Testing the red herring hypothesis on an aggregated level: ageing, time-to-death and care costs for older people in Sweden

Abstract

In this paper we test the ‘red herring’ hypothesis for expenditures on long-term care (LTC). The main contribution of this paper is to assess the ‘red herring’ hypothesis by using the probability of dying as a measure for time-to-death (TTD). In addition, we implement models that allow for age-specific TTD effects on LTC utilization as well as sex-specific effects. We also focus on total, institutional and domiciliary LTC separately. For our analysis we use high quality administrative data from Sweden. Our analysis is based on fixed effects estimates. We use our findings to project future LTC expenditures and show that, although TTD is a relevant predictor, age itself remains the main driver of LTC expenditures.

Appendices

Appendix 1

See Table 6.

Table 6 Comparison of municipalities (not) providing health care in 2008

Appendix 2

See Table 7.

Table 7 Variable definitions

Appendix 3

See Table 8.

Table 8 Sex specific summary statistics

Appendix 4

See Table 9.

Table 9 Sex-specific institutional LTC expenditures

Appendix 5

See Table 10.

Table 10 Sex-specific domiciliary LTC expenditures

Appendix 6: Changes in cost increase per 65+ life year

$$ \Delta {\text{LTC}}_{{it}} = \left\{ {\begin{array}{*{20}c} {\sum\nolimits_{{k = 65}}^{\infty } {\left( {\alpha _{j} + \beta _{j}^{k} } \right)} \left( {\frac{{s_{{2008}}^{{k - 1}} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{{k - 1}} } }} - \frac{{s_{{2008}}^{k} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{k} } }}} \right) {\text{with}} \; j = i,} \hfill & {{\text{for }}i = 1,2} \hfill \\ {\sum\nolimits_{{k = 65}}^{\infty } {\left( {\left( {\alpha _{j} + \beta _{j}^{k} } \right)\left( {\frac{{s_{{2008}}^{{k - 1}} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{{k - 1}} } }} - \frac{{s_{{2008}}^{k} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{k} } }}} \right) + \delta _{k}^{j} \left( {\frac{{TTD_{{2008}}^{{k - 1}} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{{k - 1}} } }} - \frac{{TTD_{{2008}}^{k} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{k} } }}} \right)} \right)} {\text{with}} \quad j = i,} \hfill & {{\text{for }}i = 3} \hfill \\ {\frac{{\alpha _{j} + \sum\nolimits_{{k = 65}}^{\infty } {\left( {\alpha _{j} + \beta _{j}^{k} } \right)} s_{{2008}}^{{k - 1}} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{{k - 1}} } }} - \frac{{\sum\nolimits_{{k = 65}}^{\infty } {\left( {\alpha _{j} + \beta _{j}^{k} } \right)} s_{{2008}}^{k} }}{{\sum\nolimits_{{k = 65}}^{\infty } {s_{{2008}}^{k} } }} {\text{with}} \; j = 1,} \hfill & {{\text{for }}i = 4} \hfill \\ \end{array} } \right. $$

Appendix 7

See Fig. 5.

Fig. 5

Total cost increase per capita associated with an increase of life expectancy by 1 year (various scenarios based on sex-specific estimates)

Appendix 8

See Fig. 6.

Fig. 6

Total cost increase per capita and per life year associated with an increase of life expectancy by 1 year (various scenarios based on sex-specific estimates)