"The first and most important lesson that history teaches about what monetary policy can do -- and it is a lesson of the most profound importance -- is that monetary policy can prevent money itself from being a major source of economic disturbance."
(Milton Friedman, "The Role of Monetary Policy", American Economic Review, 1968: p.12)
"I think I had best begin by making my own position clear - I regard "monetarism" as a terrible curse, a visitation of evil spirits, with particular unfortunate, one could almost say devastating consequences on our own country, Britain."
(Nicholas Kaldor, Origins of the New Monetarism, 1981: p.1)
(A) The "Chicago Plan"
(B) The Money Growth Rule
(C) The Optimum Quantity of Money
(A) The "Chicago Plan"
Monetarism was at the height of its influence on economy policy-making in the late 1970s and early 1980s and, although it has waned considerably since, many aspects of its influence still remain in the modern policy-making. As can be expected, at the heart of Monetarist economic policy recommendations is the use of monetary policy, by which we mean the conduct of open market operations, discount window restrictions, etc. by the Central Bank in order to influence output and prices. In contrast, early Keynesians have tended to stress the role of fiscal policy in stabilizing the macroeconomy - a position reinforced by the famous 1959 Radcliffe Committee report on British monetary reform, which engendered an intra-Keynesian debate on the role of monetary policy (see Kaldor (1960, 1982) for a review).
Early Chicago economists such as Henry C. Simons (1934, 1936, 1948) and Lloyd Mints (1945, 1950) had recommended the use of monetary policy for price stability - in lieu of other tools, such as fiscal policy, the gold standard, commodity reserve currencies, etc. which were being proposed in the 1930s. In particular, following the conventional monetary disequilibrium cycle theories of the day, they believed that a lot of the fluctuations in money supply intensified uncertainty and worsened the cycle. What they tended to recommend was a strong Central bank with complete control of the supply of money to the point of requiring 100% reserve requirements on deposits. They sought to endow the Fed with a monetary policy stabilization mandate restricted to stabilizing the price level -- all combined with highly flexible and competitive labor and goods markets that would permit quick adjustment to equilibrium. These policy stance was referred to by contemporaries as the "Chicago Plan" (cf. A.G. Hart, 1935).
In his early work, Milton Friedman (1948) followed on the Simons-Mints "Chicago Plan" views and recommended the use of a counter-cyclical monetary growth policy. Specifically, he proposed that the Federal Reserve set a policy of expanding the money supply during recessions and contracting it during booms in order to stabilize the price level in a "buffer stock" manner. Interestingly, like Simons before him, he did not discount the role of labor unions in affecting price inflation. In particular, Friedman (1951) noted that unions and collective bargaining could reduce price flexibility and thus force output to do most of the adjusting in response to fluctuations.
(B) The Money Growth Rule
Milton Friedman reversed his policy stance later, particularly in his famous Program for Monetary Stability (1959), where he dropped the countercyclical monetary policy rules of the Chicago Plan and opted in favor of a "constant money growth rule". Such a rule had been earlier advocated by James W. Angell (1933, 1936) and Clark Warburton (1952). Specifically, Friedman (1959, 1962) proposed that instead of trying to smooth out the cycle, the Federal Reserve should just follow a strict rule of expanding the money supply at a steady rate.
Friedman's logic can be expressed in terms of the Quantity Theory. In dynamic form, the equation of exchange implies that gM + gV = gP + gY. Assuming gV = 0 (or nearly so) and given that output has grown at a historical average of 3% per year, then in order to avoid inflation (i.e. keep gP = 0), the Federal Reserve should expand the money supply yearly by 3% (he actually recommended 3 to 5 percent per year). Friedman's suggestion was taken up by most Monetarists. Indeed, since 1973, Karl Brunner and Allan H. Meltzer have set up a "Shadow Open Market Committee" to regularly evaluate and recommend monetary policy along these lines.
Friedman's policy reversal was derived in good part from one important realization that emerged from his research: namely, his findings that discretionary countercyclical monetary policy could do more damage. He was first inspired in this direction by his analysis of the lag it takes for monetary policy decisions to affect the real economy. The analysis of lags was begun by Clark Warburton in the 1940s (see Warburton, 1966) and followed up by Friedman (1948, 1958, 1961) and Friedman and Schwartz (1963). He found the lags to be not only quite long, sometimes up to eighteen months, but also highly variable. The length and variability of the policy lag arises because, from a given output shock, it takes time for data to be collected and delivered, it takes some more time for the policymaker to be sure that there is actually a problem that needs to be remedied (to be sure there are no data errors, etc.), it then takes time to discuss and obtain agreement among the various relevant Central Bank governors on what the appropriate policy response is and, then, a little more time in getting that policy out to the trading desks to be implemented. Finally, there is the most complicated lag of all: the time it takes for the new Federal Funds interest rate to bring other interest rates into line and, more importantly, the time it takes to affect borrowing, investment and production decisions so that output is finally affected.
To forecast what output would be at such a long distance would be quite difficult: an expansionary monetary policy seeking to save a current recession might, because of this usually unpredictably long lag, actually only start impacting the real economy when the economy is already well into a boom - thus, its final result being largely inflationary. Further calculations on the lag of monetary policy showed contradictory results - at times only a few months, at other times it took almost two years, before impacting the economy. This would make countercyclical monetary policy even more difficult to conduct properly. Faced with unpredictable lags, Friedman decided that it might simply be less damaging to just stick to constant money growth rule and hope for the best.
Friedman's policy proposition also stems, in good part, from his finding in Friedman and Schwartz (1963) that, throughout history, the Federal Reserve has often been the very source of economic instability in the United States. Central Banks, Friedman argues, even highly independent ones such as the Federal Reserve, are in the hands of humans and humans, unlike rules, are not only subject to ineptitude and error, but also vulnerable to political influence - an idea clearly stemming from Simons (1936). As he argues in various places (e.g. Friedman, 1959, 1962, 1984, 1985), far too often, Central Banks' discretionary power for monetary policy can be swayed by political operators or simply public opinion, without regard for economic logic. For instance, Central Banks will often undertake expansionary monetary policy in order to provide a boost for the election efforts of their political masters - the inflationary consequences of which will only be felt after they are safely in office.
Friedman and other Monetarists have argued that the solution to the political incentives problem is not to grant the Central Bank more independence as that would increase the arbitrary power of its governors and help them cover up mistakes or serve their political ends. Rather, they have proposed that Central Banks be brought under stricter control of the legislative branch of government (not the executive!), which they view as more "politically stable" and thus better able to oversee the strict enforcement of a money supply growth rule. The ideal case, in the eyes of many Monetarists, would be to have this rule constitutionally mandated or at least formally legislated, even internationally. Several American congressmen have attempted to introduce such legislation in the 1970s in 1980s without success. However, Friedman's influence was decisive in including, in the 1978 Humphrey-Hawkins Act, the requirement that the the Chairman of the Federal Reserve Board of Governors give periodic testimony before a Congressional committee -- a practice that remains to this day. This, Monetarists felt, would at least increase the transparency and accountability of the Federal Reserve.
Friedman's lag analysis drew a lot of criticism for technical reasons (e.g. Culbertson, 1960). His recommendation to replace discretionary monetary policy with money growth rules drew even more bitter criticism. The most straighforward one was that the suggested "historical" average output growth rate of 3% rarely exists in any given year nor is it an empirically validated average. In fact, from the 1970s to the early 1990s, the average output growth rate in the United States has hugged 1.5%: if the 3-5% money supply growth rule was followed, then by quantity-theoretic reasoning, the U.S. would have had a good bout of sustained inflation for that quarter-century!
The second item to be questioned in Friedman's policy proposition was whether the Federal Reserve could actually control the supply of money and which money supply it ought to control. The Fed, of course, can only directly control the monetary base (H or M0, currency and reserves). Above this, are more fluid categories of "narrow money", M1 (currency and demand deposits) and "broad money", M2 (M1 plus some time deposits) or even broader measures such as M3, L, etc. All these aggregates are related to each other via a "money supply process", as delineated by Friedman (1959) and followed up by Brunner and Meltzer (1964), Cagan (1965) and many others since.
The central ingredient of this process are the "deposit money multipliers" which connect the different money supply aggregates with the supply of monetary base. Succinctly, because we are in a fractional-reserve banking system, only a small percentage of each bank deposit must be held in the form of reserves at the Central Bank. As such, a dollar increase in the supply of reserves (part of the monetary base, H) will increase the supply of bank deposits and thus money supply (call it M) by considerably more than a dollar. Heuristically, M = mH where m is the deposit money multiplier. The deposit money multiplier for broad money aggregates such as M2 is much greater than the deposit money multiplier for narrow money, M1 as the relevant required reserve ratios are lower for many of the components of M2.
The first step, then, is to control the monetary base, H. The Federal Reserve can potentially target it reasonably well. But even here a caveat emerges that banks often borrow reserves from the Central Bank (through what is called "discount window borrowing), thus H can be quite fluid. If it wants precise control, the Central Bank must be willing to close or severely restrict access to the discount window so that the flow of reserves into and out of the banking system remains tight. Closing this window, however, entails enormous risks: if, for instance, there is a liquidity crunch and strapped commercial banks face a closed discount window, many will go under easily. In failing to serve its role as a "lender of last resort", the Central Bank would thus exacerbate temporary shocks even further, precisely the misguided road that Friedman and Schwartz (1963) condemned! Surely it cannot be Fed policy to risk a banking crisis and all the horrors that can bring in order to stick to its money supply targets! One can suggest that in such extraordinary circumstances, the Fed can relax them and re-open the discount window. Perhaps. But then there are moral hazard problems with this: the rule would simply never be believed and the window would have to be effectively opened.
Nonetheless, suppose the Fed can and does target high-powered money base, By an appropriate calculation of the deposit money multiplier, m, then as M = mH, one can also control the money supply. However, much research suggests that the money multiplier is largely governed by the bank's own decisions about its asset and liabilities sheets, so that the money supply is at least partly endogenous -- what was known as the "New View" of J.G. Gurley and E.S. Shaw (1960) and James Tobin (1963), which was disputed by Monetarists such as Karl Brunner (1968). This is particularly pertinent if the Central Bank tries to target broader moneys such as M2 or M3, which tend to be more amenable to bank maneuvering. As a result, the deposit money multipliers are very difficult to estimate (albeit see the efforts of Johannes and Rasche (1987)). Furthemore, financial intermediaries have shown great inventiveness in overcoming reserve restrictions and modifying the deposit multiplier considerably over time -- and innovation, of course, is never really predictable. We see this already in the difference between the size of M1 and M2, which has grown enormously since the 1980s, reflecting in good part financial deregulation and innovation on the part of the banking sector and other financial intermediaries. At any rate, even if one targets the quantity of money, one cannot really control its degree of turnover, so that its velocity may be quite unstable - as it proved to be in the 1970s and 1980s.
Over the years, Friedman (1948, 1959, 1984, 1985) has suggested drastic ways to control and prevent commercial bank independence by various means, e.g. closing the discount window, raising the discount rate above the market (Federal Funds) rate, imposing 100% reserve requirements on demand deposits, removing FDIC deposit insurance, stricter supervision, imposing uniform banking laws across states, etc. in order to prevent or dissuade banks from too much manipulation or inventiveness and thus reduce the degree of endogeneity of the money supply.
In a surprising twist, the degree of almost draconian control over banks which Friedman has recommended for his Central Bank brought him directly into conflict with his fellow traveler, Friedrich A. von Hayek (1978, 1979). In Hayek's view, the Central Bank ought to be dissolved and be replaced by a "free banking" system originally proposed by Vera Smith Lutz (1936), by which he meant that commercial banks should be permitted to issue as much of their own currency as they wished. Hayek adhered to a "real bills" type of reasoning, arguing that free competition and commercial banks' concerns with reputation, etc., would ensure that no more money supply would be forthcoming than was necessary to meet the "needs of trade". Friedman and Schwartz (1986) drew themselves directly against Hayek's proposition.
[This is, of course, not the first time Friedman and Hayek have crossed swords over the issue of monetary control; back in the 1940s, when Hayek was selling the idea of a commodity reserve-backed currency (Hayek , 1943), Friedman (1951) was lukewarm on the plan, arguing that while a commodity basket-standard is much more stable and thus better than a gold standard, strict Central Bank control of an unbacked money supply would be the first-best option. He has since warmed to the idea of commodity-reserve currencies (cf. Friedman, 1985) as that would implement the money growth rule almost automatically.]
For the Keynesians, the more controversial question was not so much whether the money supply was controllable, but rather whether targeting the money supply was desirable at all (although the Radcliffe Committee Report of 1959 combined both criticisms). As William Poole (1970) has analyzed to a good extent, targeting money supply aggregates comes at the cost of allowing for interest rate volatility. If money demand is unstable and volatile, then the impact on output of following a money supply targeting rule will be enormous - at least, relatively to the alternative of setting an interest rate target and allowing the money supply to accommodate money demand. Of course, if money demand is stable - as Friedman and the Monetarist researchers tried to establish - then money supply targeting is better than an interest rate target at reducing output volatility. However, we should note that some Monetarists (e.g. Brunner and Meltzer, 1983), have disputed the existence of a trade-off between interest rate variability and money supply variability.
However, even if money supply targeting is better, this still does not establish that a constant money supply growth rule over time (which Friedman recommended in 1959) is better than attempting countercyclical money supply targets (which Friedman recommended in 1948). However, most Monetarists (e.g. Friedman, 1959, 1962, 1985; Brunner, 1981) nonetheless insist that this is the "risk-minimizing" strategy in view of the lag problem and the political incentives problem. Later work by other researchers on the "reputation" effects of money growth rules and further complications induced by monetary policy lags have lent more credence to Friedman's money supply growth rule, but strong arguments still remain against it.
One of the striking policy conclusions from the Phillips Curve debate was that attempting to reduce unemployment below the natural rate would only lead to accelerating inflation. Milton Friedman (1968, 1969, 1977) reminded everyone that there is an asymmetry in the converse policy. Government, he claimed, should not attempt to lower unemployment below the natural rate since it is costly to maintain and temporary anyway. However, government can attempt to lower inflation which is permanent and, his view, relatively costless in the long-run. In other words, he recommended that the government (the Central Bank in particular) operate a short-run employment-inflation trade-off against full employment in order to terminate inflation. If they aim for an unemployment level higher than NAIRU, inflationary expectations would be reevaluated downwards - bringing actual inflation down permanently. In other words, the cure for inflation would be to deliberately cause a recession.
Friedman's "disinflation" suggestion was met with dismay - although, in his Nobel lecture, Friedman (1977) tempered his claims somewhat. Nonetheless, economists immediately went on to calculate the "sacrifice" ratio, i.e. how much output would be foregone in an attempt to reduce inflation by a single percentage point. Arthur Okun (1978) calculated that to get a 1% drop in inflation requires approximately a 3% rise in unemployment and a 9% contraction in GDP. Okun (1978, 1981) and many other economists (e.g. Gordon, 1975), while agreeing that Friedman's proposition would, indeed, reduced inflation, nonetheless strongly recommended against it because it was far, far too costly.
Harking back to his 1951 work, Milton Friedman retorted that much of the output and unemployment costs to disinflation arise because collective bargaining arrangements tend to lock in a money wage and thus prevent quick price-side adjustment. In order to minimize the cost of disinflation, Friedman (1974) proposed the inclusion of "escalator clauses" in labor contracts that automatically corrected money wages for inflation. In this manner, he argued, the short-run Phillips Curve becomes "steeper" and thus the costs of disinflation (unemployment and output foregone) would be lower. Of course, the escalator clauses would not necessarily be a good thing in the case of aggregate demand expansion.
Despite much strenuous opposition, "Monetarist experiments" were conducted in the late 1970s and early 1980s in several Western countries - notoriously, the US and the UK. In 1979, soon after the ascendancy of Paul Volcker as chairman of the Federal Reserve in the United States and Margaret Thatcher as Prime Minister in Great Britain, interest rate targets were dropped in favor of money supply targets and "disinflation" was begun. The critics of Friedman's policy turned out to be correct: there was a long, painful recession with double-digit unemployment - by far the worse recession since the 1930s - and inflation seemed to survive. In the United States, the Federal Reserve "declared victory" in 1982, when inflation was still running at 4%, and abandoned the disinflation policy. By 1984, it abandoned money supply targets altogether. In Britain, the cost of the disinflation was even greater: output had shrunk in two years by 7.5 and a fifth of manufacturing output disappeared; unemployment soared to 10% while, surprisingly, inflation actually climbed from 10% to 22%. Faced with this result, Margaret Thatcher abandoned the disinflation attempt and, eventually, monetary targets, and laid the blame for the disasters of 1980-1 on what she publicly denounced as a misguided economic doctrine.
Many Monetarists explained the dismal results of the "Monetarist experiment" by accusing the Central Banks as not having been able to effectively control the money supply, in spite of their explicit targets -- "lack of nerve" on the part of Central Bankers was commonly cited. Keynesians, of course, had their own explanation for these results. The Keynesian argument was particularly lifted after Ronald Reagan's tax cuts and massive deficit-financed expansions in government spending in the early 1980s had a highly stimulative effect on the U.S. economy - just as textbook Keynesianism would predict.
(C) Optimum Quantity of Money
Throughout the monetary policy debate, it had not been clear whether inflation was necessarily an "evil" - or, at least, as evil as some of the Monetarists made it out be. After all, if inflationary expectations are indeed incorporated in the manner Friedman suggests, and the natural rate prevails in the long-run, then it should not actually matter either way to the real economy whether there is inflation or not. Why then so much fuss?
Part of it was ideological overhang: advocates of the earlier Chicago Plan - particularly Simons - were almost paranoid about the effects of price instability. Hyperinflationary experiences, of course, can wreak havoc on an economy. However, many macroeconomists have long believed that steady, mild inflation can be actually beneficial. For instance, modest inflation can promote desirable redistributions of income from creditors to debtors, encourage households to move away from holding money balances and accumulating capital instead (thus increasing growth), enhance labor market flexibility by allowing real wages to erode without confrontational negotiations for nominal wage reductions, etc.
The question was quick to arise: might there be an "optimal" rate of inflation? Milton Friedman (1969) had calculated that the "optimal" inflation rate was the negative of the real interest rate (i.e. zero nominal interest rates). Friedman's logic is a bit subtle, but still worth pursuing as it was one of the justifications for the disinflation.
Consider an agent who holds both bonds and money. As bonds have a return (the nominal interest rate, i) and money has none, then an agent who holds money must think that money is "useful" in some hidden way. In other words, money must provide some utility-yielding services which are peculiar to it and not shared by other assets (e.g. pure liquidity). Now, the marginal cost of holding an extra unit of cash is the interest that is foregone. The marginal benefit is the services yielded by the extra unit. How can this be measured?
Taking a leaf from Bailey (1956), we can visualize this in simple terms of the consumer surplus underneath the money demand function Md in Figure 3. The downward-sloping feature of the money demand curve reflects the opportunity cost of interest. M/P represent real money holdings. Suppose we begin with (M/P)i units of money, where the foregone interest rate is i. Suppose an agent decides to increase money holdings to (M/P)r. Notice that now the interest rate is lower at r. Thus, the cost of increasing money holdings is the interest foregone on the difference (M/P)i - (M/P)r. Now, a rational individual would not have undertaken such a move if the benefit acquired from the extra money holdings was at least as high. Alternatively stated, since the marginal benefit of money can be conceived of as the height of the money demand curve, then the total benefits in the move from (M/P)i to (M/P)r can be represented by the area under the curve between (M/P)i and (M/P)r - i.e. the darkly shaded region in Figure 3.
We can now begin to see what the welfare cost of inflation would be. Let i be the nominal interest rate and r the real interest rate. Thus, the difference i - r = p is the rate of price inflation. Now, the opportunity cost of money is the nominal rate of interest i, thus agents will choose to hold (M/P)i units of money. However, they only receive real interest rate r on their assets. If the nominal rate of interest was actually r, then they would hold (M/P)r units of money. As (M/P)i < (M/P)r, then agents are holding less money because of the pure effect of inflation. The "welfare loss" of inflation can subsequently be thought of as the darkly shaded area under the money demand curve between (M/P)i and (M/P)r.
Figure 3 - Welfare Costs of Inflation and Optimum Quantity of Money
In order to eliminate the welfare loss implied by p, all one had to do is to reduce inflation to zero so that i = r and the real money holdings are merely (M/P)r. But is this "optimal"? Friedman (1969) suggested one could do better - namely, by reducing the nominal rate of interest to zero. For this to be true, then there must be a negative rate of inflation equal to the real rate of interest, i.e. p = -r. This is shown in Figure 3 by the distance p｢ between r and zero. This negative p｢ is the "optimal" inflation rate. Notice that when the nominal interest rate is zero, desired money holdings are (M/P)0. In moving from (M/P)r to (M/P)0, we have made a welfare gain represented by the lightly shaded area under the money demand curve in Figure 3.
Alternatively stated, suppose the money supply is such that one or more agents are willing to pay a positive interest rate in order to obtain an extra unit of cash. In other words, they desire more money than is in existence. Friedman asserts that such an economy is "inefficient" because money can be printed at no cost (or all one needs is to lower all prices enough to make the existing real stock of money sufficient). There is no need to "artificially" constrain the money supply to create this "scarcity" of cash that the agents face. Consequently, the optimal amount of money, Friedman asserts, is that which would yield zero marginal opportunity cost of money, i.e. zero nominal interest rates.
As a result, Friedman (1969) proposed that the optimal monetary policy is to steadily contract the supply of money at the real rate of interest, i.e. gM = -r. In this case, by Quantity-Theoretic reasoning, as gM = p , then we obtain the desired result that the nominal rate of interest is zero, i.e. i = r + p = 0. This optimal monetary policy proposition has become known as the "Chicago Rule".
[A more elaborate general equilibrium theoretic analysis of Friedman's optimum quantity of money proposition was pursued by Jean-Michel Grandmont and Yves Youn鑚 (1973) and Truman Bewley (1980, 1983). For a contemporary critique, see Frank Hahn (1971).]
However, as noted earlier, Edmund S. Phelps (1972, 1973) made clear that, by standard cost-benefit criteria, the optimum rate of inflation might actually be positive. Phelps's reasoning relies in good part on the idea that inflation is a "tax". In other words, the reduction in consumer surplus in Figure 3 is actually transferred to the government. What the government decides to do with this extra "revenue" must be included as an item of any subsequent welfare analysis. It can easily be the case that the Chicago Rule is not the "optimal" monetary policy in this case - as was shown by Phelps.
As is evident from the admirable survey of Stanley Fischer and Franco Modigliani (1975), the only substantive cost of (modest) inflation is largely the manner in which it distorts taxation. Consequently, as Phelps (1973) argued in a famous article, optimal inflation ought to be treated as a public finance problem, effectively an "optimal taxation" issue. However, note that "optimal inflation" arguments usually yield steady, modest inflation (2-3% or so); higher rates have a tendency to become highly variable and/or spiral into hyperinflations, which brings on considerably more dangerous complications. Nonetheless, the implications are clear: as catalogued by Fischer and Modigliani (1975), the real costs of mild inflation are surprisingly few - certainly, not enough to justify the huge sacrifices disinflation requires.
Much of the policy debate between Monetarists and Keynesians was continued between New Classicals and Keynesians, with roughly the same themes. Thus, we shall suspend our discussion for now.