Topic 2: Consumption
“Estimates of the marginal propensity to consume from decade-average observations on actual consumption and actual disposable income tend to be higher than estimated based on quarter observation. This is evidence against Keynes’s theory of consumption and in favor of Friedman’s permanent income hypothesis.”
Explain and discuss.

Introduction
Consumption expenditure accounts for the largest proportion of the Gross Domestic Product in most countries. Referring to Miller (1996), consumption is the total of goods and services that people in the economy wish to purchase for the purpose of immediate consumption. More importantly, consumption is one of the main determinants of an economy’s aggregate demand and economists can therefore estimated the aggregate demand as well as evaluate the effects of fiscal policy based on the determinates of consumption. This essay will commence by explaining the Keynesian consumption function and the consumption puzzle. It will then discuss the Friedman’s permanent income hypothesis before explaining the difference of marginal propensity to consume between cross-section data and long-term time series data. Finally, it will conclude that permanent income hypothesis by Milton Friedman provides a more complete model for people to research consumption and aggregate demand.

The Basic Keynesian Function
Over the past few decades, there are lots of works have done to research consumption. The most famous study is J M Keynes’s General Theory published during America’s great depression. Keynesian consumption function associate changes in consumption with current disposable income. The simple Keynesian consumption function is given by the equation: C =c0 + c1 (Y-T),

Which shows that the consumption level is determined by an autonomous consumption (c0) and a constant fraction of disposable income (c1(Y-T)). Keynesian states that the autonomous figure (c0) would always be positive since there is a certain level of consumption is essential for an individual to stay alive, which might contain expenditure on food, heat, shelter, etc. (Miller, 1996) In addition, the coefficient c1 in the equation, which was named marginal propensity to consume by Keynesian, should lie between zero and one. The marginal propensity to consumer (MPC) shows how much consumption changes with respect to disposable income changes. According to Keynesian, “… men are disposed, as a rule and on the average, to increase their consumption, as their income increases, but not by as much as the increase in their income.” (The General Theory, 1936, 96)

Moreover, Keynesian argued that the average propensity to consume is greater than the marginal propensity to consume since the autonomous consumption does not change with income. Average propensity to consume is defined as the percentage of total income spent on consumption. (APC= CY) It can be seen from the equation that the average propensity to consume decreases as income rises, which also imply that people are increasing their saving while income grows. Keynesian’s theory can also be presented on a diagram:

Model 1: Keynesian’s Consumption Function
C0
C= y
C= C0 +C1(Y-T)
Consumption
Disposable Income
C0
C= y
C= C0 +C1(Y-T)
Consumption
Disposable Income

Figure 1: Consumption and Income in the U.S (1936)

A graphical representation of the relationship between real consumption per capital and real disposable income per capital in the US in 1936 is presented in Figure 1. The graph suggests that there is a linear relationship between consumption and disposable income. It can be seen from the graph that the as income increases, the consumption also increases as well. In addition, the marginal propensity to consume is a constant while average propensity to consumer is declining while income increases, which support Keynes’s function.

Consumption Puzzle
After the publications of The General Theory, a lot of data have been collected in order to test the validity of Keynes consumption function. However, economists discovered different sets of data based on different estimates of marginal propensity to consume. On the one hand, using the short-run data from 1929 to 1941, Keynes’ followers estimated the consumption function for the U.S: C Short Run= 26.5+ 0.75Yd, which supports the Keynes model. However, by using a long-run data from 1869 to 1938, Kuznets (1946) showed that long run time series consumption for the U.S. economy is C Long Run= 0.95Yd. Kuznet also been pointed out that the average propensity to consume stay relatively stable (ranged between 0.838 and 0.8 98) while income expanded approximately 7 times from 1869 to 1938, which conflicts Keynes’ theory. Moreover, as the data period became more long term, the autonomous constant tended to zero, and MPC tended to one. The Keynesian consumption function was confronted by the consumption puzzle.

Model 2: Consumption Puzzle
Long term
Short term
Consumption
Disposable Income

Long term
Short term
Consumption
Disposable Income

Permanent Income Hypothesis
In 1956, Milton Friedman proposed permanent income hypothesis in response to the consumption puzzle. The main assumption behind Friedman’s theory is that people prefer their consumption to be smooth rather than volatile. (Miller, 1996) It is worth mentioning that there are two different types of income under permanent income hypothesis: Permanent and transitory income. Permanent income is the amount of money people expect to get over a long period. While transitory income, is the difference between actual incomes is permanent income that is not expected to occur. It varies in every period (Miller, 1996). However, it should be noted that as the number of time periods examined increases, the transitory income would tend to zero due to the fact that good and bad spells will average out (Miller, 1996). In this case, change to current income has only a little effect on consumption unless the individual believes that the change has long-term consequences. Therefore, people consume a constant fraction of their permanent income rather than current income (Cp=KYp). Furthermore, It is also stated by Friedman that the marginal propensity to consume is equal to average propensity to consume in permanent income hypothesis.

Regarding to the difference of MPC from cross-section data and long-run time series data, it was reconciled by Friedman using “error in variable” argument. (Alimi, 2013) When economists estimate MPC base on quarter observations, most changes in income are dominated by transitory changes. Hence, the data pick up the actual income instead of permanent income, which lead to a relatively low MPC since people are trying to smooth their consumption by saving most of the temporarily high income. When the actual income is lower than temporary income, people would choose to borrow or use their accumulated saving from the past. In this case, the short-term consumption function will spread out and flatter than the long-term data on chart. However, as referred before, most changes in transitory income and consumption averages to zero when looking at a long period. Therefore, the marginal propensity to consume based on decade average data naturally tend to be higher because people have already combined the higher income into their expectations and have increased their consumption. Furthermore, the statistical evidence suggests that the MPC from a temporary tax cut is about half the MPC from a temporary tax cut, which validates Friedman’s theory. Accordingly, it can be conclude that consumption is based on permanent income in Friedman’s PIH theory rather than current income in Keynes’ consumption theory.

Figure 2. Time series data- U.S. Consumer Expenditure Survey, 2002

The graph on figure 2 shows the relationship between personal disposable income and consumption expenditure from the period 1950 to 2002 in the U.S. It is fairly clear that the consumption of U.S. has increased considerably over the past few decades. Moreover, the graph also indicates that both MPC and APC is a constant in a long run, which supports Kuznets’ finding and Friedman’s permanent income hypothesis.

Conclusion
To sum up, it can be seen that although Keynes’ theory contains many valid ideas, consumption based on current income was proved false in a long run. As shown by Milton Friedman, permanent income hypothesis is more completed either from the short run or long run data. Therefore, it can be said that Friedman’s permanent income hypothesis provides a better model for economists to research consumption and estimate aggregate demand. (Words: 1186)

Reference:
Alimi, R.S., (2013). Keynes’ Absolute Income Hypothesis and Kuznets Paradox. [Online]
Available at: http://mpra.ub.uni-muenchen.de/49310/1/MPRA_paper_49310.pdf [Accessed 13th November 2013]
Keynes, J.M., (1936). The general theory of employment, interest and money. London: Macmillan.
Miller, T., (1996). Explaning Keynes’ Theory of Consumption, and Assessing its strengths and weaknesses.[Online] Available at: http://www.economic-truth.co.uk/content-essays/bsc/keynesconsumption.pdf [Accessed 13th November 2013]