is correct, it makes a lot of sense.

I’m going to switch variables just for shorthand

T : A specific time, A: Do one thing, B : do another thing. I’ll use the same form [ Time value : thing done ]. S = Statement

So, the theorem, and for now I’ll call it a theorem instead of an axiom is that we can only know two things without observation :

S1 : [T : A] > [T+1 : A]

S2 : [T : B] > [T+1 : B]

That’s the definition of time preference in Praxeology. Above is a comment about the time relation between expected satisfaction of the same action at a nearer time and a later time. Eating happens to be a particularly bad example, but we should still be able to understand it. The problem is I think one has to introduce science.

Now, let’s take your two directional examples above :

S3 : [[T : A],[T+1 : B]] > [[T : B],[T+1 : A]]

From S1 and S2, we know that B was not deferred because of something about B. You’re asking however, if we can know S4 :

S4 : [T : A] > [T : B]

from S3. I don’t know that we can. I don’t think that Praxeology says that. It may also have been arranged this way because :

S5 : [T+1 : B] > [T+1 : A]

Now let’s take one more step. You and I both know the experience of saying… I don’t care. I’ve done it often when my wife asks if we should eat at this place or that for dinner tonight. But I question whether or not that’s the same as saying :

S6 : [T : A] = [T+1:A]

The point I was making from my time preference discussion of agents in computer simulations is that S6 doesn’t make sense. We know that the idea of an ordered stack of preferences is a thought device that’s not real in human minds. The idea however might be sufficiently accurate in representation as a logical device, to make something clear.

In programming we use stacks or queues to represent ordered data. We use sort algorithms to rearrange them. We use functions to control the ordering. But you never end up with two items in the same position in an order. Now if I’m sorting integers, there’s no reason I couldn’t actually have the same integer in two slots in the queue, but in reality the contents of each location are an [action+expected result] and I reorder based on some way of normalizing different expectations of satisfaction over time. But there is no case where I actually end up with any two instances on the stack where they are actually equal.

I think this is the point of time preference. If the human mind actually ends up struggling to order two instances on the preference stack, the ordering problem either collapses into a resolution or it locks you up and you stop functioning. Think about sitting around trying to decide which restaurant to go to, and suddenly realizing if you really wanted to go, you would have. But you just sit there and don’t do anything. The other thing that happens is that you just pick one because you’re so damned hungry… and you just wanna go get some food.

Now back to the eating example. You and I both know upon examination that even though Praxeology says how a person arrives at the action is a black box. We don’t argue about how they arrived at the result, we just know from their actions that the result was what they thought would bring the most satisfaction. But you and I can introduce experiential knowledge, and some biological science, that explains the way in which hunger works as a sensation in the body. You and I know that the value function related to food (without accounting for the food content) will result in different return values over time.

Now, I see this as a potentially confusing phenomena that seems to work at variance with the way Praxeology describes Time Preference. I presume, but have to think about it some more, that the way hunger works biologically would be considered it’s own phenomena that results in a change to the value function, and thus that phenomena breaks the “and all things are equal” clause.

At this point though you’re looking for some justification for the assertion that S1 and S2 are true.

Again, I see what you’re saying. How can we isolate an action when we can’t actually isolate an action? How can we assert S1 and S2 after isolating the action logically in a thought experiment.

I think I know where to look in this problem, but I’m also concerned that it’s actually this hard to make clear. Which means that something about it needs to be drawn out more clearly.

I’m not saying for sure this is the answer. I’m going to spend time thinking about it, I have to get to bed, but here’s a couple of questions to ponder in your mind. I’m thinking about them as I head off. I’d like to discuss them tomorrow.

  1. Think about how much someone would have to give you tomorrow to borrow 50% of your monthly budget today? You don’t need to lend it to him. He’s just asking. What’s the nature of the emergency? Why can’t he wait till tomorrow? Could you get more from him under certain circumstances?
  2. Does risk/uncertainty play into 1? Does time have a role in the uncertainty? What if he didn’t give it back to you tomorrow? What would having 300% of your budget saved do to your decision in 1?
  3. Why would I pay a finite price for a good that will last forever and will always provide the same level of service (think diamond). Is there anyway a price could be the equivalent of the area under a curve?

I’m not trying to be sneaky, you and I both know these aree the arguments made about interst, asset prices, and uncertainty. That uncertainty is viewed as the cause of time preference, and that discounted rate of future value is the source of interest and asset prices which are below the expected input value to a process.

I guess the last point, I swear I have to get to bed, is to ask, why ELSE other than time preference could question 3 or 1 and is 2 sufficient to explain the discount on future goods?

I’ll come back to the rest later, I saw some more points or questions to discuss.

86

Not quite the last point. I’m continuing to think about the examples I gave.

Think about a valuation of the diamond where you expected to consistently get greater value from it over time. Your “satisfaction” would continue to grow and never stop growing.

The area under that curve would approach infinity right?

Next question, can we think of any process where the human experience is one of growing, growing, growing satisfaction? I can think of some, but they all have a cutoff valve. The best one I can think of is “sex”, I think eating fits this description. But it seems that biology and reality have provided cutoffs that are built in to these type of positive feedback loops. The big “O” for sex, and “full” for eating. In fact, I think we can imagine the disastrous outcome of the behavior of a human being where this wasn’t true.

87

First I apologize for my tone, you are right, it was inappropriate. We are discussing the merits on the arguments themselves. The extra rhetorical banter by me was rude.

No problem. I apologize too if I’ve overreacted at times. I hope I didn’t scare Adam off in my last reply to him.

So, the theorem, and for now I’ll call it a theorem instead of an axiom is that we can only know two things without observation :

S1 : [T : A] > [T+1 : A]

S2 : [T : B] > [T+1 : B]

That’s the definition of time preference in Praxeology. Above is a comment about the time relation between expected satisfaction of the same action at a nearer time and a later time. Eating happens to be a particularly bad example, but we should still be able to understand it. The problem is I think one has to introduce science.

I agree with that being a good formulation of Mises’s claim. And yes, it is what I object to. I don’t see how we can know that without observation. In fact, given that A and B are mutually exclusive, S1 and S2 strike me as inherently contradictory propositions.

From S1 and S2, we know that B was not deferred because of something about B. You’re asking however, if we can know S4 :

S4 : [T : A] > [T : B]

from S3. I don’t know that we can. I don’t think that Praxeology says that.

I think even granting S1 and S2, B could still be deferred because of something (positive) about B–namely, that [T+1 : B] could be greater than [T+1 : A]. (Though I guess you’re trying to isolate B from its place in time–to me it doesn’t make sense for something to be greater than another when they’re removed from time. As I think Kant and Bergson noted, quantity always contains a spatial or temporal referent.). And I think there is a sense in which we can conclude S4. A is certainly chosen over B at T. However, we can’t conclude that it is preferred in a vacuum in isolation but as part of a larger interconnected network of choices.

  1. Think about how much someone would have to give you tomorrow to borrow 50% of your monthly budget today? You don’t need to lend it to him. He’s just asking. What’s the nature of the emergency? Why can’t he wait till tomorrow? Could you get more from him under certain circumstances?
  2. Does risk/uncertainty play into 1? Does time have a role in the uncertainty? What if he didn’t give it back to you tomorrow? What would having 300% of your budget saved do to your decision in 1?

I think risk certainly plays into it. If I was 100% certain that I would get all of the money back and if I was sure that I didn’t want to use that money before tomorrow, then the only other reason I could think of to not lend the money would be if I thought the person was going to spend the money on something I disapproved of. So there are three potential barriers to lending money:

  1. Risk of not getting the money back.
  2. Opportunity cost–the alternative uses of the money over the same time.
  3. Disapproval of how the borrower will spend the money.

I think 3 generally only pertains to a limited area, such as between friends and between governments. I might be less willing to lend money to a friend who wants to use it to buy heroin–regardless of whether I think I will get it back. Similarly, the US government might be willing to lend money to Israel so that it could buy arms but would be much less willing to lend money to Iran for the same purpose. But when we consider commercial, for-profit banks, I don’t think they would make decisions on this basis. Many banks are perfectly willing to lend money to oppressive dictators or tobacco manufacturers–and, I’m sure, heroin producers if it were legal.

So that leaves 1 and 2 as possible explanations for the interest rate, and I think they are both in play. Post-Keynesian economists often highlight the role of risk in determining the interest rate. This is part of their explanation as to why the interest rate goes down in a boom. A boom means that more businesses are successful and able to repay their loans, thus less risk.

I actually haven’t yet gotten to the part of Capital where Marx deals with interest rates, but I suspect his explanation will implicitly allude to opportunity cost. For Mises, interest is primary and is the explanation for why the average rate of profit is usually positive. For Marx, it’s the other way around–profit is primary and interest just represents a way in which surplus value is shared between specialized subclasses of capitalists. Profit ultimately arises due to the differing opportunity costs of the two parties in a given transaction. Ironically, I think this Marxian position is more true to the core principles of praxeology.

Why would I pay a finite price for a good that will last forever and will always provide the same level of service (think diamond). Is there anyway a price could be the equivalent of the area under a curve?

You mean, why would someone sell you the diamond for a finite price? One could always replace the diamond by spending a finite amount of labor on mining another diamond. Therefore, it doesn’t cost the diamond seller an infinite amount of satisfaction, but a finite amount of satisfaction forgone in the process of mining the diamond. (I also touched on this issue from another angle in my appendix.)

Next question, can we think of any process where the human experience is one of growing, growing, growing satisfaction? I can think of some, but they all have a cutoff valve. The best one I can think of is “sex”, I think eating fits this description. But it seems that biology and reality have provided cutoffs that are built in to these type of positive feedback loops. The big “O” for sex, and “full” for eating. In fact, I think we can imagine the disastrous outcome of the behavior of a human being where this wasn’t true.

I think what really matters when it comes to human action is the relative satisfaction. We could suppose that actions A and B both have growing satisfaction without a cutoff point. A might be greater than B at first. So one will perform A at first, but provided that the rate of potential satisfaction that B provides grows at a faster rate, eventually one will switch from A to B. If the rate could then change so that A grows faster, then A would supplant B eventually, and things could alternate in this way indefinitely.

I’m not sure if we are using the concept of satisfaction in a consistent manner though. Adam and Jon seemed to think that it isn’t the same as pleasure or enjoyment. I’m not sure Mises’s use was consistent either.

88

Disclaimer, I agree with the time preference argument, to me it seems obvious and logical, but I am struggling to express the rightness of it (and I admit there seems to be an intuitive piece to it.) Part of the difficulty in my mind is realizing that I can’t, nor can anyone actually isolate these factors in time, nor can we isolate different preferences from each other. Another point is there may in fact be a lack of clarity in the way it’s being expressed. My points about programming agents is one where time preference appears to be implicit. Where again, it’s an emergent property or a fundamental property of how the system works.

Without going into satisfaction, which I’ll come back to later, there is an assumption that satisfaction is what one’s aiming at, and satisfaction has to appear in the now in order for action to occur. Is that time preference? Is that all the statement of positive time preference means? Is it possible we’re overthinking it. In other words, are all other differences a function of the value function itself, and time preference is simply a statement that action is aimed at satisfaction actually appearing in reality? Is that perhaps all that’s meant by [T : A] > [ T+1 : A]?

Hmmm, thinking in that way, the low time preference, high time preference one might be considered as a restatement of a different metric which might be sensitivity to risk. Again however, consider no sensitivity to risk. What would that mean? Does the presence of risk, in and of itself explain time preference?

Ok, first I’ll go back to the fundamental argument about praxeology. We can’t talk about quantities when we’re talking about satisfaction. To talk about quantities is to confuse social science with the physical sciences, we can’t measure satisfaction or preference.

So, yes I think we are trying to isolate B from it’s place in time, or perhaps what we’re actually saying about time preference is that we’re considering means, end, and satisfaction and context all in one entity that we’re attaching the term B to. Then if we shift the time of the context forward (+1) then we have shifted the satisfaction lower (<) Nothing in the property of time in and of itself is not in this case sufficient to explain a rise in satisfaction. But there is a factor of time (uncertainty) that can in and of itself reduce satisfaction in and of itself. I’ll defer to Poincare and the n-body problem of physics as a description of how slight unknown differences in initial conditions result in rapidly diverging (uncertain) potential future outcomes.

So maybe the isolation part and the technical impossibility of it is simply a device to capture how the value (not a measurable quantity) that’s returned from the value function (a rhetorical device) is impacted by a shift forward into the future.

As I’ve said earlier I think risk aversion and time preference (positive) are in fact the same thing said in different ways.

Austrian’s always describe interest rates as the “market price” for money. If that makes sense. In other words as savings increase (supply for investment/credit) the rate goes down. As savings goes down the interest rates rise. So that in and of itself makes sense. Any interest rate itself overcomes a man’s desire to simply hold money as you say it overcomes opportunity cost and risk, which are both factors that must be overcome in order to get the money to move into an investment.

Well, first of all I think we have to understand that logically the anticipated rate of profit for any business must exceed interest rates. Otherwise you liquidate the capital and invest it elsewhere. As an aside, this is an emergent property. It tells you that the needs/wants of the consumers are sending signals to the market (through prices) that they want the goods produced in the market to change from what they are now. The market will adjust and produce the goods that can be profitably generated and sold in the market.

What’s your source for saying that Mises believes interest is primary? I’m not disagreeing, I’m just asking if you remember a reference.

But let me ask another question, do you embrace the idea that the micro-behavior of an individual has to contain the seeds of any market phenomena? Because the idea of profit can be seen in the micro-behaviors of a Crusoe environment. I’m interested in understanding whether or not interest or profit is primal. Consider this next sentence in the individual setting, and then extrapolate it to the social production environment. Maybe, as Mises seems to suggest, the return (profit) from a production method must exceed a threshold in order for man to engage in that production. That threshold seems to be what Mises refers to as the natural rate of interest.

I’m trying to dig these things out of the micro-behavior. For example, to engage in building a net for fishing, I have to have surplus time from meeting all other immediate needs, and then I can “invest” time in building a net. But the anticipated return must not only exceed the time invested, but must exceed it by an amount that one might call “the natural rate of interest.” Now, I know and you know that this isn’t exactly what happens, the first time a man tries something he might have an idea about the return he might expect, but the empirical evidence tells him in hindsight what his “real” satisfaction was, and what his real investment of time energy and resources was. That feedback changes the value function. The benefit expected from a production method must exceed “the natural rate of interest”. It seems to me that the 'natural rate of interest" would be a dynamic thing, it would vary from person to person, and would

I don’t think my point was well structured. The better example is probably a piece of unique artwork. However, you make the point that we look for other things that would substitute as a means to the end. That warrants more thought, but I think that’s a function of trying to minimize what we give up in relation to what we get.

I think the satisfaction concept is overblown, in that we’re reading too much into it.

If we take a couple of examples we get a sense simply of this avoidance attraction concept in the mind. We see it in our sensory input, we see it in the way we conceptualize.

We know simply that there are relationships between human experience and phenomena experienced where some phenomena repel us, we avoid them. Other phenomena attract us, we seek them.

Satisfaction/dissatisfaction, good/bad, right/wrong, love/hate, etc. It’s a binary scale, but has variance in intensity. That’s all we know. I think we can actually see and measure the differences in intensity in some phenomena, like temperature. Praxeology says that the intensity of that subjective reaction to phenomena is not quantifiable. And so we’re left with relative value only.

I think I’d like to more formally bind a bunch of factors into the input to the value function. And when I say function I mean “that thing the brain does that precedes and results in an intentional act”.

I really want to put some of this down in some kind of formal way.

89

Yes.

The way I understand this, you perform action A at time T because it is then that it appears at the top of your value stack. You demand something in return for not performing A at time T. If this potential payoff is large enough, then action B (accepting a prospect of a potential future payoff) dislodges action A from the top of your value stack at time T. Action A could be to spend your $100 at a nice restaurant. Action B could be to eat Ramen Noodle at home instead and keep the $100 in your pocket if you thought that opportunities to turn them into $130 over one year are likely, OR action B could be to eat Ramen Noodle at home and lend the $100 to another entrepreneur (sporting profit expectations similar to yours) who is willing to promise you $120 back a year later. Risk/reward and uncertainity (among other things) enter the equation when estimating the size of the potential payoff needed to dislodge action A from the top of your value stack at time T.

I had fun reading this whole thread over the last couple of days and have lots of comments to make – mostly in response to FOTH’s confusions and misunderstandings – but unfortunately I’m short on time (scarce bastard, this). Adam, Jon, and David are doing an excellent job, btw. I hope it continues.

FOTH, kudos for reading Human Action and for keeping an open mind! Looking forward to your book on the praxeological derivation of LTV.

90

Without going into satisfaction, which I’ll come back to later, there is an assumption that satisfaction is what one’s aiming at, and satisfaction has to appear in the now in order for action to occur. Is that time preference? Is that all the statement of positive time preference means? Is it possible we’re overthinking it. In other words, are all other differences a function of the value function itself, and time preference is simply a statement that action is aimed at satisfaction actually appearing in reality? Is that perhaps all that’s meant by [T : A] > [ T+1 : A]?

Why wouldn’t aiming to do A at T+1 be an example of satisfaction appearing in reality?

Ok, first I’ll go back to the fundamental argument about praxeology. We can’t talk about quantities when we’re talking about satisfaction. To talk about quantities is to confuse social science with the physical sciences, we can’t measure satisfaction or preference.

Actually, I think we have to talk about quantities when discussing praxeology and satisfaction. Remember, there are two forms of quantity, magnitude and multitude. Whenever we refer to preference or say that A > B we are expressing magnitude. In praxeology, A and B are considered identical qualities to the extent that they are both possible objects of choice. However, they also have different qualities that allow one to be chosen over the other. It is this aspect of simultaneously being identical and being different that I think is the essence of quantity. There are two choices.

So, yes I think we are trying to isolate B from it’s place in time, or perhaps what we’re actually saying about time preference is that we’re considering means, end, and satisfaction and context all in one entity that we’re attaching the term B to. Then if we shift the time of the context forward (+1) then we have shifted the satisfaction lower (<) Nothing in the property of time in and of itself is not in this case sufficient to explain a rise in satisfaction. But there is a factor of time (uncertainty) that can in and of itself reduce satisfaction in and of itself. I’ll defer to Poincare and the n-body problem of physics as a description of how slight unknown differences in initial conditions result in rapidly diverging (uncertain) potential future outcomes.

So maybe the isolation part and the technical impossibility of it is simply a device to capture how the value (not a measurable quantity) that’s returned from the value function (a rhetorical device) is impacted by a shift forward into the future.

Kant said that synthetic a priori judgments can only be made for what he called mathematical judgments (which I take to be the same thing as quantitative judgments). Any attempt to make what he calls philosophical (what I think of as qualitative) synthetic judgments a priori is erroneous. He called the former axioms and the latter dogmas. Now I actually agree that synthetic a priori judgments are the realm of praxeology*. This is because I believe praxeology to be mathematical.

Now space and time are the only means for allowing things to be both the same and different, the only means for differentiating between universals and particulars. Kant says that time allows us to attribute contradictory traits to the same object. I’m currently reading Henri Bergson’s Time and Free Will, which apparently influenced Mises. In it, Bergson explores the issue of how sensations can be said to have quantity or intensity. How can one thing be hotter than another, for example? Hot and cold are only different qualities with distinct sensations and don’t seem to have any magnitude within them. Bergson’s solution is to examine each quality in how it relates to space. He notes that as we get closer to a candle, the quality of sensation changes. Thus, the sensation of temperature can be quantified to the extent that it corresponds to spatial relations.

This is why I don’t think we can speak of satisfaction when we isolate something from time and space. When taken apart from time and space, B is merely an essence that cannot be said to be greater or lesser than something else. Outside of time and space, B does not contradict A. It is only within time and space that B contradicts A, giving rise to a quantitative of comparison between the two. When you examine [T: B] in relation to [T+1: B] while excluding everything else, B is no longer being considered as a choice. If there is no other possibility at T other than B, then B is not being chosen, is not being preferred to anything. Things can only be preferred to the extent that they could occur at the same time, because only then are they contradictory.

*It should be noted, however, that the proposition “humans act” is not an axiom, but a synthetic a posteriori judgment. I suppose it could also be an analytic judgment if you define human as a thing that acts.

As I’ve said earlier I think risk aversion and time preference (positive) are in fact the same thing said in different ways.

Austrian’s always describe interest rates as the “market price” for money. If that makes sense. In other words as savings increase (supply for investment/credit) the rate goes down. As savings goes down the interest rates rise. So that in and of itself makes sense. Any interest rate itself overcomes a man’s desire to simply hold money as you say it overcomes opportunity cost and risk, which are both factors that must be overcome in order to get the money to move into an investment.

I don’t think risk is sufficient to explaining interest rates. Let’s say that the risk is that 1 out of every 100 loans is not repaid. This means that if I lend 100 people $1 each, one of them will not pay me back. In order to cover the risk, I would have to charge each 1 cent of interest. If interest was only based on risk, I would never make any money on my loans. I would just receive the same amount that I lent. This might not be true of every lender. Some might lose and some might gain, but it would all average out to zero. This is obviously not true to the real world, where the average amount earned on interest is above zero.

Well, first of all I think we have to understand that logically the anticipated rate of profit for any business must exceed interest rates. Otherwise you liquidate the capital and invest it elsewhere. As an aside, this is an emergent property. It tells you that the needs/wants of the consumers are sending signals to the market (through prices) that they want the goods produced in the market to change from what they are now. The market will adjust and produce the goods that can be profitably generated and sold in the market.

I mean profit in a more primary sense, like the excess of revenue over costs of production. I don’t see interest as a cost of production, and thus it arises as a division of this primary profit.

What’s your source for saying that Mises believes interest is primary? I’m not disagreeing, I’m just asking if you remember a reference.

Well, I don’t think he has an explanation as to why the rate of (entrepreneurial) profit is positive without presupposing that it includes some sort of originary interest.

But let me ask another question, do you embrace the idea that the micro-behavior of an individual has to contain the seeds of any market phenomena?

I think market phenomena are emergent properties, which are absent in the things that produced them. A seed produces a tree with leaves, but the category of “leaves” isn’t very useful in describing the structure of a seed.

I’m trying to dig these things out of the micro-behavior. For example, to engage in building a net for fishing, I have to have surplus time from meeting all other immediate needs, and then I can “invest” time in building a net. But the anticipated return must not only exceed the time invested, but must exceed it by an amount that one might call “the natural rate of interest.” Now, I know and you know that this isn’t exactly what happens, the first time a man tries something he might have an idea about the return he might expect, but the empirical evidence tells him in hindsight what his “real” satisfaction was, and what his real investment of time energy and resources was. That feedback changes the value function. The benefit expected from a production method must exceed “the natural rate of interest”. It seems to me that the 'natural rate of interest" would be a dynamic thing, it would vary from person to person, and would

Yeah, I don’t think this sort of analogy is useful for understanding what are emergent properties. The concepts of profit and interest only make sense when the difference between the inputs and outputs is purely quantitative (i.e. money spent vs. money received).

I think the satisfaction concept is overblown, in that we’re reading too much into it.

The part I find confusing is whether satisfaction refers simply to there being a reason that one thing is preferred to the other or to a specific sensation. The difference can be illustrated in asking whether work is necessarily satisfying. According to the first definition, work is necessarily satisfying if one chooses it. In other words, one has a reason for doing what one does. According to the second definition, work could be dissatisfying, perhaps even necessarily dissatisfying. This is because work might entail unpleasurable sensations but might produce a desirable result nonetheless. In order to say that we always seek satisfaction–which is what Adam says Mises means by the “removal of uneasiness”–then we have to be using the first definition. But in order to conceive of time preference as a deferral of satisfaction, we must be using the second definition of satisfaction. We clearly can’t defer something that is inherent in every action. We can’t defer reason itself.

Satisfaction/dissatisfaction, good/bad, right/wrong, love/hate, etc. It’s a binary scale, but has variance in intensity. That’s all we know. I think we can actually see and measure the differences in intensity in some phenomena, like temperature. Praxeology says that the intensity of that subjective reaction to phenomena is not quantifiable. And so we’re left with relative value only.

Bergson ends the first chapter of Time and Free Will with this interesting point:

For if we grant that one sensation can be stronger than another, and that this inequality is inherent in the sensations themselves, independently of all association of ideas, of all more or less conscious consideration of number and space, it is natural to ask by how much the first sensation exceeds the second, and to set up a quantitative relation between their intensities. Nor is it any use to reply, as the opponents of psychophysics sometimes do, that all measurement implies superposition, and that there is no occasion to seek for a numerical relation between intensities, which are not superposable objects. For it will then be necessary to explain why one sensation is said to be more intense than another, and how the conceptions of greater and smaller can be applied to things which, it has just been acknowledged, do not admit among themselves of the relations of container to contained. If, in order to cut short any question of this kind, we distinguish two kinds of quantity, the one intensive, which admits only of a “more or less,” the other extensive, which lends itself to measurement, we are not far from siding with Fechner and the psychophysicists. For, as soon as a thing is acknowledged to be capable of increase and decrease, it seems natural to ask by how much it decreases or by how much it increases. And, because a measurement of this kind does not appear to be possible directly, it does not follow that science cannot successfully accomplish it by some indirect process, either by an integration of infinitely small elements, as Fechner proposes, or by any other roundabout way. Either, then, sensation is pure quality, or, if it is a magnitude, we ought to try to measure it.

91

z1235: The way I understand this, you perform action A at time T because it is then that it appears at the top of your value stack. You demand something in return for not performing A at time T. If this potential payoff is large enough, then action B (accepting a prospect of a potential future payoff) dislodges action A from the top of your value stack at time T. Action A could be to spend your $100 at a nice restaurant. Action B could be to eat Ramen Noodle at home instead and keep the $100 in your pocket if you thought that opportunities to turn them into $130 over one year are likely, OR action B could be to eat Ramen Noodle at home and lend the $100 to another entrepreneur (sporting profit expectations similar to yours) who is willing to promise you $120 back a year later. Risk/reward and uncertainity (among other things) enter the equation when estimating the size of the potential payoff needed to dislodge action A from the top of your value stack at time T.

I think I am in agreement with this. However, I don’t think this supports any notion of “sooner” being more primary than “later.” It is true that if I plan on doing A at T, providing me with an incentive could cause me to defer A to T+1. The same could be said of the reverse. I might prefer doing A at T+1 and would only do A at T if someone provides me with an incentive to do so. A deadline is an example of this.

92

The same could not be said of the reverse. At time T you only have two choices: (1) do A now, or (2) postpone A for later (i.e. hopefully T+1 if all goes according to plan). You cannot decide at time T to do A at time T-1, nor can you decide at time T+1 to do A at time T.

Furthermore, at time T you only “know” your value stack at time T. Your action A at time T reveals to the world that action A resides at the top of your value stack then (and only then). At time T you can only predict/estimate (but not know): (1) the state the universe at time T+1, and (2) your value stack at time T+1 affected by it. These predictions (estimates about the state of affairs in T+1) performed at time T affect your value stack at time T.

93

The same could not be said of the reverse. At time T you only have two choices: (1) do A now, or (2) postpone A for later (i.e. hopefully T+1 if all goes according to plan). You cannot decide at time T to do A at time T-1, nor can you decide at time T+1 to do A at time T.

Furthermore, at time T you only “know” your value stack at time T. Your action A at time T reveals to the world that action A resides at the top of your value stack then (and only then). At time T you can only predict/estimate (but not know): (1) the state the universe at time T+1, and (2) your value stack at time T+1 affected by it. These predictions (estimates about the state of affairs in T+1) performed at time T affect your value stack at time T.

By working in reverse, I didn’t mean that it works for the past but for two moments in the future that relate to each other in terms of “sooner” and “later.” Instead of providing someone with an incentive to defer something they would have done sooner to later, one might also have to provide an incentive to make someone do something they would have done later to sooner. Your previous post talked about deferring the action at T. Obviously, you were assuming that T was a point in the future since you can’t defer something that you are doing now. But here you seem to be switching the issue of “sooner” versus “later” to the “present” versus the “future.” The latter wouldn’t be any use for explaining interest rates because people don’t lend money while they are spending it.

My point stands. There’s no reason to assume that people prefer “sooner” to “later” in some sort of universal manner. The concept doesn’t even make any sense given the manner in which we experience the world.

94

You can only know your “present” value stack, and you disclose to the outside world that action A resides at its top by perfoming it then. (i.e. in the “present”). I repeat:

At time T you only “know” your value stack at time T. Your action A at time T reveals to the world that action A resides at the top of your value stack then (and only then). At time T you can only predict/estimate (but not know): (1) the state the universe at time T+1, and (2) your value stack at time T+1 affected by it. These predictions (estimates about the state of affairs in T+1) performed at time T affect your value stack at time T.

You are missing the point. People “prefer” performing action A now (vs. not-now, i.e. later) exactly because it is then that it resides at the top of their value stack, by definition. The “present” value stack is special that way, when compared to the (unknown) value stacks at any point in the future (be it “sooner”, or “later”).

95

You can only know your “present” value stack, and you disclose to the outside world that action A resides at its top by perfoming it then. (i.e. in the “present”).

You are missing the point. People “prefer” performing action A now (vs. not-now, i.e. later) exactly because it is then that it resides at the top of their value stack, by definition. The “present” value stack is special that way, when compared to the (unknown) value stacks at any point in the future (be it “sooner”, or “later”).

But how can anything general be derived from this? When I prefer to do A now, I also prefer to not do B now. Can we infer that I always prefer to do A sooner rather than later? Certainly not. We can only infer that I prefer to do A now. In the next moment, I may prefer to do B. If all we are considering is what I prefer now, then “time preference” isn’t present in the analysis at all.

But Mises says: “Satisfaction of a want in the nearer future is, other things being equal, preferred to that in the farther distant future.” Now given what you said, that we can only know what we prefer now, then how can we reach Mises conclusion that we prefer sooner satisfaction to later satisfaction? If we don’t know what we will want in the future, then how is it possible to prefer something sooner to something later? If we only know our present value stack, then how can we conclude anything about our future preferences a priori?

96

If you want to do A (i.e. if A resides at the top of your value stack), “other things being equal” (i.e. your value stack remains unchanged), you prefer to do A sooner rather than later. You prefer to do whatever is at the top of your value stack to not doing it (i.e. to postponing it, because in an uncertain world, things don’t stay equal for too long). If spending $100 on dinner (action A) is at the top of your value stack now, someone (a borrower) must induce a change in your value stack by offering you large enough interest (or a massage) by which it would displace action A from the top of your value stack.

97

Perhaps that’s why reading MES might be easier than HA. He explains there. You’re attempting to estimate your future psychic revenue in the present.

98

If you want to do A (i.e. if A resides at the top of your value stack),

That is, if I want to do A now.

“other things being equal” (i.e. your value stack remains unchanged), you prefer to do A sooner rather than later.

That is, if I can’t do A now I will then want to do it at T+1, and if I can’t do it then, I’ll want to do it T+2, etc. This qualification makes the first proposition completely unnecessary. I can simply skip it and say that if I prefer to do something at T+1 and I can’t also do it at T+2, then I prefer it sooner rather than later.

You prefer to do whatever is at the top of your value stack to not doing it (i.e. to postponing it, because in an uncertain world, things don’t stay equal for too long).

That is, I prefer to do now what I prefer to do now.

If spending $100 on dinner (action A) is at the top of your value stack now, someone (a borrower) must induce a change in your value stack by offering you large enough interest (or a massage) by which it would displace action A from the top of your value stack.

Are we then assuming that lenders prefer to spend all of their money now? If they didn’t prefer to spend their money now, then the borrower wouldn’t have to pay interest, since interest payments are merely incentives to prevent the lender from spending his money.

This doesn’t seem to hold a priori. I prefer to save money for my retirement quite regardless of whether I earn interest on it. Yet, this doesn’t prevent me from earning interest on it. I can imagine a world where every lender prefers to save their money for retirement and yet refuses to lend it out without interest. I mean, what if the borrowers vastly outnumber the lenders? Wouldn’t the competition among the borrowers drive up the interest rate quite regardless of whether the lenders wanted to spend their money now?

And does this explanation of interest as some sort of compensation hold true for other forms of “surplus value”? Do I pay rent because my landlord would prefer to sleep in my apartment if I didn’t pay him not to?

99

Imagine I have $1000 that I prefer to save for my retirement and not to spend now. Borrower A comes up to me and asks if she can borrow the $1000. Then, before I can lend it, Borrower B comes along and says that she wants to borrow it too. To prevent Borrower A from getting the money, Borrower B offers to pay 5% interest on the loan. Borrower A is unwilling to pay more than 5% for the loan, so I loan the money to B.

No recourse to the time preference of the lender then is needed to explain this interest rate. Spending the money now was never at the top of my stack. So the interest can’t be seen as compensation for the lender’s loss of spending. In fact, if the interest is supposed to serve as compensation, then it should be paid to Borrower A and not to the lender. After all, it is Borrower A who wants to spend the money now and would be spending the money now had she not been prevented from doing so by Borrower B.

Mises’s theory breaks down once we assume that the demand of borrowers exceeds the supply of lenders. This of course is true of the real world. Despite what he says, when it comes to interest, Mises in fact assumes conditions of equilibrium.

100

No recourse to the time preference of the lender then is needed to explain this interest rate. Spending the money now was never at the top of my stack. So the interest can’t be seen as compensation for the lender’s loss of spending. In fact, if the interest is supposed to serve as compensation, then it should be paid to Borrower A and not to the lender. After all, it is Borrower A who wants to spend the money now and would be spending the money now had she not been prevented from doing so by Borrower B.

But borrower A doesn’t possess the money now, and never did, and is therefore incapable of having ‘spend $1000’ at the top of her stack (At least, the $1000 that you might have lent her).

After all, A might have wanted to spend $10000 right now, whereas there might be 9 other pension-savers aside from you that preferred to keep their moolahs in the mattress. Is she now owed interest from the mattresses? What if those 9 other people lent the money to C through K? Would A through K then be owed the interest of each of those other loans that they might have wanted but were unable to secure?

You may prefer to save those $1000 for your retirement at the moment you made the loan, but that does not guarantee that those preferences will remain static during the duration of the loan, nor does it mean that A will have the ‘spend $1000’ preference as a constant for the duration of B’s loan either.

I don’t really see how Mises’ theory breaks down here. Does it assume that interest is to be paid purely due to time preference with no regard to contract?

At most I feel your example shows that in this case the interest is not only indicative of your personal preference (prefer to lend out $1000 at 5% interest rather than keeping it in the mattress or lending it out at lower rate) but also indicative of B’s preference ($1000 dollars now with the 5% rate etc.), however this would surely already be implicit from a praxeological standpoint?

Indeed, if 5% interest was not your time preference, what is your time preference, and why didn’t you follow that preference instead. (That is, why didn’t you counter-offer with a lower or higher interest rate that is your preferred rate or simply refrain from lending out your $1000?)

In the same vein, if a carrot-monger wants to sell his carrots at $1 a pound, and is offered $5 a pound by some eccentric philantropist with a serious craving for carrot, is someone owed something by the carrot-monger? How about the housewife who was willing to pay $2, but not $5? Is she then owed 3 dollars and the carrot-monger $2? She surely had a higher time preference for carrots than the carrot-salesman.