Abstract
Background
Some social protection interventions are promoted as policy tools for addressing the social determinants of health to improve individual and population health and health equity. However, there is considerable controversy over whether publicly funded financial credits have a positive, a negative or no effect on health status in adults in high-income countries. Family Tax Credit (FTC) is a publicly funded financial credit intervention designed to increase income in families living in or at risk of poverty in New Zealand (anti-poverty intervention). In-work Tax Credit (IWTC) is a financial credit intervention that provides additional income to adults in families receiving social assistance (or in low-paid employment) for taking up (or staying in) paid employment (welfare-to-work intervention). This thesis estimates the effect of the FTC and IWTC interventions on self-rated health (SRH) in parents in New Zealand. Rather than analysing cross-sectional data, this thesis uses a cohort study design and fixed effects regression methods to better infer causal effects from observational data.
Methods
Seven waves of data (Waves 1 to 7, 2002-09) were extracted from the Survey of Family, Income and Employment (N = 29,790) and restricted to a balanced panel of working-age (19 to 64 years) parents in families over two or more consecutive waves (N = 6,900). Linear fixed effects regression analyses were conducted, estimating the association of change in FTC and IWTC with change in SRH at the individual level in the study sample over the short term. These analyses controlled for all time-invariant confounding and adjusted for measured time-varying potential confounding. The exposure variables were eligibility for FTC and IWTC and the amount of FTC and IWTC that the family of an eligible participant was entitled to. The outcome variable was SRH. Potential time-varying confounding variables were the equivalised gross total annual family income (minus FTC or IWTC), family type, number of dependent children in the family and employment status. Subsidiary analyses were conducted to estimate the effect of FTC and IWTC on SRH, where the outcome variable lagged behind the exposure variable by longer periods, and on two other health outcomes (psychological distress and current tobacco smoking), as well as to test for effect modification by ethnicity and level of income.
Results
The best estimate for a change in SRH one year after becoming eligible for FTC was a small, statistically non-significant increase of 0.013 in score over the short term [95% confidence interval (CI) -0.011 to 0.037]. An increase by $1,000 in FTC amount was also not associated with any discernible change in SRH (effect estimate -0.009, 95% CI -0.057 to 0.039). Likewise, neither becoming IWTC-eligible (effect estimate 0.003, 95% CI -0.020 to 0.027), nor an increase by $1,000 in IWTC amount (effect estimate 0.000, 95% CI -0.008 to 0.008) were associated with any discernible change in SRH. Subsidiary analyses also found no effect of FTC and IWTC eligibility and amount, when SRH lagged behind the exposure variables by a longer period. No effect of FTC and IWTC was found on the two other health outcomes, psychological distress and current tobacco smoking. Finally, no evidence for effect modification by ethnicity and level of income was found.
Conclusions
This thesis found no discernible effects of FTC or IWTC eligibility and amount on SRH at the individual level, over the short term. No previous studies have investigated the effect of anti-poverty tax credits on parental health. Five studies of welfare-to-work tax credits in the United States found no evidence for an effect of these credits on health status, except for inconclusive evidence on smoking favouring a reduction. This previous evidence for the United States was generally consistent with the finding from this thesis for New Zealand. Strengths of this thesis include its relatively good classification and measurement of the exposure variables and strong control of confounding. Limitations included some risk of bias from misclassification and mismeasurement of the exposure, outcome and confounding variables, likely towards a null finding. The internal validity of the study was judged to be strong, but the thesis had risk of bias from misclassification of the exposure. This study cautions health sector policy makers against investing in anti-poverty and welfare-to-work financial credit interventions to improve adult general health status and health equity in New Zealand and comparable high-income countries. Future research should investigate the effect of FTC and IWTC over the longer term, including the effect of a regime of treatment with FTC and IWTC over several consecutive waves.