Bob Murphy linked to an an interesting paper by Mihai Topan and Cristian Paun where they are trying to provide the right arguments in defense of the pure time preference theory of interest (PTPTI). Despite their ingenuity, I still think that PTPTI is indefensible. In this post I’ll first provide a general argument against the theory, then rebut the main argument by Topan and Paun and conclude with the contours of an alternative theory of interest.
Topan and Paun realize that PTPTI can only be based on a ceteris paribus claim, i.e. that humans ceteris paribus prefer to achieve their goals earlier rather than later. The problem is that this statement implies that a choosing agent may evaluate an alternative abstracting away from the particular circumstances of time and place. For example if a person is choosing whether to save money in summer to mountain-ski in the winter the claim of Topan and Paun would be that if only it were possible for an agent choosing in summer to save money to ski or not to ski in summer, she would not save the money. This seems highly counterintuitive. Besides, there is a much stronger example against the ceteris paribus claim. This claim seems to suggest that if human beings were able to get all the satisfaction in life at a single moment they would do that. Actually, perhaps, these two counterarguments are two sides of the same coin: ceteris does not ever seem paribus.
But the argument I have provided may not sound as compelling as an argument for PTPTI that Topan and Paun marshal. They claim that accepting the negation of PTPTI leads to accepting the idea of conscious non-action which is, according to them, a contradiction in terms. First of all it is not clear whether the negation of PTPTI has such an implication. After all, not all actions are immediate, they may consist in long-term plans where it is only possible to start outward activities only at a certain point in the future. Perhaps, Tupan and Paun might respond by saying that this is covered by the ceteris paribus clause and that the problem only arises when a person chooses to achieve a goal later rather than earlier although there is nothing that makes achieving it later more preferable to achieving it earlier. Here we again face the problem with the ceteris paribus clause. The very fact that an agent chooses achievement of the goal later rather than earlier must by definition (of choice) mean that she prefers future to the present for some reason. And the upshot seems to be that both the universal statement of time preference and its negation are meaningless from the real-world perspective. Human beings never choose between present and future. They only choose between particular present goals and particular future goals.
Now, what alternatives are there to PTPTI? I think that the alternative must be a monetary one but not exactly the one that Bob Murphy seems to propose in his dissertation. We should start from a situation where an agent needs to be induced to give up a sum of money, say, X dollars. Since money can be exchanged into any other marketable good when abandoning X dollars an agent abandons the most preferable bundle of consumption goods she may acquire with X dollars (not for X dollars because the loss of X dollars may prevent the agent from using X dollars to buy something that costs X+Y dollars which is to her the most important alternative use of X dollars). In order to induce the agent to abandon X dollars, she must be proposed in exchange a future satisfaction that is more important to her than the satisfaction that she must abandon now by abandoning X dollars. But to the extent that this future satisfaction is achievable by means of a currently existing consumption good, the only way the agent may not chose this good over her current top alternative use of X dollars is if this good either costs more than X dollars at the moment of her choice or will, according to her expectation, cost more than X dollars in the future (if it is not practicable for the agent to buy the good in question immediately to store for the future). Thus, the only money-involving way to induce the agent to give up X dollars now is to to propose to pay her in the future X+Z dollars. Hence, the interest rate.