A NOTE ON THE PIGOU EFFECT AND THE UPWARD TURNING POINT*

Abstract

As soon as the increased real value of assets raises the consumption function sufficiently to start employment and output up again, prices and wages will cease falling, and the real value of assets will cease rising.… Thus the deflation of prices (which à la Pigou is supposed to continue to cause an ever increasing real value of liquid assets until full employment is reached) stops at the lower turning point. The force that is supposed to drive the economy into full employment already peters out once the lower turning point is reached.2 Hansen derived his criticism from the well-known article on the Pigou effect by Don Patinkin, “Price Flexibility and Full Employment.”3 Assuming a fixed stock of money, no adverse liquidity effects, and a flexible price level, Patinkin's argument runs, a falling price level will augment the real value of the stock of money and lead to a shifting of the consumption function to the left (less will be saved out of given alternative incomes). In effect, Hansen questions whether this shift will be large enough to re-establish a full-employment level of spending and, therefore, income. The extent to which an individual wishes to save out of current income for reasons other than the desire for future income is inversely related to the real value of his cash balances. If this is sufficiently large, all his secondary desires for saving will be fully satisfied. At this point the only reason he will continue to save out of current income is the primary one of anticipated future interest payments. In other words, if the real value of cash balances is sufficiently large, the savings function becomes zero at a positive rate of interest, regardless of the income level.4 Probably a diagram of the relevant functions at this point would clarify the arguments. In order to keep the ideas clear, let us assume straight-line functions. We start from a full-employment income, Y a , a result of consumption spending and investment spending from the functions C 1 and I 1 . The savings and investment functions are simply “rolled” 180° from their normal position and made integral with the consumption function. Thus the vertical axis in Figure 1 is either +C or —S above the xȁaxis and is +S or +I below the xȁaxis. All values are in “real” terms. The consumption function is drawn for a given real stock of money, M 0 / P 1 and the savings function is drawn for a given real stock of money and a given interest rate, r a . Thus both savings and consumption are functions of three variables: income, the rate of interest, and the real stock of money. The relationship of savings to the rate of interest is presented in Figure 2, in which savings are graphed as a function of the rate of interest, given income, and the real stock of money. The real stock of money itself is simply a given nominal stock of money with respect to some given price level. If the price level rises (falls), the real stock of money would then fall (rise). Let us start by assuming that the economy is operating at a full-employment income, Y a with an exogenous investment function, I 1 , intersecting the savings function, S 1 , at this income. Now let investment fall to I 2 —the normal Keynesian shift that is supposed to give us an equilibrium income of less-than-full employment. We should note at this point that we do not “need” the real balances effect to get an upward turning point. That is, even the pure Keynesian model will stabilize at income level Y b . Therefore, Hansen's argument uses up the real balance effect before it is even needed. Now let us assume that the price level falls to P 2 . With a fixed nominal stock of money,5 M 0 , the real stock of money will increase to M 0 / P 2 , shifting the consumption function up or to the left, simply indicating that more will be consumed at various levels of income. The rationale of this position is that the increased real stock of money is quasi-capital gains income and will thus result in greater consumption. Since a lower turning point had already been reached, may we not presume that we are now headed back toward full employment? At least we shall be at some intermediate position between Y b and Y a . Finally, we have the effect of cash-balance changes on savings. As the real stock of money increases with the falling price level, the public's “secondary” desires for savings will be satisfied by the greater stock of money it is holding. Therefore, there will be less savings at various possible interest rates, implying that the savings function in Figure 2 will shift from S 1 to S 2 .6 But a shift in the savings function must also be shown in Figure 1. Such is shown by a shift in the savings function in Figure 1 from S 2 to S 3 . Of course, the consumption must also shift up to C 3 . And income is now in equilibrium at Y a ′ , a position which appears on the diagram to be “over” full employment but which is not necessarily any specific level of income until we know the magnitudes of the variables involved. That is, we must know the actual flexibility of prices and the reactions of consumption and savings to them. If we allow ourselves to introduce an accelerator, that is, if we presume that investment is some function of consumption, we have cause for a shift in the investment function back toward its original position. But we are not limited to relying on this accelerator for an upward push after the turning point. Both the consumption and the savings functions will provide a stimulus; and it is this that Hansen failed to take account of. In brief, there are many combinations of investment (= savings) and consumption spending which could give us a full-employment level of national income. And even though we cannot say that full employment is necessarily reached, we have forces in reserve which do not “peter out once the lower turning point is reached.”7 With respect to monetary policy, the real-value-of-balances theory in effect says that we can get any price level and any level of business activity we want by manipulating the stock of money. (A priori, it would seem that if we could get a real stock of money and corresponding level of activity by altering the nominal stock of money, we should be able to get the same effect by letting the price level fluctuate around a given stock of money.) Practically, we should note, first, that we do not have a fixed stock of money and, second, that we do not have a fixed stock of “liquidity,” primarily because of the operations of the fractional reserve banking system.8 Further, our economy is not static but shows year-to-year increases in both total productivity per capita and the total number of people sharing this greater productivity. Therefore, in order that economic stability may be maintained at full employment with a reasonably stable price level, some part of the stock of liquid assets must be increased at a fairly constant rate.9 Since an increase in the real value of liquid balances may take place either by an increase in M or a decrease in P, or a combination of changes in both, we have no good reason for following the austere alternative which calls for P to do all the “work.” In fact, the analysis of the Pigou effect is simply an exercise for getting the relevant variables delineated and does not imply recourse to any particular policy. It shows only what may be expected under the most simplified conditions. At present, Federal Reserve System purchases of government securities in the open market is the principal means that our economy has for augmenting the stock of money. The alchemy, “painlessly” exercised by the Federal Reserve banks, of purchasing assets of a lower order of liquidity—i.e., government securities—and emitting liquid liabilities of a higher order—i.e., Federal Reserve Notes or member—bank reserves-is certainly capable of giving the economy in the long run, as well as in the short run, the liquidity impulses it needs to satisfy the cash-balance demands of the public without recourse to falling prices, even if other minor factors in the balance sheet behave contrariwise.10 One common charge against the effectiveness of monetary policy on the banking system is that greater reserves in the hands of the banks will result only in “bank hoarding.” We should note, however, that, although the behavior of banks is less predictable than that of people—since banks have to hold only fractional reserves against liquid liabilities—they are not likely to build up an infinite stock of liquidity either. They, too, will be faced with an inducement to disinvest cash balances for more attractive, income-earning assets. Hansen, in his discussion of policy, surprisingly seems to regard the banking system as dynamically “neutral.” He says: “Bank created money involves both assets and liabilities which balance each other off.”11 And on the next page he implies the same kind of error when he says: “A mere monetization of publicly owned securities will, however, not increase total liquid asset holdings.“12 Although Hansen is perfectly correct in holding that a Pigou effect would not be involved as the banking system expands and contracts, dramatic positive shifts in the coefficients of liquidity of the total asset structure of the economy would certainly take place. The same would hold for monetization of the public debt. These qualitative shifts in liquidity may be even more important, under some conditions, than the quantitative changes subsumed under the Pigou analysis.