Employment and unemployment — Algeria's labour force (2000-2019)

Algeria's labour force across twenty-two surveys, June 2000 to May 2019: labour force, employed, unemployed, and the breakdown of the employed by status — employers and self-employed, permanent employees, non-permanent employees and apprentices, family helpers. Broken down by sex from 2009, and by area of residence throughout.

What you cannot do without this file. Each survey is published in its own bulletin, never joined to the others, and twice a year — April and September — without the table title saying which. Two surveys from the same year therefore carry the same heading. The month was read from the body of each document, taking the first one cited whose year matches the document's: the text also cites earlier surveys for comparison, and taking the first month encountered would have dated the 2012 survey to September 2011.

Four arithmetic identities, verified in every column of every table. The four statuses sum to the employed population; employed plus unemployed give the labour force; the total column is the sum of those preceding it; and the published unemployment rate is indeed unemployed over labour force. These checks are independent: a row shift breaks the first, a column shift the third, a period mix-up the fourth. All twenty-four documents pass all four.

Two independent sources, and their confrontation. The Statistical Retrospective 1962-2020 republishes the same series years after the bulletins. Two distinct documents, two table structures, two extraction paths. Across the nineteen shared periods: nineteen agreements, zero divergences. The confirme_par_deux_sources column marks the sixty observations that benefit. The retrospective contributes June 2000 on its own — unemployment at 28.9%, the series peak — and April 2016.

The unit trap, worth a factor of a thousand. Bulletins through 2008 count in persons ("10,315,000"), later ones in thousands ("11,048"), and some say so nowhere. Every figure here is restated in persons; the unite_source column preserves the original document's unit.

What the file does not hide. One cell carries precision_faible: the office had published it in parentheses, flagging a figure estimated on too few observations. It ships with its caveat rather than being silently dropped or presented as solid.

A vocabulary change worth knowing. Through 2008 the breakdown is "aggloméré" and "épars"; from 2011 it is "urban" and "rural". The two are close but do not rest on the same definition, and they have not been merged: each row carries the label from its own publication. Whether they are comparable for your purpose is your call — we have not made it for you.

Coverage: June 2000, then September 2003 to May 2019 (22 surveys). 2001 and 2002 are absent — their document does not extract cleanly. The unemployment rate recomputed from these figures runs from 28.9% in 2000 to 11.4% in May 2019, with female unemployment at roughly twice the male rate throughout the recent period.

Our "Wages by sector and skill level" product covers the earnings side of these same years.