THE KEYNESIAN REVOLUTIONThe Beginnings of Modern Macro Economics
big question addressed by macroeconomic theory is deceptively simple, "What
determines the level of national income?" This is an important question
because there is a close correlation, of course, between the level
of national income and the well-being of the population.
The most familiar measure of economic well-being is per capita income, which is simply national income (GDP) as described in the previous topic divided by the size of the population. A high and rising level of (real, that is inflation-adjusted) per capita national income implies a high and improving level of well-being. A low level of national income implies the opposite.
Another reason for wanting to understand what it is that determines the level of national income, whether total or per capita, has to do with the economic problem as defined in Topic 1 -- the scarcity of economic resources and the limitless nature of human wants. If the level of national income is below what the economy is capable of producing, the system is operating inside the production possibilities frontier, wasting resources and leaving unsatisfied wants which could have been satisfied. This less-than-full capacity level of income/output is associated with unemployment of labour and other resources.
One of the great problems of capitalist, free enterprise economies historically has been the instability of output and employment. The waste of human resources associated with unemployment is particularly distressing, not only for those who must bear the burden of unemployment directly, but because all members of the community suffer the loss of output which would otherwise have been available. Note that such a loss is unrecoverable. Output lost today can never be regained.
Modern macroeconomic theory originated in the 1930s when the world experienced its worst modern economic depression, a depression which devastated even the most advanced economies and created widespread dissatisfaction with the performance of the capitalist, free enterprise system. The human suffering associated with mass, apparently chronic, unemployment extending over a period of years, gave rise to many movements for reform ranging from the extremes of Marxism-Leninism on the left to Fascism on the right.
For many economists the experience of the long depression of the 1930s suggested that there was something wrong with their traditional theories of how economies worked. Why were markets not making the kinds of adjustments explained in standard theory? Why did consumers not increase their spending on consumer goods and services when the prices of these goods were falling? Why did suppliers resist cutting prices far enough to eliminate surpluses of unsold goods? Why were there so few job opportunities when workers were out looking for work? Why would unemployed workers remain unemployed rather than work for low wages?
Many explanations for these failures were advanced. Some focussed attention on the role of money, suggesting that the basic problem was that the banks and other financial institutions were perversely restricting credit at a time when everyone could see that there should be more, not less, money available when business conditions were so bad. Others were cast in terms of hypotheses about long historical processes which, some suggested, had brought the system of capitalism to its final collapse.
The elegant body of neo-classical theory synthesised by Alfred Marshall seemed to have little to say that was relevant to the real-world problems afflicting the developed industrial economies in the 1930s. In particular, neo-classical theory was unable to account for the persistence of surplus labour in the labour markets and the apparent ineffectiveness of price reductions in product markets to restore consumer spending to levels which would justify expanding production again.
Indeed, to the extent the available theory had anything to say about such situations, it was to deny that they could persist for any length of time. This was based on a very old proposition known as Say’s Law, named after the French economist Jean-Baptiste Say (1767–1832). According to Say, it was impossible for the economy as a whole to produce a surplus of commodities. Although there could be surpluses and shortages of particular commodities, these would be automatically eliminated by the natural functioning of markets. Prices would fall for goods in surplus and rise for those in deficit, leading producers to switch from producing the goods in surplus to producing more of those which were excessively scarce. Overall, Say reasoned, it was impossible for more goods in general to be produced than there was a demand for simply because the production of goods creates the income with which they can be purchased. Supply in this sense creates its own demand.
This line of thinking had been incorporated into English language-economics by James Mill (1773-1836) and it was subsequently carried on, with some embellishment, into the fully-fledged neo-classical theory of Marshall and his followers. There had been some dissenters. Even in the classical era, Malthus had raised the possibility that if workers did not receive the full value of what they produced, they would be unable to buy enough to prevent gluts of commodities from developing. He had also suggested that the same thing might happen because rich people saved too much. But it was generally believed that Malthus’ views had been satisfactorily rebutted by Ricardo and others who addressed the issue. The accepted position came to be that if there was any tendency for savings to be "excessive", the price of borrowing money (the interest rate) would fall. This would cause investors to demand more savings and, consequently, any surplus of saving ("income not spent") would be eliminated automatically. On the other hand, the same equilibrating mechanism would deal with the possibility of an excess demand by investors for savings—interest rates would rise, saving would be encouraged, investment demand reduced, and equilibrium restored.
Another dissenter from the mainstream view was the Swedish economist, Knut Wicksell (1851-1926). Wicksell observed that during periods of economic depression, instead of credit being easy to obtain, it was difficult, often impossible, to obtain and the volume of investment spending and borrowing to finance it was low. Conversely, when the economy was booming and the demand for credit was high to finance new investment, interest rates were low. This appeared to be the reverse of what neo-classical theory would have predicted. In his effort to find an explanation for this, Wicksell struck upon an important idea: however an equilibrium was effected, it was due to changes in the level of economic activity (employment and output), not the rate of interest. During the 1920s economists interested in exploring such relationships were greatly aided by the development of national income accounting systems. Being able to actually measure fluctuations in the levels of output, consumption spending, saving, and investment made these large aggregative concepts tangible.
Changing economic circumstances in the interwar period combined with the growing unease over the relevance of orthodox theory created an environment within which a revolutionary change in economic thought could be brought about. The person responsible was a brilliant and well-connected member of the British establishment, John Maynard Keynes (1883-1946).
Keynes was the son of a successful British economist, John Neville Keynes. Educated at Eton and Cambridge, Keynes worked for the British Treasury, lectured at Cambridge, and was part of the gifted and more than a little self-satisfied "Bloomsbury set" in the years leading up to World War I. At the end of the war he was one of Britain’s representatives at the Versailles peace conference. Strongly opposed to the reparations package imposed on Germany, Keynes resigned from his government post and wrote a devastating criticism of British policy which he published as The Economic Consequences of the Peace. Understandably out of favour with the government of the day, Keynes returned to lecturing at Cambridge where he busied himself making a fortune for himself by speculating in the foreign currency market (and making another one for his college by handling its financial affairs) and writing on economic issues, especially monetary matters. The latter work yielded a major scholarly publication, A Tract on Monetary Reform, published in 1923. In it he argued against the wisdom of attempting to return Britain to the gold standard, recommending instead that a system of deliberate monetary management be implemented in its place.
By the end of the 1920s Keynes had established his reputation as an economist. His position on the reparations issue and the return to the gold standard was vindicated by the course of events. Both, as he had predicted, led to disastrous consequences. It is worth noting that, for all his opposition to conventional positions on matters of policy, Keynes was generally orthodox in his beliefs. Intellectually, he was part of the neo-classical tradition in economics and, more broadly, he subscribed to the traditions of liberalism and individualism. He was not opposed to free enterprise or the market system, but he did come to believe that the only way to ensure that these institutions could survive was to accept a certain kind of government intervention in economic life.
Despite being steeped in neo-classical theory, Keynes was keenly aware of the growing doubts about its relevance to the world of his time. During the 1920s he worked on a new book, A Treatise on Money, which he published in 1930, just after the great stock market crisis of 1929. This difficult, scholarly work incorporated his earlier ideas about implementing a system of managed money to replace the supposedly automatic functioning of the gold standard mechanism. But his supporting arguments were now more fully developed and they focused sharply on the problem of ensuring a balance between saving and investment. Picking up on the ideas of Wicksell and others, he argued that Say’s Law was false. The supply of savings and the demand for savings would not be automatically reconciled by adjustments in interest rates. Deliberate monetary management would be required, perhaps supplemented by direct government spending. In 1930 the importance of this break with tradition by a respected economist was not immediately evident. As the depression decade dragged on, however, this changed.
When Keynes published his next work on the subject, the impact was dramatic. Despite its complex obscurity, The General Theory of Employment, Interest, and Money, published in 1936, was immediately recognised as a revolutionary work. It was also savagely attacked by no less a figure than A.C. Pigou, who referred to it as a "macedoine of misrepresentations". Pigou correctly perceived that the thrust of the analysis in the General Theory was, as Keynes had himself pointed out, to deny the practical relevance of neo-classical economic theory, demoting it to the status of an interesting explanation of a "special case" unlikely to be encountered in the modern world. Chapter 1 of the General Theory, perhaps the only part of the book the non-specialist should attempt to read, can be quoted in its entirety. Note that Keynes uses the term "classical" to refer to what is now more commonly referred to as neo-classical theory.
Another component of total spending, sales of goods and services abroad (exports), also was an unlikely source of recovery. During the 1930s, as in every other such period of economic difficulty before and since, all countries reacted by attempting to reduce imports to protect their own domestic industries. Consequently, there was little hope that any country could experience a rise in the demand for its exports—quite the contrary. This left only one alternative, an increase in government spending to break the impasse.
The idea of using government spending to stimulate a depressed economy was not something most people, including many economists, could easily accept, even in the late 1930s, and there were many reasons for dismissing such a strategy as just another crackpot scheme (of which there were many at the time). Total government budgets in the 1930s were relatively small, accounting for a small percentage of total GDP. Therefore, it could be argued, even a large increase in government spending could have little impact on the economy as a whole. Another problem many critics of the Keynesian approach could point to was that most governments were already either broke or in debt. To increase spending seemed imprudent, unless there was also an increase in taxation, which would, of course, have negated the whole point of the exercise. Many feared that increasing the public debt to finance more government spending would only weaken the already badly shaken confidence of investors upon whom an eventual recovery would have to depend.
Keynes’ analysis in the General Theory became the foundation for modern macroeconomics and a powerful influence on public policy in the years following World War II. While Keynes himself saw his work as providing a defence for a capitalist, largely free-enterprise type of economic system, supporters of the social-democratic welfare state movement in Britain found in it intellectual support for the agenda they proposed for Britain in the period of reconstruction following the war.
The influence of Keynesian thought spread quickly throughout the English-speaking world. A generation of young economists being trained in Britain in the 1940s took Keynesianism home with them, to the universities and government departments which employed them. There were also established academic economists who found in Keynes an answer to questions they had long been troubled with, but could find no acceptable framework of theory within which to address them.
In North America, one of the most important propagators of Keynesianism was the Harvard-based economist, Alvin Hansen. He and a number of his students, who included Abba Lerner and Paul Samuelson, were responsible for incorporating Keynesian ideas into the "new economics" which dominated American academic economics from the early 1940s to the 1970s. Perhaps surprisingly, however, the direct influence of Keynesian thinking on American public policy was not explicitly recognized until the 1960s when elements of Keynesian budget policy were endorsed by the Kennedy administration.
In the Canadian academic world, the first major exposition of Keynesian thought was presented by the University of Saskatchewan economist, Mabel Timlin, whose book, Keynesian Economics, was published in 1942. In Canada, the Keynesian message made itself heard in government policy much more quickly than in the US, due largely to the work of the Queen’s University economist, A.W. Mackintosh, who developed a Keynesian theoretical framework for his draft of the Canadian federal government’s policy statement on postwar reconstruction (the "White Paper on Reconstruction") in 1945. (For more on the Canadian context visit this course site.)
Keynes himself spent
the last years of the 1930s recovering from a coronary condition which
afflicted him shortly after publication of the General Theory. During
the Second World War he worked again for the British Treasury and was an
important British representative at the Bretton Woods talks which established
the International Monetary Fund and other international economic institutions
which it was hoped would enable the world to avoid the problems which had
emerged after World War I. He did not live to see the results. When he
died in 1946, he was recognized as having been the greatest economist of
modern times, and by many as the greatest economist since Adam Smith.
—John Maynard Keynes (1935)
The first step toward understanding Keynesian macroeconomics is to recognise how this kind of analysis relates to the microeconomics studied earlier in this course. The theory of individual markets presented there implicitly assumed that changes in prices and quantities demanded or supplied took place smoothly, indeed, instantaneously. Keynesians emphasise the probability that in most markets, such changes may be sluggish and markets might never fully adjust to eliminate surpluses and shortages. In the real world, it may be difficult to change prices quickly, or to change them at all. In less than perfectly competitive markets, prices may be "sticky" because firms are reluctant to change prices, often preferring to adjust output when demand for their product rises or falls. This is reinforced by the fact that formal contracts and implicit understandings may inhibit price changes. Such influences may be particularly important in labour markets where there is strong resistance to changes in wage rates from both employers and unionised workers.
Followers of Keynes have tended to emphasise the imperfections which characterise individual markets: lack of information, concentrations of economic power, price-seeking rather than price-taking, the power of unions, the prevalence of monopoly and other forms of imperfect competition, the influence of contractual agreements and the like. (Those opposed to Keynesianism, of course, emphasise the power of market forces and prefer to believe that for the most part markets "clear" smoothly and efficiently.) The Keynesian analysis consequently focuses attention not on price changes but on changes in output levels when there are changes in demand. Indeed, the basic Keynesian theory was developed in the context of an assumed stability in general price levels. Given the Keynesian expectation that free enterprise capitalist systems had a tendency to "run down" due to a lack of spending initiatives, it was understandable that Keynesian economists were inclined to ignore the possibility of price inflation (a rise in the general price level) becoming a significant problem. While giving much attention to the instability of modern economies overall, the analysis is mainly concerned with fluctuations in income/output and employment, not changes in the general level of prices.
The microeconomic analysis studied earlier demonstrated how once the relationship between price and quantity demanded and the relationship between price and quantity supplied had been established, it was easy to find the price/quantity combination which would satisfy both suppliers and demanders and result in a market equilibrium at which there would be neither surpluses nor shortages.
Something similar can be done for the economy as a whole, but, while the concepts will appear to be similar to those used in the earlier microeconomic context, there is a very important difference between them. When the subject is the economy as a whole, rather than individual markets within it, references to "price" cannot be to some particular price, such as the price of cabbage, relative to some other price, such as the price of wheat. Price in the aggregate refers to the average level of all prices. This rather abstract concept has already been introduced in the treatment of price indices. Recall that a very broad general index, the GDP deflator, can be constructed to represent the average level of all prices in the economy. Changes in the GDP deflator indicate changes in the general level of all prices.
It is also important to recognise that moving from the micro to the macro context means that the term "supply" does not relate quantity supplied to a particular price, but to the average of all prices as represented by the GDP deflator. Secondly, aggregate supply is not the sum of what individual firms’ will be willing to produce. Recall that only perfectly competitive firms can be thought of as having supply schedules relating quantities and particular prices. Price-seeking firms, monopolies and other imperfectly competitive enterprises do not have unique supply schedules relating price and quantity which will be supplied. As will be shown in the next Topic, what aggregate supply does represent is the total real output of the economy in relation to the general level of all prices. We are now talking about the quantity of all goods and services which will be made available at different general price levels as measured, for example, by the GDP deflator.
But is this a meaningful concept? The answer depends on the period of time being contemplated. If the economy is being studied in the context of a long period of time, what difference would the overall level of prices in general make? Going back to the production possibilities frontier concept developed in Topic 1, would it be possible to increase the productive capacity of the economy by changing the level of all prices? The money value of GDP would be larger, but the real value would be the same so long as the system was operated at full employment. In this long term, historical context, then, aggregate supply would be the same whatever the level of prices in general. In other words, there would be some particular level of real output the economy was capable of producing and this would not change, whatever the level of prices was.
Prior to Keynes, this was the conventional view of the relationship between the price level and real output, even in the short term. The classical economists assumed that the economy would normally operate at full employment. If something caused the prices of goods to rise, all other prices, including wages, would rise. Firms would see their revenues going up, because output prices would be rising, but their labour and other input costs would also rise and they would consequently have no reason to alter output. Total supply would remain constant and would correspond to the full-employment capacity output of the economy.
The Keynesian view of this relationship was quite different. To begin with, Keynes was not interested in the long term, only the short term. In the short term the kind of rapid and smooth adjustment to disturbances depicted above in connection with long term supply are improbable, especially in highly imperfect markets such as labour markets. If something caused a rise in the prices of goods it would be some time before new wage contracts were negotiated to raise money wages. Workers would consequently experience a fall in their real wages (with the prices of goods higher and the same pay cheque, workers have less real spending power) but firms would experience falling production costs and expand production, hiring more workers in the process. This implies, then, that in the short run total output and the employment associated with it may vary with changes in the general price level, rather than remaining constant as the traditional analysis had assumed.
This particular feature of the Keynesian theory was obscured in the early expositions of the analysis because of the way Keynes chose to focus attention on changes in the level of real output rather than changes in the price level. What he wanted to show was that in situations where there was considerable unemployment, changes in the amount of spending in the economy would affect the level of real output (and the employment associated with it) and have a negligible impact on the level of prices. In effect, the original Keynesian theory assumed that the price level could be taken as fixed, with the level of real output being determined by the amount of spending (total demand) in the economy.
The distinguishing feature of Keynesian theory was its attempt to explain how the level of national income (GDP) is largely determined by changes in the level of planned total spending. Keep in mind why this is important. Ideally, the level of national income would always be such that all the resources currently available in the economy were being utilised to produce the current level of output. If they were not, there would be unemployment and waste. Although the early Keynesians were not much concerned about the possibility (given the circumstances of the time), trying to operate the economy in excess of such full capacity output will cause inflation —a rise in the general level of prices. In what follows, we will begin by sketching an explanation of how the determination of the level of output and employment was explained in a simple "fixed price-level" Keynesian macro-economic model. We will then look at how such a model came to be modified to incorporate the possibility of changes in the general level of prices.
The basic Keynesian
explanation of how the level of national income is determined runs in terms
of the spending intentions of consumers and private investors. For the
sake of simplicity assume for the moment that no significant spending is
done by governments and that the economy is a closed one —foreign trade
is negligible. Taken together, these spending plans (aggregate planned
expenditures) will determine the equilibrium level of national income.
Whenever you save five shillings, you put a man out of work for a day.
—John Maynard Keynes (1933)
The willingness of consumers to spend is determined in the Keynesian model mainly by the level of (real) incomes they are currently receiving. As noted earlier, a great many other influences will bear upon the spending decisions of consumers, such as their attitudes toward thrift, their expectations of the future, and the stages in life they are at. But all these other influences can be bundled up and dealt with by a ceteris paribus statement, allowing attention to centre on what the Keynesian theory holds to be the crucial variable, the level of current income. The higher the level of current income, the more consumers will plan on spending during the period of time being considered. The lower the level of consumer income, the less consumers will plan on spending.
This relationship could be expressed in terms of saving just as well as consumption because there are only two things individuals can plan to do with their current income —spend it or save it. Planned saving is equal to current income less planned consumption. If the level of current planned consumption spending is known and the level of current income is known, the level of current planned saving is simply the difference between the two. The income referred to is the current level of personal disposable income. Since it is being assumed that government is economically insignificant there are no taxes on this income to worry about.
The hypothesised relationship between consumer spending or saving intentions and the level of current income is straightforward: at low levels of income, consumers find it difficult to save and may even plan to spend more than they currently receive by drawing upon accumulated savings to maintain some level of spending they may have become accustomed to during a more prosperous recent past. Attempting to spend in excess of current income may be thought of as "dissaving".
At historically high levels of income, the proportion of income spent on consumption falls, and saving becomes positive. In the simplest Keynesian model, this attempt to withdraw income from the spending/income flow will cause the level of income to fall unless it is reintroduced as spending by private investors on capital goods —in other words, through planned investment spending. (Investors borrow savings to finance purchases of new capital goods.) If this planned investment spending is not forthcoming (and Keynes saw no reason why it necessarily would be) the level of income received by consumers will fall. At a lower level of income the amount of saving actually realised will be less than the amount planned or intended. This is why all references to consumption and saving to this point have been preceded by a word such as "planned" or "intended". If individuals or families attempt to increase their saving at some particular level of national income, ceteris paribus, they will cause the level of national income to fall to a lower level. But at a lower level of income they would be unable to save as much as they had wished to when income was what it had been when their savings plan was formed. Therefore, after the event, looking back on the period of time in which the attempt was made to increase saving, the actual saving which took place would be less than the amount they had planned to save. Realised (or "ex post") saving would be less than planned (or "ex ante") saving.
Given all the other things that affect the willingness of individuals in a particular society to spend out of their current incomes, how much consumers will plan to spend will depend on the level of their current incomes. If incomes fell, it would be expected that the amount consumers would intend spending would decline and if incomes rose, the amount of such spending would increase. But much less or more? In Keynesian jargon, the degree of responsiveness to a change in income is called the "marginal propensity to consume" (MPC). Suppose someone were to receive an additional $1 in income. Would he or she react by increasing planned spending by $1 or something less? If everyone in this situation would respond by increasing the amount he or she planned to spend by the full amount of the increase in income, whatever the level of current income, the propensity to spend on consumption out of an additional unit of income would be 1. However, if some of any additional income were intended to be saved, the propensity to spend out of additional income would be less than 1. The marginal propensity to consume is the change in planned consumption expenditure divided by the change in income.
Because of the relationship between consumption and saving, there is a kind of mirror image of the marginal propensity to consume — the "marginal propensity to save" (MPS). The marginal propensity to save is the change in planned saving divided by the change in income or, what is the same thing, it is the fraction of the last (marginal) dollar of income that is saved. Because an additional dollar of income must either be spent on consumption or saved, the marginal propensity to save plus the marginal propensity to consume always add up to 1. If the value of either is known, the value of the other can be obtained by simply subtracting the known value from 1.
This, then, is the
basic structural component of the Keynesian model, the relationship between
income and the spending or saving decisions of consumers. Now the next
component of total spending, the decisions of businesses to spend on capital
goods, that is, investment spending, is considered.
Investment spending is a relatively small part of total spending even in the rich, well developed industrial countries. But it plays an important part in the Keynesian analysis because it is so volatile a component of total spending. Unlike consumption spending, which tends to be a relatively stable percentage of total spending over time, investment spending can fluctuate dramatically and it is consequently a major cause of the variability in income levels from year to year. In the Keynesian system the amount businesses will wish to spend on new capital goods (factories, machinery and equipment, structures and inventories) is determined by two considerations: one is the expected stream of income they expect to be able to earn from such new capital goods (the amount such goods can add to the firm’s net earnings) and the other is the cost of borrowing the money needed to finance the purchase or construction of them.
The cost of investable funds is relatively easy to conceptualise in a general way: it is essentially the "going rate of interest" at which businesses could expect to borrow money (given the riskiness of the loan, the reputation of the borrower and so on). The expected return from the investment is much more difficult to generalise about, for it will be greatly influenced by the subjective judgement of the business person involved. (If the investor is optimistic about business prospects in the immediate future the estimated returns might be quite high. If the investor were pessimistic, the estimated returns would be low for the very same investment.) However, it may be safe to suggest that if (real) interest rates are low, relative to recent rate levels, businesses will plan to invest more in the current year than if interest rates were high. Thus, the level of planned investment spending in the basic Keynesian model is taken to be negatively related to the level of interest rates: the higher the going rate of interest, the lower the level of intended investment spending.
A reduction in the (real) rate of interest brings an increase in the amount of investment businesses wish to do. If there were a change in business expectations for the better, planned investment would then be greater at all levels of interest rates. A loss of investor confidence would have the opposite effect and the desired amount of investment spending would be less at all interest rates.
Again, as explained earlier, not all investment spending need be voluntary. If investment is defined to include the holding of inventories (stocks of finished goods), businesses may find themselves actually investing more than they wanted to or, in some cases, less. If sales fall short of the levels expected, firms may find themselves with inventories of finished and unsold goods piling up in their warehouses, in which case the total amount of investment in that accounting period will be greater than intended. Conversely, if sales are better than expected, inventories will be drawn down and firms may find themselves investing less than intended.
Total investment spending may include not just additions to the existing stock of capital (in real terms), but also replacement capital to repair or take the place of worn-out (depreciated) capital. There is nothing to require businesses to replace their worn-out capital. If the prospects for business profits are very gloomy, current investment plans may be to allow existing capital to wear out rather than replace it. Such capital consumption has been observed during depressions and severe recessions.
The historical record of fluctuations in the level of investment spending has been carefully studied by economists seeking to discover some pattern in it. One of many associations observed is between changes in the rate of growth in real GDP and the level of investment spending. If the growth of GDP accelerates, this often appears to induce an increase in the amount of planned investment as businesses suddenly find themselves under-equipped to cope with the rise in demand for their outputs. This "accelerator principle", as it is called in the literature, may help explain some of the surges in investment spending which are often observed in free enterprise economies, with higher levels of income inducing more investment spending and the latter generating more of the former —an "investment boom" (inevitably followed, it would seem, by an eventually collapse of business confidence in the future and a subsequent downward spiralling of planned investment spending and total income.)
In the simplest Keynesian model the equilibrium level of income in the economy is determined by the desired levels of consumption spending (and saving) and the willingness of businesses to spend enough on capital goods to offset whatever saving is attempted. There is no built-in mechanism which will ensure that the amounts business people plan to spend will be equal to the amounts income recipients will intend to save. For Keynes, these were for the most part different groups of people (although there could be some overlap) with different motivations and objectives. While some saving might be done by businesses themselves for the purpose of financing investments, the bulk of savings available and consequently the price of borrowing ("the" interest rate) would be determined by individuals attempting to provide for their old age, to save for the education of their children, etc., and not for reasons of earning income in the form of interest of profit.
Critics of the Keynesian analysis have always had trouble with this perception and particularly the implication that current savings intentions are not significantly influenced by the available return on savings. Keynes was himself a successful player in financial markets and had every reason to be familiar with the way income from savings could be maximised by skilful management of financial portfolios. Nevertheless, looking at the great mass of potential savers and investors, his theory played down the role of the rate of return in influencing the amount individuals sought to save out of their current incomes, emphasising instead the influence of other considerations, most importantly, the level of current income they had available. As for investors, the Keynesian theory clearly recognised the level of interest rates as a determinant of the amount they would be willing to invest. Even here, however, the importance of the cost of borrowing in influencing investment spending decisions was played down in the Keynesian analysis. Lower interest rates would render more investment projects attractive than would be the case at higher interest rates. More important in the Keynesian view, however, was the subjective judgement of investors with respect to the likely flow of returns from such projects. More important than variations in the cost of borrowing was the investor's state of mind, optimistic or pessimistic with respect to the business outlook for the immediate future — the "animal spirits" of members of the business community.
With the basic elements of the Keynesian system now in place, some of the rather drastic simplifying assumptions made earlier my now be removed. The economy envisioned by be made more realistic by introducing a significant role for government and by opening the system up to external influences.
Government spending on goods and services can increase the total amount of spending in the economy. While many influences may bear upon the amount governments decide to spend, the important point about this component of total spending is that it is potentially unlimited. Unlike consumers and private investors, national governments are technically unconstrained in the amount they can plan to spend simply because they have the power to create money. Governments can plan spending on goods and services in excess of their current tax revenues, and they often do. As was noted in passing in the earlier discussion of money, governments could simply print whatever amount of paper money they needed to finance their current spending plans. In practice, however, it is not so much this power to print money that frees national governments of any income constraint on their spending, as it is their ability to borrow from the central bank. Private spenders can also borrow, of course, but only to the extent that they can persuade a financial institution to advance them funds and, as we have seen, the price charged for such funds (the interest rate) may make the borrowing unattractive. National governments, however, can borrow directly from the central bank.
Whether governments exercise these powers of unlimited spending or not is "simply" a matter of policy. There may be many reasons why governments will exercise restraint in this regard, although the temptation to win votes by spending on popular services or projects which benefit politically important interest groups is always strong.
There is consequently no necessary link between the level of real GDP and the amount governments will plan to spend. Because the level of government spending is a matter of policy, further consideration of it will be deferred to the later discussion of macroeconomic policy. It may be noted here, however, that in the Keynesian view, government spending provided the main instrument for bringing the level of planned spending in an economy up to the amount needed to sustain full employment of available labour resources. Measures to achieve this based on encouraging private consumption spending or private investment spending were thought unlikely to succeed. As for the possibility of achieving the socially-desired level of total spending by manipulating an economy's international economic relations, this too seemed unlikely.
Introducing foreign trade into the Keynesian system is done by simply adding a fourth component of total spending in the form of the net effect of foreign spending on goods and services produced in that economy after deducting the spending done by its residents on imported goods and services. Subtracting the value of imports from the value of exports normally leaves a relatively small residual, which may be either positive or negative depending on whether the value of exports exceeds that of imports or vice versa.
A number of factors influence the value of a country’s exports and imports during any particular accounting period. The demand for a country’s exports will be strongly affected by the level of real GDP in the countries to which it sells goods and services. It will depend on the efficiency with which these goods and services can be produced, making them competitive or otherwise in world markets. Because, as noted earlier in the discussion of trade, foreign buyers of a country’s exports must first acquire that country’s currency with which to make such purchases, export sales will also depend on the exchange rate which will determine the price of a country’s exports expressed in foreign currency values. There is no direct connection between the demand for a country’s exports and the level of its own GDP.
The main connection between a country’s net export balance and its own GDP is the demand of its residents for imports. Obviously, a high level of (real) domestic income will encourage residents to buy more imported goods and services than when domestic income levels are low — just as a high level of real domestic income will encourage higher consumption spending on domestically-produced goods and services. Note, then, how a rise in real GDP will be associated with a rise in domestic spending on domestically produced goods and services, but that the net effect of this on total spending will be offset to some extent by an increase in spending on imports, which reduces the value of net exports. Of course, the level of demand for imports will also be affected by the same things mentioned in connection with exports, such as the competitiveness of foreign goods in the domestic market and the exchange rate.
Total, or aggregate, expenditure in a Keynesian model is the sum of the four types of spending just discussed. This total can be built up by adding together the amount (net of imports) consumers, private investors, governments and foreign purchasers wish to spend on goods and services produced in the domestic economy. Many factors impinge to make this total larger or smaller, but what is of interest here is the amount of expenditure desired at various levels of real GDP.
The basic proposition in the Keynesian analysis is that if the amount of total desired spending (private consumption spending, business spending on investment, government spending, and the net export spending) is equal to the current level of real GDP that level of real GDP and the employment associated with it will remain stable from one accounting period to the next. In other words, when the amount of total desired spending just equals the amount of income currently being produced in the economy, an equilibrium will prevail.
If in each accounting period the amount that consumers and others plan to spend is equal to the current level of national income, there will be no reason for that level to be any different in the next or succeeding accounting periods.
But is such an equilibrium level of income a desirable level of income? As we have already seen, the answer is, "not necessarily". What if this level of income can be produced using only part of the country’s productive capacity? That is, what if this is a less than full employment level of GDP? The fact that it is an equilibrium level is now not a virtue, but a fault, for such a level of income tends to persist, which is why it was defined as an equilibrium level to begin with.
Suppose the equilibrium condition, that the level of desired spending is equal to the level of current income, is satisfied at less than the full employment level of national income. The obvious policy conclusion would seem to be that the level of desired spending should be increased. How could this be done?
One possibility might be to get consumers to want to spend more — that is, save less. This would require measures to shift the propensity to consume or the propensity to save—for example, a mass campaign to convince people that it was everyone’s civic duty to spend as much as possible or that saving was a socially undesirable activity. Such measures might have some success during wartime, but under normal conditions would likely be neither effective nor politically feasible. Another possibility, of course, would be for the government to reduce taxes. If it were consistent with the government’s budgeting policies to do so, this would leave consumers with more disposable income and, given their propensity to consume, consumption spending could be expected to increase — although it is not certain that consumers’ will behave in any expected fashion.
What about investment spending? This rests, as noted earlier, on a pair of opposed influences, the expectations of business people about the probable profitability of investment, and the level of interest rates. The first of these is again probably too subjective and politically sensitive to be used successfully as an instrument of economic stimulation. Indeed, investors might react perversely to any urgings to invest, on the reasonable grounds that if there was any need to artificially encourage investment the outlook must be bleak and it would be prudent to avoid taking any further risks. Manipulating the level of interest rates is a more promising possibility. The central bank could use its power to increase the money supply and thereby make investment money easier to obtain and cheaper, that is, at a lower rate of interest. This possibility will be examined further when monetary policy will be looked at as one possible instrument of macroeconomic stabilisation policy. As noted earlier, however, in the Keynesian view investors do not react strongly to changes in interest rates, especially at times when business optimism is at a low ebb.
Is there any possibility of increasing aggregate spending by operating on exports and imports? So far as exports are concerned, under some circumstances it might be possible to encourage foreign buyers to purchase more of the country’s exports by engineering a reduction in the exchange rate. If the institutional arrangements are such as to permit a country to cause its exchange rate to fall relative to the value of its trading partners’ currencies, export sales could be stimulated. It might also be possible to increase the competitiveness of domestic producers by increasing spending on research and development, training programs, market research and a host of other promotional means. As for imports, the most direct measure would be the old Mercantilist strategy of raising tariffs, making foreign goods more expensive in the domestic market or to impose non-tariff barriers such as quotas and product standards requirements. The trouble with all these devices, however, is that they would surely lead the country’s trading partners to retaliate with similar measures of their own. As the world has moved toward freer trade and the elimination of such traditional impediments to trade, it seems unlikely that export promotion or import restriction could be used successfully to raise aggregate demand.
This leaves government spending as the obvious (certainly in the Keynesian view) remedy for deficient private spending. In the Keynesian framework, the most effective way of raising the aggregate expenditure function to a level where it will sustain a full employment equilibrium is by increasing government spending (net of taxes). Before looking at this possibility in more detail, one further structural feature of the Keynesian model must be understood.
The relationship between a change in the level of desired spending and the resulting change in the level of national income is not a proportional one. It does not take a very large increase or decrease in planned spending to have a significant effect on the equilibrium level of GDP. Even a modest change in the amount investors plan to spend can have a big impact on the level of national income. So can even a modest change in government spending, or any other component of total spending. This helps explain why the level of economic activity can be so volatile. And it also makes more plausible the possibility of using government spending as a means of offsetting the effect of unwanted changes in private spending. What is at work here is called the "multiplier". A change, positive or negative, in the level of planned spending will have a "multiplied" impact on the level of national income.
This multiplier effectcomes about because of the way an initial change in spending spreads through the economy. Imagine a firm deciding to reduce its output because it is experiencing poor sales and a build-up in inventories. It lays off workers. These workers experience a reduction in income to which they respond by reducing their spending on consumer goods and services. This reduces the incomes of those who were formerly producing these particular goods and services, causing them in turn to plan less spending. After the first few stages in such a process the marginal or incremental effects may become quite small. But if all the incremental cuts in spending are added together, they would certainly amount to far more than the initial reduction in spending that started the process off. It is this cumulative effect of reductions or increases in spending in the economy as a whole that results in the final impact of a change in spending being much larger than the initial change.
The strength of the multiplier effect depends on how much of any reduction or increase in spending is passed on as it spreads from one decision-maker to another. Suppose the economy is operating at less than full employment. Some unemployed workers are hired by a government agency to clean up an abandoned industrial site. These workers now have an increase in their incomes. Will they spend this increase? Since they were probably existing on a minimum income while unemployed, it is likely that their marginal propensity to consume will be high, probably close to 1. If they spend every dollar of additional income they get, whoever they buy goods and services from will have an increase in income of a dollar for every dollar the government lays out in wages for the clean-up crew. How will the recipients of this next stage react to their increased incomes? If their marginal propensities to spend are high they too will pass on nearly the whole amount of the original injection of spending power. If their marginal propensities to spend are low, the process will soon be checked.
Consider a simple arithmetical example. Begin with new spending amounting to $100. If all involved have high propensities to spend out of additional income and the total amount of spending is summed over 10 successive steps this might produce something like $100 + 99 + 98 + 97 + 96 + 95 + 94 + 93 + 92 + 91 = $955 in new spending. But if the propensities to spend are low the series might be $100 + 90 + 80 + 70 + 60 + 50 + 40 + 30 + 20 + 10 = $550.
Evidently the impact of some initial change in spending, then, depends strongly on the propensity to spend out of income or, what is the same thing, the propensity to not spend. The main reason for not spending is, of course, the desire people have to save some of their current income. There are two additional reasons in a modern economy why income received may not be passed on as increased spending in the domestic economy. One is the leakage of spending to imports. If there is a high propensity to spend on imports, much of any initial increase in spending in a country may be lost as recipients of the increased income buy more imported goods. This propensity to import may be quite high in countries with large trade components in the GDP. Another possible leakage is taxes. Obviously, income recipients cannot spend additional income if the government takes it away from them. If the government has a high propensity to tax additional income, the strength of the multiplier effect in the private sector of the economy will be much diminished. Note, however, that the effect of imports and taxation can be offset by the stimulative effects of exports and government spending on goods and services, both of which can initiate changes in the aggregate expenditure function, setting off multiplied effects on national income.
The value of the multiplier is equivalent to the change in real GDP divided by the change in desired spending which induces it. Mathematically the multiplier is the reciprocal of the marginal propensity to save or, what is by definition the same thing, one minus the marginal propensity to consume. Thus, if the marginal propensity to save is 1/4, the multiplier would have a value of 4, meaning that a 1 million dollar increase in desired spending (from whatever source) would raise the level of real GDP by 4 times that amount. Real world multipliers, however, are much smaller than this, sometimes not much greater than 1.
The significant policy implication of the multiplier is that manipulation of net government spending on goods and services could have powerful effects on the level of total spending. Government budget decisions could consequently be used to control the level of total spending in the economy, maintaining levels of demand in the system which were consistent with achieving equilibrium at or near the full employment level of output.
The forgoing was the foundation upon which Paul Samuelson erected the great "neo-classical synthesis" popularised in his record-breaking 1948 textbook in elementary economics which dominated teaching of the subject for at least a quarter of a century. Accepting the Keynesian macro analysis, it was possible to show how an enlightened government could use its influence in the economy to stabilise total output at the full-employment, capacity level. Once this was in place, the conventional micro-economic apparatus once again became relevant to the real world. It could be shown that with full employment equilibrium established, markets could function "normally" to allocate scarce resources among alternative employments, the institutions of private ownership and free enterprise would be preserved, income would continue to be distributed on the basis of factor market pricing of inputs with appropriate safeguards to protect the interests of the disadvantaged. The necessary ingredients were at hand with which to explain how a capitalist economy could be indirectly managed to eliminate its chief deficiencies. The reconciliation of Keynesian macro economics and conventional micro economics, the "neo-classical synthesis", thus became the theoretical base upon which a broad political or public-policy consensus appeared to be emerging in most of the industrialised countries by the end of the 1960s. This is now often referred to, somewhat nostalgically, as the "Keynesian consensus".
If our central controls succeed in establishing an aggregate volume of output corresponding to full employment as nearly as is practicable, the classical theory comes into its own again from this point onwards.
—John Maynard Keynes (1936)