Macroeconomic controversies usually center on causation. Two questions of significance for policymaking concern output and growth:

• Is output demand or supply determined?
• Is economic growth demand or supply led?

If demand is the driver of output and growth, there will be considerable scope for government spending to influence the economy’s trajectory. If, instead, supply-side factors are determining, fiscal policy will be impotent other than possibly temporarily. Recent history suggests – much like the earlier history of the Great Depression and Second World War – that fiscal policy is important to the performance of the economy, and that demand matters. In my view, the causal significance of demand follows quite strikingly from Modern Monetary Theory’s institutional analysis of sovereign currencies together with the fully compatible Post Keynesian analysis of banking and endogenous money. Before getting into that, a brief summary of opposing positions on output and growth determination is perhaps warranted.

 
Neoclassical and Keynesian approaches to output and growth

At any moment in time, there is an upper limit to output defined by available resources, labor force and technology, summed up as ‘supply-side factors’. Debate on the determination of output concerns whether the supply limit acts as an attractor, with actual output automatically tending towards potential. In the long run, either: (1) demand adjusts to potential output in response to movements in wages, interest rates and other prices; or (2) inside resource limits, actual output adjusts to a level of demand that is partly autonomous. In debates over growth, attention turns to whether the growth rate of actual output: (A) depends only on supply-side factors, with the price mechanism inducing the necessary response in demand; (B) depends primarily on supply-side factors but with a possibility of temporary demand shocks having persistent effects on the growth of potential output (hysteresis); or (C) is led by demand.

In a long-run context, modern neoclassical economists tend to take positions (1) and either (A) or (B). That is, they maintain that demand tends to adjust to potential output in response to price movements (1), though the process may well be impeded by ‘imperfections’ (such as sticky prices or sluggish adjustment of expectations). They do acknowledge a role for demand (2) in the short run before there has been sufficient time for the price mechanism to take full effect.

Neoclassical economists are divided on whether demand can matter for growth in the long run. But, within their framework, if demand does matter for growth, it is through an indirect channel in which temporary demand shocks have long-lasting (hysteretic) effects on potential output. In the absence of so-called imperfections, the tendency for demand to adjust to potential output in response to price movements is still considered operative. It is just that, in the event of hysteresis, potential output itself is not completely independent of the historical path of demand. So, in the presence of hysteresis, neoclassical theory depicts demand and actual output as tending to adjust towards a potential output that is not completely independent of the history of demand.

Heterodox economists influenced by Kalecki or Keynes, while acknowledging the likelihood of hysteretic effects on the economy’s supply-side potential, tend to take positions (2) and (C). That is, inside the economy’s resource limits, they consider both output and growth to be demand led, whether in the short or long run. Some, often influenced by Joan Robinson, have questioned the relevance of long-run analysis divorced from the short run, noting that the long run is simply a sequence of short-run periods. Kalecki also tended to take this view. When Kalecki referred to long-run factors, he had in mind key institutional features of the economy that change only slowly. The long run, in his sense, meant a period long enough that a significant institutional change could occur. But there are other heterodox economists who do attach a relevance to long-run (or long-period) analysis who have argued that both output and growth are demand led, including in the long run. The important point, for present purposes, is simply that heterodox economists following one or both of Kalecki and Keynes emphasize demand in their explanations of output and growth, irrespective of their attitudes toward long-run analysis.

 
Monetary and banking realities point to a demand-led economy

Modern Monetary Theory (MMT) has, among other things, shed light on the institutional and operational details of monetary and fiscal policy. An important contribution of Post Keynesian Economics (PKE) has been to make clear the endogeneity of private credit creation. The significance of these insights is not just the institutional and operational details uncovered. It is that these details suggest a particular causation. It is a causation that runs from spending to income, and from injections to leakages.

Private investment, for instance, is funded independently of current income, partly through bank loans. The loans create deposits, which firms then draw down to purchase investment goods. In these acts of spending, income is simultaneously created as the amounts spent flow to suppliers of investment goods. Savings are created at the same time, and are of the same amount as the income created. The savings will change hands, through subsequent market exchange, or be taxed away in each round of the multiplier process set off by the initiating act of investment.

Now, neoclassical theory suggests that autonomous spending, though exogenous of income, is ultimately induced by the price mechanism in the long run. Therefore, such spending is really endogenous, or so the theory goes, and interest rates in particular will induce the amount of private investment consistent with full-employment output and a supply-led rate of growth.

There are numerous objections to this claim. One is that there is presently no strong basis in theory for presuming interest rates can play this role. This was a lesson of the Cambridge Capital Controversy. A second objection is that empirical evidence for the claim is lacking. From a heterodox perspective this is not surprising because of a third objection, which relates to banking operations. Banks do not, and cannot, lend out savings. As the endogenous-money analysis of PKE makes clear, loans create deposits. The actual supply of loans adjusts to the (credit-worthy) demand for loans. The demand for loans is far more influenced by the state of economic activity than the interest rate, suggesting (as the available empirical evidence appears to verify) that interest-rate adjustments will have little impact on investment.

A fourth objection is that, in a sovereign-currency system, the interest rate is a policy variable, not a market-determined one. The short-term nominal rate of interest does not adjust automatically in response to supply and demand pressures. Admittedly, neoclassicals regard the relevant variable as the real rate of interest, which is not directly subject to policy (short of government imposing price controls). Even so, unless all other prices (and expectations of future prices) can be shown to adjust automatically in such a way as to defeat monetary policy and bring about the one real interest rate supposedly consistent with full-employment output, there will be little reason to expect the real interest rate to play the role claimed in neoclassical theory, even leaving aside objections 1-3. In any case, the onus would be on neoclassicals to establish such a claim in theory and/or empirically. At this point, they have not done so.

The initiating role of spending can also be seen by considering a monetary economy at its inception. In particular, we can imagine the first act of government spending.

It is clear that in a monetary economy in which government issues the currency, government spending comes first. There needs to be a system of property rights in place before monetary exchange can occur, simply because to exchange something for money, somebody first has to have ownership of that thing. Private property is a social arrangement, and it can only be recognized and enforced by some form of government, no matter how rudimentary or informal. The recognition of property rights is a collective act committed by, or on behalf of, all consenting or compelled parties. Since the economy is monetary, those performing the functions of government will be paid in money. As MMT makes clear, this is achieved by government (i) defining a unit of account, (ii) imposing taxes and other obligations in that unit of account, and (iii) issuing a currency denominated in its unit of account with the promise to accept the currency back again in fulfillment of tax and other obligations to it.

From inception, the government spends its currency into existence. Its spending goes to recipients as income. The clear implication, as before, is that spending creates income. And, again, the spending also results in leakage from the circular flow of income to taxes, saving and imports of an amount equal to the initial amount spent into the economy. It is clear here, too, that causation runs from spending to leakage. This is true no matter how high or low the marginal propensity to leak, or how large or small the multiplier.

These considerations point to the conclusion that taxes do not, and cannot, fund government spending. It is the other way round. Government spending is what makes the payment of taxes possible. A currency-issuing government is never revenue constrained. Its spending decisions are never subject to a neoclassical ‘government budget constraint’, whether in the short or long run. The real constraint pertains to resources, not to money that the government itself creates. In particular, the government can always purchase goods or services that are available for sale in its own currency.

The absence of a neoclassical government budget constraint seems significant when trying to decide between neoclassical and heterodox arguments over causation in a long-run context. The notion of demand-led growth requires that some demand be autonomous, over whatever time horizon is being considered. Currency sovereignty ensures, at minimum, that government can always play the role of autonomous spender.

For those heterodox economists who eschew long-run analysis, this point may appear less critical, but it seems important to demand-side explanations of long-run global growth. At the level of the global economy as a whole, just about any other kind of spending could be claimed endogenous in the very long run. (Of course, at the national level, we can also appeal to exogenous exports.) Private investment, even though not (in the heterodox view) automatically induced to a level consistent with full employment, can be regarded as induced by a competitive imperative to adjust capacity to demand. Credit-based consumption expenditure, though semi-autonomous over fairly long periods, is subject to income growth in the very long run, simply because currency users are financially constrained. Credit-financed consumption behavior can clearly have a permanent impact on the levels of output attainable into the future, as we are witnessing a decade after the global financial crisis, but it is less obvious that this behavior can have a permanent effect on the growth rate. For the global economy, it is arguably only currency-issuing governments that can play the role of autonomous spenders permanently.

One effect of the neoclassical ‘government budget constraint’ is to rule out (unjustifiably) the possibility of global demand-led growth. In neoclassical theory, all demand is induced in the long run. However, once it is understood that a currency-issuing government’s spending is autonomous over any time frame, the possibility of demand-led global growth becomes clear, as does the existence of many possible rates of growth and growth paths. Inside natural-resource limits, and assuming a surplus labor supply (which has been a permanent feature of the global capitalist economy), actual productive capacity at any particular point in time (reflecting existing plant and equipment) will itself be a consequence of demand. Steadily and persistently rising government spending, by adding to total demand, will tend to generate rising rates of capacity utilization and justify the installation of additional private-sector plant and equipment. In other words, autonomous demand, through its effects on capacity utilization, will tend to induce private investment. In this way, autonomous demand can be regarded as driving growth.