Key findings and background facts 2016

Background

This report highlights key dynamics of the benefit system and how they impact long-term benefit dependency. It is the fourth internal actuarial report produced in relation to the forward liability of the benefit system. The purpose of the report is for the Chief Actuary to independently:

• Review experience over the year of exit rates, numbers of new clients and clients transitioning between benefits
• Review overall performance of the welfare system and the effectiveness of investments made to reduce benefit dependency
• Review and comment on the valuation of the forward liability
• Identify areas for attention to assist in managing long-term benefit dependency

Unless stated otherwise, it covers the period from 1 July 2015 to 31 December 2016. Previous reports have only covered the year to the previous 30 June. We have extended this to 31 December for this report to ensure information is more up-to-date.

Some of the analysis in this report relies on the liability calculations performed by Taylor Fry Consulting Actuaries and detailed in their report titled Valuation of the Benefit System for Working-age Adults as at 30 June 2016^[1] (the 2016 Valuation Report) which was publicly released in May 2017.

Liability is a measure of future expected benefit cost and as such is a proxy for long-term benefit dependency.

The report does not explicitly cover social housing though makes comment on social housing aspects where relevant to benefit system clients. Social housing may be incorporated more broadly in the future, either as part of this report or separate reporting.

Key Findings

There is clear evidence of a reduction in long-term benefit dependency as shown by the valuation. At December 2016, there was the same proportion of the working-age population receiving a main benefit as there was at the height of the pre-Global Financial Crisis (GFC) economic boom (9.8%).

47% of JS-HCD clients receive JS-HCD continuously for at least six months. This is high considering that the benefit is intended to provide temporary support. Eligibility for JS-HCD is heavily influenced by medical practitioners. Given the connection between health and employment outcomes, greater connectivity between medical practitioners and case managers should be explored.

The Social Policy Evaluation and Research Unit’s (SUPERU) analysis of off-benefit outcomes shows that many people who leave benefits for tertiary education return to a benefit.

Fewer youth benefit clients are transitioning to working-age benefits than before the introduction of the Youth Service. This is despite an increase in prevalence of key risk factors, such as CYF history, amongst youth benefit clients.

Risk factors reaffirmed

Key risk factors identified in previous valuations still come through in this valuation. The most important factors are early entrance to the benefit system, intergenerational benefit dependency, evidence of poor childhood experience, low educational achievement, ethnicity and the presence of criminal convictions. Some specific factors to consider are:

• Criminal convictions are a very strong predictor of high future benefit cost amongst clients who have recently exited the benefit system. About two-thirds of people receive a main benefit within one month of leaving prison. This is closer to 80% if they were receiving a main benefit immediately before the prison spell. There is a clear pattern of imprisonment and benefit receipt. Prisoners can face considerable barriers to employment on release including negative attitudes of employers, and have high rates of mental illness and substance abuse. Two-thirds of prisoners in New Zealand have substance abuse problems.
• For clients who are not receiving a main benefit (or recently exited from the benefit system), their liability is 83% higher if they are in a social house. This is because they have a significantly higher rate of entry on to main benefits than others not in Social Housing.
• A JS-HCD client’s type of Health Condition is not particularly important for estimating their future benefit cost. Understanding the reasons for this could be useful in understanding how to best work with JS-HCD clients. Intuitively the type of health condition ought to have a bearing on how long benefit system support is required, the likelihood of repeated instances of the condition, and the likelihood of the condition becoming permanent and the client transferring to SLP.
• Risk factors are often concentrated together. This is more common for particular cohorts of clients. For example, all five of the risk factors are more prevalent for Māori. However, the relative difference between Māori and non-Māori increases as we consider larger combinations of risk factors. Māori are nearly three and half times more likely to have all five of these risk factors present than non-Māori. The liability differences are also significant.     There is a similar story for social housing clients. Clients who have lived in social housing are over three times more likely to have at least three risk factors. Again, the liability differences are significant.  
• Risk factors materialise before people first enter the benefit system. It is increasingly apparent that risk factors that correlate with likelihood of long-term benefit dependency are determined before a person enters the benefit system.The valuation does not distinguish between correlation and causality. However, there is enough for us to conclude that poor childhood outcomes are the most important determinants of long-term benefit dependency. These factors include intergenerational benefit receipt, CYF history and poor educational achievement.Logic suggests that investing to prevent or limit the emergence of risk factors during childhood delivers better value than investing in adulthood once poor adult outcomes begin to materialise. This is not to say that MSD has no part to play in working with vulnerable adults. However, there is opportunity to improve outcomes with additional investment in prevention strategies.

Benefit clients in social housing have higher average liability

Previous valuations have added Child, Youth and Family (CYF) history, criminal convictions history, education status and intergenerational benefit receipt data to improve the predictive capability of the valuation model.

This is the first year the benefit system valuation has been integrated with the social housing valuation, giving us a better understanding of people across both systems. The model has been adapted to project both future benefit receipt and future social housing usage.

Nearly half of those in the social housing cohort are in the benefit system cohort, and their average liability (excluding accommodation supplement) is $30k higher than other benefit system clients.

Combinations of risk factors are also more likely to be present. For example, 18-24 year old benefit clients who have spent time in social housing are twice as likely to have a criminal conviction and twice as likely to have had a parent on benefit for at least 80% of their teenage years.

What do Work and Income’s young, new clients look like?

The 2016 valuation explores how effectively we can predict lifetime cost at the point of entry into the benefit system.

SLP and SPS early entrants have substantially higher average liability than early entrants to other benefit categories, highlighting opportunity for investment before people become entrenched in the system.

Liability is significantly higher for Māori who are early entrants and have CYF history. As an example, of under-25 year-olds who received their first main benefit (other than SLP and SPS) in 2015/16:

• Māori clients with CYF history have an average liability of $177k
• non-Māori without CYF history have an average liability of $77k.

There has been an improvement in the proportion of clients who have exited the benefit system after the introduction of the Youth Service. From the last valuation, the projected duration on benefits has decreased from 15.3 years to 13.8 years for YP and from 15.3 years to 14.3 years for YPP. This is despite the fact that key risk factors such as CYF history and intergenerational benefit receipt have increased in prevalence amongst YP/YPP clients since 2012.

Superu Research: Off-Benefit Outcomes

Once a client leaves the benefit system they have no obligation to keep MSD informed of their circumstances or employment status. Consequently, our view of former clients’ off-benefit outcomes is relatively limited through MSD administration data alone.

In 2016 the Social Policy Evaluation and Research Unit (Superu) commissioned Taylor Fry Consulting Actuaries to use the data in Statistics New Zealand’s Integrated Data Infrastructure (IDI) to look at where people go when they move off benefit. The IDI contains linked, anonymised datasets across a number of government departments and agencies. In particular, it includes Inland Revenue data, which allows us to understand employment outcomes.

Taylor Fry looked at approximately 140,000 people who moved off benefit between 1 July 2010 and 30 June 2011, and analysed their outcomes over the subsequent two years.

Why do people leave the benefit system?

Looking at 140,000 people:

• About half of these people left to take up employment (38%) or to start an education/training course (about 11%).
• A further 13% left due to a change in life circumstances (death, retirement or going overseas), and 3% were detained in prison.
• A clear reason for exit could not be identified for a third of leavers (the two ‘other’ categories). A separate analysis of MSD’s ‘reason of exit’ code implies that about 12% were no longer eligible for a benefit (e.g. because they were supported by a partner) and another 13% left to take up employment or go overseas, but there is no record in the IDI of such activity.
• Of people moving off benefit into employment, 57% had income shortly thereafter of under $3,000 a month and 82% under $4,000 a month, implying that most are moving into relatively low-wage employment.
• After 12 months, 30% of people were in substantial employment (monthly income over $1,180) and 25% were back receiving a benefit.
• These percentages remain relatively stable between 12-24 months, though this does not mean every person stayed engaged in the same activity over that period.
• The picture varies if you isolate one particular reason for moving off benefit. For example, about 50% of those that moved off benefit into employment were engaged in substantial employment after 24 months, much higher than the 30% across all exit reasons.

What do people do after they have left?

• After 12 months, 30% of people were in substantial employment (monthly income over $1,180) and 25% were back receiving a benefit.
• These percentages remain relatively stable between 12-24 months, though this does not mean every person stayed engaged in the same activity over that period.
• The picture varies if you isolate one particular reason for moving off benefit. For example, about 50% of those that moved off benefit into employment were engaged in substantial employment after 24 months, much higher than the 30% across all exit reasons.