.
NB / the dotted line indicates it would continue through the origin, it is not 45o.
Therefore over a long period society does not consume
Secondly data suggests over the short run the APC is not constant but varies according to the business cycle. In periods of boom the APC fell below its long run average but in periods of recession the APC rose above its long run average, APC was very contra-cyclical i.e. it went in the opposite direction.
Kuznets work also questioned whether redistribution from upper to lower income groups would have any impact on the level of aggregate demand. This conflict in empirical findings provided a challenge to the generality of the absolute income hypothesis and encouraged further research directed in providing some explanation of why cross section and time series results should produce such conflicting view of the consumption function.
Duesenberry (1952), in modifications of the Keynesian consumption function highlighted in the relative income hypothesis that consumers are affected by sociological and psychological factors as well as purely economic ones. Consumers are creatures of habit and affected by social status. He argued that Keynesian analysis focused on an inappropriate concept of income, so when explaining variations in consumption, the relevant concept of income was not absolute income but relative income. There are two hypotheses involved here. First he recognised that consumers planned their consumption relative to the consumption of others and it was interdependent, this was known as the demonstration effect, this means APC will depend upon the consumers position in the income distribution. A below average income consumer will consume a larger proportion of income than the average consumer because they are trying to keep up to some national standard. Secondly if we have a consumer earning above average income they will consume a smaller proportion because their consumption
habits are pulled down by the national average and therefore we have an explanation for the cross sectional data.
Basing their consumption on what happened in a previous time frame. Once the consumer reaches a level of consumption, decreases in consumption are more difficult than decreases in saving. Once a certain lifestyle has become accustomed to it is far more difficult to revert back especially since consumers are creatures of habit. Therefore consumption will depend on current income but relative to its previous peak.
If there is a fall in income it can be argued that consumers will reduce consumption by a little as possible and adjust by letting their savings take the strain. So therefore their behaviour will be characterised by the fall in income but with consumption falling by just a small amount compared to what it was when income was Y2, so APC rises above the LR level i.e. will rise as income falls.
There is mixed support for this theory within mainstream neo-classical economists and has not been accepted as a dominant theory. There is prejudice by some that economists should not be involved with sociological and psychological factors. However it can be argued if these factors are of relevance then maybe they should be incorporated. As a consequence other theories have become more dominant, mainly from the neo-classical school.
Because both the Keynesian and Duesenberry explanations of consumption behaviour link current consumption closely to current income, they did not explain consumption behaviour of groups who have fluctuating income. The answer to this was quite straight forward, because individuals do look into the long term and do not simply just act by the short term. For example in the event of a good harvest a farmer will not drastically increase his consumption in line with his increased income and similarly in a bad harvest cut back on consumption to some minimal level. Over any given period he will be aware of his fluctuating income and therefore is aware what he can on average earn, i.e. has an idea of his normal income. As a result consumption plans are based on this rather than current income. In years where current income is in excess of normal income he will save and in years when current income falls short of normal income he will dis-save.
The normal income approach was adopted by the neo-classical school, which can be seen as anti-Keynesian and is based on microeconomic foundations by using a utility framework.
The essence of this approach is the consumer aims to maximise utility by seeking the optimal pattern of consumption expenditures through time, and it is the basis of both the permanent income hypothesis (Milton Friedman) and the lifecycle hypothesis (Ando and Modigliani). They argued that consumer behaviour should have microeconomic foundations so the consumer’s decision-making process as maximising their utility as a function of consumption over time can be represented. Therefore the normal income approach to the analysis of aggregate consumption has its roots more firmly embedded in microeconomic foundations than the current income approaches.
In the diagram below consumption and income in periods 1 and 2 are measured along the horizontal and vertical axis respectively. It is assumed that the individual receives an income of Y1 in period 1 and Y2 in period 2 and that Y1=Y2.
If current income was the determinant of current consumption then the individual would use all of Y1 in period 1 and all of Y2 in period 2, reaching point a in consumption space. But if the individual can use the capital market to deposit savings or borrow against future income to finance future consumption, the individual is open to more options. For example he could consume C1 in the first period, saving C1Y1 and consume C2 in the period 2, ending up at point d. The intercepts of the constraints are the extremes of consuming all income in the first period or in the second. In the first case the consumer would consume an amount equivalent to the present value (PV) of his income over the two periods.
PV=Y1 + Y2 _ _
(1 + r)
Income in the first period and income in the second period discounted at the current rate of interest (r ). Or he could choose to consume none of his income in period 1, invest it at the current rate of interest and consume a terminal value of
TV=Y2+Y1 (1+ r)
Two specific normal income theories will now be looked into, the lifecycle hypothesis and the permanent income hypothesis.
The development of the life cycle hypothesis is associated with Ando and Modigliani. This hypothesis argues that the individual will maximise his utility by maintaining a stable pattern of consumption over his lifetime. Income will be relatively low at the beginning of his life, high in the middle years and then low again towards retirement, as can be seen from the trend line below (c) in the diagram below.
Although income varies over the life-cycle, consumption increases at a fairly constant rate, it is implied that the very young and old will largely be dis-savers (consumption exceeding income). Whilst the middle-aged will be savers (to pay off debts incurred in earlier years and to make provisions for retirement), therefore as we can see the interest is in explaining aggregate consumption and the hypothesis assumes that consumers in a given age group have similar tastes and preferences for present and future consumption.
The main motive behind this theoretical work was to explain the data.
The reasoning here is firstly above average income classes, groups with a high proportion of consumers have high incomes because they are middle-aged. They are at their peak in income and their APC will therefore be expected to be relatively low. They have done their dis-saving and therefore APS (1 – APC) will be high. The data reveals what we would expect from the hypothesis.
Secondly below average income classes will have a high proportion of consumers who have low income because they are either young or old, and they are dis-savers and expect them to have a high APC and low APS.
The theory also argues that consumption patterns here are much more stable than those patterns predicted by linking consumption to changes in disposable income therefore the theory provided a plausible reconciliation of cross section and time series empirical studies. Typical cross section data will include individuals in the early years of work and in retirement who have a high MPC, they are dis-saving. Whereas those on higher incomes will be in the middle stage of their life cycle and have a lower MPC and higher MPS.
The life cycle hypothesis directs attention away from current income and towards a broader concept of lifetime resources, although it includes current income it places greater emphasis on expected income and wealth. There may be empirical difficulties associated with testing the hypothesis, but it is a plausible basis for analysing aggregate consumption, and has more solid microeconomic underpinnings than the Keynesian or Duesenberry approaches.
This theory was problematic in that because expected income is measured, this may prove difficult as it is not directly observable and has to be forecast
Milton Friedman (1957), put forward the permanent income hypothesis and can be regarded as the best-known critique of the Keynesian consumption function. Friedman recognised that income is not always as regular and predictable, and for some consumers it is erratic and they may experience seasonal fluctuations, just like in the life-cycle hypothesis. When Friedman made this statement he did so by looking at the macroeconomic business cycle.
This approach does not look into current income but more into long-term consumption possibilities so instead of basing income on measured income the basis here is on their expectations of permanent income. This is the part of measured income that is regarded as stable. There are three concepts of income in this hypothesis; measured income (Y-YM), permanent income (Y1 Y p) and transitory income (Y1 YTR) which can be either positive or negative.
YM=YP+YTR
CM=CP+CTR
Measured income and consumption is the sum of two components, as shown above. The principle hypothesis here is that permanent consumption is planned to be proportional to income.
CP=KYP
Where the superscripts represent measured, permanent and transitory components, and K is human capital. Measured income therefore differs from permanent income according to whether transitory income is positive or negative. Transitory income refers to gains or losses, for example an unexpected quarterly bonus would represent a transitory gain and any unforeseen expense such as emergency house repairs would represent a transitory loss.
Friedman also argued that K is independent of YP, by making this assumption he was ruling out the relative income hypothesis where it was implied high income groups are pulled down and low income groups are pulled up because of sociological factors.
However K can vary for a variety of reasons including individual preferences, rate of interest, variability of expected income – a consumer with a highly variable income, such as a gambler, would have a different APC than someone who has a stable income. The ratio of human to non-human wealth, human wealth being for example if an individual may choose to go to university to raise skills in order to raise future wealth. Non-human wealth includes factors such as cash, real assets, house, cars etc… Friedman believed this matters because if all an individual has is low human wealth then they will feel the need to build on non-human wealth.
Cross section relationship diagram
As mentioned previously going from low-income groups to high-income groups Friedman rules out the relative income hypothesis and believes it is proportional. What has been
observed in data does not represent the true underlying relationship, and believed it was statistics that distort the true picture.
The true underlying relationship is that low-income groups do not have a different APC to high-income groups, but what was observed was a falling APC. The reason we observe this is because high-income groups will have a dissapproportionatley high number of consumers who are earning above the permanent income. This due to the fact they have been pulled up into it.
Time series income changing over time diagram
The relationship in time series is proportional, long run data showing long run consumption function. The APC varies with income is what the short run explains. Consumers interpret the phases of the business cycle directly as changes in their transitory income. The neo-classical perspective is that agents act are fully informed about everything. If this were true then we would expect a short run relationship
However there are some implications of the permanent income hypothesis. Firstly the multiplier effects of any disturbance in demand will be less if the disturbance is temporary because of the multiplier, if there is a small MPC then there will be a small multiplier.
The multiplier is about the fall in income causing a reduction in consumption. If this is seen as part of the business cycle consumers will not cut back on consumption effectively thus not having much of a multiplier effect. Secondly government policy to cut taxes will not be effective in raising demand because cutting taxes will only increase spending.
To conclude it can be seen as crucial that we have a clear idea of what determines aggregate consumption, as it is the largest component of aggregate expenditure. As this essay has highlighted a lot of the empirical evidence is conflicting and needs to be harmonised in order to make fully informed and accurate predictions of consumer expenditure in changing economic environments. This would be especially beneficial to governments when making policy decisions on taxation and interest rates.
Bibliography & Sources
Greenaway D, Shaw G.K, 2nd Edition(1995), Macroeconomics, Theory & Policy in the UK. Blackwell Publishers Inc.
Brunhild gffhgG, Burton R.H, (1974), Macroeconomic Theory, Prentice-Hall Inc.
Word Count: 2,175