The Keynesian Propensity to Spend

We have all heard the claim that government spending can “kick-start” the economy when it falters. For example, governments all over the world introduced a steady stream of stimulus packages that were intended to compensate for the effects of the 2008 recession. In trying to understand the logic for such economic machinations, we can look to chapter ten of Keynes’ The General Theory [1].

Keynes postulates that a change in government spending (something he calls “investment”) is multiplied by the investment multiplier, k, to produce a change in aggregate income [2]. Related to this is a construct he calls the marginal propensity to consume (let’s call it m) which attempts to quantify the “consumption psychology” of people to consume an increased income rather than to invest it. He also describes Richard Kahn’s employment multiplier, k’, which relates a change in aggregate employment to a change in investment. He states that k and k’ do not need to be equal, but then spends most of the chapter assuming that they are because the equality, in his words, “elucidates the ideas involved”.

Keynes’ proposed relationship between the investment multiplier and the marginal propensity to consume can be written mathematically as k = 1/(1 – m) which means that the greater the marginal propensity to consume, the greater the investment multiplier (as m approaches unity), whereas the lower the marginal propensity to consume, the lower the investment multiplier (as m approaches zero). Keynes writes:

…if the marginal propensity to consume is not far short of unity, small fluctuations in investment will lead to wide fluctuations in employment; but, at the same time, a comparatively small increment of investment will lead to full employment. If, on the other hand, the marginal propensity to consume is not much above zero, small fluctuations in investment will lead to correspondingly small fluctuations in employment; but, at the same time, it may require a large increment of investment to produce full employment.

It is no wonder the investment multiplier has intrigued mainstream economists for decades, especially considering its blueprint for potentially achieving full employment with only a small increment of investment!

Keynes provides an example to illustrate the basic concept. Consider the case where people have a “consumption psychology” that induces them to spend 90% of their increased income (m = 0.9). In this case, k = k’ = 10, and therefore an expenditure by government in public works will increase both total income and employment ten times compared to that created by the original expenditure.

This assumption of the multiplier effect is commonplace today. Mainstream economists and politicians believe that spending by government can be multiplied many times over in terms of employment if people can be convinced to consume. After the economic crash of 2008, not only did economists and politicians demand increases in spending on public works but they also went about urging citizens to increase consumption through programs such as “Cash for Clunkers” and the “Home Renovation Tax Credit”. United States congressmen who introduced the personal tax rebates in 2008 “with unprecedented speed and cooperation” hoped that the citizens would “spend the money just as quickly and jolt the ailing economy to life”.

Note that Keynes’ propensity to consume is adversely affected by negative economic circumstances. Hence, central bankers and politicians will rarely admit to any serious financial concerns that might exist. For example, in 2005, Ben Bernanke (who later became Chairman of the U.S. Federal Reserve Bank) stated that the recent increase in housing prices reflected “strong economic fundamentals”. Of course that proved to be very wrong, but the message was delivered to ensure that consumers kept a high propensity to consume.

Unfortunately the investment and employment multipliers are difficult to determine in any given circumstance even though Keynes took the uncomplicated approach of reducing them to simple scalars. In reality these factors represent the economy which comprises billions of economically linked individuals and businesses whose preferences and motivations continually change. Hence, it is not possible to know precisely what magnitude or type of government expenditure or degree of consumption are required to achieve a given increase in income or employment.

Furthermore, modern economists and politicians would never admit to the possibility of the investment and employment multipliers being less than unity because that would imply that an increase in public spending would result in a change in employment and income that is less the original investment. More significant are negative investment and employment multipliers because those would imply that an increase in government spending would result in a loss of employment and income (conversely, a decrease in government spending would result in employment and income gain).

Nonetheless, the philosophy that the government can create jobs through spending persists even though they cannot know whether spending will have a positive effect on employment. Nor can they know how much, or how fast, to spend. But they feign that they do know and that they can effect the requisite changes in employment. The only real thing they do know, of course, is that any unemployment resulting from the bust period of a previous boom period needs to be reduced before the next general election, and if it is not reduced, they need to be seen as “working hard” on reducing it. Hence, the propensity is to employ more, rather than less, government spending and to spend at the greatest rate possible.

Another problem with the investment and employment multipliers is that they are used to describe a complex closed loop system with simple constants.  In a closed loop system there are continual adjustments that fine-tune the resulting outcome.  However, any such system suffers from the possibility of the output swinging well above or under the intended level, or the possibility of the loop becoming unstable.

In conclusion, Keynes’ simplistic investment multiplier, its cousin the economic multiplier and the elusive marginal propensity to consume have enchanted economists for decades with their promise of increased employment through government spending.  However, Keynes fails to consider the case where these factors become negative whereby public spending becomes detrimental to employment and income.  Furthermore, the scalar nature of these factors grossly oversimplifies the complexities of the economy which is, in essence, defined by the economic behavior of billions of people who modify their preferences and behavior on a continual basis.  There are factors other than simple fabricated ratios that determine the response of employment and income over time.


[1] John Maynard Keynes, “The General Theory of Employment, Interest and Money” (Macmillan Cambridge University Press, London, 1936).

[2] Keynes defines income, consumption and investment in terms of wage-units (where a wage-unit is the “money wage of a labour-unit” and a labour-unit is the quantity in which employment is measured).  In “A History of Economic Thought” (Wesleyan University Press, 2009), William J. Barber describes the strategy as follows:

“As an analytical device, the Keynesian wage unit has much in common with the manoeuvres performed by those classical economists who attempted to measure the value of goods in terms of labour.  They were obliged to explain how various grades and skills of labour could be reduced to a common denominator.  Normally, they treated an hour of unskilled labour as the basic unit… For practical purposes this technique of measurement – in terms of the  number of standard hours of labour employed – is unwieldy.  None of the employment statistics presently gathered lend themselves to a wage unit type of measurement without enormously time consuming adjustments. “

[3] “You could get your tax rebate by May”, (MSNBC, Jan., 2008)

[4] “Bernanke: There’s No Housing Bubble to Go Bust”, (Washington Post, Oct. 27, 2005)

Leave a Reply