2. Law of Economics: We Need a Budget Deficit Every Year

By Jan Frederik Moos

Let’s now turn to how economic models treat the development of total spending. Paul Davidson – whose name you already encountered in the first chapter – emphasised the importance of historical development, the reality in which we actually live. Each of your decisions, whether to spend money or to save, alters total spending. Mainstream economics, however, abstracts from this historical process and assumes that saving decisions leave total spending unchanged. That may sound rather absurd – and indeed it is. So, let’s begin.

„The act of saving (or abstaining from present consumption) does not lead to an equivalent amount of investment via changes in the rate of interest.“ (Thirlwall, A. P.: 1993).

This essay sets out the second of the five economic laws I have formulated for macroeconomics. The focus is a persistent assumption in mainstream economics: that household saving does not reduce aggregate demand because it is automatically transformed into investment. Aggregate demand can therefore be summarised in two components: consumption and investment, and the lower consumption is simply meant to be replaced by higher investment. The idea is commonly associated with Jean-Baptiste Say, although Say himself wrote only that "products always exchange for products" (Davidson 1994: 14).

In a barter economy (a cashless economy) this is true: goods are produced solely in order to be exchanged for other goods, and the act of production itself creates the means of demand. What is supplied on one side is, by definition, purchasing power on the other. After all, production (supply) only takes place in order to buy (demand), and the processes are therefore inseparable. With no money, there is no possibility of withholding income from circulation, and therefore no saving. Put simply: I only work (produce) in order to buy (demand) something. But once money is introduced, the doctrine unravels. Suddenly, I produce (supply) to save part of my income, so I do not demand to the same extent as I supply. Money changes everything, and we need to talk about it.

Steve Keen, a brilliant Australian economist, has highlighted in New Economics: A Manifesto that the tendency of mainstream theory to ignore the distinctive properties of money constitutes a fundamental error with far-reaching consequences. Money simply means that supply and demand do not constantly balance each other — and, as we will see, that insufficient demand is the permanent reality. At its core, mainstream economics is still dealing with Say’s moneyless barter economy, probably without being fully aware of it.

In a monetary economy, income need not be spent; it can be saved. This simple possibility poses an immediate challenge to Say’s Law—and thus to the neoclassical narrative—since private-sector net saving implies that aggregate demand lags behind total production. In other words, an increase in the propensity to save leads to an immediate fall in aggregate spending and, therefore, in aggregate demand. This simple connection is systematically excluded in the mainstream model, and not even many good critics are aware of it: Gary Stevenson often points out that mainstream economics is all about money, although money itself is not actually understood. Money is the simple reason why supply does not equal demand, and therefore the reason why all mainstream models are fundamentally wrong.

Before we continue, let us take a look at the following chart. The blue line shows that, in the US economy, the private sector—that is, the domestic economy excluding government and foreign trade—saves on a regular basis. If the blue area in the graph is positive, it shows that households and firms received more money in that year than they spent — in other words, they saved. The private sector does this consistently, year after year. And that simple fact has major implications: we cannot just leave it out of our model.

Figure 1: Annual Private Sector Savings in the US Economy

In mainstream models, the blue line would lie exactly at zero every year. The assumption is that whenever someone saves—that is, spends less than they earn—someone else would immediately go to the bank, take out a loan, and spend more than they earn (in other words, invest) to offset this. Such an assumption is clearly unrealistic, but it makes the mathematics easier to handle. The problem, however, is that this approach helps with calculation but not with understanding. In fact, by looking at the following example, you are already learning more than in a typical mainstream economics course. Let us therefore consider a simple economy to examine more closely what net saving implies.

Let us consider a simple economic system in which the state has no budget of its own and can only redistribute income. Suppose the production of one worker generates a value added of $5,000: $3,000 paid out as wages, $1,000 as profit, and $1,000 redistributed by the government to recipients of social transfers (or public service workers). Out of this total, the worker, the profit recipient (entrepreneur), and the transfer recipient together decide to save $1,000. That leaves only $4,000 spent, while $5,000 of output has been produced. When we consider the value created by a worker, any part of their income that is saved immediately creates a gap in demand relative to their output. If such behaviour prevails on average, the economy cannot reach full employment. The economy is indeed permanently ready to supply more, but since it generates its income only from its own production and saves, it is not willing to purchase the full level of output corresponding to full employment.

Now scale this up to an economy with 100 workers. In the initial year, their combined production generate $500,000 in income (value creation). But if, on average, $1,000 is saved out of the income associated with each worker — whether by the worker, the entrepreneur, or the transfer recipient — total spending falls short by $100,000. Only $400,000 worth of goods and services will actually be purchased, even though $500,000 are produced. The supply side responds through quantity adjustments, as it is unwilling to sacrifice profits per unit, and therefore contracts to the new level of expected demand. This reduction in production implies that around 20 out of 100 workers will become unemployed, since their potential output cannot be sold.

If you have grasped this point, that is already excellent. The next step is to refine our sense of time within the model. According to the logic of our simple framework, the economy — without external demand — inevitably enters a downward spiral, which is worth pausing to reflect on. Consider this example: if only 80 workers were employed (meaning 20 workers remain unemployed, as in our previous case), total income would fall to $400,000. With the same saving rate, only $320,000 would be spent, and production would again respond with another cut.

The crucial point is that this gap cannot be closed by households or entrepreneurs themselves, since their very decision to save is what creates it. The only actor able to fill the gap is the state, through deficit spending. By injecting the missing $100,000 (or $180,000) into the economy, the government restores demand and enables full employment. Note that this does not mean redistribution within the existing system — we have already accounted for government transfers. All mainstream models — including those of Stiglitz and Krugman — that focus solely on redistribution rather than on fiscal deficits are therefore fundamentally flawed.

What is required is an additional fiscal impulse in the form of deficit spending. With such spending, the state can employ idle workers and resources productively, financing schools, hospitals, infrastructure, or environmental projects. In this way, what would otherwise show up as unemployment is converted into socially valuable activity.

Before we turn to some economists of the last century—such as the well-known John Maynard Keynes with his major work The General Theory, or Michał Kalecki, a Polish economist with very similar ideas—it should be stressed that viewing the state and the private economy as separate entities is misleading. These two spheres are inseparably linked: economic crises automatically trigger higher transfer payments and reduce tax revenues, while lobbying constantly shapes government regulation. Our economy consists of people, businesses, the environment, and, of course, the state. The fiction of "free markets", described so well by the Marxist economist Ha-Joon Chang, was equally a creation of neoclassical theory. Economists of that school study a world that exists only in their textbooks. Let us now talk about a few great thinkers.

The very problem we have just analysed is precisely the one Keynes addressed in the first third of The General Theory. The perhaps most famous economics book of the last century — chaotically written and to this day either confusing most economists or simply being ignored by them. Yet amid this chaos, John Maynard Keynes expressed the core insight with great precision: "Say's law, that the aggregate demand price of output as a whole is equal to its aggregate supply price for all volumes of output, is equivalent to the proposition that there is no obstacle to full employment. If, however, this is not the true law […] there is a vitally important chapter of economic theory which remains to be written […]." That chapter was Keynes’s own. Saving is not neutral: it reduces aggregate demand immediately, and there is no automatic mechanism by which lower spending brings about an equivalent increase in investment.

Let us pause and underline the key point. I'm allowing myself a lot of redundancy here, as this is one of the most important economic relationships. Please be patient. Full employment — and profitable sales for entrepreneurs — are only possible if the demand gap is closed. This is exactly what Keynes meant by effective demand. The economy as a whole — excluding the state — derives its income solely from its own production. Yet from this very income it saves, and in doing so, persistently demands less than it produces. Only the state can close that gap. By running fiscal deficits, government keeps demand high while still allowing households and firms to save. Without public deficits, demand falls short and unemployment persists. Fiscal policy is therefore not just a temporary tool, but a permanent requirement for stability. And notice: Keynes was not arguing for occasional counter-cyclical action. He insisted that the state must constantly respond to the saving decisions of households and firms — decisions that are taken individually and can never be directly controlled.

I would like to show you one good summary of what you have learned. Joan Robinson, a close student of Keynes, offered a clear summary of his argument. She and her Student John Eatwell wrotes in the first post-keynesian textbook:

"The central point of the theory of effective demand is that the sales value of goods that are available to be sold to the public is determined by expenditure out of the incomes earned in producing them plus expenditure out of incomes earned in producing non-available output (our machines). Since less than the whole income (including profits) derived from available goods is spent on those goods, expenditure from other incomes is necessary to maintain profitable sales." (Robinson/Eatwell 1974).

We have now established that a lack of external demand is the central problem. We have also learned that this was the major concern of Keynes — and, incidentally, of Kalecki as well. This simple mechanism was not understood, which prevented economists from acting in the 1930s and continues to do so today. Both Keynes and Kalecki arrive at the same solution: the state must step in to raise demand. Yet, a crucial difference emerges in how they conceptualize the effects of state intervention.

Keynes retained the neoclassical notion of diminishing returns: as employment increases through deficit spending, firms are assumed to hire workers of progressively lower productivity, with wages tied to that productivity (Lavoie, M. 2022: 301-304). In Keynes’s model, rising employment is therefore accompanied by falling real wages—an assumption that does not align with empirical evidence. This point still causes much confusion today in the interpretation of Keynes and in the derivation of policy recommendations — all the more regrettable, since it is a point that reality simply does not support.

Kalecki rejected this restriction of diminishing returns (Lavoie, M. 2022: 312-314). Instead, he worked with conditions of constant costs: output and employment can expand up to full capacity at stable unit costs, as long as demand is strong enough. Real wages remain unchanged, since unit costs do not rise and labor productivity is assumed to stay constant. Kalecki’s approach is more consistent with advanced capitalist economies, where firms typically operate under stable returns to scale. In sum, we can adopt an optimistic view and side with Kalecki: real wages remain stable when the state ensures full employment. In short, in Keynes’s view unit costs and prices rise on the way to full employment, whereas for Kalecki they do not. If you are more interested in the differences between Keynes and Kalecki — not in terms of modelling, but in their broader distinctions — I can recommend this paper by the German economists Hein and Krämer.

The conclusion is that saving is the problem the economy cannot solve on its own: money that is saved is not spent, and no firm will produce additional output or create employment in return. The state can resolve this problem through deficit spending. Full employment is reached with constant real wages, as firms can expand output at roughly constant unit costs. In short: there is no trade-off. In the next chapter, the modelling of the supply side will become important, where the advantages of Kalecki’s approach over Keynes’s will become even clearer. Those who are interested in how neoclassical theory attempts to keep aggregate expenditure constant should take a look at the appendix. All others are advised to continue with the next chapter.

For readers who wish to explore further: Paul Davidson remains the indispensable guide to Keynes and Malcolm Sawyer offers the clearest exposition of Kalecki. Kalecki is quite readable in the original, and with good anger management, you might even enjoy taking a look at The General Theory. If you are interested in working with me as an editor, interview guest or guest lecturer, or if you have any questions, please feel free to get in touch: mmtmitjan@gmx.de

You can support my work via PayPal: https://www.paypal.com/donate/?hosted_button_id=3CDJZF82THQG8

Appendix:

The Neoclassical Mechanism for Stable Aggregate Spending: Interest Rates

It now remains to analyse how neoclassical theory has constructed this rose-tinted world. One does not have to look far. The neoclassical explanation relies on the interest rate. The claim is that higher household saving provides banks with more funds to lend, lowers interest rates, and thereby stimulates investment. Economic agents can decide whether to consume or to invest (directly or indirectly), but they cannot net save. In this way, the fall in consumption is supposedly offset, and demand remains intact. Put simply: suppose our worker saves again. In the neoclassical model, another agent immediately runs to the bank and invests this money. In reality, however, the money remains in our worker’s bank account, and nobody else can invest it — these actions are entirely separate.

The main reason for falling interest rates is therefore seen in the greater availability of customer deposits, and it is precisely this decline that is assumed to trigger investment equal in size to the savings (Samuelson/Nordhaus 2009: 574). The mechanism is elegant in theory but fails in practice. First, an increase in individual saving does not raise aggregate bank deposits; it merely redistributes the existing money stock among agents. What one household deposits was already money in an employer’s account. No new saving is created. Second, banks do not lend pre-existing deposits: they create deposits in the act of lending. Investment is not constrained by a prior pool of saving. Third, interest rates are not determined by the saving rate, but by central bank policy and banks’ assessments of credit risk. The supposed adjustment channel therefore does not exist. More saving does not reliably lower interest rates, nor does it trigger investment. What it does reduce, with certainty, is spending. The neoclassical model ends up assuming that firms increase investment precisely when their sales are falling. There is no channel that connects saving with investment in the assumed neoclassical direction. How we can better understand and work with the accounting identity I = S will be discussed in Chapter 4.