(Continuation of Lectures on Advanced Macroeconomics )
As I read more and more about effective demand, I got more and more confused — how can I explain this concept to my poor students, if I don’t understand it myself? There are a huge number of articles with different and conflicting views and interpretations of this concept, which Keynes describes as being central to his theory. Let me proceed to clarify the insights that have resulted from struggling with this material, and going through many iterations of revisions in terms of how to make sense of this theory.
Keynes and followers — both the Hicks-Hansen-Samuelson variety, as well as true blue post Keynesians — argue that it is deficiencies in the Aggregate Demand which lead to the unemployment equilibrium which is central to Keynesian economics. Stated in very simple terms, the argument can be phrased like this. The process of production generates factor incomes. These incomes are exactly the source of the demand for the product. If all the income generated is always spent on purchase of products, then the aggregate demand will exactly equal the aggregate supply — this is Say’s Law. In this case, there is no concept of shortfall in aggregate demand which could lead to unemployment.
However, Keynes and his followers deny the equality. They argue that some portion of the factor income could go into savings, thereby lowering the aggregate demand. Now the aggregate demand could be greater or lesser than the aggregate supply. An equilibrium would occur when the two are the same, but there is no guarantee that this equilibrium would occur at full employment. The standard diagram used to illustrate this idea is given below:
For our purposes, we can assume a direct proportionality between N, the amount of labor employed, and Y, the value of the output produced. Furthermore, it makes no difference for our argument whether prices are fixed or flexible, so let us assume them to be fixed for simplicity. The X-axis just measures the value of total output Y (equivalent to N, rescaled, since the two are directly proportional). The 45 degree line just measures the total factor income generated by the production of output worth Y — by definition. The consumption function is C=a+bY, where a>0 and b<1 is the marginal propensity to consume. When Y is low, the laborers/consumers demand is greater than the total product. However, since the slope b is less than 1, this consumption demand must intersect the 45 degree line. At points beyond the intersection, we have deficient demand. Both deficient and excess demand would set in motion processes to eliminate the disequilibrium so that equilibrium would occur at the point at which the demand for consumption generated by factor incomes is exactly equal to the value of the total output. This intersection can occur at a point at which full employment does not occur, demonstrating the existence of an unemployment equilibrium — which is the CENTRAL goal of Keynes.
According to Keynes, classical economics is based on Say’s Law, which requires C=Y — the consumption function has a=0 and b=1, so that the consumer demand is exactly the same as the factor income, and there is zero savings. In this case the consumer demand is exactly the same as the aggregate supply function — both are the 45 degree lines, and equilibrium can occur at any point — however much is produced, it will generate exactly enough demand required to purchase and consume it. In this case, standard economic forces will automatically drive the economy to full employment. Keynes, and post-Keynesians argue that consumer savings will disrupt the operation of Say’s law, and create a divergence between the demand for consumption at a given level of factor income Y (which will be C=a+bY) and the supply of goods at the same level (which will be Y). Because of this violation of Say’s Law, equilibrium will not occur at all possible levels of production Y. It will occur only at one particular point of intersection, and that point could easily be an unemployment equilibrium.
This idea actually works in a one period static economy, which was the mode of analysis used by Keynes. However, despite a lot of effort, I was unable to make it work in very simple dynamic extensions of the same economy. After many trials, I realize that this savings argument is much more subtle than it appears at first blush. The problem in the dynamic, multi-period setting is actually very simple. Today the consumers/laborers SAVE a portion of their income S=Y-C, reducing the aggregate demand. However TOMORROW this same saving will be available to the consumers as additional amount of money, over and above the factor income. When we talk about saving for the future, then we cannot ignore what happens in the future as a result of this saving. Furthermore, in a dynamic setting, except for the very first period, every period has a past as well as a future. Let M(T) be the total money in the hands of the consumers in period T. This amount is split into C(T) and S(T), the consumption and savings in period T. Now, in period T+1, the money holdings in hands of consumers will be M(T+1)=F(T+1)+S(T) — the factor income PLUS the amount the consumers saved. The consumption function should be defined using THIS amount, and not just the factor income amount — if savings plays no role (as it does not in a one period static analysis) then the model is not logically consistent.
Once we realize that for logical consistency, savings MUST enter into any dynamic analysis, we are led to the realization that it MUST also enter into even a static one-period analysis. This is because we must have M(T)=F(T)+S(T-1). The money in the hands of the consumers today must be the factor incomes in the current period PLUS the savings from the previous period. It is logically inconsistent for a model which has savings in current period to ignore the savings from the previous period. But, as soon we put in the CRUCIAL missing variable S(T-1) into the analysis of the current period, the savings gap appears in a very different light — as we shall soon see. To the extent that today’s saving reduce current demand, yesterdays savings offset this by adding to demand. Let S* be a normal level of savings. Then in long run equilibrium S(T-1) = S* = S(T), and so the savings gap created by present savings will be exactly made up for by savings coming in from period T-1.
As this analysis demonstrates, normal levels of savings S* do not create deficiencies in aggregate demand, and hence cannot create unemployment. However, abnormal levels can indeed create such deficiencies, and lead to the kind of unemployment that Keynes wanted to explain. To see how this can happen, suppose we are at a full employment equilibrium, with shortfall in aggregate demand from saving S* being exactly compensated for by the increased income due to savings S* from previous period. Now suppose that there is a catastrophic crop failure. The Landlords have cushions, but the laborers draw down their savings and go into debt to survive. Come planting season, the landlords note the general misery in the land, and how the general population is in debt, and times are tight. They wonder whether they should go to the trouble and expense of producing a huge amount of crop. Who will buy it? Would it not become surplus, and drive down the price of crops? In this situation, they are likely to restrict production to have enough for self-consumption, and a moderate surplus for sale. The concept of restricting production to get good prices on crops is well known. The aggregate demand has gone down because the normal savings cushion S* of the laborers has been wiped out. Now, when they are paid wages, instead of using it to buy corn, they will use it repay debts, and to build up their savings back to the normal levels. This abnormally high savings is what causes the collapse of aggregate demand. This corresponds exactly to the analysis of Atif Mian and Amir Sufi in the House of Debt. See my brief analysis and summary in “Why does Aggregate Demand Collapse?“.
The relation between debts and depression was emphasized in the debt-deflation theory of Irving Fisher. Even though Keynes was aware of the links, he chose not to emphasize this in his General Theory. Making the connection explains why “helicopter money” would go a long way towards ending the depression. A moratorium on debt repayments, and an infusion of money to rebuild savings and restore aggregate demand create full employment. This would then create prosperity which would allow people to pay off debts. The Aggregate Demand Depression leads to loss of jobs which further inhibits the ability of people to pay of debts, creating a vicious cycle, as has been noted by many.