6: Tips and Traps When Using Formal LMI
Tips
Use the NOC to Connect Real-World Job Titles to Formal Data
Both Employment and Social Development Canada (ESDC) and Statistics Canada have lists of example job titles for every detailed occupation. The NOC system is the official way job information is organized in Canada, however, the names of many of the occupations can seem disconnected from the real world.
For example:
NOC 51111
This NOC code is named Authors and writers (except technical), and includes job titles like Advertising copywriter, Novelist, and Speech writer.
NOC 84110
This NOC code is named Chain saw and skidder operators, and includes job titles like titles such as Bucker, Faller, Feller, and Grapple skidder operator.
Use the NOC to help you associate real-world job titles with the way statistical information is structured so that you can find reliable information related to average earnings, hours worked, or employment levels for the jobs in question.
Use the following tools to help guide this:
- NOC matrix structure and hierarchy
- NOC profiles (search by job title)
- Statistics Canada NOC Index of Titles
The Rule of 2,000 (Interpreting Wages and Earnings)
Most people want to know how much they can earn in different jobs. Different sources of information report earnings in different ways, but typically we see either annual salary or an hourly wage.
To quickly convert from hourly wage to annual salary, you can use the rule of 2,000. Double the wage rate and add 3 zeros. To go from yearly salary to hourly wage, do the reverse (drop 3 zeros and divide by 2).
Hourly Wage → Annual Salary
$25/hour
× 2
────────
50
+ 3 zeros
────────
≈ $50,000/year
Annual Salary → Hourly Wage
$50,000/year
− 3 zeros
────────
50
÷ 2
────────
≈ $25/hour
This quick and dirty calculation works by assuming someone works 40 hours a week and 50 weeks a year, meaning the hourly wage worker has two weeks of unpaid leave during the year.
For more on how people make use of LMI information like wage information, download LMIC’s article Socio-Demographic Differences in Labour Market Information Use, Sources and Challenges.
Understand Seasonal Adjustments for Any Season
A lot of labour market information is presented as “seasonally adjusted” data. This includes the headline unemployment rate and monthly changes in employment reported in the media.
Seasonally Adjusted Data
Seasonally adjusted data is useful when you want to look at long-term trends. This is because the adjustment smooths out regular, known fluctuations such as few construction jobs in the winter compared to summer, fewer workdays in February versus other months, and the jump in retail employment during the Christmas shopping season.
Raw or Non-adjusted Data
Know the Limits of Sample Size
All data has its limits. The critical limiting factor is always the number of underlying observations (e.g., sample size). This is true of survey data and administrative data like tax files.
For example, the Labour Force Survey (LFS) surveys 60,000 households each month. LFS reports employment information by occupation or by industry – but not both. In principle, it is possible to cut the LFS data by both industry and occupation (and region, and gender, etc.), but the smaller the group, the less reliable the information.
Understand the Many Shapes and Sizes of Canadian Geographies
Canada is organized into many different geographic categories, all of which are built up from the Census Dissemination blocks. Census Dissemination blocks are the smallest geographic unit used by Statistics Canada for the purpose of collecting census data. There are nearly half a million Census Dissemination blocks in Canada, LMI at this level is not readily available as the sample sizes are simply too small.
These blocks get rolled up into two parallel tracks.
Track One
Track One organizes Canada into larger regional divisions used for reporting official data. These include Census Subdivisions (CSD), Census Divisions (CDs), and Economic Regions (ERs).
Economic Regions are commonly used in labour market reporting and cover every part of Canada, including large urban centres and remote northern areas. For example, Yukon is a single Economic Region, while the Toronto Economic Region includes about 6.5 million people.
Economic Regions never cross provincial or territorial borders.
Track Two
Track Two focuses on towns and cities. Large urban centres are called Census Metropolitan Areas (CMAs), while smaller urban centres are called Census Agglomerations (CAs).
Unlike Economic Regions, CMAs and CAs are not geographically exhaustive because they only represent populated urban areas. However, about 83% of Canadians live in one of these regions.
CMAs and CAs can also cross provincial boundaries, such as the Ottawa–Gatineau CMA.
To look up a city, town, or region and find out what geographic classification it falls into, you can use the Census Population Tool from Statistics Canada.
Understanding which geographic unit a data source uses helps prevent misinterpretation when clients say “my area” or “my community.”
Traps
Be Wary of Forecasts
Forecasting employment levels by occupation is prevalent – usually called an “occupational outlook.” Forecasting, in general, is incredibly challenging to do with accuracy, and economic shocks cannot be reliably predicted. Think, for example, of the enormous changes in the labour market that occurred in the first year of the COVID-19 global pandemic.
That said, forecasts or persistent long-run trends can and do offer useful insights. We know, for example, baby boomers are starting to retire, and there are few young cohorts to replace them – so forecasts of growing employment in healthcare are pretty reliable.
Measuring Labour Market Shortages
Labour shortages and skills shortages are consistently hot topics, but there’s a lot of disagreement about how they should be identified and measured. Media reports of shortages typically rely on two sources:
- Employers struggling to hire
- Differences between forecasts of future supply and demand in an occupation
If employers are reporting difficulty finding people to hire, then regardless of what one calls it, it is an indication of opportunities for job seekers… but it may not take into account the capacity of employers to find or screen candidates well, suggesting more of an issue with recruitment methods rather than a growth in openings. If, on the other hand, the reason for reporting a shortage is because of forecasts, then it is less clear if there are opportunities for jobs.
Online Job Postings Missing Implicit Requirements
Assuming the Level of Skills Required
Whether you are using data sourced from online job postings or a formal taxonomy of skills such as ESDC’s Taxonomy, there are significant limits to knowing the depth of skills required in any particular job.
Prioritizing Specifics
While people often want very local, granular LMI, sometimes bigger is better. Small sample sizes might limit what’s available, and the most local, granular data available may not be the best data.
Take, for instance, average wages by occupational group reported in the Labour Force Survey (LFS). The data are available, but because of the small number of observations, they are very noisy – meaning each month’s wage estimate jumps up and down a lot. The larger the group, the less noisy and more reliable the information. Think about what dimensions don’t need to be very detailed for your purpose and aggregate across those categories.
For example, if you need wage information by occupation-level, you might not need it by month, in which case using annual wage data by occupation will be better.
By aggregating across 12 months, you’ll significantly reduce the volatility (noisiness) of the information and have much more reliable observations.
The following resources can help further guide you.
- LMIC – Local Granular Data for Beginners
- Statistics Canada – Average Wages by Occupation
