A Thought on Intertemporal Coordination

I haven’t ironed out my thoughts on intertemporal coordination, so I can’t offer a complete theory — even if the theory is wrong —, but I have some thoughts on the contention that since an act of saving is separate from an act of investment, there’s no process by which the former must necessarily lead to the latter. The simplest explanation for intertemporal coordination is that the rate of interest is the central lever: an increase in savings will decrease the rate of interest, inducing an increase in investment. The problem is that the relationship between savings and the rate of interest may not be so straightforward. At the extreme, for example, if all savings were to be in the form of cash, the rate of interest on loans through the banking system may be unaffected, leaving the quantity of credit demanded the same.

Quickly, the reason we care about intertemporal coordination is for solving the paradox of thrift. Economists generally agree that an increase in investment requires an increase in savings. Suppose that there is a fixed stock of means of production that can be used to produce either producers’ or consumers’ goods. What an increase in savings allow producers to do is to reallocate a certain amount of the means of production towards the manufacturing of producers’ goods, and away from consumers’ goods. But, the act of saving and the act of investment are committed to by two different people, and there has to be a process by which one is connected to the other. The paradox of thrift argues that, despite the necessity of savings for the sake of funding investment and economic growth, if there isn’t a process of intertemporal coordination, savings may precipitate a drop in income: essentially analogous to a business cycle. In fact, this is Keynes’ business cycle theory.

To some extent, most economists acknowledge that the problem of intertemporal coordination is not as simple as an adjustment of the rate of interest. If it were, there would be no need for the various institutions that have been developed to increase the efficiency of the intermediation of savings. Admittedly, some of these developments are meant to decrease variations from the mean — what we can call an equilibrium rate of interest —, including reducing risk premia and similar factors which may affect interest rates apart from their direct determinants. Other factors of this kind are the transaction costs faced when borrowers are looking for lenders, explaining the development of financial institutions and the various financial instruments. But, still, alternatives to an interest-focused theory of intertemporal coordination are hard to find, other than the belief that there is no (or that there is only a weak) process of intertemporal coordination.

One alternative is Hayek’s Ricardo effect. Rather than through interest rates, the Ricardo effect works through changes in relative prices. The basic theory is that a fall in consumption will increase the relative wage rate in the consumers’ good industries (which tend to be relatively labor intensive), therefore these firms will demand labor-saving capital goods.This will inspire the profit motive that leads to a growth of investment in the production of labor-saving capital goods. (For those familiar with trade theory, the Ricardo effect actually reminds me of the Samuelson–Jones specific factors model.) I think this theory has a lot of merit, and I certainly invoke it as one of my explanations for intertemporal coordination, but this isn’t the thought I currently have in mind.

I wonder if there is another, separate process of intertemporal coordination. Maybe because of the dominance of the loanable funds theory, we usually think of the process as going from an increase in saving to an increase in investment. Could it happen in the other direction?

Consider what interest is. The pure time preference theory of interest stipulates that interest is a ratio between the some amount of saved income and the value of future income necessary to accrue to these savings to induce sacrificing said amount of present consumption. I think about it like this: some given amount of income divided by the rate of interest tells us the amount of money that a borrower has to offer some lender to induce the latter to part with present consumption. If you prefer a liquidity interest theory of interest, the concept is still similar. This theory argues that the rate of interest is a ratio of some given present amount of money and the amount of money a borrower must pay a lender to induce the latter to part with liquidity. An important of the pure time preference theory of interest (I haven’t read anything to this effect in the literature on the liquidity preference theory) is that the rate of interest doesn’t decide the rate savings. Rather, interest is the rate, on the margin, at which we have to pay potential lenders to part with present consumption (or, more ambiguously, to part with present liquidity).

Interest is an emergent phenomenon, implying that it forms in the process of temporal intermediation. If we conceptualize this process in the form of a supply and demand graph, the loanable funds theory looks at shifts of, and movements along, the supply of savings. It doesn’t deal much with shifts of, and movements along, the demand curve. But, if we think of interest as the rate at which we have to pay someone to “part with present liquidity,” I think it makes sense to focus on the demand for savings. Changes in the returns that entrepreneurs expect in return for investment will change their willingness to pay for savings. In other words, it’s not the saver. This interpretation of the process, though, makes the rate of interest almost unremarkable.

One potential problem is that this seems to contradict the fact that the an increase in the stock of capital leads to a decreasing rate of return. But, an increase in the stock of capital is only possible if this additional capital is produced. The way I conceptualize it is that the current stock of capital factors into people’s time preference (we generally believe that as income grows, the marginal propensity to save will increase), and so while changes in the rate of profit in disequilibrium will induce lenders to change their willingness to pay, the “reference point” will always be the equilibrium rate of interest (the rate achieved at a stable intertemporal equilibrium, where profits are zero). As production increases and capital accumulates the rate of return falls, but the implied increase standards of livings leads to a fall in time preference, and therefore a fall in the equilibrium rate of interest (the “reference point”). The average rate of profit that accrues to each stage of production will also fall, as the number of stages increases with changes in time preference.

I think this reasoning complements Hayek’s Ricardo effect.  The changes in relative prices will create profitable opportunities of investment in the production of labor-saving machinery. These will lead willing and able investors to increase the rate at which they’re willing to pay lenders, allowing them to outbid competing borrowers (in a non-equilibrium world, no one is a pure price taker). But, if the Ricardo effect is already an alternative to the loanable funds theory, my thoughts here imply that the market rate of interest is almost a residual of the process of intertemporal coordination.