Extrapolaholics Anonymous

Abstract

Criticisms of demographers by other demographers have become frequent in scientific literature, generally consisting of accusations that trends observed in the recent past have been extrapolated unjustifiably into the future. Demographers, along with their colleagues in the actuarial profession, are in an invidious position in this regard, knowing full well that extrapolation is almost always only minimally justifiable, but knowing also that their readers, colleagues, and sources of funding tend to be much more interested in the future than in the past. It is unfortunate that, while actuaries typically resolve this dilemma by emphasizing the limitations of their methods and thereby lowering expectations that their predictions will be accurately fulfilled, demographers are more prone to respond combatively, attempting to reinforce the credibility of their extrapolations by recourse to data from areas in which their expertise is less tested, such as biology. This is valuable in that it raises the profile of the debate on the likely rate of scientific progress relevant to mortality rates, but it also runs the risk of lowering the technical quality of that debate, by telling policy makers and the public what they want to hear and thereby entrenching their expectations without recourse to the relevant biological facts. Extrapolations based on plausible sequences of scientific advances and the sociopolitical responses to them, summarized in this article, have led to the prediction of four-digit life expectancies of cohorts born in the 21st century and possibly even in the 20th. This prediction has attracted inevitable ridicule from prominent demographers, but being founded on science and sociology rather than on history it may be much more reliable than the extrapolations that those demographers presently prefer.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 14 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.