Hulsmann's 'A Theory of Interest'

Economics Economics Questions
21

I. Ryan,

Putting the drinking example aside for a moment, there are clear cases where means and ends are distinguishable, and there are clear cases where means and ends are indistinguishable. So it makes sense to give names to these two different types of actions: labor and leisure. Ironing and dancing are the two clearest examples. It makes no sense to distinguish the end of “to be dancing” from the means of “to dance”. But we can distinguish between the end of “to have crease-free shirts” and the means “to iron my shirts”. If I could “click my figures”, i.e. just “will it to be so”, that my shirts are crease-free, I will take this option; I will choose the end over the means.

This is entirely independent of time and in a sense independent of uncertainty as well. Suppose ironing my shirts will take 10 minutes and have a 95% chance of success. Then if I were given the option to click my figures, wait 10 minutes, and then there is a 95% chance that my shirts will be crease-free, I will take this option. It saves me the effort of performing the action; I get my end ‘for free’ that way. Why act if you can get your end without acting?

The following might help explain where time and uncertainty come in: define a future-oriented action as one where the end is achieved at the end of the action, and define a present-oriented action as one where the end is achieved during the action. Now there are four possible sets of actions:

  1. Present-oriented leisure. [Dancing, etc]
  2. Future-oriented leisure. [Empty]
  3. Present-oriented labor. [“Having you hear my voice”]
  4. Future-oriented labor. [Ironing, etc]

Since with leisure, the means and ends are indistinguishable, future-oriented leisure is a contradiction in terms, so set 2 is empty. Clearly, there are many examples of actions that fit in sets 1 and 4. The question is: does anything fit in set 3? This is the significance of Hulsmann’s example of the end of “having you hear my voice”. It’s labor, but its also present-oriented. Does this action give rise to originary interest? If yes, then time must be irrelevant for originary interest, with labor being the key factor.

What about uncertainty? These are my own thoughts and they are still quite tentative…

I do not think that all actions are subject to uncertainty (uncertainty being defined as being unsure whether the action will result in the end being achieved). Leisure actions are not. How can I possibly fail to achieve my end of “to be dancing”? Either I dance, and my end is achieved, or I don’t dance and I never acted towards my goal in the first place. I’ve either achieved my end, or not acted towards it; I can’t possibly act towards my end and fail to achieve it. The same with whistling, or reading (to the extent that these are considered leisure activities); nothing can go wrong with them.

So uncertainty doesn’t apply to leisure. It definitely does apply to future-oriented labor. I could easily fail to achieve an end of this kind. But what about present-oriented labor? Can I fail in my goal of “having you hear my voice”? I would say that I can. My goal is not guaranteed to be achieved as soon as I begin the action (as it is with leisure). So uncertainty does apply to this kind of action, even though it is present-oriented. So uncertainty, like originary interest, only applies to cases of labor, and applies to all cases of labor, independent of the aspect of time.

So I think in a way you are right that uncertainty is key here. But I make two claims about uncertainty that seem to be at odds with Mises and Rothbard: (1) that uncertainty only applies to labor actions, and (2) there are some labor actions for which time is irrelevant, but these are still subject to uncertainty, and give rise to originary interest. In other words, “actions which are subject to uncertainty” is equivalent to “labor actions”, but uncertainty is only one of a number of features of labor actions; uncertainty plays a part in the means-ends value spread, but is not the whole of it.

22

Don’t neglect the uncertainty about what ends you may (or may not) have over the next year. Higher uncertainty would drive you to demand higher interest on the current means ($100) you are lending out for the year.

Z.

23

Then why are you calling one the means and the other the ends?

But then wouldn’t clicking your fingers become the means?

But doesn’t clicking your fingers come before the crease-free shirt, and wouldn’t it be possible for clicking your fingers to not lead to the crease-free shirt?

But that’s not picking the end over the means; it’s picking one means over a different means.

Not for free, but for less.

But how could there even be an end without any action?

It all boils down to the fact that it doesn’t make any sense to talk about choosing something without there being a means. It’s always a choice between one means and another, never between a means and an end. In your example, you fell back into talking about choosing one means over another (clicking your fingers over ironing). I can’t figure out how to interpret the words “choose the end over its means”. As far as I know, it’s totally meaningless.

What if you imagine yourself dancing in front of your chair, but then fall on your face as you try to get up?

But don’t you hear my voice slightly after I start trying to speak?

(I can’t figure out how to interpret the difference between the “present-oriented” labor and the “future-oriented” labor without simply seeing the former as simply being a quicker production process than the latter.)

What’s the other part?

(Please forgive me if you already spelled it out.)

24

The end is an imagined state of affairs, or arrangement of the environment. The means is the element of the environment that can be used to alter the environment in such a way as to bring about the end. By the way, how would you describe the means and ends of this action?

By clicking my fingers I was talking metaphorically. For normal actions, we 1) consider various ends and pick one, 2) consider various means to achieving the end and pick one, 3) perform the action, then 4) attain the end. I am saying what if we could cut out 2 and 3 and just jump straight from 1 to 4: just will something to be so. You may object that in the real world we can never do this, but there must be some sense in which, prior to any action, we consider each of our ends as if we could attain them costlessly. We must say, ‘well if I could will end A into existence, or end B, I will pick end A’, and then we go on to consider the means we will have to use to bring about A or B, which is the necessary secondary step since we cannot really just will either of them into existence costlessly.

Maybe you’re right to say that Hulsmann and I are still really just comparing two means. Attaining crease-free shirts by means of ironing, or by means of “willing it so”, i.e. costlessly. But this has now become semantic. When we say “choose the end over its means” you could translate that as “choose the end using costless means over the end using costly means”. In fact, when you put it like that, it becomes obvious that one would always choose the least costly means. This is really another way of wording Hulsmann’s insight.

Ends exist in our heads before action.

Then I would never have started the action, so I didn’t fail to achieve it. I failed to achieve my end of “be in position to start dancing”, which is a labor action and hence subject to failure.

You are right. But there is still a sense in which that example is different from both dancing and ironing.

It shares some characteristics with dancing - namely that it is in a sense “continuous”, the end is achieved during the action, and if we say that in fact there is a momentary gap during which time my voice travels through the air, then we would have to say that I was making “continuous actions”. I sing a bit, achieve my goal, sing a bit more, achieve my goal again, etc. This is possibly a more precise definition of what I mean by present-oriented. It contrasts sharply with a future-oriented act like ironing, where the goal is only achieved once, at the end of the action.

And yet my goal of having you hear my voice is also like ironing in that we can make sense of achieving it costlessly, and would pick that option if it were available.

Perhaps my terminology is objectionable, but I think there are some really important concepts here. Do you have names for the three distinct kinds of actions I am talking about?

25

I’ve re-watched Murphy’s lecture and re-read his paper. Most of it is him criticising time-preference theory, some of which overlap with Hulsmann’s criticisms. Murphy considers the pure theory of time preference to be “either trivial or false”, in that it is either false as an explanation of interest rates, or a triviality which explains nothing, depending on how time-preference is formulated. I have no complaints about his criticisms.

He does not go into much detail about his own explanation for interest, except to say that he doesn’t think interest makes much sense without money. At the end of the lecture, he is asked: “Are you saying that interest is just a monetary phenomenon? What happens in a barter economy?” He answers tentatively with “I’m not sure, I don’t want to deny that [interest exists without money], but it works through money” and that (paraphrasing) without money there would be no “one interest rate” but instead an “apple interest rate”, an “orange interest rate”, etc. In other words, a seperate interest rate for each commodity. So if by interest we mean the interest rate on money, of course there’s no such thing as interest in a barter economy.

I don’t think he’s wrong necessarily. I think he has hit on the point that, like profit and loss and calculation, which can only really be done with money, a single interest rate is also something that requires money to exist. However Murphy doesn’t really have an explanation for why interest profits are not arbitraged away altogether, which is the whole point of an interest theory.

26

If this were the whole explanation for interest (i.e. opportunity costs, uncertainty, risks) then interest would not exist in the Evenly-Rotating Economy (ERE). Interest profits would have been arbitraged away. These things are certainly components of the interest rate, but they cannot be the whole thing, or the “root cause”. For a root cause, we need something that exists even in the ERE, and we have two candidates: time-preferences, and means-ends value spread.

27

This is why a pure-time-preference theory of interest ultimately fails to explain all phenomena, and why a comprehensive theory of interest must incorporate productivity alongside time preference. For example, technological innovation increases the demand for investment and therefore loanable funds, which puts upward pressure on the interest rate. But I fail to see how this corresponds to a higher time preference. The fact of the matter is that it elevates the marginal productivity of capital. This doesn’t mean that Bohm-Bawerk and Mises’ critique of the productivity theory is entirely invalid; it’s still true that productivity alone does not guarantee a positive net return on investment.

Bob Murphy’s criticisms are entirely spurious and he merely regurgitates Piero Sraffa’s incoherent argument. Time preference doesn’t vary amongst different economic goods; it expresses the ratio of exchange/demand between two classes of goods, namely producer and consumer goods. It refers to the fact that if I value a good at all, then I must necessarily prefer possessing it now rather than later, all other things equal. If you drop the ceteris paribus condition then it will appear as though, but only superficially so, that different goods have different rates of interest.

Additionally, it is not true that a uniform rate of interest is inconceivable in a barter economy (though it is impossible). If we add the assumption of perfect information then we should expect arbitrage to equalize/normalize interest rates across the board. Of course, arbitrage is entirely impossible in a nonmonetary (barter) economy, where economic calculation simply doesn’t exit. Thus, we should expect varying rates of interest in a barter economy amongst different economic goods, but we should also expect varying rates of interest for producers that produce physically identical goods. In other words, we should expect two Macintosh apple producers to have different rates of return in an economy where economic calculation is impossible. The fact that economic calculation is impossible in a barter economy does not pose a problem for Austrian capital theory.

And finally, it has long been understood by Austrian economists that the market, i.e., money rate of interest, is partially determined by monetary conditions (liquidity preference). This is not an anomaly for Austrian capital theory. In other words, the fact that the money (market) rate of interest rises during a liquidity crisis is no more problematic for Austrian capital theory than the fact that the money rate of interest falls during an artificial credit expansion. Hayek writes about this extensively; he calls it “secondary deflation/phenomena,” where the money rate of interest actually rises above the natural rate due to an elevated demand for money. But the natural rate of interest is entirely independent of all monetary phenomena.

Thus, it is obviously true that you can’t have money prices (explicit expression of opportunity costs) in a barter economy, but the opportunity cost, in the case between current and producer goods, still exists. It is this purely theoretical distinction between the natural rate of interest (unseen) and the market rate of interest (seen) which allowed the Austrians to deduce their business cycle theory, which is the only theory that can explain what we see today.

28

Well the problem is that the loanable funds market does not determine the rate of interest, its only one the many markets that reflect the discount on future goods.The interest rate is determined by the supply of P.G for F.G and the net demand for P.G by supplying F.G (the original factors who work on the circulating capital). Their rivalrous bidding that eventually equalizes in the ERE determines the discount on P.G and the price spread. If this price spread has not changed, then neither interest rates nor time preference have changed. If the price spread between stages (the interest rate) has not changed, what is the incentive for producers to borrow at higher rates of interest?

Additionally, the “entrepreneurs” demanding the loanable funds aren’t really the savers/investers here, the actual people who save (the people who buy stocks or loan out money) are the capitalist investors (and the true entrepreneurs in the real world), the “entrepreneurs” who borrow the money are really just hired agents of the savers

29

Yes, that would be a more precise summary.

30

Esuric,

It sounds like you agree with the pure theory of time-preference as an explanation for originary interest (i.e. time-preferences are the primary determinant of the natural interest rate), but that the theory is incomplete as an explanation for the market interest rate. That it needs to be supplemented with discussion of secondary phenomena, like changes in the demand for loanable funds. So that in the ERE, where they are no technological innovations, and no changes in liquidity preference, the market rate of interest will equal the natural rate of interest, which is determined solely by time-preferences (i.e. the ratio of demand between producer goods and consumer goods). Is that a fair summary of your position?

Assuming it is, I would ask that we leave all considerations of secondary phenomena out of this discussion, and just focus on what determines the natural rate of interest, i.e. what causes originary interest. Do you find ALL of Hulsmann’s and Murphy’s criticisms of PTTP to be spurious? Maybe you could respond to I Ryan’s critique in the first thread that I linked to?

31

Actually, I don’t find the terms “means” and “ends” very useful, but I can try to explain how the action plays out anyway.

Okay, let’s say that I’m sitting in a chair, but then get up and start dancing in front of it. Still in the chair, I would imagine myself getting up and then dancing in front of it (the sequence of 2D images of the scene in 1st or 3rd person, the bodily sensations of the way that I would contract my muscles during the action, and so on) all very vividly. But then what would follow? If I was correct about everything, I would simply get up and go through all of that stuff. I imagine it (the end?), and then do it (?).

Now, what would it be like to prefer the end to its means? In the ironing example, I do the same thing as in the dancing one. I imagine myself going over to the ironing board, firing up the iron, putting the shirt on it, ironing it, and so on all very vividly. And if I was correct about everything, I would simply go over and do all that. But wait. I think that I see a difference. So let’s start over. In the ironing example, I might start off by simply imagining the end result (the crease-free shirt). I don’t have to imagine the whole action right from the beginning. In the dancing example, because the whole thing is the end, I must imagine it all right from the start. But in the ironing one, the end is only the last part. Anyway, let’s say that I begin by merely visualizing the end result of the whole thing (the crease-free shirt). And after that, I might go back and imagine the last 1/8. But that’s not far enough back to start the action yet! So then I might go back even further and imagine the whole thing. Now that would be enough to get me to act, right?

So for the ironing example, the end is me imagining the crease-free shirt, and the means are me imagining the rest of the stuff which would lead up to that, right? But for the dancing one, it’s not clear what the means and ends are. It’s as if the whole thing is the means and ends. Or maybe it’s that they alternate extremely quickly (like what you were calling “continuous actions”). But, then again, maybe this is simply an example where the means-ends framework breaks down? I mean, there’s no need to hold onto it as a fetish or anything (not saying that you’re doing that!). If it doesn’t live up to our expectations, we could simply scrap it. Well, I’m not sure yet whether we should or not, but I’m certainly thinking about it. It seems to oversimplify our action or something. Or maybe not. Whatever.

So to get back to the original question, what would it mean to prefer the end to its means (to achieve it costlessly)? As I simply imagine moving my arm upward and it happens, I would simply imagine the shirt go crease-free and it would. I wouldn’t have to go through all of the other stuff (the means?) because I would simply imagine the end result and there it would be. I could cut out the whole rest of it. But would we say that the end result would be the means? Maybe that me imagining the end result would be the ends, and me witnessing the end result would be the means? Well, who cares anyway? The whole means-ends thing seems broken anyway. If we stick to simply talking in more realistic terms (imagining stuff, things happening, and so on), then it all gets a lot clearer.

Anyway, I think that the distinction between the ironing example and the dancing example is this. In the ironing one, you only care about the ironing because you care about the crease-free shirt. If you didn’t associate the ironing with the crease-free shirt, and you’re imagining the crease-free shirt, your mind would never transition to imagining the ironing. But in the dancing, you care about the whole thing. Or at least that was the supposition. But could I really care about the whole thing? Couldn’t we say that I only want to dance because I expect that to lead to something else? Are there really any situations where you keep something up for a long period of time simply because you’re caught in a stream of seeing each part of the event as an end in itself? No idea.

And what about the example where I’m trying to get the other person to hear my voice? Well, let’s say that I gauge whether the other person hears my voice by whether they move in a certain way. I imagine me speaking for a split second, them moving like that, speaking for another split second, them moving like that again, and so on. It’s like the means and ends (if we insist on using those terms) alternate in a super fast pace. As you said, it’s like it’s a continuous action. So I only see the 2 kinds of actions: Continuous and discrete (if I may use that term). And that distinction is just a matter of degree. If there’s a lot of alternating between means and ends (if you know what I mean), then it’s continuous. If not, it’s discrete. But perhaps I still haven’t managed to grasp the significance.

32

I. Ryan,

I have been considering whether means and ends are appropriate terms with respect to continuous (I prefer this term as a replacement for how I was using “present-oriented”) actions ever since Z said that “the means/end and sooner/later paradigms are different aspects of one and the same thing”. I have been arguing that with the example of dancing, means/end are not sooner/later, but inextricably intertwined, occuring in the same physical event. So I can either disagree with Z, and say that means/end is not “one and the same” as sooner/later, or I can agree with him, and abandon the terms means/end when talking about continuous actions. At the moment I am tending towards the former, keeping the terms means/end as applicable to all actions, and saying that the means/end concept and the sooner/later concept are not the same when discussing continuous actions.

How you’ve described it above is actually different from the continuous/discrete dichotomy. Here you are talking about the labor/leisure dichotomy. Ironing is labor because you care only for the product of the action: the crease-free shirt. Dancing is leisure because “you care about the whole thing”. Ironing is also discrete in the sense we’ve discussed, and dancing is continuous. So here we are comparing a labor/discrete action with a leisure/continuous action. Now if it is the case that the set of labor actions is identical to the set of discrete actions, and the set of leisure actions is identical to the set of continuous actions, then there would be no need for this discussion: time would be bound up with means/ends, and PTTP and Hulsmann’s theory would be equivalent. But my thesis is that this is not the case.

The example that shows they are independent of each other is the singing example.

You have agreed that it is a continuous action. So if it is a labor action, then I have shown the existence of a labor/continuous action. It clearly is an example of a labor action, because I don’t “care about the whole thing”, all I care about is that “you hear my voice”. We can conceptually seperate the end from the means; I would rather have the end using costless means, i.e. I wish that you could hear my voice, without me having to sing, just like I wish I had crease-free shirts, without me having to iron. So in this sense its equivalent to ironing, a labor action, but also a continuous action, like dancing.

What’s the significance of all this? We are trying to explain originary interest. There are three kinds of actions: leisure/continuous, labor/continuous and labor/discrete. What kinds of actions give rise to originary interest? If originary interest is about value inequality over time (i.e. time preferences), then it applies to the set of discrete actions. If originary interest is about value inequality between means and ends (i.e. Hulsmann’s theory), then it applies to the set of labor actions. The existence of a labor/continuous action shows that these sets - discrete actions and labor actions - are not identical. Their common coincidence is merely incidental, although this does explain why time-preference theory seems to explain originary interest so well. It also reassures us that things like ABCT are not in doubt; time-preferences can be taken as a kind of shorthand meaning “the preference for ends over the means to obtain those ends”. But at the level of fundamental principles, PTTP should be abandoned as faulty and replaced with means-ends value spread theory, which is firmly rooted in the action axiom.

33

trulib,

I think that you’re still taking Hulmann’s idea that there’s a “value inequality between means and ends” for granted. I challenged that idea in my first post in this thread, but regardless of whether I was successful, I still have no idea what the hell it would even mean. I know that you admitted that you and he might simply be comparing 2 different kinds of means, but in your post above, you go back to the original way of saying it (“value inequality between means and ends”). I’m not sure how the theory would hold up if you really took it seriously that it’s meaningless to talk about preferring the ends to it means. I know what it would mean to prefer one means over another (maybe clicking your fingers rather than having to stand there ironing it for 10 minutes), but I have absolutely no idea what it would mean to prefer the end over its means. Again, I know that you already conceded that you and he might be merely comparing 2 means, but then I don’t know why that applies to anything. Sure, you would probably prefer clicking your fingers to standing there ironing for 10 minutes (or would you?), but I’m totally confused about how that would explain where originary interest comes from.

Actually, let’s get back to the dancing example. Why would I dance anyway? Maybe I would feel better afterwards or something? I think that you’re taking it for granted that we could somehow see every part of it as an end in itself. What if it’s simply that engaging in those movements would make you happier once your done? But wait, aren’t you happy the whole time? Yeah, but that happens in all kinds of “labor actions”. I mean, I might really enjoy cooking even if I’m only doing it for the meal, and it might be the same way with the meal itself even if I’m only doing it for the fuel. If you’re simply dancing for the happiness of having danced, why wouldn’t it make sense to say that you wish that you could be happy without having to dance? And the same for anything else. Or it might be that each submovement in the dance makes you happy, which would mean that it’s simply an action where the means and ends alternate in a super fast pace.

But enough of that. If there’s anything here, certainly you could re-state it in a more realistic manner? I find that the terms “means” and “ends” are totally confusing and make everything sound a lot simpler than it really is. In fact, I’m not even sure what the hell they’re supposed to mean. AE often uses the word “means” to refer to something like a hammer, but then turns around and uses that term to refer to something like ironing? Well, is the ironing the means, or is it the iron? It would clear up a lot of stuff for me if you could try to re-state Hulsmann’s theory in some more realistic terms (“imagine”, “see”, “pleasure”, “pain”, “feel desire”, and so on). I’m still not sure where the value inequality is. I mean, I’m even starting to wonder why I would pick willing the shirt to go crease-free instead of having to iron it. Aren’t there a bunch of considerations that we’re missing? If we were in a world where longer production processes were safer, didn’t use any extra energy or anything, and so on, wouldn’t I pick the ironing? I’m starting to question whether it’s even self-evident that we would pick the one over the other. Can’t we imagine a world where I really would pick the longer production process (the ironing instead of willing)? And then what?

Anyway, I still think that the answer is time preference. I think that the PTTP is broken, but that doesn’t mean that I abandoned time preference or anything. But what is time preference? How would I reduce it to my perceptions? I think that the answer is buried somewhere in this post that I wrote to Azure. It’s where I discount the pain/pleasure for one of my further future selves in comparison to one of my nearer ones. But how would that happen? And why? Well, all that I can think of is where you’re not certain that your further future selves are even going to exist. If you don’t expect to make it through the night, your time preference will shoot through the roof. Maybe normally you might want to postpone eating until the next morning to sleep better or something, but with your uncertainty that you’re even gonna be there in the morning, you indulge right then. That’s really all that I can think of. Do you have any ideas? Can your time preference ever come from anything but being uncertain about whether you’re still gonna be around? (And by the way, that would explain why your time preference can’t be negative. You’re either 100% certain or less. You’re never 101% certain or anything like that.)

34

Ends are the final states aimed at.(ironed clothes)

Means are utilized to achieve them. be it objects (irons) , or transformations/labour(hand movements), often both in combination(ironing)

Often it takes time for the process of using the means to achieve the ends to bear fruit, but as many have commented, in the case such as dancing where it is a pleasure in itself to be doing it, there need not be a lapse in time between means and ends, yet this does not invalidate the theory of means and ends which does not rely on seperation in time, only that a strict ordering must adhere if the means and ends are seperated in time.

35

What’s that even mean?

(I would understand if you said that dancing could be a means to the end of being happy or something, but I can’t figure out what it would be like for dancing to be “a pleasure in itself”.)

36

it just means that ‘dancing’ and ‘feeling pleasure’ are coincident for the person at that time.

This doesnt do very much more than simply distinguish it from other activities where ‘feeling pleasure’ lags behind necessary transformations required to bring about pleasure. i.e. the act of exercising so as to be fit for when one starts dancing…

‘pleasure in itself’ is just referencing the strong association between pleasure and the acts that bring them about concurrently.

for example, to say that you feel pleasure whilst dancing but that dancing is not your end, but happiness that dancing produces in you is just so much talk in which it is supposed that the dancing is seperable from the pleasure, what merits does such an approach offfer? doesn’t this rather assume that pleasure is just one thing, and that there are no varieties of pleasure, that pleasure brought about by different activities/events/objects are not differentiated but are entirely the same thing. I’m not confident thats right, and I don’t see the point of doing away with the language that draws attention to when means and ends are qualitatively more tightly bound then other cases when they are more greatly seperated.

37

Seems to me that all this talk of ‘pleasure’ etc is just a dead end, and not really relevant to what Hulsmann is trying to say. I think the distinction between an activity that is an ‘end in itself’ and another activity that is a ‘means to an end’ can be defined.

End in itself = When performed, there is no further effect arising from that activiity necessary to impute it with value, according to a subjective valuation. It is self-contained.

Means to an end = Would not be performed at all in the absence of a further effect arising from it, because it does not have sufficient value in itself (according to a subjective valuation of the performer) to warrant its performance.

I’m sure someone can come up with a better worded definition, but does that sound reasonable?

To answer the OP I think that Hulsmann’s theory is profound, but that a lot more work needs to be done on precisely defining ‘means’ and ‘ends’ and also on its implications for economic theory as a whole. I have always had severe reservations about ‘time preference’ as it never seemed to have the force of certainty that other praxeological insights have.

I found it interesting that Hulsmann brought up Mises use of the word ‘pathological’ when speaking about people who appear not to have time preference, I spent a long time pondering that section of HA, and trying to understand how someone as intellectually precise as Mises could dismiss this problem with such weak reasoning.