Download - Economic Growth and Distribution:On the

26/06/17
1
The Capital Critique in Pension. Saving, Investment and Growth in the Debate on
Pension Reform.
Sergio Cesaratto
Department of Political Economy
University of Siena
[email protected]
Paper prepared for the Conference on Economic Growth and Distribution: on the Nature and
Causes of the Wealth of Nations
Lucca 16-19 June 2004
Abstract
This paper will show the relevance of Keynes’ paradox of thrift, reinforced by the controversy on capital
theory suggested by Sraffa, to three themes related to the current debates on pension reforms. KeynesSraffa’s criticism is indeed particularly biting with regard to the relation between saving, investment and
output which is central in the pension debate. The paper is critical not only of the neoclassical approach,
but also of some positions expressed by less orthodox economists. More specifically it is argued that: (i) A
reform aimed to create a Fully Funded (FF) pension scheme (based on the accumulation of real assets),
even if successful at raising the marginal propensity to save (which should not be taken for granted), would
fail since, according to Keynes-Sraffa’s criticism, it does not follow that the larger potential saving supply
is necessarily translated into an increased amount of investment (more precisely in a higher per-capita
capital endowment). (ii) This is also true in an open economy: the criticism of neoclassical capital theory
and the theory of effective demand also dispose of the conventional idea that a higher saving supply from
the northern countries will naturally flow to the capital-poor southern countries stimulating investment
there. In this light, it is surprising to hear from less mainstream oriented economists that if a country opts
for pension funds, the risk is that savings are ‘exported’ instead of being invested at home. The risk is just
deflation at home. (iii) A more sympathetic view to a FF pension reform is expressed by Foley & Michl
who base it on a ‘Classical growth model’. However, their support of a transition reform towards a FF
scheme does not appear analytically robust since it lacks any plausible saving-based explanation of
accumulation (alternative to the neoclassical one that, after the capital controversy, has to be rejected).
26/06/17
2
Introduction*
The saving-investment nexus is central in the pension debate, an important aspect of which
concerns the impact of the different pension schemes, unfunded or funded, on economic growth. The
dominant, neoclassical approach considers transfer-based pay-as-you-go programs (PAYG) as injurious to
capital accumulation and favours the adoption of saving-based fully funded schemes (FF) that would instead
encourage it. An alternative approach (let us label it ‘Classical-Keynesian’) based on the extension to the
long period of the Keynesian postulate of the independence of investment from saving regards PAYG as
favourable (or at least neutral) with respect to accumulation and any reform aimed to encourage saving as
detrimental to aggregate demand and growth. This approach is based on Keynes’ paradox of thrift, reinforced
by the controversy on capital theory suggested by Sraffa which, as we shall see, is valid both in the short and
in the long run, in a closed as well as in an open economy.1 A third view, based on a ‘Classical growth
model’ is more sympathetic to a FF reform by taking investment as determined by saving presumably
assuming the validity of Say’s Law at least in the long period.
Section 1 will summarise some results of the attempt conducted by a number of Sraffian
economists to reinforce the implications of Keynes’ theory of effective demand for the explanation of
accumulation (that is with regard to the long period) in a direction that, however, is divergent from that taken
by the economists that work in the tradition of Joan Robinson and Pasinetti. Section 2 will single out the
nature of autonomous expenditure of PAYG’s transfers in the theory of effective demand. Section 3 will
criticise the proposal of a wider adoption of a FF scheme based on a ‘Classical’ growth model stance
advanced by Foley & Michl. Section 4 will cast-off the feasibility of an analogous proposal based on the
neoclassical growth model. The two final sections will consider the saving-investment nexus in the pension
debate in the open economy. Section 5 rebukes the mainstream argument that investment in the Southern
countries or the mechanism singled out by the Theory of Endogenous Growth are such as to assure the
profitability of a larger amount of old-age saving. Section 6 will consider the preoccupation expressed by
some critics of FF reforms that these are bad because the pension funds would ‘take domestic saving
abroad’.
1. The short and long period theory of Effective Demand
*
Financial support from the Ministry of Education & University is gratefully acknowledged.
26/06/17
3
In his General Theory, Keynes showed that, within the limits of full utilisation of output capacity,
a larger amount of investment does not require a prior reduction in consumption, and that the higher level of
output and income generated by the greater utilisation of capacity generates savings equal to decisions to
invest. Michael Kalecki proposed a similar approach. The ‘Neoclassical synthesis’ limited this mechanism to
short-period situations of low business and financial confidence, arguing that in those circumstances active
fiscal and monetary policies were required to reach full employment. This was the conventional wisdom
until, in the late 1960’s, the Monetarist revolution began to reaffirm the pre-Keynesian doctrines. Although
the ‘new classical economics’ subsequently receded, the currently prevailing conventional wisdom still
fundamentally reflects pre-Keynesian views. In the meantime, however, a number of non-orthodox
economists have tried the opposite strategy of extending Keynes’ analysis to the long period. Accordingly,
in the long run (when productive capacity may vary considerably), even more than in the short term,
investment is independent of, and, indeed, determines saving through, increases in output.2 Thus aggregate
output is determined by effective demand, defined as aggregate expenditures forthcoming at normal prices of
production.3
According to the theory of long period effective demand the engine of growth is represented by
the autonomous or final components of aggregate demand, those components that do not depend on the
actual or expected level of real income generated by firms’ decisions to produce. These include autonomous
consumption, government spending and exports. Cesaratto et al. (2003) have argued that gross investment
should be excluded from the autonomous components and rather be considered as induced by expected
demand.4 Autonomous expenditure embraces also social transfers from the State that contain PAYG’s
pension payments. Far from being a mere transfer of income, the theory of long period effective demand
regards the payment of pensions as an autonomous decision to spend by the government, as shown below. In
2
Cf. Garegnani 1983, 1992; Garegnani and Palumbo 1998; Serrano, 1996, Cesaratto et al. 2003; Palumbo
and Trezzini 2003.
3
Classical-Keynesian economists stress the distinction between normal and full-capacity utilisation. They
argue that firms tend normally to install enough capacity in order to be able to respond to sudden peaks of
demand, so as not to lose customers to competitors. This implies that a speed-up in the accumulation rate can
be accommodated by increasing the degree of capacity utilisation over the normal level, without the need to
cut wages, as implied by the models based upon the ‘Cambridge equation’ (cf. Ciccone 1990)
26/06/17
4
the meanwhile, it should be appreciated that in the Classical-Keynesian approach social spending can
represent an engine of growth, so that there is no necessary contradiction between greater social equity and
growth.5 Welfare State expansion, although favourable to stable economic growth may however encounter
an obstacle in the political acceptance of distribution changes – for instance higher taxation on profits - by
the most affluent classes. In post-Second World War historical conditions the Western economies
accommodated this expansion, but since the late 1970s the distribution role of the Welfare State has been
under constant attack (although at the price of weaker and more unstable growth). By contrast, in the
Marginal theory the contradiction between ‘equity’ and ‘efficiency’ is more mechanical. For instance,
according to this theory, pension transfers diminish disposable income and may reduce savings decisions and
accumulation.
The rescue of the long-period implications of Keynes’ theory of effective demand is reinforced by
Sraffa’s criticism of the Marginal capital theory. This criticism undermines the very possibility of deriving
the demand curves for factors in a rigorous and general way. This derivation relied on two alternative
mechanisms (e.g. Solow, 1970, ch.1): (a) direct factor substitution in production; and/or, in the case there are
fixed production coefficients, by (b) indirect factor substitution, which operate through adjustments in the
composition of consumers’ optimal consumption baskets in response to relative price changes. Sraffa not
only showed that they have analytical flaws, but also suggested that these substitution mechanisms were
absent in the Classical approach. This stimulated the controversy surrounding the neoclassical theory of
capital in the 1960s and 1970s, which refers precisely to the analytical plausibility of the decreasing ratio
between demand for factors and their remuneration.
The first substitution mechanism predicts that when, for instance, the wage rate falls, more ‘labour
intensive’ methods of production will become more profitable and labour demand will rise. The controversy
has shown that when there are a multiplicity of techniques and more than one type of capital good, the
possibility of reswitching techniques undermines this neoclassical prediction (Sraffa, 1960, pp.81-84;
Garegnani, 1970). Indirect substitution by reshuffling consumption is also precluded. According to
4
Palumbo and Trezzini (2003) are of the opposite opinion
5
The role of the Welfare State and PAYG’s pensions in the Classical distribution theory are discussed in
Cesaratto (2004c).
26/06/17
5
marginalist theory, a fall in the relative price of any factor leads to a fall in the relative price of and a rise in
the demand for the goods in the production of which the factor is used relatively more intensely. But as the
real wage varies from zero to maximum, the price of any commodity A produced using a given technique
may alternately fall and rise with respect to the price of another commodity B, produced using a different
given technique, so that no a priori expectations as to the direction of the change, based on the ‘factor
intensity’ in the production of the two commodities, are justified (Sraffa, 1960, pp. 37-38).6
This criticism is damaging both with regard to the marginalist theory of distribution, based on the
simultaneous determination of the rate of profit and of the wage rate as reflecting the relative capital and
labour scarcity; and with respect to its theory of output and employment, based on the tendency towards full
employment of both capital and labour - with the sole proviso that the financial and labour markets are
competitive enough not to obstruct the full adjustment of factor demand and supply. If the Classical
approach provided Sraffa with an alternative distribution theory, Keynes’, Kalecki’s and (the latter) Kaldor’s
criticism of the neoclassical theory of output and employment have been taken as the starting point for the
construction of an alternative theory of capital accumulation.
2.
Pensions, effective demand and economic growth
According to prevailing wisdom, the creation of a PAYG system is tantamount to an original and
reiterated sin: potential savings out of viz. wage growth that could be used for capital accumulation are
wasted in sustaining the old generation.7 This was not the prevailing view in the US and other countries at
the time of the inception of the public pensions programmes, well into the middle of the Keynesian
revolution. In those times the dominant view was that social transfer programmes were not only beneficial as
such, but also favourable to effective demand and accumulation. This reflected Keynes’s view that ‘measures
for the redistribution of incomes in a way likely to raise the propensity to consume may prove positively
favourable to the growth of capital’ (1936, p.373). Beveridge (1942) viewed full employment as a prerequisite for the Welfare State. This view took a more explicit Keynesian flavour in Beveridge (1944) in
6
General equilibrium models seem to avoid the capital measurement problems by taking the vector of capital
goods listed in physical terms as exogenously known data. This approach, however, opens up new problems
(cf. Garegnani, 1990).
7
Mainstream economists justify the existence of PAYG in various ways critically examined in Cesaratto
(2004c).
26/06/17
6
which social transfers were seen as a policy instrument for reaching full employment. Beveridge regarded
‘State action in re-distributing income by measures of Social Security, and by progressive taxation’ as
favourable to the growth of ‘private consumption outlay’ (1944, p.30), since ‘the income provided by the
scheme to persons who are sick, unemployed, injured or past work, will almost invariably be spent to the
full’ (ibid, p.160).8 In 1943 Metzler drew a distinction between the ‘Investment Multiplier’ and the
‘Redistribution Multiplier’. The latter referred to the effects of a subsidy levied on the ‘low propensity to
consume group’ in favour of the ‘high propensity group’. Subsequent discussion intertwined with that of the
famous Balance Budget Multiplier. In this debate PAYG transfers were considered with greater clarity as an
autonomous expenditure. When, one year after Metzler’s contribution, Wallich published his important
paper on the ‘Income-Generating Effects of a Balanced Budget’9, he started by taking as ‘generally assumed
that the income-generating effect of a balanced budget depends upon its “progressiveness” i.e. upon the
extent to which taxes and expenditures lead to a redistribution of monetary income from high-saving to lowsaving groups and thus to a rise in the average propensity to consume’ (1944, p.78). The Balanced Budget
Theorem actually set out to show that economic expansion was possible with a balanced government budget,
even with a ‘non-progressive budget’. The analysis was reconsidered later by Musgrave (1945; 1959, chap.
18), who pointed out that a full employment policy based upon the Balanced Budget Theorem could take
place either through an increase in public expenditure, or by an increase in public transfers that take
advantage of the differing marginal propensity to consume of different income groups, those taxed and the
beneficiaries of the transfers.
8
Beveridge regarded his proposed policy of full employment as a ‘policy of socialising demand rather than
production’ (ibid, p.190), and one that had to ‘be undertaken if free society is to survive’ (ibid, p.192).
Clearly, the failure of the market economies before the war and the hopes raised in vast sectors of the
population by the Soviet experiment, explain why the best minds in the bourgeoisie urged designing an
alternative both to extreme laissez-faire and to the socialization of the means of production.
9
As Wallich himself points out, the Balanced Budget Multiplier was independently discovered by a number
of economists. He mentions A.Hansen, P.Samuelson and W.Salant. Haavelmo (1945, p.311) mentions
P.Samuelson, A.Hansen, N.Kaldor and H.Wallich.
26/06/17
7
Suppose that, in a closed economy, the transfer beneficiaries have a marginal propensity to
consume equal to cr higher than the marginal propensity of income earners c. Using a standard textbook
notation, if transfers are increased by TR and taxation by TA  TR , we have:
Y 
1
1
c r TR 
cTA , that is:
1  c(1  t )
1  c(1  t )
Y 
1
(c r  c)T
1  c(1  t )
(1)
which is positive.10
In the case of pension transfers in a Balanced Budget context, the actual value of the multiplier
may be assessed in the light of the empirical circumstances that in each historical period govern the
propensity to consume of the various social groups. It may be thought that the propensity to consume of
retirees is higher than average, but not necessarily higher than the labour classes, and if the system is
financed precisely by payroll taxes on the latter’s incomes, then the value of the multiplier may be small.
Progressive taxation would tend to increase the value of the multiplier. If pension transfers are financed out
of deficit spending, of course, they would generally have a positive effect on Effective Demand. Given the
expected low value of the Balanced Budget Multiplier in equation (5.1), deficit financing appears to have
more marked income-generating effects.11
Only later, during the Monetarist revolution of the seventies, following the seminal contribution by
Feldstein (1974), did PAYG begin to be seen as detrimental to capital accumulation, since workers would
interpret contributions as a substitute for private savings whereas the olds consume most of the transfers they
receive.12 For instance, in his classic paper, Feldstein remarks that: ‘The evidence presented in this paper
seems ... consistent with the Keynesian view that the aggregate saving would increase as income rose if there
10
Note that in the multiplier we have used the marginal propensity of income earners, since the multiplier
effects on output do not affect transfer beneficiaries but only income-earners (cf. Musgrave, 1945, p. 390).
11
By the Keynesian multiplier process, a Social Security deficit can be financed out of the increase in private
saving. If these are mainly motivated by the ‘foresight’ motive, we may have created what I have elsewhere
called a ‘privatised PAYG’, that is a pension scheme in which pension transfers are financed (at least in part)
by issuing government bonds bought by workers through the pension funds (Cesaratto 2004b).
12
It is generally recognised that the elderly have an higher propensity to consume than workers (e.g. Engen
& Gale, 1997, p.111).
26/06/17
8
were no offsetting government policies’ (1974, p. 922), that is payroll taxes (‘offsetting government
policies’) would crowd out net saving out of rising wages. This suggestion is picked up, for instance, by the
World Bank which states: ‘suppose the government introduces a mandatory pay-as-you-go old age security
plan that requires young people to contribute payroll taxes (equivalent to their previous saving) to the plan
and pays them a pension (equivalent to their previous dissaving) later on. Introducing the system reduces
national saving initially, because the first group to benefit from the program … receive a windfall gain.
…fewer resources are left to be saved and invested during the initial period, permanently reducing capital
stock and national income’ (1994, p.307). The Bank, however, is forced to conclude that: ‘Despite the logic
of this argument, numerous empirical investigations (…) have been unable to prove conclusively that saving
did, indeed, drop once pay-as-you-go programs were established’ (ibid).13
To sum up, in the theory of Effective Demand, pension transfers are an autonomous component of
Effective Demand that, either by changing income distribution in favour of the social groups characterised
by a propensity to consume higher than those who are called to finance them, or by being deficit-financed,
may positively affect the level of the social product. Only in the case that the propensity to consume both of
the beneficiaries and of the contributors is the same, and the social security budget is kept in balance, do
PAYG pensions appear as a mere transfers without any effect on the level of National Income. From the
opposite view, according to the dominant theory PAYG’s transfers may displace saving and investment, but
the relevance of this displacement effect is questioned even by mainstream economists. Anyhow, the results
of the capital controversy give us enough self-confidence to rebut the neoclassical stances and reaffirm the
Keynesian interpretation.
13
To appreciate this uncertain result, suppose that the old receive a transfer Tr . The consequences on the
saving rate (the share of full employment income which is saved) hinges upon the combination of four
possible saving behaviours of workers and of three possible bequest reactions of the olds. On the one hand,
workers may be too poor or little provident to save, and in this case Tr levied on their wage does not
displace any saving. If savers, workers may cut their savings by Tr , or by sTr (where s is the marginal
propensity to save of the young), or not cut them at all. The old, on the other hand, who see their receipts rise
by Tr , may altruistically react by leaving their consumption level untouched, bequeathing, therefore, Tr .
This is of course the much discussed Barro’s Ricardian Equivalence case (cf.Engen & Gale, 1997, p.112).
Alternatively, the olds may leave just sTr by way of a bequest (perhaps because they see their children’s
saving falling by that amount) or, ultimately, they may greedily fully increase their consumption.
26/06/17
9
3. A ‘ Classical’ model of FF reforms
The Classical economists were quite open to diverse approaches to the theory of accumulation
(Garegnani, 1983, pp.24-28). In particular, mechanical forces leading to full employment are not to be found
either in Ricardo or in Marx. Ricardo believed in Say’s Law, which was instead rejected by Marx. The Long
Period Theory of Effective Demand suggested above combines the rejection of automatic tendencies to full
employment with the rebuff of Say’s Law (basing this rejection on Keynes, Kalecki and the results of the
capital controversy rather than on Marx). An alternative stance combining the rejection of automatic full
employment but accepting Say’s Law is pursued by Foley & Michl (2002; 2004; cf. also Michl, 2001) who
apply it to the pensions issue. We shall consider Foley and Michl’s (F&M for short): (i) distinction between
Classical and neoclassical theory, (ii) their ‘Classical’ model, (iii) their social security reform proposal and,
lastly, (iv) the role that they attribute to macroeconomic policy in the context of the reform.
(i) F&M consider the ‘theory of population’ as the main difference between the ‘Classical’ and the
neoclassical growth theory. The former would regard ‘population growth as an endogenous response to the
accumulation of capital’ (F&M, 2004, p.2; Michl, 2001, p.88). In neoclassical theory, on the contrary, the
rate of accumulation is adapted to the rate of growth of population. In this approach , labour as an exogenous
constraint to growth would frustrate any attempt to speed growth up by increasing the saving rate - in the
neoclassical context the saving supply out of full employment income - because of the operation of the law
of decreasing returns to capital accumulation. This would not happen in the ‘Classical’ model. More
precisely, in neoclassical growth theory a rise in the saving rate, given the labour supply growth rate, would
bring about three frustrating consequences: (a) it has only level effects on the output per-worker but not on
the rate of growth which is exogenously determined by population and technical progress; (b) the increase of
output per-worker is progressively less and less than proportional to (is this what you mean?) the rise in the
per-capita capital endowment - as shown by the standard decreasing slope of the well-behaved per-capita
production function -, and (c) the marginal product of capital tends to fall with the progressive abundance of
capital with respect to labour.14 In F&M view these implications would discourage any reform aimed at
14
No doubt that these dismal implications of a larger saving supply when labour supply is binding are
present in the mainstream literature on pensions. They are, for instance, mentioned in the following passages
by a conventional economist in which he assesses the ‘impact of ageing on macroeconomic performance’:
26/06/17
10
fostering old-age saving through a FF scheme reform (Michl, 2001, p.88). However, they argue, once the
idea of labour supply as an exogenous constraint to growth is dismissed, also the question of the decreasing
returns would be discarded. There are two aspects of F&M’s position to be discussed, one theoretical and the
other empirical.
Beginning from the first aspect, F&M do not criticise the analytical consistency of the marginal
theory. In the first place, they do not note that the mainstream conclusions that the rate of accumulation is
conditioned by the rate of growth of the labour supply - that acts, therefore, as a constraint -, depends on the
belief that in a market economy competition brings about a tendency to full employment. This tendency is in
turn based on the idea of decreasing demand functions for productive factors. The absence of these demand
functions has instead allowed the Classical economists to regard the determination of the accumulation rate
and of labour demand as independent of labour supply: labour supply is not mechanically binding in
Classical theory because there is not full employment. The observation of reality suggested to them that
capitalism did not suffer from labour scarcity, and that the labour supply tended to be responsive to periods
of faster accumulation, although the relation between labour demand and population developments was far
from the mechanical views that have been often attributed to them. 15 In the second place, F&M do not note
‘As regard to growth, it is widely considered that it will decelerate as ageing proceeds, principally because of
lower labour-force growth. There will also be lower growth in living standards (…) than has been the case in
recent decades, reflecting the accompanying increase in the dependency ratio. Effects on growth of a fall in
labour force growth are unlikely to be offset by higher investment. Indeed, investment is itself likely to
decline given a lesser need for capital widening, while capital deepening is likely to be limited by
diminishing returns. Moreover, slower growth will tend to reduce returns on capital directly, thus again
putting downward pressure on investment’ (Davis, 2004, p.2).
15
Classical economics displays a primary interest in the relations between population, wages and
accumulation. On the one hand, the centrality, in their theories, of the notion of wage rate as the amount of
commodities that labourers and their families must receive in order to survive suggests a linkage with the
circumstances that regulate population developments; on the other hand, the determination of the wage rate
on the basis of the bargaining power of labour suggests a connection between the speed of the accumulation
process and population developments. Having said so, with the exception of Malthus, the Classical
economist were far from the simplistic view of these inter-relations that have been often attributed to them.
In particular: (i) labour demand, and not just the wage level, was seen as inducing changes of population; (ii)
the level of wages could influence population in either direction; finally, (iii) unfavorable population
developments and the accumulation rates had not mechanical and immediate negative influences on the wage
26/06/17
11
that the notion of decreasing returns to capital accumulation is also a theoretical result derived from the same
marginal apparatus. The two authors seem to argue as if decreasing returns were not an a theoretical
construct, but a factual result similar to Ricardo’s classification of lands according to their different fertility.
F&M appear thus to limit themselves to take issue with the empirical side of the hypothesis of labour
constraint, that is whether de facto a situation labour scarcity is actually biting or it will in the near future.16
Given the present fertility trends in developed countries, the two authors regard, and with some
reason, immigration as the most responsive source of labour supply to the U.S. accumulation patterns, in the
past as well as into the future. There are three problems here, however. First, this may not be true in the long
run for a number of developed not-settlement countries such as Europe and Japan. This opens new problems
that cannot be addressed here.17 More importantly, if labour supply is not a constraint to accumulation in the
Classical framework (for a given capital-labour ratio), this would also be true in a neoclassical setting. This
shows that it is not enough to distance themselves from marginal theory in the way F&M do. Finally, even if
rate, that depended on lasting social habits and conventions (cf. Stirati, 1994, Chapter 4 and pp.175-176).
Marx regarded the existence of an excess of labour population over the job opportunities as a necessary
check on wages, but he considered population movements too slow to be behind the preservation of a labour
reserve army over the cycle and during the process of accumulation –a role that was not, however, excluded
in the long run. The modalities of the accumulation process itself would rather take care of the industrial
reserve army, in particular by the weakening of the accumulation process, for a given ‘organic composition
of capital’, and by the progressive substitution of variable capital (direct labour) with constant capital in the
production process that diminished labour demand in the course of the accumulation process (Marx, 1867,
vol. I, Chapter XXV, sections 1-3).
16
F&M also argue that once the real wage rate is exogenously given, ‘the capital intensity of the technique of
production and the profit rate are invariant in the short as well as the long run’ (F&M, 2004, p.2). But in this
way they base the absence, in the Classical framework, of decreasing returns to capital accumulation on the
hypothesis of a given wage rate – an hypothesis which is certainly close to the Classical economists – and
not on the (weaker) hypothesis that in the Classical context labour is not binding, but, once more, this is not
accompanied by a more general criticism of the marginal apparatus. There is a similarity to what happens in
mainstream theory with the non-substitution theorem according to which when the wage rate is exogenously
given and there are constant returns to scale, then distribution is not affected by output levels (Garegnani
[1985, p.128] shows that this is not enough to criticise neoclassical theory).
17
Unless population reaches explosive levels, in most traditional non-settlement countries immigration can at
best help to stabilise the population in the working-age (cf. UN, 2000). The big question to be explored is
how capitalism can operate with a limited labour reserve army in an ageing society (Cesaratto 2004a).
26/06/17
12
the supply of material labour in developed countries was binding, investment in developing labour-rich
regions and Endogenous Growth Theory (EGT) have represented for neoclassical economists a safety belt to
the three discouraging results enlisted above. Yet, F&M do not discuss (at least in the pension context)
foreign investment and EGT, which we shall do in section 5.
(ii) Let us now consider F&M’s ‘Classical’ model on which they base their proposal of social
security reform. This model is simply based on the assumption, poorly justified, that in the long run capital
accumulation and employment growth are determined by saving. 18 The authors do not go much further than
pronouncements like: ‘The Classical model we employ in this analysis assumes full utilisation of the capital
stock, and thus excludes consideration of aggregate demand effects of social security’ (F&M 2004, p.3); or:
‘the mere possibility of a problem with aggregate demand does not absolve us from making decisions that
are fundamentally long-run in nature’ (Michl 2001, p.87). In their book F&M argue that since in ‘the long
run, the economy would be attracted to the full utilization of capacity’ (u = 1), then an independent
investment function ‘becomes superfluous at u = 1’, and endorse the idea that ‘the economy is Keynesian in
the short run but Classical in the long run’ (F&M, 1999, p.192). Both the Classical and neoclassical
traditions, they specify, ‘see saving as the engine of capital accumulation and assume that saving decisions
always lead to a corresponding decision to invest. In these models Say’s Law holds, and there can be no
discrepancy between aggregate demand and supply’ (ibid, p.193, italics in the text). There are two problems
18
According to F&M in the ‘Classical’ model (as well as in the neoclassical model) ‘aggregate accumulation
will be determined by the saving of the most “patient” class of households’, the class that the Classical
economist would identify with ‘capitalists’ (F&M, 2004, p.2). This suggests that, in spite of the authors’
contention that their model is in the tradition of the ‘Cambridge equation’ (F&M 2004, p.1; Michl 2001,
p.90) - in which are the ‘animal spirits’ that determine the rate of accumulation g and, as a consequence, also
the profit rate according to the well-known equation g  sc r (in which sc is the saving propensity of the
capitalists) -their model resembles more the EGT’s AK model (e.g. Rebelo 1991). Indeed, the mentioned
assumption of an exogenous wage rate (F&M 2004, p.2 and p.6) is at odds with the endogenous
determination of distribution in the Cambridge equation. As suggested by Solow (1992), the AK model is
nothing other than a resumption of the Harrod model in which it is the saving rate that determines the rate of
capital accumulation according to the famous equation of the ‘warranted’ rate of growth: g w  s vd (in
which v d is the desired capital-output ratio) and distribution is exogenous. A consistent neoclassical
26/06/17
13
here. First, neoclassical economists advance a sophisticated argument to show the adequacy of investment
decisions to full employment saving decisions, which is lacking in the Classical theory endorsed by F&M.
Secondly, it was certainly Keynes himself in the General Theory who first repeatedly suggested that
although his theory was valid when resources were less than full utilised, traditional theory resumed its
validity once full employment was restored.19 However, the extension to the long run of the Keynesian
principle of the autonomy of investment from saving (see section 1) does not depend on denying the
tendency of capacity to adjust, in the long run, to long period aggregate demand - as in the neo-Kaleckian
models that F&M (1999, chapter 10) take as the only example of long period Keynesian models. 20 Whereas
in the short run the adjustment of saving to investment takes place through variations of output, it can be
argued that, in the long run, the tendency of saving to adjust to the long run level of investment takes
precisely place through the variations of capacity (cf. Garegnani 1992; Garegnani and Palumbo 1998, and
for a criticism of the neo-Kaleckian models see Trezzini, forthcoming). In conclusion, a long run tendency
toward normal utilisation does not imply that, in the long run, increases in saving will translate into increased
accumulation, as F&M seem to believe.
(iii) On these fragile foundations F&M suggest a social security reform based, in synthesis, either
on an additional payroll tax on wages, or on capitalists’ profits or wealth or, finally, on the utilisation of a
possible on-budget government surplus, which should preferably be the outcome of progressive taxation, to
constitute a reserve fund held by the social security administration that should be used to foster (private)
foundation of the AK model is presented in an unfortunately forgotten paper by Frankel (1962) (cf. Cesaratto
1999a, 1999b, Serrano & Cesaratto 2001).
19
For instance he wrote: ‘If the rate of interest were so governed as to maintain continuous full employment,
Virtue would resume her sway; - the rate of capital accumulation would depend on the weakness of the
propensity to consume’ (Keynes, 1936, p.112).
20
In this regard Michl (2002, p.114) writes: ‘[our] analytical choice (shared by economists across the
theoretical spectrum) makes good sense in describing an economy whose behavior is Keynesian in the short
run but classical in the long run. Macroeconomic models, such as that of Badhuri and Marglin (1990), that
attempt to give meaning to demand-constrained growth require firms to operate with unplanned excess
capacity even in a long-run equilibrium when they are continuing to add new capital. Many find this to be
intellectually unappealing. …these models are more convincingly interpreted as representations of the shortrun behavior of a system that has classical characteristics in the long run. Social security, fundamentally a
long-run macroeconomic policy, demands long-run analysis’.
26/06/17
14
capital accumulation (F&M 2004, p.10; Michl 2001). Supposing that the reserve fund is financed out of
increasing payroll taxes – and assuming that workers do not cut their voluntary old-age saving –then its
constitution would demand ‘a period of primitive accumulation, during which taxes exceed benefits’ (2004,
p.10). ‘If this surplus’ – they continue –‘ is built up from the payroll taxes of workers, it implies a kind of
forced saving in which one or more generations of workers contribute to a fund which benefits future
generations’ (ibid). Since the rate of profit r is considered to be higher than the rate of income growth g –
which, given the share of wages in income and the PAYG’s. contribution rate, is the rate of growth of the
wage bill, often regarded as the rate of return on PAYG’s contributions –, then this accumulation would
enable labour to receive some of the results of capital growth and even to permit a reduction of the
contribution rates necessary to obtain a final target replacement rate (the initial pension benefit over the last
wage). The suggestion is that the sacrifice of some generations, that are called upon to increase their
mandatory saving, and provided that they do not cut their discretionary saving, is invested by the social
security fund in private capital accumulation, thus creating a fund of real capital assets that can later finance
pensions out of profits: ‘With returns from the reserve fund subsidizing the system, a given level of benefits
can be supported with lower taxes, increasing the money’s worth of the payroll tax’ (Michl, 2001, p.85).21
21
The idea that the higher ‘profitability’ of a FF scheme and the continuous reinvestment of profits would
consent a reduction of the contribution rates necessary to reach a final target replacement rate is common
with the reforms proposals of Feldstein (1996) and of Modigliani et al.(1999) (cf. Cesaratto 2002). If the
creation of the reserve fund is financed out of additional payroll taxes (additional to those necessary to
finance the current payment of existing PAYG’s pensions), then although workers see their net wage initially
falling, the same workers or the subsequent generations of workers will later see their net wage increasing.
This is because the higher returns from the invested reserve funds will allow a reduction of the contribution
rate. There are also some differences between F&M and the neoclassical Modigliani/Feldstein reform
proposals that may help to clarify what F&M have in mind. According to Modigliani/Feldstein a net increase
in old-age mandatory saving does lead, given a fully employed labour supply, to a higher capital-labour ratio
and, consequently, to a higher per-capita product. The advantage of the reform lies therefore in the workers’
higher endowment of real capital and productivity (cf. section 4 below). F&M assume instead a constant
capital-labour coefficient and the existence of a perfectly elastic labour supply. In this case, the mandatory
net saving is used to provide extra workers with the average capital endowment, but there is no increase in
labour productivity. Profits out of the new capital, however, now pertain to the labour class through the
public fund. So, although the product per worker has remained the same, this is not true for the total income
accruing to the working-class. The extra income is used to finance pension expenditure and reduces the
26/06/17
15
Moreover, thanks to the sacrifice of some previous generations, the working class (in particular retired
workers) will end up, through Social Security possessing a share of the larger capital stock and earning
profits. This would explain the opposition of conservatives to this kind of pension reform.
As seen, F&M endorse Say’s Law in a rather naïve way, assuming that saving decisions determine
investment. F&M are well acquainted with the capital theory critique (cf. F&M 2004; Foley 2001), so they
avoid the marginalist decreasing demand curve for capital, but do not suggest any robust alternative
adjustment mechanism between supply and demand for capital.
Ricardo’s original view of accumulation
was based on the coincidence of the saving and investment decisions that, in a view that was perhaps not so
unjustified for that time, were taken by an entrepreneurial capitalist class that saved in order to invest. The
Ricardian view is echoed by F&M when they argue that ‘growth depends critically on the existence of a
capitalist class which accumulates wealth for bequest purposes’ (2004, p.19). Bequest may well explain
saving decisions, as in some neoclassical models that F&M quote but, as noted above, the mainstream
authors put forward a sophisticated mechanism – the capital demand and supply apparatus which is the
object of the destructive capital theory critique - whereby saving decisions are translated into investment
decisions. F&M have none, and the decisions to buy a machine tool cannot be easily justified by the will to
leave a bequest.
(iv) F&M limit the role of aggregate demand in the determination of the growth rate to the short
run (Michl 2001, p.87) and rely on fiscal and monetary policy to compensate the deflationary effects of a
rising propensity to save (F&M 2004, p.3). The employment of fiscal policy is self-contradictory since it is
equivalent to destroying with the right hand (the creation of a budget deficit), what the left hand has just
done (the constitution of a budget surplus to finance the social security reserve fund). More consistently
Michl concedes that ‘the fiscal surpluses needed to fund social security represent a deflationary policy, in so
far as the short run is concerned’, and relies particularly on monetary policy as an effective instrument to
avoid deflation in the short run (2002, p.113).22 So the curious policy mix that he proposed is a fiscal surplus
pressure on active workers. Alternatively, a higher labour productivity can be obtained in this class of
models by introducing some sort of externality from capital accumulation, as in the mentioned AK models of
endogenous growth theory, or else some Kaldorian technical progress function.
22
Michl is here replying to the Keynesian criticism by Palley (2001,) who argued: ‘Michl’s vision is
predicated on a “corn” model approach to economic process, whereby increased saving (corn not consumed)
26/06/17
16
to stimulate investment in the long run (ibid, p.115), plus an expansionary monetary policy to avoid the
deflationary effects of the same fiscal surplus in the short run. Traditionally, however, Keynesian policies
have regarded monetary policy as rather ineffectual on investment, given their lack of elasticity to the
interest rate (although this is not the case of consumers’ credit or residential investment), and expansionary
fiscal policy as the most effective. To rely on monetary policy to sustain private investment in order to
absorb the higher potential saving supply that, according to F&M’s wishes, would be brought about by the
social security reform appears, frankly, unsustainable.23
Incidentally, Say’s Law also leads F&M to regard PAYG as detrimental to capital accumulation
‘by discouraging life cycle saving’ [2004, p.11]. In the first place this displacement effect is not so obvious
since not necessarily would workers have saved the mandatory contributions to PAYG (cf. section 2 above).
Secondly, there is no reason why these saving are necessarily invested.
To sum up, F&M recommend creating a Social Security fund to be used to foster capital
accumulation. Our criticism has pointed in three directions. To begin with, F&M criticism of neoclassical
theory appears rather weak and not analytical, being based on a factual observation that the US will not
suffer from labour scarcity in the near future (or at least will have less troubles than other developed
countries). But if this is true in their ‘Classical’ context it will also be true in the neoclassical framework that,
is immediately translated into investment (corn planted in the ground) to yield more income (corn) in the
future. This is the implicit logic behind the claim that buying equities automatically results in investment.
This vision of the investment process is at odds with the reality of a monetary economy. Buying equities
already in issue simply transfers money from the bank account of the buyer to the bank account of the seller,
and does not result in more investment (…). Dissecting the transaction in this fashion makes clear that it is
investment that drives saving, rather than the other way around as assumed in corn model economics’ (Palley
2002, p.106). Michl [2001, p. 114] recognises that he is using a corn model. The capital theory critique
allows us to extend this criticism to the more sophisticated neoclassical explanation of the relation between
saving and investment.
23
It might be thought that if the budget surplus is obtained by taxing profits or capitalists’ capital assets, the
deflationary effects are negligible since taxes will mainly hit savings or financed out of decumulation of
wealth. But in this case the only effect would be a reshuffling in the property of capital but not higher growth
(even if saving caused investment). In the case of a tax on profits, ceteris paribus, the larger government
saving would compensate the lower capitalists’s saving. In the case of a tax on capital wealth, the reserve
fund would just buy the capital assets sold by the capitalist to pay the new taxes.
26/06/17
17
therefore, will not meet the mentioned ‘decreasing returns’ to capital accumulation. Secondly, F&M’s effort
to support a transition reform towards a FF scheme by basing it on a Classical model does not appear
analytically robust since it lacks any plausible saving-based explanation of accumulation. Finally, F&M
admit the necessity to compensate the deflationary effects of the proposed reform by using fiscal and
monetary policy. But an expansionary budget contradicts the very essence of their reform which is based on
increasing government saving, while monetary policy is ineffectual with respect to investment.
4. The Marginalist foundations to FF reforms
Whereas F&M mainly regard a FF reform as a tool to increase labour’s share of capital wealth,
mainstream economists see it as an instrument to deal with the demographic developments which are taking
place – a possible fall in the full-employed labour-supply and higher longevity. According to marginal
theory, capital consists of a fund of consumption goods through which consumption can be postponed. The
existence of a decreasing demand schedule for capital elastic to the interest rate assures that a rise in
‘foresight’ saving is matched by additional capital accumulation, in particular, given the labour supply, by a
higher per-capita capital endowment. In the case of a fall in the labour supply, the accumulated capital could
be either absorbed by a higher capital-labour ratio, or reconverted into consumption goods. Increasing
longevity can be dealt with by increasing the individual saving rate, working longer, spreading the
consumption of the accumulated capital over a longer period.24 Synthesising a longer argument (I refer the
reader to Cesaratto 2004b), I have envisaged three problems with the mainstream suggestion of a FF reform.
To begin with, the reform may fail to tangibly raise the marginal propensity to save even in
economies in which there is no mandatory pension scheme in operation, since the mandatory old-age saving
may substitute similarly motivated discretionary saving.
Secondly,reforms that rely on public debt to finance the transition in economies in which the
government is committed to honour the current PAYG’s pension promises, are doomed to fail, since the
higher private saving is compensated by a lower government saving. Prima facie, what happens is that once
workers’ contributions to PAYG are diverted to the pension funds, these use the funds to finance the public
debt that the government has in the meantime issued to finance the pension commitments. In practice the
24
The limits of the capacity that the neoclassical economists attribute to a FF scheme to cope with the
demographic developments are discussed in Cesaratto (2003).
26/06/17
18
reform is a privatisation of PAYG. Although this is well known, these reforms are still proposed both by
individual scholars and by international organizations as genuine shifts towards FF programs. The World
Bank itself acknowledges that the significance of privatising PAYG lies in the social alarm it may generate
about its sustainability, creating a climate more favourable to its dismantling.
Thirdly, the attempt to increase the average community propensity to save by imposing extra
contributions on workers in favour of an FF scheme will have deflationary effects. The criticism of the
neoclassical view of the saving-investment relation (see above section 1) is the ultimate challenge to the
conventional view of capitalisation reform. So, not only is it difficult for policy makers to raise the
propensity to save, but even if successful, the effects on investment will be nil and the reform prove abortive.
5. The saving-investment nexus in open economies and in endogenous growth theory
Another standard argument often heard is that Southern countries will be the natural outlets for the
abundant capital supply from the ‘grey’ developed countries. According to Reisen (2000), for instance,
current investment in southern countries will raise the rate of return on pension saving and help to meet the
expected abundance of supply of financial assets when the ‘baby-boom generation’ retires. Although
demographic changes are expected also in southern countries, they will be much less dramatic and much
slower there, so that a market for the capital assets supplied by the retiring Northern baby-boom generation
will not be missing in them.25 This theoretical prescription is warmly subscribed to by the international
institutions. Let us explore this argument.
25
Reisen is worth quoting for his clear exposition of this standard argument: ‘In the absence of foreign
pension investment into younger economies, what should we expect to happen to capital returns on funded
pensions once the OECD baby boomers have started to retire? As the labour force declines, the existing
capital stock becomes oversized relative to the labour force. The change in relative factor proportion reduces
the rental return on capital relative to wages; this effect is reinforced if fully-funded pensions indeed
stimulate savings. Simultaneously, the prior phase of asset accumulation would give way to a long period of
asset decumulation, as the baby boomers start to draw on their pension assets to finance their retirement.
Clearly, therefore, a fully funded pension scheme is bound to get under stress by population ageing, very
much like an unfunded scheme. But the funded pensions, unlike the unfunded schemes, can partly beat
demography in an open economy. The asset decumulation during the retirement period will not be confined
to home assets, but to emerging-market assets that still will be benefiting from net pension contributions of
the underlying younger population. And capital returns, unlike in a closed economy, will not be lowered by a
26/06/17
19
First run back over some basic Balance of Payments relations. As we know from national
accounting, using the standard textbook notation, the Gross National Product of an open economy is equal
to:
GNP = C + I + G + (E – M) + X
(2)
where X represents net income from abroad (that includes, for instance, emigrants’ remittances and
interests payments on capital lending) and (E – M) + X is the current account. Let us assume, according to
the neoclassical theory examined in this Chapter, that the GNP is at its full-employment or natural level.
Subtracting private consumption and public outlay from both sides of the equation we obtain the following
expression for national saving:
S N = I + (E – M) + X
(3)
(E – M) + X can be defined also as foreign saving (or foreign investment), S F , so that
S N  I  S F . If S N  I the country is using foreign saving – actually is importing financial capital. If
S N  I the country is lending abroad.26
As known, in Solow’s model of growth a decline in the rate of growth of population leads to a new
secular equilibrium characterised by a higher per capita capital endowment and a lower rate of interest.
According to the dominant opinion, Southern countries are generally in an opposite situation: a younger and
abundant labour force accompanied by a scarce capital endowment. Hence, these countries offer an higher
rate of return on capital investment than developed countries. As a result, if capital funds can move freely,
the abundant saving supply from developed countries will tend to flow to more profitable Southern countries,
thus helping the latter to avoid the low saving trap, and the Northern countries to cope with demographic
imbalances without incurring too low interest rates on the pension funds.
Having said this, let us now examine the fears sometimes expressed that investment in foreign
countries by pension funds may lead to the loss of ‘national savings’ that are useful for national investment.27
declining labour force, but by the world capital market and the demand for capital by the younger nonOECD area’ (2000, pp.3-4).
26
Clearly, in the short period in the former case the country can melt away its net financial wealth by
financing the current account deficit through the foreign reserves, or hoard the current account surplus
improving its foreign reserves.
26/06/17
20
Conceding that the introduction of a capitalisation scheme leads to a rise in national saving, according to the
traditional theory this might be partially offset by a rise of domestic I, that is, of endogenous absorption, but
also by a rise of foreign lending to Southern countries, if this is convenient to do so. Mainstream economists
may thus confidently argue that, on the one hand, a positive net foreign lending of pensions savings to
‘emerging markets’ increases the global welfare and, on the other hand, this is not associated, in the long run,
with a loss of domestic investment (cf. Reisen, 2000, cap.2 and 3). Aptly, in their textbook Krugman and
Obstfeld (1994) deal with foreign lending by Northern to Southern countries as intertemporal trade, that is a
channel by which the former nations can conveniently postpone their consumption exporting capital to those
countries where this is relatively scarcer.
It should be appreciated that in the case of foreign investment, no less than in the domestic case,
the idea that domestic saving finds an automatic debouche in investment in Southern countries depends on
the neoclassical saving –investment relationship and it is therefore subject to the critiques offered or made
above. In the non-orthodox context there is no automatic mechanism that would translate the larger
(potential) saving supply into domestic or foreign investment since a fall in the rate of interest does not affect
investment either in the domestic economy or in that of other countries. The result, therefore, is a fall in
national income and employment, as we shall see in the next section.
As an alternative to lending to Southern countries, conventional economists have picked up the
outcomes of Endogenous Growth Theory as a source of a solution to the excessive supply of capital and
consequent decline in its marginal productivity that may result from demographic shocks. The results of
EGT are not, however, very robust. It is only by employing an impressive number of ad hoc devices, that this
approach assumes away the traditional causes of a falling marginal product of capital. The main idea is that
any rise in the full employment saving rate, for a given labour supply, is an opportunity for technological
advancement – for instance externalities from capital accumulation or increasing resources for R&D
activities – that ex hypothesis takes the form of ‘labour augmenting’ technical change. This keeps the ratio
between capital and labour - the latter measured in ‘efficiency units’ – constant, so that the marginal product
of capital is unvaried. The present author has not only demonstrated elsewhere that Endogenous Growth
27
Diamond, for instance, observes in passing that while ‘economists generally favor having part of the
portfolio invested abroad, this would be controversial in the United States today’ (1996, p.78).
26/06/17
21
Theory is far from being a new approach, since it was already well known in the 'Sixties and discarded at the
time for being too simplistic and ad hoc (cf. Cesaratto, 1999a, 1999b where other criticism by distinguished
neoclassical authors, e.g. by Solow, Stiglitz and Hahn, is quoted), but has also added new analytical bases to
reject it.28
6. A Keynesian view of international capital flows
28
Serrano and Cesaratto (2002) condenses the analytical results of my own work over neoclassical growth
literature. It shows the high price that those theories pay for being based on the neoclassical theory of
distribution, which in these models is used to ensure the full employment of the labour force and the full
utilisation of the capital equipment. It is shown that it is the need to be consistent with these conditions (by
themselves quite unrealistic) of full employment of factors that forces these theories to rely on extreme and
highly artificial assumptions concerning technology and technical progress. In short, it is the neoclassical
theory of distribution and of factors’ utilisation that requires decreasing marginal returns for capital
accumulation (and in general for the increase of the use of any factor of production). The decreasing
marginal returns for capital accumulation is what underlies the widespread dissatisfaction with the
exogenous growth models such as that of Solow, that make the steady state rate of growth of the economy
independent of the saving ratio. Decreasing returns to capital accumulation are, in general, retained even
when externalities and economies of learning are included (together with embodied technical progress). To
worsen the situation, these variants of the Solow model with externalities (and more generally any
neoclassical model of growth with increasing returns to scale) are shown to unavoidably generate the
particularly implausible result of a strong positive causal relation between the rate of growth of population
and the rate of growth of labour productivity, a relationship that seems to suggest that a permanent
demographic explosion represents the fastest way for, say, Southern countries to catch up with the first
world. It is therefore concluded that the reason why, in general, Solow's model must introduce the
assumption of constant returns to scale is not due to analytical difficulties, but to avoid results that are even
more implausible than those of the original model. The striking characteristic of recent theories of
endogenous growth is the complete elimination of the decreasing returns both to physical capital (‘AK
models’) and to stock of ‘knowledge’ (a la Lucas), so that the factor that can be accumulated has constant
marginal returns. As mentioned in the main text, these models, which already rely on very artificial
hypotheses about technology when the labour force is assumed to be constant, require even more extreme
and artificial hypotheses in the case in which the labour force grows. These additional hypotheses must bring
down to zero the marginal product of labour, since otherwise the economy, with constant returns either for
physical capital or ‘knowledge’, will display ever accelerating growth rates for any positive rate of growth of
the labour force. A more sympathetic view of Endogenous Growth Theory has been expressed in various
papers by Kurz & Salvadori (e.g. 2003). Their approach to accumulation resembles that of F&M, but a
detailed criticism must be deferred to another paper.
26/06/17
22
In the preceding section we used a neoclassical perspective to discuss the objection to a
capitalisation reform that the pension funds would divert national savings abroad that could alternatively be
used to finance domestic investment. We argued that according to the conventional approach a positive net
foreign lending from Northern to Southern countries, on the one hand, reflects the relative abundance of
capital in the former nations and its relative scarcity in the latter and it is therefore economically convenient
(that is the increase of per-capita income due to the domestic employment of that saving is lower than the
rise due to the interest revenue from the foreign investment). In this context, a capitalisation reform that
succeeds in raising the domestic propensity to save, that is the saving supply out of a given national income,
can be used depending on its relative profitability for domestic or foreign investment. From the point of view
of aggregate demand the choice is equivalent, since the larger saving supply is matched in the first case by a
higher domestic investment and in the latter case by a greater amount of exports. It is important to appreciate
that also in an open economy it is the conventional relation between saving and investment, based on the
inverse relation between investment and the interest rate - this time that determined in the international
financial markets -, that assures that the greater saving supply (from the Northern nations) is translated into a
greater investment level in the Southern countries. In other words, an increase in the saving supply from
Northern countries determines a fall in the interest rate in the international financial markets that induces an
increased amount of investment in the Southern countries.29
In the light of the criticism of neoclassical capital theory these theoretical outcomes cannot be
accepted. Conceding that a capitalisation reform induces a rise in the marginal propensity to save, in the
non-orthodox context there is no automatic mechanism that would translate the larger (potential) saving
supply into domestic or foreign investment since a fall in the rate of interest does not affect investment either
in the domestic economy or in other countries. The result, therefore, is a fall in national income and
employment.
29
It is assumed that in the Northern countries capital is so abundant relative to labour that the demand for
investment has become relatively insensitive to a fall in the rate of interest; the opposite is true in the
Southern countries where the relative scarcity of capital renders investment demand highly reactive to a fall
of the interest rate.
26/06/17
23
This can be seen in terms of the national account identity S  I  (G  T )  ( X  M ) , the
neoclassical economists would argue that a rise in S , given domestic I , causes a rise of I in the south
followed by a rise in domestic X . According to Keynesian economists, a rise in the propensity to save in the
north, given domestic I , causes a fall in national income accompanied by a fall in M (incidentally, this
produces a fall of income also in the south). North’s foreign saving has increased, and so perhaps pension
funds created, but at the cost of a lower national income. This is shown by this simple example: with
c  0.8 , m  0.2 , E  200 , output is Y  500 , and S  X  M  100 . If c  0.7 , we get Y  400 and
S  X  M  120 . This disappointing result is an extension of Keynes’ saving paradox to an open
economy: an increased amount of foreign saving is realised at the price of a lower domestic income. In an
enlightening paper – that was brought to my attention after this section was written –, Dalziel & Harcourt
(1997) also extend Keynes’s multiplier analysis to an open economy. The capital theory critique reinforces
their point of view.
In this light, it is surprising to hear from less mainstream oriented economists that if a country opts
for pension funds, the risk is that savings are ‘exported’ instead of being invested at home. 30 In the light of
the theory of effective demand this assertion is wrong and a non-Keynesian argument can be envisaged
behind it. As argued above there is no automatic mechanism that guarantees that domestic or foreign
investment will increase, so that the potential greater saving supply does not actually materialise. The
adjustment is brought about by a fall in national income. True, this fall will result in a reduction of import
and an improvement of the current account, so that the country improves its foreign lending capacity.
However, this does not resemble what the economists mentioned earlier seemed to have in mind. They are
referring rather to a more direct causality from a capitalisation reform, through a rise in national saving, to a
rise in foreign lending. They miss the Keynesian point that a country cannot achieve an increase in its saving
30
In Italy, for instance, Pizzuti has advanced this thesis: ‘a very real risk for the Italian economy has been
noted. A recent survey in the daily financial newspaper of the Italian industrialists’ association, analysing
investment opportunities for Italian pension funds, found “high performance and low risk in foreign stock
markets”, while “the Italian market is outperformed even by those in emerging countries”. The conclusion is
that “in the long run, the highest-yielding asset is foreign shares”. In short, the development of private
pension funds could result in an additional outflow of Italian savings’ (1998, p.58).
26/06/17
24
supply for domestic or foreign uses at will, by the mere institutional design of a new mandatory pension
scheme. 31
Final remarks
In the light of the theory of Effective Demand, PAYG is an institution that has a double linkage
with full employment policies: its sustainability both depends on them and is itself an instrument of those
policies. As we have seen, this was in fact recognised very early on by William Beveridge, whose famous
report on the welfare state (1942) ‘called for the maintenance of full employment’ in order to finance social
insurance, although it did not initially rely on Keynesian instruments to achieve that aim (Dimand, 1999, p.
232). Beveridge soon came to see that the ‘policy of full employment’ he was advocating was ‘a policy of
socializing demand’ through Social Security and progressive taxation aimed at directly attacking ‘the central
weakness of the unplanned market economy,’ that is, its ‘failure to generate a steady state effective demand’
(Beveridge, 1944, pp. 190–191). Whereas in a neoclassical perspective PAYG is a mere sharing of full
employment income, and can be detrimental to capital accumulation, in a Classical-Keynesian perspective
31
The conventional approach to international capital flows echoes the time honoured Robertsonian Loanable
Fund theory. Accordingly, behind capital flows there are surplus countries – those which present an excess
of domestic saving over domestic investment – that lend this excess to borrowing countries – those who
invest more than the domestic saving supply. An alternative approach might be tentatively sketched as
follows. In the absence of controls over capital inflows, the international private financial institutions are
able to create credit facilities for Southern countries t to increase their imports – only in the most fortunate
cases of investment goods, more often of consumption goods. It is this increasing demand for exports from
the Northern countries that generates the ‘twin deficits’ (both negative foreign saving and trade balance) in
the Southern countries, and the corresponding ‘twin surpluses’ in the North, in which the increasing exports
determines a rise of income and saving. In this alternative account, ‘credit precedes investment and both
precede savings’, as Kriesler and Halevi (1996, p.309) effectively put it, and the saving-investment gap is
‘nothing but the ex post accounting result of the operation’. In other words, at the beginning the borrowing
country uses foreign liquidity to increase domestic investment or consumption - from the point of view of the
lending economy, it is irrelevant which of the two. What is relevant is that the larger investment or
consumption is accompanied by a greater amount of imports from the lending country. If this happens, in the
lending country, given the amount of domestic investment I 0l and imports M 0l , there is an upsurge in the
saving supply S 1l over I 0l that is matched by the equivalent increase of export X 1l over
M 0l , so that
26/06/17
25
income transfers to older generations may positively affect aggregate demand and income by increasing the
average propensity to consume. On the contrary, the attempt to raise the average community propensity to
save by cutting PAYG and extending the FF schemes will have deflationary effects (Palley, 1998, pp.99100; Cesaratto, 2004a). An incomplete appreciation of the implications of the Keynesian-Sraffian critique of
the neoclassical saving-investment nexus with regard to the secular determinants of capital accumulation,
associated with a wrong inference that a tendency of capacity to adjust to aggregate demand would restore
the traditional causal relation between saving and investment, is, perhaps, behind the support given by nonconventional economists, such as F&M, to a FF reform.
The criticism of neoclassical capital theory and the theory of Effective Demand also dispose of the
conventional idea that an higher old-age saving supply from the northern countries will naturally flow to the
capital-poor southern countries stimulating investment there. On the contrary, from a Keynesian point of
view it is an investment decision (or autonomous consumption or public spending decisions) in the south
region, independent from any saving decisions in the north, that, by raising imports from the north region,
generates an higher output and savings in the north. Ex post, this saving will appear as a surplus in the
north’s trade balance matched, in the balance of payments, by a capital outflow to the south. This may give
the optical illusion – often also to non orthodox economists - that foreign savings in the north has generated
investment in the south. It is enough to think of the globe as a closed economy and the same Keynesian
causality, if valid in a single country, will necessarily apply to the world as well, as James Meade put it in a
letter to Dalziel and Harcourt (1997, p.628).
To sum up, demographic developments are a serious fact (UN 2002) and pose new challenges, but
saving-led models of economic growth provide an analytically incorrect recipe for dealing with them.32
References
Beveridge, W. 1942, Social Insurance and Allied Services, New York, Macmillan.
S1l  I 0l  X 1l  M 0l (the famous ‘twin surpluses’). The origin of foreign saving is not, however, in an higher
saving supply ‘which is lent abroad’, but is the result of the process described above.
32
The ability of PAYG’s reforms to deal with the ageing problem are dealt with in Cesaratto 2004a. It is
argued there that an ageing society does necessarily imply a larger share of GDP going to the old, if decent
pensions are to be paid, but that this is not in principle a problem in affluent societies, although it may
generate a political conflict over income distribution.
26/06/17
26
Beveridge, W. 1944, Full Employment in a Free Society, London, George Allen & Unwin.
Cesaratto, S. 1999a, New and old neoclassical growth theory: A critical assessment, in:
G.Mongiovi and F.Petri (eds.), Value, Distribution and Capital: Essays in Honour of Pierangelo Garegnani,
London, Routledge.
Cesaratto, S. 1999b, Savings and economic growth in neoclassical theory: A critical survey,
Cambridge Journal of Economics, 23, pp.771-793.
Cesaratto, S. 2002, The economics of pensions: a non conventional approach, Review of Political
Economy, 14, pp,149-177.
Cesaratto S. 2004a, Notes on the pros and cons of the Notional Defined Contribution reforms,
Journal of Income Distribution, forthcoming.
Cesaratto S. 2004b, The Transition to Fully Funded Pension Schemes: A non-Orthodox Criticism,
Cambridge Journal of Economics, forthcoming (working paper version: Quaderni del Dipartimento di
Economia Politica, n.382, www.econ-pol.unisi.it).
Cesaratto S. 2004c, The Classical 'Surplus' Approach and the Theory of the Welfare State and
Public Pensions, mimeo.
Cesaratto S., Serrano F., Stirati A. 2003, Technical change, employment and Effective Demand,
Review of Political Economy, 15, pp.33-52.
Ciccone, R.1990, Accumulation and Capacity Utilisation: Some Critical Considerations on Joan
Robinson’s Theory of Distribution, in Bharadwaj K. and Schefold B., 1990, Essays on Piero Sraffa –
Critical Perspectives on the Revival of Classical Theory, Unwin Hyman, London.
Davis, E.P. 2004, Demographic and Pension-System Challenges to Financial and Monetary
Stability, paper prepared for the Austrian Central Bank conference, Vienna May.
Dalziel, P.C. and Harcourt, G.C. 1997, A Note on ‘Mr Meade’s Relation’ and International
Capital Movements, Cambridge Journal of Economics, 21, pp.621-631.
Diamond, P. 1996, Proposals to restructure Social Security, Journal of Economic Perspectives,
10, pp. 67–88.
Engen E.M. & W.G.Gale, 1997, Effects of Social Security Reform on Private and National
Saving, in S.A.Sass & R.K.Triest, Social Security Reform, Federal Reserve Bank of Boston, Conference
Series no.41.
Feldstein, M. 1974, Social Security, induced retirement, and aggregate capital accumulation,
Journal of Political Economy, 82, pp. 905–926.
Feldstein, M. 1996, The missing piece in policy analysis: Social Security reform, American
Economic Review, Papers and Proceedings, 86, pp. 1–14.
Foley, D.K. 2001, Value, Distribution and Capital: A Review Essay, Review of Political Economy,
13, pp. 365–381.
Foley D.K. and Michl T.R.1999, Growth and Distribution, Harvard University Press, Cambridge
(Mass.).
26/06/17
27
Frankel M. 1962, The Production Function in Allocation and Growth: A Synthesis, American
Economic Review, 52, pp.995-1002.
Garegnani, P. 1970, Heterogeneous capital, the production function and the theory of distribution,
Review of Economic Studies, 37, 407–436.
Garegnani P. 1983, Notes on Consumption, Investment and Effective Demand, in J.Eatwell and
M.Milgate, Keynes’s Economics and the Theory of Value and Distribution, Duckworth, London.
Garegnani, P. 1990, Quantity of capital, in: J. Eatwell, M. Milgate and P. Newman (Eds) Capital
Theory, London, Macmillan.
Garegnani P., 1992, Some Notes for an Analysis of Accumulation, in J.Halevi, D.Laibman,
E.Nell, (eds), Beyond the Steady State, Macmillan, London.
Garegnani, P. and Palumbo A. 1998, Accumulation of capital, in: H.Kurz and N.Salvadori, Elgar
Companion to Classical Economics, Cheltenham, Edgar Elgar.
Haavelmo, T. 1945, Multiplier Effects of a Balanced Budget, Econometrica, 13, pp.311-318.
Keynes, J.M. 1936, The General Theory of Employment, Interest, and Money, London,
Macmillan.
Keynes, J.M. 1942, The Beveridge Report, in: The Collected Writings of John Maynard Keynes,
Vol. XXVII (D. Moggridge, Ed.), London, Macmillan.
Kriesler P. and Halevi J. 1996, Asia, Japan and the Internalization of Effective Demand,
Economies and Sociétés, Monnaie et production, Séerie M.P., n°10, pp.301-320.
Krugman P.R. and Obstfeld M., 1994, International Macroeconomics, third ed., HarperCollins
College Publishers, New York.
Kurz H.D. and Salvadori N.2003, Theories of Economic Growth: Old and New, in Salvadori N.,
The Theory of Growth. A ‘Classical’ Perspective, Edward Elgar, Cheltenham.
Marx K. 1867, Capital, A Critical analysis of the capitalist production, Lawrence and Wishart,
London, 1974.
Metzler, L.A. 1943,Effects of Income Redistribution, Review of Economics and Statistics, ,
pp.49-57.
Michl, T.R. 2001, Why We Should Fund Social Security Permanently, Challenge, 44-6, pp.78-92.
Michl, T.R. 2002, Prefunding Is Still the Answer, Challenge, 45-3, pp.112-116.
Michl T.R. and Foley D.K. 2004, Social Security in a Classical Growth Model, Cambridge
Journal of Economics, 28, pp.1-20.
Modigliani F., Ceprini, M.L. and Muralidhar, A.S. 1999, A Solution to the Social Security Crisis,
Sloan Working Paper, n. 4951.
Musgrave, R.A. 1945, Alternative Budget Policies for Full Employment, American Economic
Review, 35, pp.387-399.
Palley, T.I. 1998, The Economics of Social Security: an old Keynesian Perspective, Journal of
Post Keynesian Economics, 21, pp.93-110.
26/06/17
28
Palley, T.I. 2002, Social Security: Prefunding is not the Answer!, Challenge, 45-3, pp.97-118.
Palumbo A. and Trezzini A., 2003, Growth without Normal Capacity Utilisation, European
Journal of Economic Thought, 10, pp.109-135
Pizzuti, F.R.1998, Pension Reform and Economic Policy Constraints in Italy, Labour, 12, pp.4566.
Rebelo S. 1991, Long-run Policy Analysis and Long-run Growth, Journal of Political Economy,
99, pp.500-21.
Reisen, H.2000, Pensions, Savings and Capital Flows, Edward Elgars (in association with the
OECD), Cheltenham.
Serrano, F.L. 1996, The Sraffian Supermultiplier, unpublished Ph.D. thesis, University of
Cambridge.
Serrano F.L. and Cesaratto S., 2001, As Leis de rendimento nas teorias neoclasica de crescimiento:
Una critica Sraffiana, Revista Ensaios FEE, vol.23 (English version: www.networkideas.org)
Solow, R.M.1956, A Contribution to the Theory of Economic Growth, Quarterly Journal of
Economics, 70, pp.65-94.
Solow R.M 1992, Siena Lectures on Endogenous Growth Theory, vol. 6, Collana del
Dipartimento di economia Politica, Università di Siena.
Solow, R.M., 1970, Growth Theory: an Exposition, Oxford, Clarendon Press.
Sraffa, P.1960, Production of Commodities by Means of Commodities. Prelude to a critique of
economic theory, Cambridge, Cambridge University Press.
Stirati, A. 1994, The Theory of Wages in Classical Economics, Aldershot, Edward Elgar.
Trezzini A. forthcoming, Steady State and the Analysis of Long-Run Tendencies: the Case of
Kaleckian Models, in Ciccone R. et al, Proceedings of the Conference Sraffa and Modern Economics,
Routledge, London.
UN-Population Division, 2000, Replacement Migration: Is It a Solution to Declining and Ageing
Population?, New York.
UN-Population Division, 2002, World Population Ageing: 1950-2050, New York.
Wallich H.C., 1944, Income-Generationg Effects of a Balanced Budget, Quarterly Journal of
Economics, 59, pp.78-91.
World Bank 1994, Averting the Old Age Crisis, Policies to protect the old and promote growth,
Oxford, Oxford University Press.