The Treasury

Global Navigation

Personal tools


New Zealand’s Social Assistance System: Financial Incentives to Work - WP 03/18

2  Financial Returns to Work

Social assistance programmes largely influence labour supply through influencing financial incentives to change hours of work and to participate in the labour market.[5] As well as decisions to supply labour, the financial returns from social assistance programmes also influence other important decisions, such as whether or not to participate in education and training and whether or not to enter into or remain in a relationship in the nature of marriage. In the discussion below the focus is on financial incentives to supply labour.

People’s decisions to supply labour are influenced by more than the financial incentives associated with social assistance programmes, however. As well as these financial incentives the changes in people’s labour supply reflect the uncertainty that they feel about the level of their likely work or social assistance income, non-financial considerations for entering or remaining in the work force (including self-esteem and fear of damage to future employment prospects), and social assistance programmes’ administrative incentives (such as work tests) [Barr, 1999, p. 13; Wilson, 1996, pp 13-15].[6] Yet, as the analysis in the following section demonstrates, financial incentives from social assistance programmes do, nevertheless, play an important role in influencing the labour supply of different demographic groups.

In section 2.1 this paper discusses ways of thinking about financial incentives and heterogeneity within the population. This section is then followed by discussions on the distribution of individuals by Effective Marginal Tax Rates (EMTRs) and benefit types in section 2.2. Section 2.3 then discusses EMTR profiles, family budget constraints, and the frequency distribution of hours of work for five different family types at two different wage rates.

2.1  Financial Incentives and Population Heterogeneity

The financial incentives associated with social assistance programmes influence people’s decisions in two ways. By changing relative prices of goods (such as labour and leisure) a social assistance programme may induce the person facing those incentives to substitute consumption of one good for another (the substitution effect). The social assistance programme may also alter the person’s real income (and the ability to reach a desired level of consumption) and consequently his or her demand for certain goods (the income effect). Whether these two effects reinforce or offset each other depends on the case at hand and requires empirical analysis [Rosen, 1988, pp. 29-30].

Financial incentives can be difficult to measure. One source of difficulty is that financial incentives not only reflect the design of a particular social assistance programme but also the (often complex) interaction of the programme with other social assistance and taxation programmes. Different programmes may use different definitions of what counts as income, income units (individual, family, and household), income periods (annual, fortnightly, or weekly), and implementation agencies (the Inland Revenue Department and the Ministry of Social Development) and be earned and abated in different ways.

Population heterogeneity is a further source of difficulty in measuring financial incentives. People differ in the circumstances that they face. People’s circumstances do not, furthermore, remain fixed over time but change in line with social and economic changes. Often social assistance programmes aim to recognise this heterogeneity by allocating differing levels and types of assistance to people in different family types and with differing levels of attachment to the workforce. Failing to account for the heterogeneity in the population when measuring financial incentives would be likely to lead to a misleading picture of the nature of policy problems and the effectiveness of government responses.

The financial incentives resulting from an income support programme differ among people with different characteristics, depending on factors such as hours of work, wage rates received, marital status, number and ages of children, availability of childcare, accommodation needs, and receipt of other assistance.[7] People also differ in the degree to which financial incentives lead to changes in their behaviour. Some people are more sensitive to financial incentives than other people. For instance, New Zealand studies have shown that the labour supply of sole parents, part-time workers, secondary-income earners, and teenage men tends to be more responsive to financial incentives than that of primary earners and prime aged males [Brosnan et al, 1989, p. 31; Maani, 1989; Maloney, 1997; Prebble et al (eds.), 1992]. Since these studies were completed the labour market in New Zealand has undergone notable change with, for example, deregulation of employment relationships and increasing labour market participation rates of secondary earners.

While thresholds in the income tax and social assistance systems are usually defined in terms of income, evaluating changes to people’s incentives to supply labour requires looking at relationships between changes in hours of work and changes in net incomes. Effective marginal tax rate (EMTR) profiles and budget constraints can be used to show such relationships.

EMTRs show the proportion by which a dollar increase in gross income is reduced by taxes and the abatement of social assistance benefits. An EMTR is one minus the change in net income (after tax deductions and abatement of social assistance payments) resulting from earning an additional gross dollar [Prebble et al (eds.), 1992, p. 7]. When people are faced with a decision on whether or not to make relatively small changes in income (e.g., from working a few extra hours) EMTRs are likely to illustrate the financial incentives applying to their decisions. However, when people are considering relatively large changes in income (e.g., whether to work part-time or full-time) or when they are constrained in the degree to which they can change their hours of work (e.g., when they have employment contracts containing fixed hours) EMTRs are less likely to illustrate the financial incentives applying to their decisions.

Budget constraints can be used to show the net income (after taxation and the payment of abated assistance) that is received at different hours of paid employment [Prebble et al (eds.), 1992, p. 11]. EMTRs are reflected in the slope of the budget constraint. As the EMTRs facing a person increase the budget constraint of that person becomes more flat [Prebble et al (eds.), 1992, p. 13].

Estimating EMTR profiles and budget constraints of people in partnered families poses particular challenges. In partnered households the labour supply decisions are joint decisions. In these households when either the primary or secondary earner changes his or her supply of labour the total household income changes. As social welfare benefits and Family Assistance programmes abate against total household income, the individual incentives facing the primary and secondary earners are influenced by the earnings of the other person. However, due to the difficulty of modelling joint decisions, in this paper the changes in household income are modelled as individual decisions, where only one person per family makes a labour supply decision and the rest of the labour supply decisions in the family are held constant [Prebble et al (eds.), 1992, pp. 37-38].

Findings based on different family types should be assessed against the degree to which these family types are representative of the general population or of those receiving income support. With this in mind this paper also considers effective marginal tax rate profiles and budget constraints in the light of data on the frequency distribution of hours for various wage rates of various demographic groups. Because there are small numbers of some family types it is likely that there will be few observations of these family types at certain wage rates. Thus in order to provide a useful number of observations it has been necessary to use wage bands (such as below and above median wages for different family types) rather than single wage rates in this paper.

The interaction of financial incentives with the income distribution influences the number of people who face certain financial incentives and which decisions are influenced by these incentives. A person’s hourly wage rate influences the number of hours of work for which they face particular financial incentives (or disincentives). If a person’s hourly wage rate fell from $15 to, say, $10 the number of hours of work over which income support abates would increase. In contrast, if the person’s hourly wage rate increased the abatement of income support would occur over a shorter range of hours of work. The hourly wage rate thus impacts on the incentives facing the person. Further, when a programme is targeted by income the proportion of the income distribution over which it

applies also reflects the level of the payment and the rate at which this payment abates. Either increasing the payment or decreasing the rate of abatement would increase the coverage of the programme.[8]

The interaction of financial incentives with the income distribution also influences which decisions are influenced by these incentives. Low-income groups, for example, contain a relatively high proportion of sole parents and secondary income earners facing decisions regarding whether or not to participate in the labour market. (Note that some of these low-income individuals may be in high-income households.)

This paper draws on estimates calculated with TaxMod (which is a micro-simulation model of New Zealand’s income tax and social assistance systems) [Prebble et al (eds.), 1992, pp. 29-44]. TaxMod calculates income tax liabilities and social assistance entitlement based upon characteristics of the population and rules regarding eligibility and abatement of income tax and social assistance programmes. A population of families is derived from demographic, income, and expenditure data contained in the Household Economic Survey (HES) [Gordon, 1997]. The HES was established to measure the Consumers’ Price Index and was conducted annually from 1983-4 to 1997-98. The HES is now conducted every three years with the most recent survey being completed in June 2000-01. (TaxMod and the Household Economic Survey are discussed in greater detail in appendix 4.)

In TaxMod each surveyed household is given a weighting representing the degree to which households of that type occur in the total population. This allows the HES sample to be weighted up to estimate the entire New Zealand population. TaxMod also re-weights the HES sample to allow for changing rates of unemployment and adjusts income data for inflation (with separate inflators for wage, self-employed, and interest income). In order to estimate financial incentives for the 2003-04-income year, TaxMod takes data on observed hours of work from the most recent HES survey (2000-01) and applies this data to the 2003-04 income tax and social assistance systems. However, these hours of work are likely to reflect the income tax and transfer systems that were in existence in the year of the survey. There may thus be some expected variation between the observed hours of work and the hours that actually correspond to the 2003-04 social assistance system.[9]

As the analysis in the following sections demonstrates, the financial incentives from social assistance programmes do, nevertheless, play an important role in influencing the labour supply of different demographic groups. Financial incentives matter.


  • [5]Social assistance programmes can also contain administrative incentives to work though measures such as work testing. Financial incentives and administrative incentives should be seen in conjunction. For instance, a weakening of administrative incentives for encouraging labour supply could increase the significance of financial incentives for encouraging labour supply.
  • [6]Responses to financial incentives may also reflect the timing of payments or abatement, as people may not be aware of or may discount the impact of such incentives when making decisions that, for instance, effect entitlements paid on an annual rather than a more frequent basis (such as fortnightly) [Barr, 1999, p. 18].
  • [7]A number of people who may appear to not face particular financial incentives (e.g., people located above or below the income levels at which incentives occur) may still have been affected by these incentives when making decisions (e.g., by being discouraged from locating at income levels associated with high disincentives to work).
  • [8]Thus, for example, while paying assistance at a flat rate and then abating assistance fully at a particular threshold would lead to high EMTRs at this threshold, the range of incomes over which these EMTRs would occur would be narrow. The relative merits of such a regime would depend on whether it would be better to have high EMTRs over a narrow range of incomes or more moderate EMTRs over a larger range of incomes.
  • [9]In contrast a behavioural model would estimate financial incentives using a two-step procedure. First, changes in the hours of work between 2000-01 and 2003-04 would be estimated. Second, the effects of a policy initiative would then be modelled using these new estimated hours.
Page top