Counting employment – which figures to use
One of the most important bits of economic information a council can have is about employment. However there are different ways of measuring employment and it can be confusing figuring out which one to use.This article is designed to help demystify them all and give you an idea of which ones to use when, and what their limitations are.
Understanding employment figures
There are several different ways of counting employed people, including Census data based on where people live and where they work, population data from the Labour Force survey, and modelled employment estimates. id’s economic profile (economy.id) sets out a number of measures of employment including:
- Resident employment
- Census employment
- Total estimated employment
- Full-time equivalent employment
Using the Western Australian City of Joondalup’s economy.id site, we’ll explain the differences between them and when to use which measure.
When analysing the employment of an area, it’s important to distinguish between resident employment (people who live in the area and may work elsewhere) and the workforce (those who work in the area but may live elsewhere).
Firstly, have a look at the Local labour force. This is about the employed people who live in Joondalup. Once you remove the unemployed, this total is 81,380. Of course many of them commute out of Joondalup to work, and you can see where they go by industry in Residents by place of work.
However, most information in economy.id is related to where people work, as this directly relates to the industries which are located in the area. This is where it gets a bit tricky. There are three main measures of how many work in an area.
Census employment figures
The first is Census data. This is found under Workers by industry. This shows that the number of people who worked in Joondalup in the week before Census 2006 was 32,782. Census is a fantastic resource for a wide range of population characteristics, but for total employment by work destination, there are some people missing in a Census count. Census employment captures the number who:
- Were counted in the Census;
- Answered the Census question on employment;
- Answered the Census question on workplace destination;
- Had a defined workplace address, not “No usual place of work”, and
- That workplace address could be coded accurately by the ABS to an address inside the City of Joondalup.
What that means is that the Census employment count is an undercount of actual employment. There is an undercount in all Census data, but it’s usually only about 1.5% of the population. For employment by work destination, the undercount is a bit higher, due to the number of ways it’s possible for a person to not be coded to the correct area.
This is not to say the Census isn’t useful – it’s still a brilliant resource for understanding the characteristics of employed people, and the great thing about the Census data in the new economy.id site is that you can expand each of the industry categories, into more than 200 subcategories with detailed industry data available.
However as a count of total employment it’s a little limited.
Estimated total employment (NIEIR)
To adjust for the Census undercount, a new dataset from the National Institute for Economic and Industry Research (NIEIR) has been included, which uses modelled economic data to estimate the employment by industry, using supplementary datasets such as the quarterly Labour Force Survey employment estimates, tax office and Centrelink data. The modelled data are also averaged over the year, while the Census count is a snaphshot at a point in time.
This dataset is found in Employment by industry.
This shows total Census employment of 41,046 in Joondalup and an increase to 44,117 in 2010 – the other great thing about the modelled data is that it’s updated annually. It is also available for a wide range of industry categories – 49 subgroupings.
This dataset is a better measure of actual employment in the area at any point in time, as it factors in the Census undercount and is updated regularly. It does differ significantly from the Census results, particularly in a few industries. The chart below shows the differences. The industries, such as Construction, and Transport, Postal and Warehousing, which are likely to have large numbers of employees with no usual workplace address, have the greatest increase when comparing the modelled data to Census data. Industries like Finance and Manufacturing, which generally have a fixed place of work, show smaller differences.
Full time equivalent employment (NIEIR)
The other useful measure to come out of modelling employment is an estimate of total hours worked in each industry, which translates into Full Time Equivalent Employment. This is really useful if you’re looking at the total amount of work done in each industry or the relative employment size of each industry.Industries such as Retail Trade and Accommodation and Food Services tend to have the largest drops in employment when viewed on a FTE basis, as they have a high proportion of part-time and casual workers.
Joondalup’s FTE employment in 2006 was 33,415.
So which one to use?
- If you need to know the total employment in a particular industry, use the NIEIR estimated total employment.
- If you need to compare the size of industry’s workforces on an equal basis, or worker productivity, use FTE modelled employment.
- If you need to compare the demographic characteristics of the workers in different industries, or breakdown to very detailed industry subgroupings, use Census counts of employment.
- If you need to compare resident employed to those who commute out, use Census counts of employment (journey to work).
All these options are available in .id’s economic profiles (economy.id). When you’re in a topic, the table clearly shows whether you’re viewing Census or modelled data.
If you enjoyed this article, visit our website at id.com.au or sign up to get updates via email and twitter (above), and feel free to share it. Leave us a comment to share your own experience.