Content uploaded by Tim Congdon
Author content
All content in this area was uploaded by Tim Congdon on Nov 16, 2017
Content may be subject to copyright.
8. Friedman (1948), Friedman (1996) and the effectiveness of
fiscal policy in the USA
As is well-known, views on the effectiveness of fiscal policy vary between
economists. One position is widely-held and, superficially, appears to be little
more than organized common sense. Since government expenditure is part of
aggregate demand, and since Keynes taught in The General Theory that demand
generates output and that output requires employment, it follows (or seems to
follow) that an increase in government expenditure is certain to boost demand,
output and employment. As Keynes’ biographer, Robert Skidelsky, said in an
interview in December 2009, if the aim is ‘to get aggregate spending up’, ‘the
surest way to do this is by the government spending the money itself’.1 The
simplicity of this line of argument lends it plausibility and, at first acquaintance,
it goes a long way to establish a case for the effectiveness of fiscal policy.
Nevertheless, many economists deny the effectiveness of fiscal policy as a
means of influencing demand, while a high proportion of these fiscal sceptics
believe monetary policy to be more important. Their problem is partly one of
public relations, that the argument for monetary policy is harder to formulate
and project.
This essay has two purposes. The first is to show not only that different
economists have sharply divergent views at any one time, but that the same
economist can have contrasting views at different times. The economist chosen
for the exercise, Milton Friedman, is accepted as having been one of the greatest
of the twentieth century. The inconsistency between his standpoints in 1948
(when he said fiscal policy mattered enormously) and 1996 (when he said fiscal
policy didn’t matter at all) is so extreme that someone new to his work might
ask questions about his intellectual integrity. The second purpose is to suggest
that the Friedman of 1996 had learned from ideas and events. His position then
had evolved over years of thinking and research, and does need to be respected.
Indeed, he knew that the views he held about fiscal policy later in his career put
him in a small minority of economists and said so. Some data presented here
will show that the USA’s experience towards the end of the twentieth century
did validate the mature Friedman’s dismissal of the effectiveness of fiscal
policy. The notion that extra government spending boosts demand and output
may appeal to common sense, but it is not necessarily true.
I
Friedman, born in 1912, was 17 when the American stock market crashed in
October 1929. He was then an undergraduate at Rutgers University, but moved
on to Chicago in 1932 and Columbia in 1933 for postgraduate studies.2 His
student years coincided with the most traumatic cyclical upheaval ever inflicted
on the USA, its Great Depression of 1929 to 1933. It is perhaps unsurprising
that Friedman should devote most of his subsequent career to trying to
understand what happened to the American economy in those years. According
to a biographer, in the early 1930s ‘Friedman was a hardworking student and
took copious notes in class and of works that he read.’ At Chicago he ‘took, for
example, eight-seven pages of notes on Keynes’ Treatise on Money.’3 In the
Treatise Keynes was optimistic about the effectiveness of monetary policy as a
means of combating depression, with his favourite prescription for a slump
being large-scale purchases of securities by the central bank which he termed
‘monetary policy a outrance’ (or ‘monetary policy to the uttermost’).4 More
specifically, the operations Keynes had in mind in the Treatise were the
combination of ‘a very low level of the short-term rate of interest’ with central
bank purchases of ‘long-dated securities either against an expansion of central
bank money or against the sale of short-dated securities until the short-term
market is saturated’. (Is it going too far to suggest that the ‘very low’ short-term
rates were to be maintained by ‘money market operations’ in the terminology of
Essay 4 in this collection, while the purchases of long-dated securities were to
be by means of what are called there ‘debt market operations’? Essay 2 above
considers the relative merits of Keynes’ Treatise and his General Theory.)
Given the obvious keenness of Friedman’s interest in monetary economics in
his first postgraduate year, it might be expected that his earliest writings as a
professional economist would highlight the advantages of monetary policy.
Instead his June 1948 paper on ‘A monetary and fiscal framework for economic
stability’ had very different emphases. For a start it was less than flattering
towards banks and the role of banking in a market economy.5 In this respect it
reflected the influence of Henry Simons, who had been professor of economics
at Chicago until his early death at the age of 47 in 1946. Simons claimed that
the ability of banks to create money by extending loans to the private sector
could result in large and untoward fluctuations in the quantity of money and,
hence, to great and unnecessary macroeconomic instability.
Friedman latched onto this thinking, advocating that banks should not have the
power to expand or contract money by altering credit to the private sector.6 The
aim should be rather that the state regulates directly the quantity of money. In
Friedman’s judgement, ‘the chief function of the monetary authorities’ should
be ‘the creation of money to meet government deficits or the retirement of
money when the government has a surplus’. An implication was that – in the
middle, at a desired high-employment level of national income – the
government’s budget should be balanced and the quantity of money should not
change. It followed that officialdom should estimate two budget concepts, ‘the
stable budget’ in which the figures referred to the hypothetical desired level of
national income, and the actual budget. Consequently, in Friedman’s word, ‘The
principle of balancing outlays and receipts at a hypothetical income level would
be substituted for the principle of balancing actual outlays and receipts’.7
Evidently Friedman (1948) did think that the equilibrium level of national
income was a function of the quantity of money, in line with his later
convictions. But – compared with his subsequent beliefs – the 1948 paper had
two remarkable features. First, instead of favouring stable growth in the
quantity of money to match increasing real output, he wanted the quantity of
money to vary cyclically around a level that was to be kept constant across
cycles. Secondly, fiscal policy was to be harnessed in order to deliver the
desired contra-cyclical variations in the quantity of money. Booms would
increase tax revenues and reduce unemployment costs, leading to budget
surpluses. Such surpluses would allow ‘the retirement of money’.8 Conversely,
recessions wound undermine tax revenues and add to unemployment
expenditure, resulting in budget deficits. The deficits could be financed by
borrowing from the banks and ‘the creation of money’. In other words,
monetary policy was to dance to the tune of fiscal policy. To quote, ‘Deficits or
surpluses in the government budget would be reflected dollar for dollar in
changes in the quantity of money.’ In this sense Friedman (1948) not only
accepted the potency of fiscal policy, but also put fiscal policy before monetary
policy.9
Now roll the camera forward almost 48 years. In January 1996 two British
economists, Brian Snowdon and Howard Vane, were interviewing Friedman in
the study of his apartment in San Francisco. (Coincidentally the American
Economic Association was holding its annual conference in the same city.)
Friedman had won the Nobel prize for economics in 1976 and was by now a
massively distinguished figure whose every utterance on economics was of
wide interest. Snowdon and Vane asked Friedman, ‘in the light of your work on
the consumption function and monetary economics in general, what role do you
see for fiscal policy in a macroeconomic context?’. Here is his answer,
None. I believe that fiscal policy will contribute most if it doesn’t try to
offset short-term movements in the economy. I’m expressing a minority
view here, but it’s my belief that fiscal policy is not an effective instrument
for controlling short-term movements in the economy. One of the things I
have tried to do over the years is to find cases where fiscal policy is going
in one direction and monetary policy is going in the opposite. In every case
the actual course of events follows monetary policy. I have never found a
case in which fiscal policy dominated monetary policy and I suggest to you
as a test to find a counter-example.10
The contrast between this answer and the position stated in 1948 was sharp and
obvious. In 1948 he saw the variations in the budget position which arose from
the economy’s cyclical fluctuations as worth exploiting by means of deliberate
contra-cyclical policy adjustments to the quantity of money. So at that stage the
budgetary position figured centrally in his favoured approach to monetary
management. But in 1996 Friedman favoured a regime in which the budget
balance did not have even a bit part in the larger drama of macroeconomic
decision-taking.
Why had Friedman changed his mind? And was he right to have rejected fiscal
policy so emphatically?
II
In the Snowdon and Vane interview Friedman offered two informal theoretical
arguments for the virtual irrelevance of fiscal policy, as he saw it. The second
was that fiscal policy is ‘much harder’ to adjust in ‘a sensitive short-term way’
than monetary policy. But the first was the more telling and deserves detailed
discussion. It was a direct challenge to hundreds of textbooks and the received
wisdom of the majority of academic economists. In Friedman’s words, ‘I
believe it to be true…that the Keynesian view that a government deficit is
stimulating is simply wrong.’ The explanation was the wider effects of the way
the budget deficit is financed. To quote again, ‘A deficit is not stimulating
because it has to be financed, and the negative effects of financing it
counterbalance the positive effects, if there are any, on spending.’11
His interviewers seem to have been startled by the forcefulness of Friedman’s
rebuttal of fiscal policy. Implicitly, Friedman repudiated a view held almost
universally among Keynesians, that the best policy response to the Great
Depression of the 1930s would have been a large programme of public works
and an associated increase in the budget deficit. When Snowdon and Vane went
on to ask whether Friedman would not have advocated ‘expansionary fiscal
policy’ in the 1930s, his reply was trenchant. ‘It wasn’t fiscal policy, it was
monetary policy that dominated. There was nothing you could do with fiscal
policy that was going to offset a decline of one third in the quantity of money.’
Bringing the discussion up-to-date, Friedman went on to condemn as futile the
increases in the budget deficit on which Japan’s government had embarked in
the 1990s, in an attempt to break out of its macroeconomic doldrums. In his
judgement, the Japanese were ‘wasting time and money in trying to have an
expansive fiscal policy without an expansive monetary policy’.12
Friedman’s denigration of fiscal policy in the Snowdon and Vane interview
relied, then, on his claim that monetary policy – by which he meant changes in
the quantity of money – dominated fiscal policy, and the related implicit
assumption that changes in the budget balance did not necessarily have any
impact on the rate of money growth. It has to be said, in criticism, that this
implicit assumption in Friedman (1996) is difficult to reconcile with Friedman
(1948). In Friedman (1948) the creation of money by the private sector has been
abolished by official fiat, and all money creation and destruction is the result,
‘dollar for dollar’, of fiscal deficits and surpluses; in Friedman (1996) fiscal
policy is irrelevant, because money dominates in the determination of
macroeconomic outcomes and the rate of money growth is not a function of the
budget balance.13 However, consistency could be restored to Friedman’s
position by appealing to an undoubted fact about the financing of budget
deficits in most industrial nations in the two or three decades before 1996 (and
in fact subsequently). In those two or three decades governments almost
invariably tried to finance budget deficits outside banking systems, partly to
avoid difficulties in rolling over public debt as it came up for redemption, but
more fundamentally in order to avoid the inflationary consequences of the
monetization of budget deficits.
Friedman’s position in the Snowdon and Vane interview then became
intellectually coherent and rested on two points. First, if budget deficits are
financed from non-banks and have no effect on the rate of money growth,
changes in the budget deficit (i.e., ‘fiscal policy’, as usually understood)
similarly have no effect on the rate of money growth. Secondly, if national
income is a function of the quantity of money and not of the level of public
debt, the unresponsiveness of money growth to fiscal policy means that fiscal
policy is unimportant or even irrelevant to the macroeconomic situation. That
was why – when asked ‘what role do you see for fiscal policy?’ – Friedman
could say ‘none’.
The mature Friedman’s scorn for fiscal policy therefore hinged on governments
deciding, in the late twentieth century, to finance budget deficits responsibly by
medium- and long-term debt issuance (i.e., by the issuance of debt with an
initial maturity of at least five years). (Friedman never denied that monetary
financing of budget deficits would have all the macroeconomic consequences
attributable to changes in the quantity of money.) The importance of the budget
financing pattern was also central in the so-called ‘Treasury view’ of Britain in
the inter-war period. The Treasury resisted pleas by Keynes and others for extra
public works expenditure, on the grounds that long-term debt issued to finance
public works would to some extent replace debt that might have been issued by
the private sector. Public spending might be higher, but private spending would
be lower and the net impact on spending was not necessarily positive.14 The
Treasury view remains one of the most effective critiques of naïve Keynesian
confidence that increases in public spending and the budget deficit translate,
almost automatically, into increases in national income. By the 1990s this sort
of objection to Keynesian fiscal activism had in fact become rather
unfashionable within the economics profession, having lost ground to the so-
called ‘Ricardian equivalence theorem’ proposed by Barro in a celebrated 1974
article.15 Friedman seems to have appealed only infrequently to the neo-
Ricardian ideas in the tendency to downplay fiscal play which marked the later
stages of his career.16
III
The Snowdon and Vane 1996 interview hinted that Friedman had been working
for some time on research to compare the effectiveness of fiscal and monetary
policy. That was confirmed in another interview at Stanford University four
years later, with the questions now being asked by John Taylor. The 1948 paper
was again a topic. Friedman admitted that this paper implicated the budget
balance in the contra-cyclical variation of the money supply, but said that in the
1950s he came to the conclusion that the contra-cyclical money ‘policy rule was
more complicated than necessary’. Instead his verdict was ‘you really didn’t
need to worry too much about what was happening on the fiscal end, that you
should concentrate on just keeping the money supply rising at a constant rate’.
He was driven to this position by ‘the empirical evidence’. Examination of the
data persuaded him that ‘the link from fiscal policy to the economy was no
use’.17
But Friedman published nothing of any substance in his later career about fiscal
policy.18 (It needs to be remembered that he was in his late eighties when he
gave the interview to Taylor. He did publish further heavily empirical work, but
not on fiscal policy.)19 So what was ‘the empirical evidence’ he mentioned?
Without access to his personal papers, it is impossible to say. At any rate, the
rest of this essay uses a database made available by the International Monetary
Fund to see whether Friedman was talking sense or nonsense. The IMF website
has data, covering the period since 1980, for the cyclically-adjusted budget
balance and the level of the output gap for a high proportion of its members,
including the USA. From the data it is easy to calculate changes in the
cyclically-adjusted budget balance and the output gap, and to test whether the
output gap responds to fiscal policy. As explained in more detail in Essay 5,
fiscal policy would ‘work’, in the Keynesian manner, if increases in the
cyclically-adjusted budget deficit were accompanied by above-trend growth
(i.e., by increases in the output gap, where ‘the output gap’ is the monetarist
concept whose evolution is discussed below in Essay 7). Was this the pattern
observed in the USA from 1981 to 2008? The table below sets out the key
numbers.
Some ‘naïve Keynesians’ may be confident that that increases in government
spending and the budget deficit lead – with little or no further ado, and with
next to no qualification – to similar or larger to increases in aggregate demand.
They may believe, in other words, the fable scripted in chapter 10 of Keynes’
General Theory, and then rehearsed and replayed many thousands of times in
textbooks and university lectures. For them Table 8.1 may come as a profound
shock. If naïve Keynesianism were correct, an inverse relationship would be
expected to hold between changes in the budget balance and the rate of growth
of output. (That is, as the budget balance becomes more negative [i.e., the
budget deficit increases], the rate of output growth should rise and that ought to
make above-trend growth more likely.) The table shows that – over a 28-year
period in the world’s largest and most well-studied economy – an inverse
relationship of this kind held only a minority of the time. The American
economy did not behave in the way expected and proselytized in one of the
most influential chapters in The General Theory. On the contrary, as Table 8.2
below shows, in the 28-year period under review there were twice as many
years in which the economy responded in anti-Keynesian fashion to fiscal
policy than in Keynesian fashion. (The economy behaves in anti-Keynesian
fashion if an increase in the budget surplus or a decrease in the deficit is
associated with above-trend growth, and decreases in a surplus or an increase in
the deficit is associated with beneath-trend growth.)
Various criticisms might be made of this survey of evidence on fiscal policy
effectiveness. The approach is of course ‘casual empiricism’, with a
straightforward table of outcomes inviting quick appraisal, rather than
sophisticated empiricism, with elaborate econometric techniques. But – bluntly
– the exercise benefits from its simplicity and the clarity of the resulting
answers. There is, of course, the objection that influences on demand were of
two kinds, from fiscal policy and from all other factors, and the most
satisfactory analysis would be one in which the role of all the other factors
could be isolated. So – the Keynesian might say – the conjunction of ‘restrictive
fiscal policy’ and above-trend growth in Clinton’s second term proves nothing.
The Keynesian might claim that, in Clinton’s second term, restrictive fiscal
policy by itself was holding back demand growth, while ‘all the other factors at
work’ were motivating strongly above-trend growth. The Keynesian might then
insist on the building of a full-scale ‘structural model’ of the economy, so that
the different sources of demand growth could be identified and measured.
Friedman was always suspicious of intellectual arguments of this kind. While a
strong believer in confronting theories with evidence, he was sceptical about
elaborate models and the incorporation of too many variables in a purportedly
explanatory framework.20 In building a structural model to test the relative
effectiveness of fiscal and monetary policy, it is essential that the inherent
properties of the model are not such as to bias the answer. But most models at
present available are expansions of Keynesian income-expenditure analysis.
Models of this kind, by their very method of construction, imply that an
increase in the cyclically-adjusted budget deficit boosts demand and output.
Since the evidence on the effects of fiscal policy in the real world is murky and
unimpressive, the income-expenditure models need to be handled with care. In
any case the message of Table 8.1 is that – even if fiscal policy could be shown
to work in a properly-specified structural model – in the real world the effects of
fiscal policy must be largely overwhelmed by the effects of ‘all the other factors
at work’. If these factors include agents’ adjustments of portfolios and
expenditure to excess or deficient money holdings (see Essays 15 and 16
below), there is at least a possibility that money is very important in ‘all the
other factors’ at play. As Friedman showed in a large body of work, the power
of monetary forces – in the sense of the power of changes in the quantity of
money to affect changes in demand, output and the price level – was an
enduring feature of the American economy over many decades.
1 ‘What would Keynes say now?’, dialogue between Robert Skidelsky and Tim Congdon, December 2009 issue
Standpoint (London: Social Affairs Unit Magazines), pp. 30 – 35. The quotation is from p. 32.
2 Lanny Ebenstein Milton Friedman: a biography (New York and Basingstoke, England: Palgrave Macmillan, 2007),
pp. 13 – 25.
3 Ebenstein Friedman, p. 24. These notes still survive. One of Friedman’s teachers at Chicago in 1932 – 3 was Lloyd
Mints, who gave a lecture course numbered ‘Economics 330’ which was organized around The Treatise on Money. The
first words that Friedman wrote in his notes were ‘Econ 330 Keynes’. He then wrote that Mints’ judgement was that
‘General framework of Keynes likely to endure much longer than details’. (Robert Leeson ‘From Keynes to Friedman
via Mints’, pp. 483 – 525, in Robert Leeson [ed.] Keynes, Chicago and Friedman (London: Pickering & Chatto, 2003),
vol. 2. The quotation is from p. 485.)
4 Elizabeth Johnson and Donald Moggridge (eds.) The Collected Writings of John Maynard Keynes, vol. VI, A Treatise
on Money: 2. The Applied Theory of Money (London and Basingstoke: Macmillan, 1971, originally published in 1930),
p. 347.
5 The essay appeared originally in the June 1948 issue of the American Economic Review, vol. XXXVIII, pp. 245 – 64.
6 The argument was a theme throughout Simons’ Economic Policy in a Free Society (London and Chicago: University
of Chicago Press, 1948), which was published after its author’s death. Simons followed Irving Fisher in believing that,
by limiting the quantity of money to the cash reserves created by the central bank, macroeconomic instability could be
sharply reduced. (See Irving Fisher 100% Money [New York: Adelphi, 1935] for more detail.) However, it is clear that
banks which are forbidden from extending loans to the private sector (and so from making profits on such loans) will
need to charge for the money transmission services they provide to depositors. For the argument that the cost of banks’
services to non-banks declines with falls in the ratios of banks’ cash and capital to their assets, see pp. 39 – 84 of Tim
Congdon Central Banking in a Free Society (London: Institute of Economic Affairs, 2009). One point of Central
Banking in a Free Society is to reject the 100 per cent cash reserve notions as illiberal and inefficient.
7 See Milton Friedman’s essay on ‘A monetary and fiscal framework for economic stability’, pp. 133 – 56, in his Essays
in Positive Economics (Chicago and London: University of Chicago Press, 1953). The quotations are from p. 136 and p.
138 respectively.
8 When a private individual pays more in tax than he or she receives in government expenditure, his or her bank
deposits falls. So there is less money in the economy. Of course the government’s bank deposit rises, but the level of a
government’s deposit has no bearing on its behaviour. See footnote (10) to Essay 4 for more on this important
asymmetry.
9 The quotation is from p. 140. A closer reading of the 1948 paper supports this interpretation. It included a section (pp.
144 – 8 of Essays in Positive Economics) on ‘lags in response’ which devoted much more space to lags between fiscal
changes and expenditure than to those between changes in the quantity of money and expenditure. The whole
performance, from Friedman already in his mid-30s, was – in a word – ‘fiscalist’.
10 See ‘Interview with Milton Friedman’, pp. 198 – 218, in Brian Snowdon and Howard R. Vane Modern
Macroeconomics: its Origins, Development and Current State (Cheltenham, UK, and Northampton, USA: Edward
Elgar, 2005). The quotation is from p. 213.
11 Snowdon and Vane Modern Macroeconomics, p. 213.
12 Snowdon and Vane Modern Macroeconomics, p. 213.
13 Friedman & evidence to the Treasury & Civil Service Committee
14 The Treasury view is associated with Ralph Hawtrey, who was director of financial enquiries (in effect, chief
economic adviser) at the Treasury from 1919 to 1945. The definitive statement is usually regarded as his 1925 article in
Economica. (R. G. Hawtrey ‘Public expenditure and the demand for labour’, Economica, vol. 5 [1925], pp. 38 – 48).
But as late as 1939 Hawtrey continued to object in internal Treasury memoranda to Keynes’ neglect of the importance
of a budget deficit’s financing pattern to its macroeconomic effect. ‘…[T]he form in which money is raised [to finance
a budget deficit] may make all the difference.’ (The quotation is from p. 183 of G. C. Peden [ed.] Keynes and his
Critics: Treasury responses to the Keynesian Revolution 1925 – 46 [Oxford and New York: Oxford University Press,
for the British Academy, 2004].)
15 The underlying thought in the Ricardian equivalence theorem is that, if government spending is financed by debt
issuance rather than taxation, people understand that they will have to service and repay public debt in due course. So
debt issuance does not make them better-off. As they are no wealthier, total expenditure is unaffected by running a
budget deficit rather than by financing public spending entirely by taxation. (Robert J. Barro ‘Are government bonds net
wealth?’, Journal of Political Economy (Chicago: University of Chicago Press), 1974, vol. 82, issue 6, pp. 1095 –
1117.)
16 The exchange with Tobin in his 1972 Journal of Political Economy article, ‘A reply to the critics’, was partly neo-
Ricardian in character, before the 1974 Barro paper usually cited as the beginning of the neo-Ricardian argument. (See
pp. 914 – 5 in Friedman ‘A reply to the critics’, Journal of Political Economy, vol. 80, no. 5 [Sep. - Oct., 1972], pp.
906-950. Edward Nelson of the Federal Reserve has also brought my attention to two journalism pieces with a neo-
Ricardian slant. (‘Closet Keynesianism’ in Newsweek, 27 July, 1981, and an article in The Wall Street Journal of 26
April, 1984.)
17 Paul A. Samuelson and William A. Barnett Inside the Economist’s Mind: Conversations with Eminent Economists
(Oxford: Blackwell Publishing, 2007), p. 130.
18 Friedman believed that economists had come to accept the important role of money in the economy as a result of a
debate between himself and Meiselman on the one side, and Ando and Modigliani on the other, that had taken place in
the mid-1960s. This debate was nicknamed “the radio debate” after the initials (AM and FM) of the protagonists. See
Samuelson and Barnett Inside the Economist’s Mind, p. 131.
19 Milton Friedman ‘A natural experiment in monetary policy covering three episodes of growth and decline in the
economy and the stock market’, Journal of Economic Perspectives, 2005, vol. 19, no. 4, pp. 145 – 50.
20 On the need to confront theories with evidence in order to obtain a serviceable body of generalizations, the opening
essay (‘The methodology of positive economics’, pp. 3 – 43) of Friedman’s collection, Essays in Positive Economics
(Chicago and London: University of Chicago Press, 1953) is central. For Friedman’s cynicism about multiple
regressions, see Snowdon and Vane Modern Macroeconomics, p. 209, and Samuelson and Barnett Inside the
Economist’s Mind, p. 133 – 4.
