C4SE ©2007 The Centre for Spatial Economics
Developing an Occupation Supply Modelling Strategy
FLMM Meeting, VancouverOctober 18, 2007
Ernie StokesManaging Director
C4SE ©2007 The Centre for Spatial Economics
Presentation Objectives
• Introduction• Occupation system components• Current Canadian approaches (Fairholm)• Key factors in choosing an approach• A suggested approach (CSC experience)
C4SE ©2007 The Centre for Spatial Economics
Projection System Components
Data Models
Analysts
Clients
Projection Process
C4SE ©2007 The Centre for Spatial Economics
Occupation Model Components
• Demand – total employment/hours requirements of the economy
(expansion/contraction demand)
• Supply– outflows associated with deaths, retirements, and other
factors such as migration (replacement demand)– inflows including new entrants into the labour force such
as school leavers, entrants from other occupations, and migration
– changes in hours worked on the supply side
• Education linkage• Demand-Supply interaction
C4SE ©2007 The Centre for Spatial Economics
COPS Approach
Strengths– Detailed labour market projections at National level– Includes several aspects of education– Includes some Provincial aspects– Most detailed inclusion of immigration in the World
Weaknesses– D/S imbalances have no impact on demand or supply– Incomplete replacement demand– Incomplete supply side– Provincial aspects incomplete
C4SE ©2007 The Centre for Spatial Economics
Provinces Approaches
Provinces Monitor Labour Force Developments– Many Provinces monitor labour markets using simple
statistics and charts– Some have developed labour market indicators– Most meet with representatives from other government
departments, industry and associations
Larger Provinces Tend to Have Models– Ontario models/forecasts labour force withdrawals– Quebec has a detailed and sophisticated model/forecast– Alberta has a detailed bottom up model/forecast
C4SE ©2007 The Centre for Spatial Economics
Provinces Approaches - More
Smaller Provinces Do Not Have Models
– Lack of models and general quantitative techniques– Rely more on external providers– Conference Board and COPS– Rely more on qualitative techniques– Ad hoc analysis
C4SE ©2007 The Centre for Spatial Economics
Key Choice Factors
• Purpose• Usefulness
– Accuracy– Consistency– Transparency– Education
• Data• Resources
C4SE ©2007 The Centre for Spatial Economics
Purpose
Forecasting and scenario analysis
Governments • Policies and planning
– Labour market programs and education– Time horizon: medium to long term– Require great detail
Private sector• Human resource planning
– Time horizon: short to medium term
C4SE ©2007 The Centre for Spatial Economics
Usefulness
• The purpose of forecasting and scenario analysis is to reduce uncertainty about the future
• A projection or forecasting system is said to be “useful” if it reduces the level of uncertainty about the future to below that which existed before the system is used
• This definition, of course, should be thought of in terms of the condition that the marginal benefits associated with reducing the uncertainty is greater than or equal to the costs of developing and using the system
C4SE ©2007 The Centre for Spatial Economics
Accuracy
• Accuracy refers to the proximity of the projection to the actual results (a more accurate forecast is generally a more useful one)
• The ability of a projection technique to achieve a high degree of predictive accuracy depends on such factors as the– time horizon of the projection (long term more difficult
than short term)– cyclical volatility of the economic indicators to be
projected (food expenditures or housing starts)
• In scenario analysis accuracy is important in the sense that the model should predict well if the assumptions are correct
C4SE ©2007 The Centre for Spatial Economics
Consistency
• Is the approach “consistent” with the theory behind the phenomenon in question?
• A model is a description of the process (data generating mechanism) that generates the actual labour market outcomes
• The approach must use a model that produces results consistent with the views in this regard held by the clients (otherwise it will not be useful)
• Since there are always differences in individual views, Occam’s Razor can be employed (limited consensus)
C4SE ©2007 The Centre for Spatial Economics
Transparency
• The transparency of the projection and its methodology is important if one expects to get the consumers of the projections to buy into it (otherwise it will not be useful)
• To be transparent the projections and methodology have to be documented and presented in such a manner that projection users have a good understanding of how the methodology works and what assumptions have been employed to produce the projection
• In this case, analysts could, if desired, duplicate the results obtained by other analysts
C4SE ©2007 The Centre for Spatial Economics
Education
• Decision making characterized by uncertainty requires a good understanding of the issues surrounding the decision. Numbers are not enough!
• What are the key drivers behind future labour market developments?
• What are the major risks associated with the future performance of labour markets?
• A system that delivers this information is useful, particularly in cases where forecasting is difficult (when isn’t it?)
C4SE ©2007 The Centre for Spatial Economics
Data
• The type and amount of data that are available for the development of a system have a very important impact on its development and use
• If there are few data for the desired model inputs and outputs, then it is not possible or very difficult to develop a detailed system
• If the analysts are required to create estimates of the data, then additional resources will be required to develop and use the system on an on-going basis
• Most of us working in LMI are working with very limited data
C4SE ©2007 The Centre for Spatial Economics
Resources
• The real, financial, and time resources available to the forecasting organization are important factors when choosing a projection approach
• More complicated and detailed approaches require greater resources to develop, maintain, and use
• Organizations can use internal and/or external resources, as evidenced by current labour market forecasting organizations
C4SE ©2007 The Centre for Spatial Economics
A Suggested Development Process
• Set up an LMI team– Clients and analysts– Need to involve an LMI expert (someone with
experience in developing LMI systems)
• Identify needs and priorities– What is needed in general and what should be given
the highest priority for development?
• Survey LMI approaches• Assess existing resources and resource needs
– Can the system be developed with existing staff?– Is there sufficient funds to hire additional staff or
external help?
• Set up a plan for “on-going” system development
C4SE ©2007 The Centre for Spatial Economics
A Modest Beginning
• Develop a relatively simple, not too detailed, but still theoretically acceptable model (Occam’s Razor) to start
• Must demonstrate the usefulness of the system to clients as soon as possible
• Supply models require a lot of detail (age/sex groups for the components), start with fewer occupations (140 occupations rather than 520)
• Can use a top down approach, estimate total labour force change and some components leaving a residual component
C4SE ©2007 The Centre for Spatial Economics
CSC Experience
• The CSC started modelling employment and labour force for about 30 occupations/trades
• The components of labour force change were not explicitly modelled (reduced-form model) and there was no age-sex detail
• This approach was sufficient to meet the initial needs of clients and to get them involved in the process
• It also led to capacity building for the analysts and the clients
C4SE ©2007 The Centre for Spatial Economics
CSC Labour Force Model
Possible Labour Force =
Trade Requirements/Total Occupation Requirements * Total Labour Force
Labour Force Change =
Adjustment Coefficient *(Possible Labour Force – Labour Force Previous Year)
Adjustment coefficient reflects time required for new entry, exits, and mobility (geographic and inter-occupational)
C4SE ©2007 The Centre for Spatial Economics
Build Capacity Over Time
• Once clients become familiar with the process they demand more information from the system
• Clients will identify what is needed and additions to the system can be added to meet their needs
• After the first year, CSC clients identified additional trades that should be included in the system and also expressed the need to examine replacement demand
• The CSC responded by adding new trades and developing a model that could explain the retirements and deaths parts of replacement demand
C4SE ©2007 The Centre for Spatial Economics
Retirements and Deaths
Deaths (age,sex) = Labour Force (age,sex)*
Death Rate (age,sex)
Retirements (age,sex) =
Labour Force(age,sex)*Retirement Rate(age,sex)
C4SE ©2007 The Centre for Spatial Economics
Developing an Occupation Supply Modelling Strategy
QUESTIONS
Ernie StokesManaging Director