Bongaarts 1999 — Fertility Decline in the Developed World Where Will It End

Summary

Bongaarts argues that observed period total fertility rates (TFRs) in the developed world overstate the actual decline in completed fertility because of tempo distortion: as women postpone births, the mean age at childbearing rises, mechanically depressing the period TFR below true quantum. Applying the Bongaarts-Feeney (B-F, 1998) tempo adjustment to seven countries over 1985–89 yields an adjusted TFR averaging $2.03$ versus an observed $1.78$, a gap of $+0.25$ births. Survey evidence on expected family size is consistent with the adjusted figure: across 15 developed countries and all age groups, expected births remain near or above $2.0$. The paper concludes that fertility is unlikely to fall much further, and may rebound once the postponement wave ends.

Key Claims

• Tempo distortion inflates the apparent fertility decline: Period TFR falls when women delay births (rising mean age at childbearing), because births are shifted out of the observation window. This does not reflect a change in lifetime fertility intentions.

• Table 1 — B-F tempo adjustment, 7 countries (1985–89 avg):

  Country	Observed TFR	Adjusted TFR*	Gap
  France	—	—	$+0.40$
  Netherlands	—	—	$+0.40$
  United States	—	—	$+0.08$
  7-country avg	$1.78$	$2.03$	$+0.25$

  Countries with the most rapid postponement (France, Netherlands) show the largest gap. The US, with comparatively stable timing, shows the smallest.

• Table 2 — Expected family size near/above 2.0: Survey data from $\approx 15$ developed countries show that expected ultimate births are close to or above $2.0$ across all age groups (teens, 20s, 30s), reinforcing that reproductive preferences have not collapsed.

• Forecast implication: If postponement eventually stabilizes, $r \to 0$ and observed TFR will converge toward tempo-adjusted TFR ($\approx 2.0$). The long-run fertility level is more likely to stabilize near replacement than to continue declining.

• Caveat: The paper does not distinguish parity-specific tempo effects and uses the aggregate B-F adjustment, which is later shown by Kohler and Ortega (K-O, 2002) to be subject to compositional bias from the parity distribution.

Concepts Introduced or Extended

• Tempo Effects in Fertility
• Demographic Transition

Entities Mentioned

• John Bongaarts
• Griffith Feeney

Quotes

"The quantum of fertility has declined only slightly and remains well above the replacement level when estimated properly."

"Fertility in the developed world is unlikely to decline much further, and may, in fact, increase somewhat in the near future."

My Take

This is the key empirical application of the Bongaarts-Feeney (1998) framework to cross-national developed-world fertility data. Its main strength is packaging the tempo-correction argument in a concise American Economic Review (AER) format accessible to economists. Its main limitation — acknowledged by subsequent literature — is that the aggregate B-F adjustment used here conflates parity-distribution effects with true tempo effects; the Kohler-Ortega (2002) refinement is needed to cleanly separate them. The survey evidence on expected family size is suggestive but does not address the well-known gap between stated reproductive intentions and realized fertility. Despite these caveats, the paper successfully established the "tempo distortion" argument in the mainstream economics literature and set the stage for the K-O critique.