Modern Money Theory: A Critical Assessment and a Proposal for the State As Innovator of First Resort

ABSTRACT

Modern Money Theory (MMT) describes the functioning of a pure credit economy, assuming that the state can finance public spending via monetisation on the part of the central bank: in this light MMT proponents maintain that taxation and bond issues are irrelevant to public deficit financing. Another feature peculiar to MMT is the belief that expansionary fiscal policies can guarantee full employment in a condition where the state acts as an employer of last resort (ELR). The aim of this paper is threefold: (a) to understand and rationalise the logical framework of MMT (Section 2.1); (b) to address some of the controversial issues of this approach, with particular regard to the ELR programme proposal and to the actual role played by taxation and bonds in public deficit financing (Section 2.2); (c) to propose an extension of the ELR programme, arguing that it can be used for the application of innovations by the state (Section Three).

KEYWORDS:

• Endogenous money
• fiscal policy
• innovations
• modern money theory
• monetary policy
• neo-Keynesian dynamic portfolio models

JEL CLASSIFICATIONS:

• C62
• C63
• E32
• E62

Acknowledgements

The authors kindly thank two anonymous referees for their very helpful comments and suggestions that allowed to significantly improve this article. Any remaining errors are our own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 It may be interesting to point out that this belief marks a further difference from MTP, since circuitist scholars maintain that expansionary fiscal policies generate inflationary pressures (cf. Graziani, 2003, p. 108).

2 On these matters, see also Palley (2019, p. 7 and ff ), where the author explicitly refers to the stock-flow models proposed by the ‘Yale school macroeconomics research programme developed by James Tobin in the 1960s’. See also Palley (2015b, p. 51).

3 However, at the end of this section we will touch upon these matters.

4 In what follows, the time argument is suppressed whenever this does not cause any confusion. All variables are specified at time t. As to the time unit, since we are dealing here with a short-run model, it may be set arbitrarily. As usual, a dot over a variable indicates differentiation with respect to time t. Sometimes, for the sake of readability, we will use the operator g(•) as the growth rate of a variable, i.e., g(x) = $\dot{x}/x$. Furthermore, when it does not cause confusion, we will denote the first total derivative with a prime mark, and the first partial derivative with a subscript, that is if y=f(x) and $z=g(x_1,x_2,\dots ,x_i,\dots x_n)$ are two functions of respectively one and n variables, we put: $\mathrm{d}y/\mathrm{d}x=f^{\prime }$ and $\mathrm{\partial }z/\mathrm{\partial }{x}_i=g_i$.

5 In the light of what we have just said about the banking sector, i is the sole interest rate of the economy.

6 For the sake of simplicity we ignore the relationship between consumption demand and the interest rate.

7 For the moment we do not distinguish between monetary base and money supply. Furthermore, the important issue of exogenous vs. endogenous money supply will be addressed below.

8 The last term in the definition of disposable income, $\pi W$, stands for expected capital losses on wealth.

9 At the end of Section 2.3 we will come to the sustainability of increasing public debt (and of the related increasing flux of interest to be paid on it), where we will show that under assumptions consistent with MMT, with particular regard to the actual existence of a ‘consolidated sector’, these matters could not represent, in principle, an effective source of instability.

10 This particular feature of MMT brings us back to Friedman's statement that Keynes's system is characterised by a ‘missing equation’, that is, the equation that relates the real with the monetary part of the economic system. See Friedman (1975).

11 We point out that there may be alternative formulations of the ‘on-off’ rule. For instance, the following one: $\dot{p}=0\mathrm{i}\mathrm{f}Y<Y_{FE};\dot{p}=\gamma (Y-Y_{FE})+p^e,\mathrm{i}\mathrm{f}Y>Y_{FE}$, is still consistent with Palley's suggestion. However we will follow formulation (7), since it allows us to easily compare the MMT model with the one described by system (6).

12 As Blinder and Solow (1973, p. 323) stated (the reference is to the dispute between Keynesians and Monetarists): ‘There is no controversy over government spending financed by printing money. Both sides agree that it will be expansionary; but one group likes to call it fiscal policy, while the other prefers to call it monetary policy’.

13 The chartalist approach is sometimes taken to extremes by MMT authors, leading to an oversimplification of money demand, which is traced back only to the need to pay taxes. See for instance the recent handbook by Mitchell, Watts, and Wray (2019, p. 137, italics added), where in a section entitled ‘Taxes drive the demand for money’ we read:

government's currency is the main (and usually the only) thing accepted by government in payment of taxes and other obligations to the government. It is true of course that government currency can be used for other purposes: coins can be used to make purchases from vending machines; private debts can be settled by offering government paper currency; and government money can be hoarded in piggy banks for future spending. However, these other uses of currency are all subsidiary, deriving from government's willingness to accept its currency in tax payments. It is because anyone with tax obligations can use currency to eliminate these liabilities that government currency is in demand, and thus can be used in purchases or in payment of private obligations.

14 The scepticism about the real balance effect is often justified on the grounds of empirical evidence. For instance (Mitchell, Watts, and Wray, 2019, p. 211) state that the

empirical evidence is that in normal price movement ranges, the measured real balance or wealth effect is very small and clearly insufficient to remedy a major shortfall in aggregate demand. So, while the Pigou effect presented a logical possibility, it did not provide the Classical employment theory with the support it required to negate the damaging critique made by Keynes.

Taylor (2019, p. 22), for his part, argues that: ‘The famous real balance effect, much touted by Neoclassically inclined economists, is irrelevant today. A broad monetary aggregate is around $6 trillion, or six per cent of household wealth. The propensity to consume from wealth is around four per cent so a small decrease in real balances will have a negligible impact on consumption demand’. However, since we are dealing here with the effects of (expansionary) economic policies, the effectiveness of the Pigou effect should be considered in the light of those policies proposed by MMT, with particular regard to the ELR programme (which we will deal with in the next section). For instance, Palley (2019, p. 149 and ff.) has estimated that the implementation of the ELR programme in the US would imply a direct aggregate demand injection of about 12.3 per cent of GDP: clearly, with these magnitudes the irrelevance of real balance effect cannot be taken for granted.

15 It may be interesting to point out to the reader that Abba Lerner, in his famous 1943 contribution on functional finance —which, as is well-known, represents a remarkable source of inspiration for the development of MMT—acknowledged the relevance of the wealth effect on agents' consumption behaviour. For instance, as Lerner (1943, p. 49) maintained referring to the increase in the stock of public debt:

The greater the national debt the greater is the quantity of private debt. [ … ] The greater the private fortunes the less is the incentive to add to them by saving out of current income. As current saving is thus discouraged by the great accumulation of past savings, spending out of current income increases (since spending is the only alternative to saving income).

16 For a similar statement see Turnovsky (1977, p. 31)

17 We advise the reader that we will come back to these matters in Section 2.3. Furthermore, it may be interesting to mention that at the beginning of the critical debate that followed early MMT contributions, the endogeneity of money supply in MMT was not always taken for granted. For instance, Rochon and Vernengo (2003, p. 61) stated, for ‘chartalists, state money is exogenous, and credit money is a multiple of the former. [ … ] For chartalists, money is primarily verticalist in nature, with bank money playing a secondary role. For post-Keynesians, it is the other way around: credit money takes precedence and is the primary money creation force, whereas fiat money takes a secondary role’. The reader is also referred to Rochon's recent 2019 contribution, where the author goes back to his former strong criticism, recognising that since then: ‘MMT has gone on to develop quite a following. Indeed, whereas MMT was often consumed with internal post-Keynesian debates it [ … ] has became a true force to be reckoned with, leaving many of us green with envy’ (Rochon, 2019, p. 160).

18 See for instance (Tymoigne and Wray, 2015, p. 28 and ff.).

19 As Seccareccia (2004) has maintained, the choice of $w_G$ is a very critical matter, given the possibility of the existence of multiple equilibria. In particular Seccareccia shows that a low $w_G$ could lead to a ‘bad’ equilibrium characterised by low-wage full employment and higher disguised unemployment, while a higher (and higher) $w_G$ would move the system toward a ‘better’ equilibrium characterised by higher wages, lower disguised unemployment and a lower number of involuntary unemployed to be absorbed by the ELR programme. The reader is referred to Seccareccia's paper for further details about these matters.

20 The measurement unit of employment, of course, may be expressed in hours or equally in number of workers.

21 As we will see in what follows, $v^d$ may differ from $v^{\ast }$; in any case, $v^{\ast }$ must be considered the ‘natural employment rate’, that is, $\overline{N}-L^{\ast }$ stands as the ‘natural unemployment’ level.

22 In other words and, perhaps, more clearly, $G_1$ may be defined as the part of G associated with all the other policy objectives except the ELR full-employment (and other ELR) policy objectives.

23 See also Wray (2018, p. 2, first italics added), where the author states—with reference to the ELR programme—that an ‘important feature of our proposal is that, other than establishing a minimum wage and benefit package, projects would not compete with the private sector. We have focused our programme on public service employment because we expect most of the jobs to involve provision of care service’. The reader is also referred to Tcherneva (2002)– as far as we know, as we have already said, one of the few attempts at building a formal MMT model—where the author assumes that there is only one service produced in the economic system, ‘fighting fires’, and that the only government expense is paying wages to the ‘fire fighters’.

24 With regard to workers employed in the ELR programme, at least two alternative assumptions could be made, both consistent with a single-commodity economy and with the MMT logical framework: (a) their productivity is positive but lower than that of workers employed in the private sector, i.e., $Y/L_G={\alpha }^{\prime }=\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{n}\mathrm{t},{\alpha }^{\prime }<\alpha$; (b) they produce services that support the private sector of the economy. With regard to the former alternative, from what we have just said it does not seem to be in line with the MMT logical framework, since the output produced by workers employed in the ELR programme should not in any way compete with that produced by workers employed in the private sector (see above, fn. 23). The latter alternative assumption, instead, can be consistent with MMT prescriptions and may also imply the possibility that $L_G$ output may affect, e.g., increase, α. However, as will be clear in the following, we can safely avoid these complications with no major consequence for the rest of our investigation.

25 The first inequality is obvious, implying that the greater the income multiplier, the smaller $G_{elr}^{\prime }$ must be in order to guarantee full employment $L^{\ast }=L_p+L_G$. The second inequality may be less intuitive. Its explanation lies in the fact that the greater the labour productivity α, the smaller will be the increase in private employment $\mathrm{\Delta }{L}_p$ in response to the increase in public expenditure $G_{elr}^{\prime }$: as a result, as α increases, $G_{elr}^{\prime }$ must increase too—that is, the proportion of $L_G$ with respect to $L_p$ must increase—in order to guarantee full employment.

26 Seccareccia (2004, p. 33 and ff.), in asking the same ‘obvious question’, argues that the answer should be found in the belief of some ‘defenders of the ELR programme’—the author refers to Mitchell and Wray—according to whom, unlike the ELR programme, this ‘pure’ Keynesian policy would have inflationary effects.

27 This kind of mismatches are well-known and are at the center of economic investigation by policy-makers, see for instance (Nikolov et al., 2018).

28 Wray (2018, p. 3) too recognises that this may be an objective of the ELR programme, stating that: ‘there should be room in the JG for time-limited training and education. While on-the-job training should be a part of every project, proposals can be solicited for specific training and basic education programmes that will prepare workers for jobs in the JG and, eventually, for work outside the JG’. Clearly, in the implementation of the ELR programme the institutional and social context must also be taken into due account. Clearly, in the implementation of the ELR programme the institutional and social context must be taken into due account. See for instance Sovilla's proposal for Mexico (Sovilla, 2018).

29 We stress the point that the inflationary effects of ELR are referred to what we have called in the previous section the ‘extreme case’, i.e., the case for which all the unemployed are absorbed by the ELR programme. The impression one gets is that even some MMTers recognise the possibility of this inflationary outcome. For instance, coming back at a greater length to a Tymoigne and Wray (2015, pp. 41–42) statement we have already quoted above, ‘the presumed problem is that while JG workers get wages (and thus consume) they do not contribute any production that is sold (hence do not absorb wages). The “excess” wages from newly employed workers makes spending rise. What could government do? It would have at its disposal the usual macroeconomic policy tools: raise taxes, lower government spending on programmes other than JG, and tighten monetary policy’. On this matter see, for a similar reasoning in line with ours (Druedahl, 2019, p. 17 and ff.).

30 For the sake of simplicity in what follows we ignore the relationship between money demand and the wealth stock: we point out that this omission is of no consequence in our discussion, since we are mainly interested to the ‘impact effect’ of the perturbation.

31 For a similar position, see Aspromourgos (2000). The belief that the task may be accomplished by taxes, as Tymoigne and Wray (2015) seem to suggest, is very hard to conceive, since it would imply that tax rates should continuously vary in order to promote price stability and a given targeted equilibrium on the financial market.

32 Asada and Ouchi (2013), for instance, have investigated the long-run dynamics of a model characterised by the presence of a ‘consolidated sector’ that follows a monetary policy based on the Taylor rule.

33 One only has to think of the determination of loan interest rates. It is usually assumed that, following the seminal contributions of Monti (1971) and Klein (1971), the interest rate on loans is determined by adding a mark-up—related to the monopolistic degree of the banking sector—upon the overnight interest rate. Furthermore—in line with the so-called credit view (see Bernanke and Blinder, 1988)—this mark up should also reflect the riskiness of the loan, as expressed for instance by firms debt to capital ratio. As is evident, the right choice of a given targeted interest rate by the Central Banks may determine—via loan interest rates—the stock of firms indebtedness, thus preventing, for instance, potential speculative bubbles. See Colacchio (2014) for a long-run dynamic model where all these interrelations are taken into account in the long-run

34 However, as Druedahl (2019, p. 8) has pointed out, ‘the proponents of MMT rarely clarify whether they are talking about the short-run [ … ] and the long-run’.

35 As has been shown, this proposition holds even if we assume that public bonds are entirely and passively bought by the private (and/or the banking) sector: all it needs, once again, is that the dynamics of the primary public deficit is bounded. For this curiosum see Colacchio (2014).

36 It may be interesting to recall that Lerner's functional finance proposal was strictly based on the combined use of both instruments. For instance, according to Lerner (1943, p. 41), functional finance, among others, prescribes ‘the adjustment of public holdings of money and of government bonds, by government borrowing or debt repayment, in order to achieve the rate of interest which results in the most desirable level of investment’.

37 However this is a well-known bone of contention, see for instance (De Grauwe, 2011, 2018). It goes without saying that these issues are a dramatic example of economic policy trade-off. In order to avoid the temptation for governments to use, and increase, the inflation tax in an opportunistic manner, there has been a progressive shift toward a greater central banks independence. The other side of the coin is that this increased independence has exacerbated the fragility of governments, with particular regard to economies characterised by high public debt-to-output ratios. A classical example of this trade-off is the Eurozone. For an authentic ‘interpretation’ of ECB independence, see for instance (Bini, 2007).

38 On these matters the reader is referred to the strong criticism of MMT by Fiebiger (2012a, 2012b). Interestingly, Lavoie (2019, p. 105, italics added) argues that ‘the consolidation of the central bank and the government into a single unity should enter the policy debate as an objective through institutional change, and not as an actual feature of the economy upon which policy advice could be offered’. On the consolidated sector issue, see also (Kregel, 2019, p. 94).

39 For a database of Sovereign Defaults, see Beers and Mavalwalla (2017).

40 The book recently edited by Foggi (2019) collects a number of contributions also on the implementation of an ELR programme in the Italian economy.

41 See: (OECD, 2019)

42 The so-called Manuale Frascati states that ‘Researchers are professionals engaged in the conception or creation of new knowledge, products, processes, methods and systems, as well as in the management of the projects concerned’. See: https://data.oecd.org/rd/researchers.htm#indicator-chart.

43 This indicator shows the employment rates of people according to their education levels: below upper secondary, upper secondary non-tertiary, or tertiary. The employment rate refers to the number of persons in employment as a percentage of the population of working age. The employed are defined as those who work for pay or profit for at least one hour a week, or who have a job but are temporarily not at work due to illness, leave or industrial action. This indicator measures the percentage of employed 25–64 year-olds among all 25–64 year-olds.

44 See: https://data.oecd.org/emp/employment-by-education-level.htm#indicator-chart.

45 SVIMEZ (2014) shows that the Fondo di Finanziamento Ordinario (FFO) for state universities fell from 7 billion 250 million euros in 2008 to around 6 billion 500 million in 2014, with a reduction of 14 per cent. This measure appears designed to reduce the quality of the workforce, since Italian firms do not express a great demand for high-skilled workers. The outcomes are unemployed and underemployed intellectuals and, above all, migration of high-skilled workers from Italy to other countries (in the absence of high-skilled workers attracted from abroad to Italy), with the consequent loss of labour productivity in Italy.

46 On the relation between aging and productivity, see for instance (Aiyar, Ebeke, and Shao, 2016).

47 On the relation between aging and consumption, see for instance (Kwon and Oh, 2014).

48 The state cannot only finance R&D but also directly produce innovations, via public firms. As stressed by proponents of the so-called National system of innovations, this could be a case where innovations spread via imitation.