Presentation is loading. Please wait.

Presentation is loading. Please wait.

Spatial microsimulation approaches to population forecasting Dimitris Ballas Department of Geography, University of Sheffield

Similar presentations


Presentation on theme: "Spatial microsimulation approaches to population forecasting Dimitris Ballas Department of Geography, University of Sheffield"— Presentation transcript:

1 Spatial microsimulation approaches to population forecasting Dimitris Ballas Department of Geography, University of Sheffield ESRC Research Methods Festival Oxford, 1-3 July 2008 RES

2 Outline What is microsimulation? What is spatial microsimulation? Dynamic spatial microsimulation Projecting small area statistics into the future Projecting small area microdata into the future Available software Concluding comments

3 What is microsimulation? A technique aiming at building large scale data sets Modelling at the microscale A means of modelling real life events by simulating the characteristics and actions of the individual units that make up the system where the events occur

4 What is microsimulation? PERSONAHIDPIDAAGE12SEXAJBSTAT…AHLLTAQFVOCATENUREAJLSEG… …1169… …207-8… …207-8… …217-8… …202-8… …212-8… …213-8… …2 3 … … 3 … …202-8… …202-8… … 2 …

5 Some examples of microsimulation applications in Economics PENSIM. This was a microsimulation model for the simulation of pensioners’ incomes up to the year Hancock et al. (1992) Sutherland and Piachaud (The Economic Journal, 2001) developed and used a microsimulation methodology for the assessment of British government policies for the reduction of child poverty in the period Results suggest that the number of children in poverty will be reduced by approximately one-third in the short term and that there is a trend towards further reductions

6 Microsimulation in Geography and Regional Science First study by Hägerstrand (1967) – spatial diffusion of innovation Foundations for spatial microsimulation of populations laid by Wilson and Pownall (1976): building small area microdata Clarke et al. (1979 onwards) extended the theoretical framework of Wilson and Pownall

7 Spatial microsimulation applications Static ‘What-if’ simulations –impacts of alternative policy scenarios on the population can be estimated if a factory closed what are the impacts on the local economy if we close a school where will the pupils be re- distributed “Static updating” –update a basic micro-dataset and future-oriented what-if simulations if local taxes are raised today what would the redistributive effects have been between different socio- economic groups and between areas of the city by 2007?

8 Examples of spatial microsimulation (1) Birkin and Clarke (1988 & 1989) SYNTHESIS model Williamson (1992) OLDCARE model Williamson (1996) and Williamson et al. (1998) first ever application of combinatorial optimisation for static microsimulation Holm et al. (1996), Vencatasawmy et al. (1999) – SVERIGE model (Spatial Modelling Centre – Sweden – the first comprehensive spatial microsimulation model in the world! (http://www.smc.kiruna.se/)

9 Examples of spatial microsimulation (2) Caldwell et al. (1996) CORSIM model Wegener and Spiekermann (1996) Urban models: land-use and travel Veldhuisen et al. (2000) RAMBLAS – daily activity patterns Ballas (2001), Ballas and Clarke (2000, 2001a & 2001b) SimLeeds model Ballas, Clarke and Commins (2001) SMILE – model of the Irish rural economy Ballas et al. (2005) SimBritain model Birkin and colleagues, MoSeS model

10 Spatial microsimulation procedures The construction of a micro-dataset from samples and surveys Static What-if simulations, in which the impacts of alternative policy scenarios on the population are estimated: for instance if there is a taxation policy change today, what would be the “morning after” effect? Which areas would be most affected? Dynamic modelling, to update a basic micro-dataset and future-oriented what-if simulations: for instance if the current government had raised income taxes this year what would the redistributive effects have been between different socio-economic groups and between central cities and their suburbs by 2011?

11 Reweighting approaches (1) PERSONAHIDPIDAAGE12SEXAJBSTAT…AHLLTAQFVOCATENUREAJLSEG… …1169… …207-8… …207-8… …217-8… …202-8… …212-8… …213-8… …2 3 … … 3 … …202-8… …202-8… … 2 …

12 Reweighting approaches (2) Small area table 1 (household type) Small area table 2 (economic activity of household head) Small area table 3 (tenure status) Area 1 60 "married couple households" 80 employed/self- employed 60 owner occupier 20 "Single-person households" 10 unemployed20 Local Authority or Housing association 20 "Other"10 other20 Rented privately Area 2 40 "married couple households" 60 employed/self- employed 60 owner occupier 20 "Single-person households" 20 unemployed20 Local Authority or Housing association 40 "Other"20 other20 Rented privately

13 Tenure and car ownership example Household car ownership characteristics Household tenure characteristics 1 car 2+ cars No car Owner- occupier LA/HA rented Other Simulation Census Absolute error

14 Deterministic Reweighting the BHPS - a simple example (1) A hypothetical sample of individuals (list format) In tabular format: Hypothetical Census data for a small area:

15 Reweighting the BHPS - a simple example (2) Calculating a new weight, so that the sample will fit into the Census table In tabular format: Hypothetical Census data for a small area:

16 Probabilistic synthetic reconstruction After Birkin, M., Clarke, M. (1988), SYNTHESIS – a synthetic spatial information system for urban and regional analysis: methods and examples, Environment and Planning A, 20,

17 Dynamic spatial microsimulation Probabilistic dynamic models, which use event probabilities to project each individual in the simulated database into the future (e.g. using event conditional probabilities). Implicitly dynamic models, which use independent small area projections and then apply the static simulation methodologies to create small area microdata statically

18 Probabilistic dynamic models after Ballas D, Clarke, G P, Wiemers, E, (2005) Building a dynamic spatial microsimulation model for Ireland, Population, Space and Place, 11, 157–172 (http://dx.doi.org/ /psp.359)http://dx.doi.org/ /psp.359

19 Event modelling  Demographic transitions 1.Age all individuals 2.Change marital status (marriage & divorce rates: trends & assumptions) 3.Birth (fertility rate: trends & assumptions) 4.Death 5.(use 5-year survival rates  deaths/pop at risk) 6.Migration  Socio-economic transitions Education (Enter school, university, etc.) Labour market (become employed/unemployed etc.)

20 Simulating migration, education and social mobility “It is well known that mobility rates are substantially higher among renters than among homeowners. Similarly, the age structure of migrants to and from neighborhoods is likely to be quite different in a neighborhood comprised primarily of homeowners in comparison with a renter-dominated neighborhood.” (Rogerson and Plane, 1998: 1468) “During their lifetimes, the simulated individuals have to change their educational and employment status. They will enter school with different probabilities when they are between 14 and 20 years old, they will be employed in different jobs, lose their jobs, earn an income which depends on their type of job, and eventually retire with different probabilities depending on their ages.” (Gilbert and Troitzch, 1998)

21 Determining inter- dependencies …while a woman’s labour force status can depend on the number of children she has and on her marital status, it cannot also influence the probability of the woman having a child in any year. The ordering of the modules necessarily involves making assumptions about the direction of causality in relationships between variables. (Falkingham and Lessof, 1992: 9)

22 The SimBritain model Funded by: –Joseph Rowntree Foundation –BT –Welsh Assembly Government Aimed at creating small area microdata for the years 1991, 2001, 2011 and 2021 (at electoral ward and parliamentary consistency level) for the whole of Britain by combining the Census small area statistics and the British Household Panel Survey Extrapolate constraint values and re-populate each area anew at the- yearly intervals using the original samples Simulate this population for the years 2001, 2011, 2021 (“groundhog day” scenario) What-if policy analysis

23 SimBritain: combining Census data with the BHPS Census of UK population: 100% coverage fine geographical detail Small area data available only in tabular format with limited variables to preserve confidentiality cross-sectional British Household Panel Survey: sample size: more than 5,000 households Annual surveys (waves) since 1991 Coarse geography Household attrition Ballas, D., Clarke, G.P., Dorling, D., Eyre, H. and Rossiter, D., Thomas, B (2005) SimBritain: a spatial microsimulation approach to population dynamics, Population, Space and Place 11, 13–34 (http://dx.doi.org/ /psp.351)http://dx.doi.org/ /psp.351

24 How do we make SimBritain dynamic? Original strategy: model the ageing death and creation of households (from the panel nature of the BHPS) and the geographic movement of households (using migration data from the Census and other sources). This was abandoned when migration data proved to be of insufficient quality. Intermediate strategy: extrapolate constraint values and re-populate each area anew at the-yearly intervals using the original samples Future strategy: create synthetic household histories from the panel data. Methods are also being developed to allow for inflation of values over time (e.g. income, pc ownership etc) and for changing geographical composition (via projected constraint values)

25 Projecting small area statistics into the future (1) where u, v and w are the smoothed proportions in 1971, 81 and 91 respectively, W is the observed ward proportion in 1991 and A is the projected ward proportion in 2001.

26 Projecting small area statistics into the future (2) where Lt and bt are respectively (exponentially smoothed) estimates of the level and linear trend of the series at time t, whilst Ft+m is the linear forecast from t onwards

27 Projecting small area statistics into the future (3) where W = ward proportion w = smoothed ward proportion t = census year

28 SimBritain: spatial distribution of “poor” households, 1991

29 SimBritain: spatial distribution of “poor” households, 2001

30 Spatial distribution of “poor” households, 2011

31 Spatial distribution of “poor” households, 2021

32 SimBritain: spatial distribution of “retired” households, 1991

33 SimBritain: spatial distribution of “retired” households, 2001

34 SimBritain: spatial distribution of “retired” households, 2011

35 SimBritain: spatial distribution of “retired” households, 2021

36 Census data Year Predicted proportion for 1991 Difference between projection and actual data Class I & II19%21%24%28%34% 0% Class III51%50%49%47%43%44%1% ClassIV & V30%29%27%25%24%22%-2% How do we know it makes sense? Comparing Census data to projected data for 1991 (projection based on data from the Censuses of 1961, 1971 and 1981)

37

38 “Projecting” small area microdata into the future 1.Establish a set of constraints 2.Choose a spatially defined source population 3.Repeatedly sample from source 4.Adjust weightings to match first constraint 5.Adjust weightings to match second constraint 6.… 7.Adjust weightings to match final constraint 8.Go back to step 4 and repeat loop until results converge 9.Save weightings which define membership of SimBritain

39 CONSTRAINT TABLES

40 How do we know it makes sense?

41

42 The potential of dynamic spatial microsimulation for policy analysis Classifying households Very poor: all households with income below 50% of the median York income Poor: all households with income more than 50% of the median but lower than 75% of the median Below-average: all households living on incomes higher than 75% of the median but less than or equal to the median Above-average: all households living on incomes higher than the median and lower than 125% of the median Affluent: all households living on incomes above 125% of the median Ballas, D., Clarke, G P, Dorling D, Rossiter, D. (2007), Using SimBritain to Model the Geographical Impact of National Government Policies, Geographical Analysis 39, pp (doi: /j x)doi: /j x

43 SimBritain results in York

44 SimBritain results, York: children in households

45 Very poor households Households (% of all households in York)17.2%17.3%17.8%21.3% Individuals (% of all individuals in York)14.7%13.3%13.7%20.5% Children (% of all children in York)21.8%17.7%18.6%38.5% LLTI (as a % of all individuals in group)9.0%7.3%5.4%7.9% Elderly (over 64 years as a % of all individuals in group)30.1%32.0%33.3%44.2% Individuals in group with father's occupation: unskilled (%)10.5%6.8%3.3%15.1% Reporting anxiety and depression (% of all individuals in group)10.6%10.3%7.4%3.1% Reporting health problems with alcohol or drugs (% of all individuals in group)0.9%1.1%0.3%0.0% Individuals who reported that they have no one to talk to19.9%23.8%31.1%31.5% Living standards of very poor households

46 Causes of poverty Very poor households Unemployed (as a % of economically active in group) 45.4%25.7%16.7%9.6% Economically active (%) 18.3%17.1%16.8%17.7% Vocational qualifications (% of all adult individuals in group) 20.9%20.7%18.9%12.2% Full-time job (% of economically active in group) 43.1%65.9%80.7%90.1% Adults with no qualifications (% of all adult individuals in group) 58.4%65.2%72.3%78.9%

47 Very poor households: sources of income An analysis of persons in the city who are below the “primary” poverty line shows that more than one half of these are members of families whose wage-earner is in work but in receipt of insufficient wages. Rowntree (2000: 114)

48 Future challenges: modelling income and substitution effects A substitution effect making leisure more attractive than work An income effect, encouraging people to work more to make up the loss of income “Different taxes have different effects, and affect people at different levels of income or in different household circumstances in different ways.” (Hill and Bramley, 1986: 85)

49 Simulating geographies of happiness (Ballas, 2008)

50 Estimated geography of happiness in Wales (%) happy more than usual, 1991

51

52 Estimated geography of happiness in Wales (%) happy more than usual, 2001

53

54 Estimated geography of happiness in Wales (%) happy more than usual, 2011

55

56 Estimated geography of happiness in Wales (%) happy more than usual, 2021

57

58 Spatial Microsimulation software: Micro-MaPPAS Micro-simulation Modelling and Predictive Policy Analysis System Turning academic research in to a usable tool for the real world Ballas, D., Kingston, R., Stillwell, J., Jin, J. (2007) A microsimulation-based spatial decision support system, Environment and Planning A 39(10), 2482 – 2499

59 Income < £10k, 3 children, LA home

60 Conclusions Tackle calibration problems Create a socio-economic atlas of the future for Britain Policy spatial micro-modelling - income and substitution effect Include more regional subsystems (labour demand, schools, hospitals, etc.) Small area multiplier analysis What-if, what-will-happen-if and What-would- have-happened-if analysis

61 More information on microsimulation: … and free book downloads:


Download ppt "Spatial microsimulation approaches to population forecasting Dimitris Ballas Department of Geography, University of Sheffield"

Similar presentations


Ads by Google