Real wages and purchasing power — Algeria (2009-2022)

Algerian nominal wages rose 73.5% between 2009 and 2022. Real wages fell 10.2%.

This file shows the calculation, year by year, sector by sector: 1,564 observations across 13 years, 12 economic sectors, 3 legal sectors and 4 skill levels. For each row: the published nominal wage, that year's price index, the real wage in 2009 dinars, and both variations — nominal and real — side by side.

What the statistics office does not publish. It publishes wages. It publishes prices. It never publishes their ratio, and that ratio is what answers the only question that matters to an economist, a union, an HR department or a journalist: has purchasing power improved?

The sector detail says what the aggregate hides. Only one sector gains purchasing power over the period — transport and communications, +1.3%. Hotels and restaurants lose 17.9%, community services 15.3%, extractive industries 13.7%, construction 10.3%. And the decline accelerates: −4.9% in 2021, −7.5% in 2022.

The deflator is the national index, never the Algiers city one. The office publishes both, in the same format, distinguished only by a column. Our first version mixed them: it reported −9.1% instead of −10.2%. A result of the right order of magnitude, and wrong. That is exactly the kind of error a buyer cannot spot in a delivered file, which is why the deflator's scope is written in the documentation as much as in the code.

You can redo the calculation. The price index is published to four decimals — not out of fussiness, but because rounded to two it reproduced the real wage only to within two dinars, and anyone recalculating must be able to tell a flaw in their method from a discrepancy in our file. At four decimals, all 1,564 rows reproduce to within a tenth of a dinar. The formula is any deflation's: real = nominal × reference-year index ÷ that year's index.

The number of months behind each annual average is in the file. An index average over three months is not worth one over twelve, and nothing in a flat file would say so. Every year kept here has twelve; incomplete years are dropped rather than averaged — 2026, in progress, among them.

One operation we did NOT perform. The price series carries two bases, 1989 = 100 and 2001 = 100, and we expected to have to chain them. It does not require it: the recent base covers 2002 to 2026 on its own, while wages run only from 2009 to 2022. Chaining would have been a pointless operation, hence a free risk. We verified it anyway, for anyone wishing to extend the series before 2002: the two bases overlap across five years with a ratio of 0.1748 and a spread of 2.8%.

Reference year: 2009, the wage survey's first vintage. Real amounts therefore read in 2009 dinars, and 2009 carries real values identical to its nominal ones — a check you can run in one line.

This product derives from our "Wages by sector and skill level" and "Consumer Price Index — continuous monthly series" datasets. Buying those separately gives you the ingredients; this one gives you the dish.