Goldstein 2011 — A Secular Trend toward Earlier Male Sexual Maturity

Summary

Goldstein (2011) uses the peak age of the young adult "accident hump" in male mortality as an indirect proxy for male sexual maturity, arguing that risk-taking behavior reaches its apex at a developmentally fixed lag after pubertal onset. Fitting the Gompertz residual method to Human Mortality Database (HMD) data for Sweden (1751+), Denmark, Norway, England/Wales, and Italy, the paper documents a secular decline of ~0.19 years per decade in hump-peak age from the mid-18th century to ~1950, after which the trend halted. This parallels the well-documented slowdown in the secular decline of female menarche age, and is consistent with historical corroboration from voice-change records in Bach's Leipzig choir.

Key Claims

• The accident hump peak age (Gompertz residual method) is a valid indirect proxy for the timing of male puberty-linked risk-taking behavior.
• Secular decline in hump-peak age: ~0.19 y/decade from mid-18th century through mid-20th century, then halted ~1950.
• Results are consistent across all five HMD countries analyzed (Sweden, Denmark, Norway, England/Wales, Italy).
• England/Wales is a delayed outlier: hump timing declined later, tracking a similarly delayed secular gain in male stature due to urbanization effects on nutrition.
• Voice-change age in Bach's Leipzig choir (18 y in the 1720s) vs. 20th-century London choirboys (13 y) corroborates the ~0.19 y/decade estimate and the post-1950 halt.
• Male rate of secular change (0.19 y/decade) is slower than female menarche decline (0.3 y/decade), consistent with life-history theory: female reproductive timing is more phenotypically plastic than male.
• The decline of biological maturity has been accompanied by a secular delay in social adulthood (marriage, career, economic independence), producing a widening mismatch between biological and social maturity.
• Proposed mechanism: improved food/disease environment → earlier hormonal development → earlier onset of risk-taking behavior → earlier peak in accident-type mortality.

Method

• Gompertz residual method: Fit Gompertz $M(x) = a\exp(bx)$ to ages 40–80 (senescent regime); compute residuals at ages 10–40; fit cubic spline to residuals; identify peak residual age as hump timing.
• Data: Human Mortality Database; 10-year period averages to reduce noise; Sweden from 1751, others from mid-19th century.
• Corroboration: Historical choir records (Bach Leipzig 1727–1749, Ley 1954); comparison with female menarche secular trend.

Concepts Introduced or Extended

• Young Adult Mortality Hump — extended with a developmental-timing interpretation: hump peak as a clock for male sexual maturity onset; documents secular shift in timing; Gompertz residual method as an alternative to Sum of Smooth Exponentials (SSE) decomposition.
• Sexual Maturity Secular Trend — core concept of the paper; ~0.19 y/decade male decline paralleling menarche trend.
• Life-History Theory — invoked to explain sex differential in secular plasticity (female more responsive than male to environmental improvement).

Entities Mentioned

• Joshua R. Goldstein
• Human Mortality Database

Quotes

"We find a secular decline in the age of the accident hump of about 0.19 years per decade, with a halt in the trend since about 1950."

"The accident hump can be thought of as a behavioral signature of puberty, in particular the risk-taking behavior associated with the surge in testosterone that accompanies male maturation."

"The timing of male voice change provides another indirect measure of male sexual maturity ... the average age of voice change in Bach's choir (1727–1749) was about 18 years old compared with about 13 years in a 20th-century London choir."

My Take

The Gompertz residual method is an elegant workaround for the near-total absence of historical male puberty data, but the key assumption — that the lag between pubertal onset and peak accident-hump mortality is constant across centuries and countries — is asserted rather than tested. If risk-taking behavior is also affected by social institutions (compulsory schooling, military service, urbanization), the hump peak would shift for reasons unrelated to biology. Goldstein acknowledges this partly via the England/Wales exception but does not rule out confounding from cohort-level behavioral change. The post-1950 halt is the most striking and robust finding; the pre-1950 trend is noisier (especially for smaller countries). The voice-change corroboration is compelling but thin (two data points). The paper's framing of a "social mismatch" between biological and social maturity is speculative but fertile for future research.