Heather Booth

Overview

Heather Booth is a demographer at the Demography and Sociology Program, Research School of Social Sciences, Australian National University. Her research focuses on mortality forecasting, population projection methodology, and the extension of the Lee-Carter model.

Key Contributions / Features

• Co-developed the R(S)/RD(S) fitting-period selection criterion for the Lee-Carter model, providing an objective, data-driven procedure for identifying structural breaks in k(t).
• Demonstrated that the Lee-Miller "1950 starting year" recommendation is among the worst choices for Australia and that 1968 is the optimal fitting period.
• Showed that k(t) non-linearity and b(x) instability are jointly resolved by restricting the fitting period to the structurally homogeneous recent regime.
• Co-authored Booth, Tickle, and Smith (2005), a 10-country out-of-sample evaluation establishing that jump-off bias is the dominant source of LC forecast error (57% female, 42% male share) and that BMS and LM substantially correct for it.
• Authored Booth (2006), a 25-year survey of demographic forecasting (1980–2005) that catalogued the LC variants landscape (LM, BMS, Li-Lee, Hyndman-Ullah, De Jong-Tickle), synthesized the three complementary approaches to forecast uncertainty (model-based, expert-based, ex post error), and documented Keilman's (1997) finding that forecast accuracy has not improved despite methodological advances. First paper to synthesize Ashley's impossibility theorem as a principled explanation for why structural demographic models underperform extrapolation at long horizons.
• Co-authored Booth, Hyndman, Tickle, and De Jong (2006), which extended the BTS (2005) evaluation to include HU and DJT: confirmed that all four non-original methods beat LC on log death rates but found no significant LE differences (p = 0.21); demonstrated via factorial decomposition that fitting period length and jump-off bias dominate adjustment method as error sources; established that LE accuracy is "largely a matter of luck" given the nonlinear life-table transformation.

Related

• Lee-Carter Model
• John Maindonald
• Len Smith
• Ronald Lee
• Lawrence Carter
• Probabilistic Population Forecasting
• Rob J. Hyndman
• Piet de Jong

Sources

• Booth Maindonald and Smith 2002 — Age-Time Interactions in Mortality Projections Applying Lee-Carter to Australia
• Booth Tickle and Smith 2005 — Evaluation of the Variants of the Lee-Carter Method of Forecasting Mortality A Multi-Country Comparison
• Booth 2006 — Demographic Forecasting 1980 to 2005 in Review
• Booth Hyndman Tickle and De Jong 2006 — Lee-Carter Mortality Forecasting A Multi-Country Comparison of Variants and Extensions