# The Economics of Living Together

Identifier: GSX:HH-001
Series: Household Economics
Version: 1.0
Status: published
Schema: GreyResearchObject 1.2

## Primary question

How do early cohabitation, delayed cohabitation, and separate households compare in a Gauteng lifetime scenario?

## Summary

A Gauteng lifetime scenario comparing early cohabitation, delayed cohabitation and separate households.

The complete source manuscript is available as Markdown. Published claim objects are a claim-led extraction, not an exhaustive list of every statement in the paper. Read each claim with its assumptions and main limitation; the catalog summary is not a finding.

## Conditional answer

For two Gauteng adults bringing home R30,000 and R20,000 a month, combining households at 25 rather than 35 leaves about R3.972 million more combined household wealth at 75 in the average scenario, in constant 2026 rand. The direct living-cost saving is about R1.059 million. The larger gap comes from investing that earlier saving and from avoiding expensive negative balances.

- Evidence status: modelled
- Depends on: GSX:HH-001:A01, GSX:HH-001:A02, GSX:HH-001:A03, GSX:HH-001:A04
- Derived from: GSX:HH-001:R01, GSX:HH-001:R02
- Sensitivity: high
- Main limitation: This is a modelled scenario, not a forecast, and it is household wealth rather than each person's wealth. The manuscript keeps the relationship intact and invests the surplus. The advantage can disappear or reverse if the shared household costs almost as much as two homes, if the saving is consumed, or if separation costs exceed the accumulated benefit. Cohabitation here is a shared household, not a marriage result.

## Related questions

- What is the lifetime household comparison between living together and maintaining separate households?
- How does the timing of cohabitation change the cost of living and saving?

## Epistemic status

- Study type: Exploratory simulation
- Evidence status: Modelled
- Peer review: None
- External critique: None recorded
- AI assistance: GreyScienx manuscripts are prepared with substantial AI assistance in research, drafting, and computation. This record is not peer-reviewed. The published manuscript does not itemise AI participation by activity.
- Record date: 2026-09-28. Date of this machine record. It is not a separately catalogued manuscript issue date.

## Topics

- [cohabitation](https://greyscienx.com/topics/cohabitation)
- [lifetime-wealth](https://greyscienx.com/topics/lifetime-wealth)
- [housing](https://greyscienx.com/topics/housing)
- [south-africa](https://greyscienx.com/topics/south-africa)

## Geographies

- South Africa
- Gauteng

## Assumptions

- [A01](https://greyscienx.com/research/household-economics/economics-of-living-together/assumptions/A01) Two Gauteng earners. The household is two adults in Gauteng with take-home pay of R30,000 and R20,000 a month. The results are combined household wealth, not each person's legal ownership.
- [A02](https://greyscienx.com/research/household-economics/economics-of-living-together/assumptions/A02) Ten-year difference in household timing. The Early couple combines adequate households at 25. The Later couple combines them at 35. From 35 onward the model gives both the same housing arrangement, income, expenses, shocks, and retirement assumptions.
- [A03](https://greyscienx.com/research/household-economics/economics-of-living-together/assumptions/A03) Average-case returns and costs. In the average case, real salary growth and real living-cost growth are both 0.75% a year. The net real investment return is 3% a year after fees and tax drag. The initial monthly cohabitation dividend is R8,500.
- [A04](https://greyscienx.com/research/household-economics/economics-of-living-together/assumptions/A04) Surplus is invested. The model invests the available surplus. It does not spend that surplus on a higher lifestyle.
- [A05](https://greyscienx.com/research/household-economics/economics-of-living-together/assumptions/A05) The relationship remains intact. Both relationships remain intact through the simulation. Separation, custody, maintenance, and a second replacement home are not given a numeric cost.

## Model

- [M01](https://greyscienx.com/research/household-economics/economics-of-living-together/models/M01) Lifetime household-timing model. The model follows two household paths from age 25 to 75 in constant 2026 rand. They differ in when they combine households. After 35 their housing arrangement matches, so the remaining gap comes from returns on the earlier balance or from the cost of a negative balance. Best, average, and worst cases change growth, returns, shocks, children, and housing.

## Results

- [R01](https://greyscienx.com/research/household-economics/economics-of-living-together/results/R01) In the average case, the Early household holds R3.972 million more combined wealth at 75 than the Later household, in constant 2026 rand. (3972000 constant 2026 rand; Average case, age 75, early consolidation minus later consolidation).
- [R02](https://greyscienx.com/research/household-economics/economics-of-living-together/results/R02) In the average case, the Early household spends about R1.059 million less on living costs over the model. At 35, before the later years are financially identical, its lead is R1.188 million. (1059000 constant 2026 rand of living costs avoided; Average case, lifetime living costs and age-35 lead).
- [R03](https://greyscienx.com/research/household-economics/economics-of-living-together/results/R03) At 75 the Early household's lead is R11.287 million in the best case and R1.882 million in the worst case. In the worst case both positions are large funding gaps: about -R16.346 million and -R18.228 million. (11287000 constant 2026 rand, best-case lead at 75; Best case and worst case at age 75).

## Sources

- GSX:HH-001:PDF [The Economics of Living Together](https://greyscienx.com/research/household-economics/greyscienx-economics-of-living-together.pdf) — Published manuscript used for this semantic-twin extraction. External references remain represented in the manuscript unless separately registered.
- GSX:HH-001:SRC01 [Review of Income and Wealth study](https://onlinelibrary.wiley.com/doi/10.1111/roiw.12483) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC02 [NBER research on wage risk and spousal insurance](https://www.nber.org/papers/w28294) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC03 [World Bank poverty-measurement guidance](https://documents1.worldbank.org/curated/en/156931468138883186/pdf/2980000182131497813.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC04 [Journal of Urban Economics study](https://www.sciencedirect.com/science/article/pii/S0094119013000284) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC05 [Ecological Economics study](https://www.sciencedirect.com/science/article/pii/S092180091500213X) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC06 [Stats SA General Household Survey 2024](https://www.statssa.gov.za/publications/P0318/P03182024.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC07 [Statistics South Africa](https://www.statssa.gov.za/?p=17283) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC08 [PayProp Rental Index coverage](https://www.citizen.co.za/business/personal-finance/average-rent-in-gauteng-tops-r9k-how-do-other-provinces-measure-up/) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC09 [Stats SA Income and Expenditure Survey](https://www.statssa.gov.za/?p=17995) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC10 [UCT Retirement Fund investment objectives](https://uctrf.co.za/systems/uctrf/useruploads/UCTRF%20Summary%20of%20Investment%20returns%20for%202025-06%20SA.pdf) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.
- GSX:HH-001:SRC11 [South African Reserve Bank](https://www.sarb.co.za/en/home/what-we-do/monetary-policy) — The manuscript cites this resource. Inclusion does not independently validate it or assign it to a specific claim.

## Manuscript

- [The Economics of Living Together (PDF, 19 pages)](https://greyscienx.com/research/household-economics/greyscienx-economics-of-living-together.pdf)
- Datasets: https://greyscienx.com/research/household-economics/economics-of-living-together/data
- Reproducibility: https://greyscienx.com/research/household-economics/economics-of-living-together/reproduce

## Related studies

- [The First Three Years Advantage](https://greyscienx.com/research/household-economics/the-first-three-years-advantage) (GSX:HH-002)

## Machine-readable record

- JSON: https://greyscienx.com/api/v1/research/economics-of-living-together
- Full-text manuscript API: https://greyscienx.com/api/v1/research/economics-of-living-together/manuscript
- Canonical page: https://greyscienx.com/research/household-economics/economics-of-living-together
- Citation: GreyScienx. The Economics of Living Together. GSX:HH-001. Version 1.0.

## Claim-level record

### GSX:HH-001:C01

For two Gauteng adults bringing home R30,000 and R20,000 a month, combining households at 25 rather than 35 leaves about R3.972 million more combined household wealth at 75 in the average scenario, in constant 2026 rand. The direct living-cost saving is about R1.059 million. The larger gap comes from investing that earlier saving and from avoiding expensive negative balances.

- Status: conditional
- Evidence status: modelled
- Depends on: GSX:HH-001:A01, GSX:HH-001:A02, GSX:HH-001:A03, GSX:HH-001:A04
- Result: GSX:HH-001:R01
- Sensitivity: high
- Main limitation: This is a modelled scenario, not a forecast, and it is household wealth rather than each person's wealth. The manuscript keeps the relationship intact and invests the surplus. The advantage can disappear or reverse if the shared household costs almost as much as two homes, if the saving is consumed, or if separation costs exceed the accumulated benefit. Cohabitation here is a shared household, not a marriage result.
- Page: https://greyscienx.com/research/household-economics/economics-of-living-together/claims/C01

### GSX:HH-001:C02

Across the manuscript's three scenarios, earlier household combination still leaves the stronger position at 75. The lead is R11.287 million in the best case, R3.972 million in the average case, and R1.882 million in the worst case. In the worst case that lead only narrows two large funding gaps.

- Status: conditional
- Evidence status: modelled
- Depends on: GSX:HH-001:A01, GSX:HH-001:A02, GSX:HH-001:A04, GSX:HH-001:A05
- Result: GSX:HH-001:R01
- Sensitivity: high
- Main limitation: Negative amounts are unmet consumption if the household does not change course. They are not a statement that a bank would finance the shortfall. The worst case cannot repair stagnant earnings, repeated unemployment, weak returns, or inadequate retirement income.
- Page: https://greyscienx.com/research/household-economics/economics-of-living-together/claims/C02

5 assumptions, 3 results, and 1 model descriptions are included in the JSON record.

## Full manuscript

Source checksum: 9ac1edb15cae5cfb673bc18ecd7c32eff09620cd1d64aa6bd3ba91ed8b8f8d54

The text below is the complete Markdown source used to produce the manuscript, preserved separately from the semantic claim layer above.

---

# The Economics of Living Together

## From the cohabitation dividend to a hypothetical lifetime model

**GreyScienx research note — version 0.5**  
**Evidence status:** Armchair model. The people, budgets, and futures are hypothetical. The model is designed to expose mechanisms, not predict anyone's actual wealth.

# Part I: The economic foundation

The central economic effect is a **cohabitation dividend**: two people can maintain a given standard of living with fewer total resources when they share housing and household infrastructure. But the gain may be distributed unequally, and higher household welfare does not necessarily produce higher measured GDP.

## 1. Define the experiment

Start with the cleanest counterfactual:

> Two adults who would otherwise maintain separate one-person households consolidate into one two-adult household, while their incomes and employment initially remain unchanged.

This isolates consolidation from marriage, children, romantic satisfaction and legal status.

Separate living requires two sets of fixed or shareable expenses, such as housing, internet, appliances and basic utilities. It also includes partly shareable expenses, such as food and household services, and costs that remain personal regardless of the living arrangement.

The immediate financial dividend is simply the difference between the cost of maintaining the two separate homes and the cost of the shared home. The complete economic surplus is broader. It includes risk pooling, time savings and improved access to assets, then subtracts dependency, conflict, moving costs and the expected cost of separation.

This distinction separates **money saved** from the total economic value—or cost—of living together.

## 2. Microeconomic effects

### Lower cost of maintaining a standard of living

The largest savings should occur in goods that are substantially non-rival:

- One dwelling instead of two
- Shared kitchen, bathroom and living space
- One internet connection
- Fewer appliances, furniture sets and household tools
- Shared security, cleaning and maintenance
- Lower electricity and water use per person
- Some food purchasing and cooking economies

Housing should generate the largest component because the second person usually costs much less than a second dwelling. Research consistently finds meaningful household consumption economies, although their magnitude varies by country, income and methodology. One collective-household study estimated that couple members effectively faced a 35% lower price for non-assignable goods than comparable singles—but this is evidence from the Netherlands, not a South African estimate. [Review of Income and Wealth study](https://onlinelibrary.wiley.com/doi/10.1111/roiw.12483)

### Greater saving and wealth accumulation

If income remains constant, the cohabitation dividend can be:

- Saved or invested
- Used to repay debt
- Applied toward a house deposit
- Used to acquire insurance
- Redirected toward education or business formation

As a generic illustration, separate from the R30,000/R20,000 lifetime case in Part II, suppose two people each earn R30,000 after tax:

| Monthly expenditure | Living separately | Cohabiting |
|---|---:|---:|
| Housing | R16,000 | R12,000 |
| Utilities and internet | R4,600 | R3,100 |
| Food | R8,000 | R7,000 |
| Transport | R6,000 | R6,000 |
| Private consumption | R10,000 | R10,000 |
| **Total** | **R44,600** | **R38,100** |

The hypothetical dividend is R6,500 per month. Combined saving rises from R15,400 to R21,900—a 42% increase—even though income has not changed.

This is illustrative, not yet an empirical South African estimate.

### Risk pooling

Two incomes provide partial insurance against:

- Unemployment
- Illness or disability
- Variable self-employment income
- Unexpected expenses
- Study leave or parental leave

One partner may temporarily support the other without immediately losing the dwelling or defaulting on debt. Economic models describe this as **spousal insurance**, supplemented by household saving and adjustments in the other partner’s labour supply. [NBER research on wage risk and spousal insurance](https://www.nber.org/papers/w28294)

However, correlated shocks—such as a regional recession—reduce this protection.

### Household production and time

Cohabitation can reduce the time required for cooking, shopping, cleaning and maintenance. Tasks may also be allocated according to ability or availability.

But specialization is not automatically a gain for both people. If one partner absorbs most unpaid work, sacrifices employment or loses career mobility, the household’s measured surplus can conceal an individual loss.

### Increased access to assets

Combined resources may allow the household to obtain:

- Better-quality housing
- Safer neighbourhoods
- More reliable transport
- Larger loans
- Insurance that neither individual could afford alone
- Durable assets with high upfront costs

Cohabitation can therefore relax liquidity and credit constraints, not merely lower monthly expenses.

### Bargaining, dependency and unequal welfare

The household is not a single person. A positive aggregate dividend does not prove that both partners benefit.

The division of the surplus depends on:

- Relative income and asset ownership
- Whose name appears on the lease or title
- Access to personal bank accounts
- Unpaid household labour
- The credibility of leaving
- Social and legal protections
- Control over shared expenditure

Household-level income can consequently classify both people as financially secure even when one lacks meaningful access to the resources. The World Bank warns that household-level measures can conceal substantial intrahousehold inequality. [World Bank poverty-measurement guidance](https://documents1.worldbank.org/curated/en/156931468138883186/pdf/2980000182131497813.pdf)

### Costs and possible diseconomies

Cohabitation can also create:

- Moving and furnishing costs
- Overcrowding and loss of privacy
- Longer commutes or reduced geographic mobility
- Financial exposure to a partner’s debts or behaviour
- Loss of means-tested benefits
- Relationship-specific investments
- Housing insecurity after separation
- Legal and administrative disputes
- Search costs if the relationship dissolves

The relevant outcome is therefore the **expected net dividend**, including the probability and cost of dissolution.

## 3. Macroeconomic effects

### Fewer households and lower housing demand

For an unchanged population, the number of housing units required falls as average household size rises. Population alone therefore does not determine housing demand; household formation does.

More consolidation means fewer independent households and therefore less demand for separate dwellings—especially studios and one-bedroom units. Demand may shift toward larger units rather than disappearing completely.

The reverse relationship is also important: unemployment and recessions can induce people to combine households, reducing household formation and housing demand. [Journal of Urban Economics study](https://www.sciencedirect.com/science/article/pii/S0094119013000284)

Possible consequences include:

- Lower rents or slower rent growth
- Higher vacancy rates in small units
- Reduced residential construction
- Conversion of small units to other uses
- Greater demand for two-bedroom and flexible housing
- Lower land and infrastructure requirements per person

### Less duplicated consumption

Industries supplying household fixed goods could face weaker demand:

- Appliances
- Furniture
- Internet subscriptions
- Household insurance
- Security services
- Basic utilities
- Small rental units

Other sectors might gain because disposable income is released:

- Travel and restaurants
- Financial services
- Education
- Childcare
- Larger homes
- Recreation
- Investment products

Cohabitation therefore changes the **composition** of consumption, not only its level.

### The welfare–GDP paradox

Suppose two households spend R16,000 maintaining separate dwellings but only R12,000 after consolidating.

The household has released R4,000 of resources and is materially better off. Yet if the R4,000 is saved or used to repay debt, measured current consumption falls and short-run GDP may also fall.

Thus:

> A society can become materially more efficient while producing less measured market activity.

Long-run GDP depends on what happens to the released income and resources. If additional saving finances productive investment, later output may rise. If construction and durable-goods production contract without an offsetting expansion elsewhere, output may remain lower even though household welfare is higher.

### Labour supply

The aggregate effect is ambiguous:

- Risk pooling may allow one partner to reduce working hours.
- Financial security may enable education, entrepreneurship or better job search.
- Greater dependency may reduce labour-force participation.
- Shared childcare and domestic work may increase participation.
- Relocation compromises may worsen job matching for one partner.

Consequently, we should not assume that cohabitation mechanically raises or lowers total labour supplied.

### Financial stability

At scale, consolidation could:

- Increase household saving
- Reduce consumer-credit dependence
- Lower arrears and defaults
- Improve mortgage-deposit accumulation
- Reduce demand for very small mortgages or rental units

But jointly leveraged households could also become more exposed to relationship breakdown or correlated employment shocks.

### Energy, infrastructure and emissions

Larger households generally use less energy and material per person because members share heating, lighting, appliances and buildings. One US study estimated that a representative person living with others had a carbon footprint approximately 23% below the same person living alone, although that figure should not be transferred directly to South Africa. [Ecological Economics study](https://www.sciencedirect.com/science/article/pii/S092180091500213X)

Possible public effects include:

- Lower residential electricity demand per capita
- Lower water and waste-service requirements per person
- Fewer buildings and appliances
- Reduced construction materials and embodied carbon
- More efficient use of urban land

### Government and municipal finance

Effects could include:

- Changes in benefit eligibility and household means-testing
- Lower demand for housing assistance
- Fewer municipal accounts and basic service charges
- Lower infrastructure costs per resident
- A narrower property-tax base if fewer dwellings are required
- Difficulty identifying financially vulnerable individuals hidden inside apparently solvent households

Policy that treats household income as equally available to every member may overstate individual security.

## 4. Why South Africa is an important case

Statistics South Africa estimated 63.2 million people and 19.6 million households in 2024—about 3.23 people per household. In 2002, the corresponding figures implied roughly 4.10 people per household. Household numbers therefore grew considerably faster than population over this period. [Stats SA General Household Survey 2024](https://www.statssa.gov.za/publications/P0318/P03182024.pdf)

In 2023, 26.5% of South African households consisted of only one person. That does **not** mean those households are all candidates for romantic cohabitation, but it shows that household fixed costs are being incurred separately across a large part of the housing stock. [Statistics South Africa](https://www.statssa.gov.za/?p=17283)

South Africa also combines:

- High unemployment
- Severe income inequality
- Expensive formal housing relative to many incomes
- Extended-family households
- Informal dwellings
- Unequal access to services
- Significant geographic separation between jobs and affordable housing

The effect will therefore differ substantially between two professionals combining apartments, two low-income workers sharing a room, and someone joining an already crowded multigenerational household.

## Initial thesis

Our provisional answer is:

> Household consolidation normally produces a private economic surplus through shared fixed costs, risk pooling, household production and improved access to assets. At the macro level it reduces the number of dwellings and duplicated goods required to support a given population, shifts consumption toward discretionary uses or saving, and lowers resource use per person. Its effect on GDP, employment and public revenue is ambiguous, while its effect on individual welfare depends critically on bargaining power, unpaid labour, crowding and separation risk.

The first major empirical question should be:

> How large is the South African cohabitation dividend, by income, location and housing arrangement?

Everything else in the research programme can build outward from that estimate. Part II now applies this foundation to the requested Gauteng lifetime thought experiment.

---

# Part II: The ten-year head start

## The question

Consider two otherwise identical couples in Gauteng. In each couple, one person brings home R30,000 a month and the other brings home R20,000. Both couples remain together from age 25 to age 75.

The only initial difference is where they live.

- **The Early couple** moves into one home at age 25.
- **The Later couple** maintains two separate homes from 25 to 34 and moves in together at 35.

What does that ten-year difference do over a lifetime?

The short answer is that it does more than save ten years of rent. The Early couple receives ten years of shared-cost savings and gives every saved rand more time to compound. In a difficult life, that head start can prevent debt from snowballing. In an average life, it can produce a multimillion-rand difference by 75. In a highly favourable life, it can become very large because a small early advantage compounds for four additional decades.

This does not mean that people should move in together for financial reasons. It means that household formation has a measurable economic price and benefit, even when love, compatibility, privacy, and personal freedom are deliberately left outside the calculation.

## Meet the two couples

Both couples begin at age 25 in 2026. One partner has R30,000 in monthly disposable income and the other R20,000, giving each couple R50,000 a month after tax and payroll deductions.

Treating the salaries as disposable income is a modelling choice. It keeps the exercise focused on household cash flow. If the amounts were gross salaries, tax and retirement-fund deductions would first have to be removed, and every result would be lower.

Both couples begin with no financial wealth or debt and intend to work until 65. The expanded model then gives them a recognisable lifetime rather than treating them as permanent childless renters:

- they have children at 36 and 39 and support each child through age 22;
- they provide recurring financial support to extended family;
- employer pension contributions accumulate while salaries are being earned;
- investment returns are reduced by fees and tax drag;
- they try to buy a Gauteng home from age 40, but may do so only if the deposit and costs leave a six-month emergency buffer;
- care work can reduce the R20,000 earner's paid income;
- unemployment, disability, emergencies, inheritance and death vary by scenario;
- retirement is funded from accumulated investments, pension assets and housing wealth rather than an assumed replacement salary.

Both relationships remain intact. In the worst case the higher earner dies at 63, after the couple has stayed together for 38 years, and the model follows the surviving partner to 75.

The Later couple is not single. They are in a relationship but maintain two homes for ten years. This matters because the model is comparing **household consolidation**, not the economic value of having a partner.

## The starting Gauteng budget

The average-case budget uses a rent of approximately R9,500 for each separate home and R13,000 for one adequate shared home. This is broadly anchored to PayProp's reported Gauteng average rent of R9,201 in the first quarter of 2025, with a small allowance for the later price base. The shared home is more expensive than either separate home, but much cheaper than two homes. [PayProp Rental Index coverage](https://www.citizen.co.za/business/personal-finance/average-rent-in-gauteng-tops-r9k-how-do-other-provinces-measure-up/)

| Monthly household cost at 25 | Two separate homes | One shared home | Amount released |
|---|---:|---:|---:|
| Rent | R19,000 | R13,000 | R6,000 |
| Utilities and internet | R4,400 | R2,900 | R1,500 |
| Food | R8,000 | R7,000 | R1,000 |
| Transport | R6,000 | R6,000 | R0 |
| Medical, personal, insurance and leisure | R7,400 | R7,400 | R0 |
| **Total** | **R44,800** | **R36,300** | **R8,500** |

At age 25, the Later couple can save R5,200 a month while maintaining two homes. The Early couple can save R13,700. Cohabitation therefore releases R8,500 a month without changing either salary.

The model is not describing the average Gauteng household. Statistics South Africa estimated Gauteng's annual average household consumption at R170,628 and median consumption at R96,933 in the 2022/23 Income and Expenditure Survey. Our hypothetical household spends much more and should be read as an urban, employed, middle-income renter scenario. [Stats SA Income and Expenditure Survey](https://www.statssa.gov.za/?p=17995)

## How the lifetime model works

The simulation runs monthly from the twenty-fifth birthday to the seventy-fifth. Every figure is expressed in **constant 2026 rand**. This means the results describe purchasing power rather than the much larger nominal rand amounts that inflation would produce by 2076.

Income and living costs can grow faster or slower than inflation depending on the scenario. Investments and pension assets earn a return after inflation, fees and estimated tax drag. A purchased home is included at estimated resale value after selling costs, less the outstanding mortgage.

When a household falls below zero, the negative value is treated as an unfunded balance exposed to a financing cost. After retirement, a negative result should be read as a cumulative funding shortfall—the amount the household would need to eliminate through downsizing, family or government support, more work, asset sales or lower consumption—not as a prediction that a lender would advance that amount.

Both households pay a moving cost when they consolidate. The Early couple pays at 25; the Later couple pays at 35. From age 35 onward they have the same housing arrangement, income, expenses, shocks, and retirement assumptions. Their later wealth gap is therefore the accumulated consequence of the first ten years.

The conceptual assumptions from Part I enter the lifetime model as follows:

| Economic idea | How it appears in the simulation |
|---|---|
| Shared fixed costs | Separate and shared budgets have different rent, utilities, internet and food costs. |
| Risk pooling | Employment and disability shocks remove one income while the shared household keeps one set of overheads. |
| Household production and unpaid work | Care responsibilities reduce the lower earner's salary in the average and worst cases. |
| Access to assets | A home can be purchased only when the deposit, transaction costs and a six-month buffer are affordable. |
| Saving and debt | Positive balances earn investment returns; funding deficits face financing costs. |
| Taxes and investment fees | Gross investment returns are reduced before being applied to wealth. |
| Children and extended family | Two children and recurring family support are explicit monthly cash flows. |
| Retirement | Employer pension contributions accumulate to 65, after which all assets fund living costs. |
| Illness, disability, inheritance and death | Their timing and financial impact differ across the three scenarios. |
| Bargaining and personal autonomy | Discussed separately because combined net worth cannot reveal who controls the assets. |
| Separation and relationship quality | Kept outside the numeric path because custody, maintenance, replacement housing, asset division, legal cost and personal welfare require a separate model rather than one invented deduction. |

## Three possible lives

The point of the scenarios is not to choose one future. It is to see whether the cohabitation advantage survives very different futures.

| Assumption | Best case | Average case | Worst case |
|---|---:|---:|---:|
| Monthly cost while separate at 25 | R44,800 | R44,800 | R47,000 |
| Monthly cost while cohabiting at 25 | R34,800 | R36,300 | R42,500 |
| Initial monthly cohabitation dividend | R10,000 | R8,500 | R4,500 |
| Real salary growth | 1.5% a year | 0.75% a year | 0% |
| Real living-cost growth | 1.0% a year | 0.75% a year | 0.5% a year |
| Gross real investment return | 5.5% a year | 4.5% a year | 2.5% a year |
| Fees and tax drag | 0.5% a year | 1.5% a year | 1.5% a year |
| Net real investment return | 5% a year | 3% a year | 1% a year |
| Cost applied to funding deficits | 7% a year | 8% a year | 3% a year |
| Moving cost | R20,000 | R30,000 | R50,000 |
| Employer pension contribution | 10% of current salary | 8% | 4% |
| Children | At 36 and 39; R3,500 each monthly | At 36 and 39; R5,000 each | At 36 and 39; R6,500 each |
| Extended-family support | R500 monthly | R1,500 monthly | R2,500 monthly |
| Care-related salary effect | None | Lower earner loses 20% from 36–42 | Lower earner loses 40% from 36–45 |
| Employment and disability | None | Six months unemployed at 30; 50% disability loss for a year at 50 | Unemployment at 28 and 43; 50% disability loss from 58 |
| Home purchase | Both qualify and buy at 40 | Both qualify and buy at 40 | Neither qualifies |
| Inheritance | R500,000 at 55 | R200,000 at 55 | None |
| Extraordinary event | None | R100,000 at 50 | R250,000 at 58; death and R100,000 funeral at 63 |
| Retirement living costs | 70% of pre-retirement costs | 80% | 90% |

The 5% best-case real investment return is ambitious but not invented from a nominal headline return: a South African retirement-fund balanced portfolio has used CPI plus 5% as a long-horizon objective. It should be read as a favourable outcome, not a guarantee. [UCT Retirement Fund investment objectives](https://uctrf.co.za/systems/uctrf/useruploads/UCTRF%20Summary%20of%20Investment%20returns%20for%202025-06%20SA.pdf)

South Africa's current inflation target is 3%, with a one-percentage-point tolerance band. Because the model is in constant rand, inflation itself is removed; only returns and growth above inflation appear in the assumptions. [South African Reserve Bank](https://www.sarb.co.za/en/home/what-we-do/monetary-policy)

![Lifetime net-worth trajectories for the Early and Later couples in all three scenarios](lifetime-wealth-trajectories-v0.5.png)

*Figure 1. Total net worth includes liquid investments, pension assets and net housing equity. Negative values are unfunded continuation of the modelled lifestyle.*

## Best case: the early saving becomes its own source of wealth

In the best case, income rises slightly faster than living costs, the couple shares expenses efficiently, neither partner experiences unemployment, and invested savings earn 5% a year after inflation, fees and tax drag. Both couples have two children, buy the same R1.8 million home at 40, support relatives, receive a R500,000 inheritance at 55 and accumulate employer-funded pensions.

| Age | Early couple | Later couple | Early couple's lead |
|---:|---:|---:|---:|
| 25 | R0 | R0 | R0 |
| 35 | R3.402 million | R1.818 million | R1.583 million |
| 45 | R9.240 million | R6.628 million | R2.612 million |
| 55 | R20.252 million | R15.998 million | R4.254 million |
| 65 | R41.047 million | R34.117 million | R6.929 million |
| 75 | **R61.382 million** | **R50.095 million** | **R11.287 million** |

The Early couple spends approximately R1.261 million less on living costs over the full model. Almost all of that direct saving is created between 25 and 35. Yet the final wealth difference is R11.287 million.

That does not mean cohabitation magically multiplies R1.261 million into R11.287 million. The Early couple invests each saving when it occurs. The earliest rand can grow for nearly fifty years, while the Later couple's equivalent rand was already spent on maintaining a second household. Once investment returns themselves begin earning returns, the original housing decision becomes less important than the asset base it created.

By retirement, both couples are financially comfortable in this scenario. The difference is not survival but optionality. The Early couple has more capacity to retire earlier, buy a better property, support relatives, absorb medical costs, or leave an inheritance.

Children consume approximately R2.1 million across the model and family support another R264,000. Employer pension contributions add roughly R3.3 million before returns. Investment fees and tax drag cost the Early couple an estimated R4.0 million relative to the gross-return path. The result is deliberately a best case: stable employment, favourable returns, uninterrupted care arrangements, an inheritance and restrained lifestyle growth. Its purpose is to show the upper mechanism, not a likely promise.

## Average case: a ten-year housing decision becomes a R4 million gap

The average scenario is less generous. Income and living costs both rise by 0.75% a year in real terms, so raises do not continually make life cheaper. Investments earn 3% after inflation, fees and tax drag. One partner loses six months of earnings at 30. At 50, the higher earner loses half of one year's income through disability and the household pays a R100,000 emergency. The lower earner also gives up 20% of salary between 36 and 42 because of care work.

The employment interruptions are deliberately placed on both sides of age 35. At 30, only the Early couple has consolidated. At 50, both couples have. This shows how shared fixed costs matter most when income is interrupted before the Later couple combines households.

| Age | Early couple | Later couple | Early couple's lead |
|---:|---:|---:|---:|
| 25 | R0 | R0 | R0 |
| 35 | R2.260 million | R1.072 million | R1.188 million |
| 45 | R4.138 million | R2.501 million | R1.636 million |
| 55 | R7.260 million | R5.060 million | R2.199 million |
| 65 | R13.743 million | R10.787 million | R2.956 million |
| 75 | **R13.833 million** | **R9.861 million** | **R3.972 million** |

The Early couple spends about R1.059 million less on living costs over the lifetime model. That difference compounds into R3.972 million of additional wealth at 75.

The important milestone is age 35. The Early couple has accumulated R2.260 million, including pension saving; the Later couple has R1.072 million after paying to move. From that point onward their lives are financially identical, but their starting positions are not. The extra R1.188 million keeps earning returns through every subsequent decade.

At 65, the Early couple enters retirement with almost R3 million more. Both couples have bought the same home at 40 and raised the same children, so the remaining gap still comes from the earlier household consolidation. The Early household finishes close to R13.8 million while the Later household declines toward R9.9 million as retirement spending is drawn from the smaller portfolio. This is a model result, not a withdrawal recommendation.

The average scenario produces the paper's clearest finding:

> The first ten years create roughly R1.1 million of direct living-cost savings, but the lifetime value of those early savings approaches R4 million in constant purchasing power.

Across the average life, the two children cost approximately R2.8 million, family support costs R747,000, employment and disability remove R342,000 of income, and care work removes another R320,000. Pension contributions total about R2.2 million before returns. Fees and tax drag reduce the Early couple's potential investment value by roughly R3.3 million.

## Worst case: cohabitation narrows the damage but cannot repair the retirement system

The worst case is intentionally severe. The couple remains together, but their economic environment is hostile and the higher earner dies at 63. The model then follows the surviving partner to 75.

Real salaries never grow. Living costs rise 0.5% a year above inflation. Sharing is inefficient, releasing only R4,500 a month at 25. Investments earn only 1% after inflation, fees and tax drag. Moving costs R50,000. The R20,000 earner loses twelve months of income at 28; the R30,000 earner loses nine months at 43; and disability removes half the lower earner's income from 58. Care responsibilities reduce that person's income by 40% between 36 and 45. A R250,000 emergency arrives at 58, followed by a R100,000 funeral at 63. Neither household qualifies to buy the modelled home.

| Age | Early couple | Later couple | Early couple's lead |
|---:|---:|---:|---:|
| 25 | R0 | R0 | R0 |
| 35 | R584,000 | R41,000 | R543,000 |
| 45 | **–R1.565 million** | **–R2.340 million** | R775,000 |
| 55 | **–R4.096 million** | **–R5.137 million** | R1.042 million |
| 65 | **–R8.159 million** | **–R9.559 million** | R1.400 million |
| 75 | **–R16.346 million** | **–R18.228 million** | **R1.882 million** |

Negative figures are funding gaps. At 35, both households remain above zero, although the Later couple has only R41,000 after maintaining two homes through the first employment interruption and paying to move. By 45, the combination of two children, remittances, care-related lost income and another unemployment shock pushes both paths below zero.

The Early couple spends only about R554,000 less on ordinary living costs across the model—far less than in the other scenarios because its shared household is not much cheaper. By 75, however, its shortfall is R1.882 million smaller. Avoiding early deficits matters because financing costs compound against the household rather than for it.

Neither couple reaches retirement on the modelled path without adjustment. In reality they could not finance an R16–18 million shortfall indefinitely. The graph shows unmet consumption demand if they refuse to change course. They would have to reduce spending much earlier, have fewer or less costly dependants, provide less family support, work longer, receive insurance or government support, or generate substantially more pension income.

This is an important counterweight to the optimistic story:

> Cohabitation can improve resilience, but it cannot compensate indefinitely for stagnant earnings, repeated unemployment, weak investment returns, and inadequate retirement income.

The two children cost roughly R3.5 million, family support R1.2 million, employment and disability R1.1 million, and care-related lost earnings R864,000. Employer pension contributions total only R852,000 before returns. The worst case shows why household consolidation is a buffer, not a substitute for income growth, insurance and a viable retirement plan.

## The comparison in one table

| Lifetime result in constant 2026 rand | Best case | Average case | Worst case |
|---|---:|---:|---:|
| Living costs avoided by moving in at 25 rather than 35 | R1.261m | R1.059m | R554k |
| Early couple's position at 65 | R41.047m | R13.743m | –R8.159m |
| Later couple's position at 65 | R34.117m | R10.787m | –R9.559m |
| Early couple's position at 75 | R61.382m | R13.833m | –R16.346m |
| Later couple's position at 75 | R50.095m | R9.861m | –R18.228m |
| **Lifetime advantage from the ten-year head start** | **R11.287m** | **R3.972m** | **R1.882m** |

The signs do not change. In every scenario, moving in earlier leaves the household in a stronger position at 75. What changes is what the advantage means.

- In the best case it creates substantially more optional wealth.
- In the average case it creates a meaningful retirement and inheritance difference.
- In the worst case it reduces an unfunded retirement gap but does not eliminate it.

![The Early couple's financial lead from age 25 to 75](ten-year-head-start-v0.5.png)

*Figure 2. From 35 onward, the direct housing arrangements are identical. The remaining gap changes through investment returns or deficit-financing costs.*

## Why the difference keeps growing after both couples move in

At 35, the Later couple has caught up in living arrangement but not in balance sheet.

From then on, both couples pay the same living costs. The Early couple is not receiving a new cohabitation dividend that the Later couple lacks. Its advantage grows because the earlier saving has become capital.

This creates two paths.

When balances are positive, the Early couple earns returns on a larger portfolio. The difference compounds in its favour. When balances are negative, the Later couple reaches debt or an unfunded shortfall sooner. Financing costs then compound against it.

The model therefore separates three effects that are often mixed together:

1. **The direct effect:** one shared household costs less than two separate households.
2. **The timing effect:** saving at 25 has longer to grow than saving at 35.
3. **The resilience effect:** a larger buffer reduces the probability and duration of expensive negative balances.

The direct saving is the smallest of the three in a long horizon.

## What the expanded model now includes

The earlier simple version treated both couples as permanent childless renters. This version includes two children, support for extended family, employer pensions, investment fees and tax drag, a care-related career penalty, unemployment and disability, emergencies, inheritance, a conditional home purchase and death in the worst case.

![Cumulative lifetime pressures included in the expanded model](expanded-model-lifetime-pressures-v0.5.png)

*Figure 3. The bars show the cumulative value of several newly modelled pressures. Fees and tax drag is an opportunity cost; the other bars are spending or income not received.*

These assumptions change the absolute outcome considerably. In the average case, adding children, care-related lost income, disability and family support reduces age-75 wealth even though both couples buy a home and receive pension contributions. The best-case wealth rises relative to the earlier renter-only version because pensions, housing equity and inheritance outweigh the new family costs. The worst case deteriorates sharply because dependants, remittances, weak earnings, disability and death arrive without enough income growth or investment return.

They do not remove the timing effect. Most added events occur after 35 and affect both couples equally. The Early couple enters those events with a larger buffer, so the lifetime gap survives.

## What still cannot be reduced to a credible rand value

The model still does not price love, companionship, relationship quality, privacy, personal safety or the value of learning whether a relationship is stable before sharing a home. It assumes both relationships remain intact, so it describes separation risk but does not impose an actual breakup. A breakup scenario would require assumptions about custody, maintenance, two replacement homes, asset division and legal costs; inserting one arbitrary number would give a false appearance of precision.

The simulation also assumes every available surplus is invested. Real households may spend some of it on lifestyle improvements. Taxes, benefits and investment treatment are compressed into disposable salary and an investment-return drag rather than modelled from individual tax returns. Children are represented as expenditure and care-related income loss, not as sources of welfare. Home values, mortgage terms, and inheritance remain hypothetical rather than property- or family-specific.

Most importantly, the model does not say that the Later couple made a mistake. Separate homes may purchase privacy, independence, shorter commutes, safety, or time to test compatibility. Those are real benefits even when they do not appear in a financial account. The R8,500 average monthly difference can be understood as the price of those benefits.

## The household and the individuals

The results describe combined household wealth. They do not prove that both people own half.

The R30,000 earner contributes 60% of income and the R20,000 earner 40%. They may split expenses equally, proportionally, or through a completely different arrangement. The lower earner could gain more from sharing fixed costs but become more dependent. The higher earner could contribute more cash while gaining more unpaid household labour.

A real version of the model would need two additional ledgers:

- assets and liabilities legally owned by each person;
- paid work, unpaid work, and leisure performed by each person.

Without those ledgers, “the couple has R16 million” may conceal one person with legal ownership and another with only informal access. The model therefore establishes a household-level opportunity, not an automatic distributional outcome.

## What this means outside the household

The model contains a small macroeconomic story.

For ten years, the Later couple supports two rental units, two sets of utility connections, and more duplicated household infrastructure. The Early couple supports one larger unit. If many similar couples moved in earlier, demand would shift away from small separate rentals toward shared homes. Furniture, appliances, internet subscriptions, and residential energy use per person would also tend to fall.

The released money would not disappear. In this model it is invested, which is why the wealth gap becomes large. Real households might instead spend it on restaurants, travel, education, children, better housing, or debt repayment. The wider economy would lose some demand for duplicated household goods and gain demand wherever the dividend is redirected.

Measured consumption can therefore fall while welfare rises. The Early couple needs fewer market transactions to maintain its household. GDP does not record “a second rent avoided” as a positive output, even though the household is better able to save and withstand shocks.

## The research claim

The model supports a narrower and more defensible conclusion than “cohabitation makes people rich.”

> For two Gauteng adults bringing home R30,000 and R20,000 a month, combining adequate households ten years earlier can create a persistent lifetime financial advantage. The direct saving is significant, but the larger effect comes from decades of investment compounding and avoiding expensive negative balances.

Under the assumptions used here, the advantage at 75 ranges from approximately R1.9 million in the worst case to R11.3 million in the best case, with an average scenario of about R4.0 million—all in constant 2026 purchasing power.

The result is conditional. It disappears or reverses if the shared household is almost as expensive as two homes, if the relationship creates large personal or career costs, if the saving is entirely consumed, or if separation costs exceed the accumulated benefit. Those are not footnotes. They are the next questions the research must investigate.

## Final observation

The largest economic decision in this model is not the R30,000 salary, the R20,000 salary, or even the monthly rent. It is time.

The Early couple does not merely save more. It gives the saving forty additional years to become something else: a deposit, an investment portfolio, an emergency buffer, or protection from debt. The Later couple buys ten years of independent living and gives up that compounding period in exchange.

That is the economic price of maintaining two households. Whether it is worth paying remains a personal decision.