This page shows the difference in costs over the six years outcomes were monitored (from 2005/06) – following application for, or the granting of, social housing – for the two comparison groups: those in social housing and those not.
Charts 1 and 2 compare the average difference in total household costs between the two comparison groups over the six years.
The comparisons average the difference in total costs towards households who were placed in a social house, compared to households who were not.
The average costs are represented by the circles: lower costs; higher costs.
Where the confidence interval crosses the zero axis, it can't confidently be stated whether the actual change in costs is either positive or negative. These are represented by the dashed circles and grey bars .
The confidence intervals were estimated using bootstrapping methods involving random selection of samples. Small variations can occur and the sum of items by spending type per agency may not add up exactly to the total reported in Chart 1.
A negative cost indicates, on average, a household in social housing costs less over the six years, in comparison to a similar household not in social housing. The only exception is the negative value reported for tax revenue - this means social house tenants have, on average, paid less tax from their wages and salary (W&S).
See the full Technical Report for information on the methodology used and for the social housing context.
Compares the average difference in costs (by agency), i.e., the average saving or extra cost per household receiving social housing – compared to the control group.
* The negative value reported for tax revenue means social house tenants have, on average, paid less taxes from their W&S.
Key:
Lower costs
Higher costs
Lower confidence
Details the differences in costs (by the various spending types measured).
* The negative value reported for tax revenue means social house tenants have, on average, paid less taxes from their W&S.
** Due to data availability issues, the two comparison groups have not been able to be balanced on prior PRIMHD and Pharmaceutical costs.
Key:
Lower costs
Higher costs
Lower confidence
Changes were only able to be assessed over six years, which is not long enough to see some of the benefits of this investment.
For example, education costs for enrolments increased for households in social housing. However, the potential effect of increased education on future earnings and other social outcomes weren’t able to be measured.
While data on these various costs will provide a better indicator of return, it’s important to note cost is still only a proxy for outcomes and there are confounding effects.
For example, people in social housing may be more likely to access social services, which would increase the cost to government though improving social outcomes.
This table contains links to descriptions and guidance about the data source for each spending type used for the analysis.
The data used for the chart above is available for download as a CSV file:
cost-difference-per-agency.csv
The file contains these fields:
| Column name | Description |
|---|---|
| subject_area | Abbreviation of the spending type, which is classed as a "subject area" in the dataset. The "subject area" describes individuals' interactions with government, which are recorded by agencies and are available in the IDI and Social Investment Analytical Layer (SIAL). |
| subject_area_fullname | Full name of the spending type or "subject area" - see above. |
| agency | Abbreviation of the agency's name. The subject areas are grouped by the agency primarily responsible for funding the interaction. |
| agency_displayname | Name of the agency used in the chart. |
| meanavgcost | The average difference in costs over the six years, between households that were provided social housing in comparison to households that weren't. |
| lowCI | The lower limit for the 95% confidence interval for the average difference in costs (see "meanavgcost" above), calculated using the bootstrap confidence interval technique. |
| highCI | The upper limit for the 95% confidence interval for the average difference in costs (see "meanavgcost" above), calculated using the bootstrap confidence interval technique. |
For further information about these fields, including the methodology for the calculations, please refer to the Social Housing Technical Report.