Working with NHS and Taxfiler data to measure income and poverty in Toronto neighbourhoods

Transcription

1 Working with NHS and Taxfiler data to measure income and poverty in Toronto neighbourhoods Wayne Chu Planning Analyst Social Development, Finance & Administration, City of Toronto CCSD Community Data Canada Roundtable, June 18, 2014

2 Overview 1. How has the City of Toronto used neighbourhood income data in the past? 2. The new reality: NHS and Taxfiler income data 3. How is the City of Toronto moving forward with the new data sources?

3 Explore City of Toronto demographic data at toronto.ca/wellbeing

4

5 Past Examples of Income Data Use 1999 Recreation Priority Centres Designed to increase access to recreation in neighbourhoods with high incidence of low income by eliminating program fees. Areas identified using LICO from the long-form Census Strong Neighbourhoods Priority Areas Strong Neighbourhoods Task Force identified 13 Priority Areas for targeted investment and program coordination based on poor service coverage and high need, based partly on LICO Child Care Fee Subsidy Children s Services provides fee subsidies to families where low income is a significant barriers. Funds are allocated by Ward based on the concentration of families under LICO.

6 The New Reality: NHS and Taxfiler National Household Survey LICO is no longer offered as a standard product LIM-AT does not consider relationship between income and necessities Voluntary and therefore not comparable to previous Censuses T1FF (Taxfiler) Many users unfamiliar with this dataset Data cannot be cross-tabbed with other sociodemographic variables beyond age, sex and family structure Taxfiler families are derived and therefore may double count some individuals (e.g. children in lone parent families)

7 DEFINITION Low-Income Cut-off (LICO) LICOs identify those who are substantially worse off than the average income thresholds are determined by analysing expenditure data threshold = HH spending 20% or more of their income on necessities than the average HH HH that devote a larger share of gross income to basic necessities than the average, would fall into the category of straightened circumstance as the name implies, it is a low income cut-off not a poverty line

8 Low-Income Measure (LIM) DEFINITION those living in families that have an after-tax income lower than 50% of the median income for all families in a given year Personal income is adjusted for household size. as the name implies, it is a low income measure not a poverty line

9 Key Evaluation Question How do the two data sources compare when it comes to estimating absolute and relative income levels in Toronto? Absolute: Relative: The actual estimate of income or poverty in a local area such as a Census Tract. How does a local area differ from others in the same region? (e.g. spatial distribution)

10

11

12

13

14 Summary Issue NHS Taxfiler Absolute income Higher median income Lower median income Relative income Absolute low-income (LIM-AT) Relative low-income (LIM- AT) Similar spatial pattern Lower inequality Fewer Census Tracts with high LIM-AT (>=25%) Similar spatial pattern in the City of Toronto Fewer low-income areas in surrounding municipalities Similar spatial pattern Higher inequality More Census Tracts with high LIM-AT (>=25%) Similar spatial pattern in the City of Toronto More low-income areas in surrounding municipalities Between the two, NHS tends to tell a more positive story.

15 Takeaways Use Taxfiler data if cross-time comparisons of income data are needed. NHS data should not be compared to previous Censuses. NHS may not be reliable for establishing hard income thresholds for program delivery. NHS is more useful for regional and spatial comparison in cases when the absolute figures are less important. Consider using broader indices of socioeconomic vulnerability to smooth out errors (e.g. Neighbourhood Financial Health Index) Ultimately, data limitations depend on your research question (What is your acceptable margin of error. Always provide users with caveats or guidance on how to interpret data.

16 DRAFT City of Toronto Caveat There have been changes in the way information has been collected for portions of the 2011 Census. This will impact the extent to which comparisons can be made with previous Census periods on some Census variables. In general, 2011 Census data on population, dwelling counts, age, sex, families, household living arrangements, marital status, structural types of dwellings types and language can be compared to the data from previous Censuses, with due regard for changing definitions of individual variables. Information on Aboriginal peoples, immigration, ethnocultural diversity, education, labour, income and housing was previously collected by the longform Census questionnaire. In 2011 this information was collected as part of the new voluntary National Household Survey (NHS) by Statistics Canada. In general, the 2011 NHS data is less comparable to that of the previous Censuses due to non-response bias inherent in voluntary surveys. The risk of a voluntary survey is that the results may only reflect the kinds of individuals who are inclined to participate in a survey in the first place. As the National Household Survey User Guide notes, "because non-respondents tend to have different characteristics from respondents. As a result, there is a risk that the results will not be representative of the actual population." Comparisons between the 2011 NHS and previous Censuses should not be considered reliable. Users should also use caution when considering data at a local level of geography, such as Census Tracts and Dissemination Areas. The smaller the geographic area or population group, the greater the risk of non-response and non-comparable data. The National Household Survey User Guide states: "Caution must be exercised when NHS estimates are compared with estimates produced from the 2006 Census long form, especially when the analysis involves small geographies. Users are asked to use the NHS's main quality indicator, the global non-response rate in assessing the quality of the NHS estimates and determining the extent to which the estimates can be compared with the estimates from the 2006 Census long form. Users are also asked to read any quality notes that may be included in dissemination products." For more information, please see the National Household Survey User Guide Catalogue No x released May 8, The risk of non-comparable data for population groups grows when two variables are cross-tabulated or tabulated for small geographic areas, such as variables relating to minority populations, income and shelter costs. Seventy-four percent of the Census Tracts in the City of Toronto reported by the 2011 NHS have a Global Non-Response Rate above 25%, the threshold for suppression for the 2011 Census. Comparisons may be more suitable where the Global Non-Response Rates of the geographies are similar and the Coefficients of Variation of the variables are low. Use of NHS data involves a careful understanding of the significant differences and higher non-response inherent in the NHS by comparison to the Census.

17 How have we moved forward? Parks and Recreation Priority Centres has transitioned from LICO to National Household Survey LIM-AT with a more inclusive threshold for free service (30% LICO vs. 28% LIM-AT) and have added new Priority Centres. Toronto Strong Neighbourhoods Strategy 2020 incorporates Taxfiler LIM-AT as part of a larger 15 indicator index of socioeconomic inequity. Children s Services likely to use NHS LIM-AT as a replacement for Census LICO in ensuring geographical equity.

18 City of Toronto Parks, Forestry & Recreation Community Centres where Programs are Free Source:

19

20 Further Resources Wellbeing Toronto City of Toronto Demographics Portal Community Data Program FCM Quality of Life Reporting System Statistics Canada National Household Survey

21 For more information contact: Wayne Chu Planning Analyst City of Toronto Social Development, Finance and Administration Social Policy, Analysis and Research Telephone: Fax: