Research Note 16 1998-99

The Risk of Unemployment

Tony Kryger
Statistics Group
23 March 1999

Introduction

It is commonly held that youth, the less well educated, residents of rural areas, individuals with limited English skills and persons born outside the main English speaking countries have relatively high unemployment rates (Table 1). What is not known is the extent to which each of these factors contributes to the probability that a person will become unemployed.

Table 1. Unemployment Rate, January 1999*
	%
Birthplace
Born in Australia	7.9
Born in other main English speaking country	7.4
Born in non-English speaking country	9.9
Age
Aged 20+	7.0
Aged 15-19	20.3
Place of residence
Live in capital city	7.5
Not live in capital city	9.2
Educational Qualifications
With post-school qualifications	5.0
Without post-school qualification	10.5
* Except educational qualifications which is at May-98.

Factors used in the unemployment model and the contribution which each makes (assuming all other factors are held constant) to the unemployment outcome are discussed below.

Factors Influencing Unemployment

• Educational Attainment

This is one of the most important indicators of unemployment outcomes. Compared to a benchmark group of persons who had left school at age 15 years or younger, holders of higher degrees and holders of diplomas were found to be at least nine percentage points less likely to be unemployed than the benchmark group. Individuals who had only attended the highest level of secondary school were six percentage points less likely to be unemployed than the benchmark group.

• Age

Age is a proxy for work experience. The model discovered that the probability of being unemployed initially decreases with age but then increases as the individual becomes older. Thus at 20 years, for each extra year of age, the probability of being unemployed decreases by 0.7 of a percentage point. At 40 it decreases by 0.1 of a percentage point, and at 45 years the unemployment probability starts to increase a small amount with age.

• English Proficiency

Individuals who do not speak English at home and whose command of English is 'poor' are 10 percentage points more likely to be unemployed than monolingual English speakers.

• Disabilities

Persons with a disability are six percentage points more likely to be unemployed than labour market participants with no disability.

• Marital Status

Persons who are not married were found to be six percentage points more likely to be unemployed than those who are married. Marital status is often regarded as a proxy for labour market stability and motivation.

• Place of Residence, Place of Birth and Indigenous Status

Place of residence (capital city, major urban, other urban or rural), place of birth (born in main English speaking country or other country) and indigenous (Aboriginal or Torres Strait Islander) origin were all shown not to have a significant impact on unemployment outcomes. These findings, however, are at variance with other studies and suggest that further investigation is required. Miller and Le note that the finding with regard to indigenous status may be attributable to the small number of indigenous people in the sample for the SEUP.

Risk Index

Model results can be represented as a points index in which the higher the number of points assigned to a particular characteristic the higher the probability of becoming unemployed (Table 2).

Table 2. Points Index
Characteristics	Points
Limited English proficiency	15
Disability	10
Never married	10
Age 15-19 years	10
Left school at 15 or less, or never attended school	10
Left school at 16 or higher before finishing secondary school	5
Speaks a language other than English at home; good English skills	5
Born in non-English speaking country	5
Live outside major metropolitan region	5

Other Factors Influencing Unemployment

The model described above, however, contains considerable prediction errors. That is, some individuals predicted on the basis of SEUP data to be at high risk of unemployment were found not to experience any time looking for work, while others predicted to have little risk of unemployment were found to experience lengthy periods of unemployment. To improve the predictive capability of the model some new variables were added.

One variable found to have a strong influence on the probability of a person becoming unemployed is his/her labour market history. In other words, being unemployed today increases the likelihood of being unemployed tomorrow. This link between past adverse labour market experiences and current labour market activity is generally termed a scar effect.

Another variable affecting labour market outcomes is family background. Belonging to a family where one or more members is unemployed increases the probability of unemployment compared to that of a family where no members are unemployed.

Despite the addition of variables to the model, there is still a need to better understand the influence that other factors (contactability, availability of transport, personal factors) have on labour market outcomes.

1. See Australian Bureau of Statistics, A Risk Index Approach to Unemployment, Occasional Paper (Cat. No. 6293.0.00.001).

Comments to: web.library@aph.gov.au
Last reviewed 17 October 2001 by the Parliamentary Library Web Manager
© Commonwealth of Australia
Parliament of Australia Web Site Privacy Statement
Images courtesy of AUSPIC