Economics•Chapter 5•6 min read•Updated September 24, 2026

Macroeconomics — The Consumption Function: From Keynes to Permanent Income

O
OiyoContributor
5/23

Macroeconomics — Consumption Depends on Lifetime Resources, Not Just This Year’s Income

Consumption is the largest component of GDP, and how strongly it responds to income (the marginal propensity to consume) sets the size of the multiplier. This chapter computes the multiplier with Keynes’s simple consumption function and then turns to theories that ask which income the marginal propensity to consume refers to. A one-off payment and a permanent tax cut are not the same number.

1. The Keynesian consumption function and cross

Keynesian consumption function
C=c0+c Yd,0<c<1C = c_0 + c\,Y_d,\qquad 0<c<1
c_0 is autonomous consumption, c the marginal propensity to consume (MPC) and Y_d = Y − T disposable income. The average propensity to consume C/Y_d falls as income rises.

With C=20+0.8(Y−10)C=20+0.8(Y-10) and investment plus government spending I+G=50I+G=50 (trillion won), planned expenditure is AE=62+0.8YAE=62+0.8Y. Equilibrium is where planned expenditure equals output: Y=62+0.8YY=62+0.8Y, so Y=310Y=310 trillion won.

The Keynesian cross and the multiplier
Y AE 310360 45° AE AE' E E'

When government spending rises by 10 trillion won and AE shifts up by 10, equilibrium output rises by 50 trillion won, because 80% of each increase in income is spent again and again.

If government spending rises by 10 trillion won, the equilibrium is Y=72+0.8YY=72+0.8Y, or 360 trillion won. The multiplier is 1/(1−0.8)=51/(1-0.8)=5. With an MPC of 0.6 the multiplier falls to 2.5. The size of the multiplier is very sensitive to the MPC.

2. The Kuznets puzzle

The Keynesian consumption function fitted cross-sections of households and short time series well: higher-income households consumed a smaller share of their income. But when Kuznets examined decades of long-run data, the average propensity to consume was almost constant. The mismatch — a falling average propensity in the short run but not in the long run — is the Kuznets puzzle. The next two theories answered it.

3. The permanent-income and life-cycle hypotheses

Friedman’s permanent-income hypothesis: income consists of permanent income, expected over the long run, and transitory income, and consumption responds mainly to permanent income. A household whose income is temporarily high in a given year has a low measured consumption ratio that year. This is why high-income households appear to have a low average propensity in cross-sections. In the long run transitory income averages out and the average propensity is constant.

Modigliani’s life-cycle hypothesis: people try to spread their lifetime resources evenly over their lives. They save while working and run down savings in retirement.

Life-cycle consumption (simple form with zero interest)
C=W+R YTC = \frac{W + R\,Y}{T}
W is current wealth, R remaining working years, Y annual income and T remaining lifetime. With wealth of 200 million won, 20 years of work at 50 million won a year and 40 years of life left, C = (200 million + 1 billion)/40 = 30 million won.

In the example, this person saves 20 million won a year while working. If wealth rises by 100 million won (say, from a rise in house prices), annual consumption rises by 100 million/40=2.5100\text{ million}/40=2.5 million won: the marginal propensity to consume out of wealth is 0.025.

The MPC depends on the kind of income change
Income changeKeynesPermanent income / life cycle
Permanent tax cutLargeLarge
One-off paymentLargeSmall (a small share of lifetime income)
Announcement of higher future incomeNone (current income unchanged)Consumption rises now
Temporary income for a liquidity-constrained householdLargeLarge (it would like to smooth but cannot borrow)

4. Random-walk consumption and liquidity constraints

Hall showed that combining the permanent-income hypothesis with rational expectations implies that changes in consumption should be unpredictable, because predictable income changes are already reflected in today’s consumption (random-walk consumption). In the data, consumption often responds even to predictable income increases, such as scheduled tax refunds — excess sensitivity.

One reason is liquidity constraints. Even if future income is high, a household that cannot borrow now cannot smooth consumption. In a study of the 2008 US tax rebates (Parker and co-authors), households spent a substantial share of the rebate within a quarter, and those with few liquid assets responded more. Credit card limits and mortgage terms are channels that change the aggregate MPC. The role of present bias and mental accounting is covered in Behavioural Economics, Chapters 4 and 6.

Check your understanding

In this economy a one-off transfer of 10 trillion won is paid, and recipients’ MPC is 0.3. What is the first-round increase in consumption? 3 trillion won. If the households whose incomes then rise have an MPC of 0.8, the total effect is 3×5=153\times 5=15 trillion won. Applying Keynes’s 0.8 from the start, 10×0.8×5=4010\times 0.8\times 5=40 trillion won, overstates it by more than twice.

References

  • N. Gregory Mankiw, Macroeconomics, ch. 16
  • Milton Friedman, A Theory of the Consumption Function (1957)
  • Robert Hall, “Stochastic Implications of the Life Cycle–Permanent Income Hypothesis,” Journal of Political Economy (1978)
  • Jonathan Parker et al., “Consumer Spending and the Economic Stimulus Payments of 2008,” American Economic Review (2013)
O

Oiyo

Editorial Desk

The OIYO editorial desk researches money, law, lifestyle, and self-understanding topics against primary sources and public statistics. Every piece carries source notes and is reviewed on a regular cycle for accuracy and usefulness.