Permanent Income Hypothesis
Table of Contents
The permanent income hypothesis explains why a household might save most of a one-time bonus yet increase its regular spending after a lasting improvement in earnings. The relevant question is how much the household can sustainably consume across time, given its resources and expectations. Understanding that distinction helps connect today’s budget with future needs, while showing why uncertainty and limited access to credit can make smooth spending difficult.
This guide develops the idea with checkable hypothetical household budgets. For the broader process of setting goals and organizing your finances, see personal financial planning basics.
What Is the Permanent Income Hypothesis?
Milton Friedman’s permanent income hypothesis links consumption to the resources people expect to sustain, rather than treating each period’s measured income as the sole determinant of spending. Permanent income reflects expected earning capacity and wealth; transitory income captures a temporary departure from that underlying position. These are economic concepts inferred from expectations and behavior, not separate lines on a paycheck.[1]
Consumption smoothing means using saving, asset drawdowns and, where available and affordable, borrowing to keep consumption from moving one-for-one with temporary income fluctuations. It spreads the use of resources across time; it does not create additional resources.
Here, consumption means the goods and services a household uses. Moving cash into an investment account is saving, while repaying loan principal reduces a liability. Neither is itself current consumption. Spending on a durable item, such as a car, also differs from the stream of transportation services it provides over several years. That distinction matters when interpreting empirical studies of spending.[3]
Permanent Does Not Mean Certain
A temporary overtime payment may leave expected future earnings largely unchanged. A promotion that raises expected after-tax pay for many years can change the resources available for regular spending. But a promotion can be reversed, and a job loss that first looks temporary can become persistent. The classification depends on what the household learns about its future prospects.
Permanent income is therefore neither guaranteed income nor simply the arithmetic average of all paychecks eventually received. Friedman’s original discussion explicitly distinguishes expectations from realized outcomes and leaves room for those expectations to change.[1]
The Household Resource Budget
A resource budget makes the smoothing idea concrete. The following deterministic framework assumes a known income path and one constant real interest rate. It describes what a plan can afford before asking which consumption path the household prefers.
Use these timing and measurement conventions:
- Time: the plan covers T years; year-t income arrives at year-end immediately before that year’s consumption is paid.
- Initial assets, A0: net financial wealth available at the start of the plan.
- Income, Yt: after-tax income received during year t, excluding interest already counted on assets.
- Consumption, Ct: the year’s consumption, booked at the same year-end date.
- End wealth, At: assets minus debt after year-t consumption. Negative wealth means borrowing.
- Real rate, r: the annual return after inflation, assumed identical for saving and borrowing. All dollar amounts have constant purchasing power.
The target AT is the net wealth that must remain after the final year’s consumption. Borrowing limits apply separately in every year; a lifetime total cannot override them.
Carrying this relationship through all T years gives the present-value budget. Each sum below runs from t = 1 through T, with all future amounts discounted to the start of the plan.
This uses the same discounting principle as the time value of money. With a positive rate, dollars arriving at different dates cannot simply be added as though they were available together. The sequential budget, borrowing constraints and final-wealth condition are central features of modern household-consumption models.[3]
When a Constant Consumption Benchmark Applies
Suppose the household values consumption equally in each year and receives diminishing additional satisfaction from extra consumption. With no uncertainty, a 0% real rate and sufficient access to funds, an equal spending path avoids concentrating consumption in high-income years. At that zero rate, the affordable constant annual amount is:
At a nonzero rate, a chosen constant consumption amount is instead the present-value resources divided by the present value of a $1 annual payment over T years. A constant path need not be the household’s preferred path when interest rates, impatience, household needs or uncertainty change. The budget is an accounting constraint, not the complete permanent income hypothesis.
Consumption Smoothing: A Three-Year Example
Consider a hypothetical household with a three-year planning horizon. It starts with $0 net wealth, receives after-tax income of $30,000, $60,000 and $30,000, and must finish year 3 with $0 net wealth. All amounts are constant-purchasing-power dollars. Saving and borrowing both carry a 0% real rate, with no fees or uncertainty; borrowing up to a $10,000 outstanding balance is available. Income arrives immediately before each year-end consumption payment, and the household values consumption equally across years.
The household’s income varies sharply, but its full three-year resources can support a level consumption plan. Compare the two annual flows:
| Time (years) | Income (USD/year) | Consumption (USD/year) |
|---|---|---|
| 1 | 30000 | 40000 |
| 2 | 60000 | 40000 |
| 3 | 30000 | 40000 |
Hypothetical; 0% real saving and borrowing rates, $0 initial and final wealth, and up to $10,000 debt allowed. The reference marks zero annual flow.
Constant annual consumption is ($30,000 + $60,000 + $30,000) / 3 = $40,000. The flat spending line is financed by changing net wealth, rather than by making each year’s spending equal that year’s pay.
| Year | Income minus consumption | End net wealth |
|---|---|---|
| 1 | −$10,000 | −$10,000 |
| 2 | $20,000 | $10,000 |
| 3 | −$10,000 | $0 |
In year 2, the $20,000 excess of income over consumption first clears the $10,000 debt, leaving $10,000 in assets: −$10,000 + $20,000 = $10,000. The final year’s $10,000 drawdown then brings net wealth back to zero. Every dollar consumed is accounted for.
This example also reveals the financing assumption. Without initial assets or access to that first-year loan, the household cannot spend $40,000 in year 1, even though its three-year income is sufficient in total. The assumed loan is a modeling device, not advice to borrow against an optimistic earnings forecast.
Temporary vs. Persistent Income Changes
Reset to a different hypothetical household with five years remaining, $0 initial and required final net wealth, and baseline after-tax income of $50,000 every year. Retain constant-purchasing-power dollars, a 0% real rate, no fees or uncertainty, equal valuation of consumption across years, and year-end income immediately before spending. None of the positive-gain plans below requires borrowing.
Before any year-1 spending is chosen, the household learns about one of two changes: a $6,000 payment only in year 1, or a $6,000 increase in each of all five years. Both changes raise first-year income by the same amount. The second adds $30,000 of total resources, compared with $6,000 for the first.
| Income scenario | Annual consumption increase (USD/year) |
|---|---|
| One-time gain | 1200 |
| Gain each year | 6000 |
Hypothetical five-year plans; same $6,000 first-year income gain and 0% real rate. Bars show the increase above the $50,000 annual baseline consumption.
The one-time payment supports an extra $6,000 / 5 = $1,200 per year, taking annual consumption to $51,200. Receiving $6,000 in each year supports an extra (5 × $6,000) / 5 = $6,000 per year, taking consumption to $56,000. The larger response comes from the larger increase in resources across the same horizon.
With the one-time gain, year-1 saving is $56,000 − $51,200 = $4,800. Drawing $1,200 in each of the four later years uses that balance exactly. Most of the receipt is initially saved, but all of it eventually finances consumption within the stated plan.
What This Says About the Marginal Propensity to Consume
The marginal propensity to consume, or MPC, describes the consumption response to additional income. Specify both the shock and the response period. In this example, the first-year response to the one-time payment is $1,200 / $6,000 = 20%. Over all five years, the household consumes the entire extra $6,000.
That 20% is a result of this five-year, zero-rate, equal-spending example. It is not an empirical estimate for all households. A different remaining horizon, end-wealth target, financing constraint or preference can change the response. A payment anticipated well in advance can also affect plans before the cash arrives; new information and receipt timing are separate events.
Permanent Income vs. Life-Cycle Planning
The permanent income and life-cycle approaches share a focus on resources across time. Their emphasis differs: the former distinguishes underlying income from temporary deviations, while the latter highlights a finite life, changing needs and the transition from earning to retirement.[2]
| Framework | Central question | Main caution |
|---|---|---|
| Simple current-income spending rule | How much of today’s pay is spent today? | A fixed share of current income can miss changing wealth and future prospects. |
| Permanent income hypothesis | How does income news change the resources expected to sustain consumption? | Permanent income is estimated and expectations can be wrong. |
| Life-cycle planning | How are resources allocated over working years, retirement and other life stages? | Longevity, needs, income and desired bequests are uncertain. |
The broader foundation is consumer choice: a budget determines what is affordable, and preferences help determine what is chosen. Choosing consumption across dates extends that logic through time.
A Stripped-Down Retirement Illustration
Consider a separate household starting with $0 net wealth and planning for exactly 40 years: 30 working years at $60,000 annual after-tax income, then 10 retirement years with no income. Assume year-end flows, 0% real interest, no fees or uncertainty, equal valuation of consumption across years and $0 required ending wealth. No borrowing is needed.
Level annual consumption is (30 × $60,000) / 40 = $45,000. The household saves $60,000 − $45,000 = $15,000 in each working year, accumulating 30 × $15,000 = $450,000. Retirement consumption then uses 10 × $45,000 = $450,000.
This illustrates why a planned income decline at retirement need not require an equally abrupt consumption decline: earlier saving can finance later spending. It is not a safe withdrawal-rate calculation. Real plans must consider uncertain longevity, benefits or pensions, changing needs, returns and a suitable ending reserve. For the separate question of choosing portfolio investments through life stages, see investor lifecycle portfolio planning.
Borrowing Constraints and Uncertainty
A Lifetime Budget Does Not Pay Today’s Bill
Return to the five-year household with baseline annual income of $50,000, $0 initial and final net wealth, year-end flows and a 0% real rate. Now suppose it learns before choosing year-1 consumption about a one-time $6,000 income loss. Income is $44,000 in year 1 and $50,000 in each later year. Other assumptions remain unchanged.
With adequate borrowing access, level annual consumption would be ($44,000 + 4 × $50,000) / 5 = $48,800. That requires borrowing $48,800 − $44,000 = $4,800 in year 1 and repaying $1,200 in each of the four later years.
Impose a no-borrowing constraint instead. With no initial assets, year-1 consumption cannot exceed $44,000. One feasible plan is $44,000 in year 1 and $50,000 in each later year. It uses the same $244,000 of total resources but cannot achieve the frictionless level path. A household may therefore respond strongly to a temporary income loss without ignoring its future.
Expected earnings are not a bank balance or a guaranteed credit line. Loan availability, interest, fees and repayment obligations can make a mathematically smooth plan unavailable or unsuitable. Borrowing shifts spending across dates and adds costs when rates are positive; it cannot replace income that is permanently lost.
Accessible savings can help bridge a temporary disruption without relying on new credit. The appropriate reserve depends on the household’s obligations and risks. The CFPB emphasizes both accessibility and household-specific savings needs, rather than one universal amount.[7] Our guide to savings accounts and financial services explains the vehicles used to hold cash reserves.
Why Uncertainty Can Increase Saving
The worked examples know future income in advance. Actual households must choose while future employment, expenses and returns remain uncertain. In models with a precautionary-saving motive, the risk of a severe shortfall can encourage a buffer of assets even when average expected income is unchanged. Buffer-stock models describe this behavior under specified preferences and income-risk conditions.[4]
Consequently, replacing uncertain income with its average and dividing by years is not a complete stochastic consumption model. Households also need a rule for adjusting spending when outcomes differ from expectations. Risk, access to insurance and borrowing limits can influence both reserves and the response to new income.
Smoothing also does not require identical dollar consumption in every year. Household size, health, childcare and other needs change. A plan can aim for a stable standard of living while allowing spending to vary with those needs.
What Does the Evidence Show?
Evidence does not support using a single MPC for every household and every income change. Studies differ in the shock observed, households sampled, response horizon and measure of consumption. Two examples illustrate why those details matter.
| Study and setting | Finding | Interpretation limit |
|---|---|---|
| Agarwal, Liu and Souleles; 2001 U.S. tax rebates; paper revised 2007 | Credit-card data showed initial debt paydown followed by additional spending; likely constrained consumers had stronger spending responses.[5] | Transactions on sampled card accounts do not measure every household’s total consumption. |
| Dutch survey described in the ECB’s March 2018 Research Bulletin | For hypothetical unexpected changes equal to one month’s income, respondents reported an average nondurable-consumption response of 19.6% of a gain versus a 23.8% reduction for an equal-sized loss.[6] | These are stated responses in a Dutch survey, not observed U.S. spending or a rule for an individual household. |
These findings make liquidity, direction of the shock and measurement essential to interpretation. The Federal Reserve’s 2024 research review similarly emphasizes differences across research designs and estimates.[3] Saving a receipt initially does not imply it will never finance spending, and paying down debt does not mean the household consumed that amount immediately.
The household analysis here also leaves economy-wide income feedbacks outside the calculation. Connecting household spending responses to a change in GDP requires additional assumptions about investment, government, trade and policy. Those belong in the separate discussion of the fiscal multiplier effect.
How to Evaluate an Income Change
- Identify what has changed. Separate cash already received from new information about future after-tax income. Ask how persistent the change is likely to be and what evidence supports that expectation.
- Use a consistent horizon and purchasing-power basis. Include relevant future needs and income, and distinguish inflation-adjusted amounts from nominal dollars. A higher paycheck need not buy more goods if prices have also risen.
- Check resources and timing. List available assets, debts, required payments and desired ending reserves. Reconcile the annual or monthly cash path as well as the total. See budgeting and personal financial statements for the operational records.
- Allow for uncertainty and financing limits. Consider a delayed recovery, a smaller raise or an additional expense. A plan that requires credit should reflect actual terms and a credible repayment path.
- Review the plan when information changes. A windfall, income loss or change in household needs can justify a revised spending path. Avoid treating a favorable surprise as permanently spendable before assessing its implications.
Money’s source can influence behavior as well as expectations. Labeling a bonus as “free money” can encourage a different decision from receiving the same resources through wages. That is related to mental accounting bias, which asks how account labels affect choices; PIH asks how consumption relates to underlying resources.
Common Mistakes
- Treating temporary income as irrelevant. A one-time payment still adds resources. The benchmark spreads its consumption effect rather than making the effect disappear.
- Comparing inconsistent horizons. A one-year bonus and a five-year raise add different totals. State what is held equal: first-year income, lifetime resources or the payment’s present value.
- Calling every pay increase permanent. Higher gross pay can be offset by taxes, lost benefits, prices or a short-lived change in hours. Estimate the change in sustainable purchasing power.
- Forgetting how smoothing is financed. An average-income calculation can conceal a negative balance before future income arrives. Check the borrowing limit period by period.
- Turning one MPC into a universal constant. A first-month card-spending response and a multi-year consumption response answer different questions.
- Using the model as blanket borrowing advice. Equal spending is only one benchmark, and its assumed credit terms may bear little resemblance to actual household borrowing.
Limitations of the Permanent Income Framework
Expected resources are difficult to measure. Two households with equal current pay can have very different career prospects, wealth, obligations and access to support. Permanent income is not directly observable from a single year’s tax return.
Preferences and needs matter. The same budget can support multiple spending paths. Impatience, changing household circumstances and a desire to leave assets can all affect the choice.
Markets have frictions. Borrowing and saving rates differ, income risks cannot always be insured, and illiquid wealth may not be available to meet an immediate bill.
A known endpoint is a simplification. Ending wealth of zero makes the examples easy to audit. It is not a general retirement objective, particularly when longevity and later expenses are unknown.
Observed behavior has several explanations. A strong spending response can reflect limited liquidity, changing expectations, durable purchases or behavioral influences. One response pattern alone does not uniquely identify its cause.
Frequently Asked Questions
It is the idea that consumption depends on the resources a household expects to sustain, rather than only its current paycheck. A temporary income gain may therefore produce a smaller immediate spending change than a lasting improvement in expected earnings. The response still depends on assets, needs, uncertainty and access to funds.
No. A fixed percentage of current income makes spending rise and fall with current pay. Smoothing can involve saving more in a high-income year and drawing on assets in a low-income year. It does not require identical spending every year, especially when household needs change.
No. A bonus increases resources, so some extra consumption can be consistent with smoothing. In the article’s five-year zero-rate benchmark, the $6,000 payment supports $1,200 of extra consumption in each year. That allocation follows the example’s assumptions; it is not a prescribed savings percentage.
A household may lack accessible savings and be unable or unwilling to borrow on acceptable terms. Future income does not make current cash available. The household may also be uncertain whether the loss will remain temporary, leading it to revise its longer-term spending plan.
Both connect consumption to resources across time. Life-cycle analysis explicitly emphasizes a finite horizon and predictable transitions such as retirement, while permanent-income analysis emphasizes underlying income and temporary deviations. They can be combined, with uncertainty, borrowing constraints and changing needs added to a more realistic household model.
No. An empirical MPC describes a response measured for particular households, a particular income change and a particular period. It does not establish a suitable spending rule for your circumstances. A practical plan must also account for obligations, reserves, credit terms and the risk that expected income will not arrive.
Sources
- Milton Friedman, “The Permanent Income Hypothesis,” chapter III of A Theory of the Consumption Function (1957), especially pp. 21–26.
- Franco Modigliani, “Life Cycle, Individual Thrift, and the Wealth of Nations,” Nobel lecture (1985), especially pp. 153–154.
- Edmund Crawley and Alexandros Theloudis, “Income Shocks and Their Transmission into Consumption,” FEDS 2024-038 (2024), sections 1 and 3. Research authors’ views.
- Christopher D. Carroll, “Theoretical Foundations of Buffer Stock Saving” (July 30, 2021 version), introduction and sections 2.1 and 2.10.
- Sumit Agarwal, Chunlin Liu and Nicholas S. Souleles, “The Reaction of Consumer Spending and Debt to Tax Rebates—Evidence from Consumer Credit Data,” WP 2007-10 (revised December 2007). Research authors’ views.
- Dimitris Georgarakos, “Do consumers respond symmetrically to positive and negative income shocks?” ECB Research Bulletin No. 44 (March 19, 2018). Research author’s views.
- Consumer Financial Protection Bureau, “An essential guide to building an emergency fund” (last modified October 29, 2025).
Disclaimer
This article is for educational and informational purposes only and does not constitute financial, investment or borrowing advice. All worked household examples are hypothetical and use simplified assumptions. Actual income, expenses, returns and credit availability are uncertain. Consult a qualified financial professional about decisions involving your circumstances.
Keep reading
More in Economics
-
Supply and Demand: How Markets Determine Prices
Supply and demand is the foundational economic model explaining how prices are set, how markets reach equilibrium, and what causes surpluses and shortages.
-
Gross Domestic Product (GDP)
Gross domestic product (GDP) is the total market value of all final goods and services produced within a country.
-
Inflation & Consumer Price Index (CPI) Explained: How It Works & Why It Matters
Learn what inflation is, how the Consumer Price Index (CPI) measures it, and why it matters for investment decisions.
