What were the official U.S. unemployment rates by year from 1929 to 1936 according to BLS and historical reconstructions?
Executive summary
The Bureau of Labor Statistics’ reconstructed annual unemployment series for 1929–1939—published in a 1948 technical note and later loaded into BLS databases—provides the “official” annual unemployment rates used by most historians; independent reconstructions (notably Lebergott, Darby and later NBER compilations) produce very similar but sometimes materially different peak values for the early 1930s (the worst years of the Great Depression) [1] [2] [3]. Below are the BLS annual estimates for 1929–1936 followed by a note on alternative historical reconstructions and their interpretation.
1. Official BLS reconstructed annual unemployment rates, 1929–1936
The BLS’s reconstructed annual unemployment series (the estimates produced in the 1948 Monthly Labor Review article and later published online) lists the annual unemployment rates as follows: 1929 — 3.2%; 1930 — 8.7%; 1931 — 15.9%; 1932 — 23.6%; 1933 — 24.9%; 1934 — 21.7%; 1935 — 20.1%; 1936 — 16.9% (these figures come from the BLS historical reconstruction that became the official pre‑1940 series) [1] [4].
2. Why these are “reconstructed” and what that means for comparability
Those percentages are not direct outputs of the Current Population Survey (CPS) household method that became the official monthly series in 1940; instead, the BLS in the 1940s synthesized multiple administrative and survey sources to approximate CPS concepts back to 1929, and the 1948 article documents the estimation methods and assumptions used to make the series comparable to later CPS‑based measures [1] [2]. The BLS explicitly warns that pre‑1948 estimates refer to a different age population (14 and over in earlier years) and are not seasonally adjusted in the same way as post‑1948 data, so year‑to‑year comparisons across the 1940 boundary require caution [2].
3. Independent reconstructions and scholarly alternatives
Scholars have produced alternate annual unemployment reconstructions—Stanley Lebergott and Michael Darby among the most-cited—using different source mixes and adjustments; those series broadly mirror the timing of the BLS peaks and troughs but can differ in peak magnitudes and in exactly how fast unemployment falls after 1933, because of differing definitions of “unemployed” and denominator choices for the labor force [3] [5]. NBER and FRED host historical series that compile BLS, NBER and other authors’ estimates; these repositories make comparisons straightforward and reveal that while the qualitative story (massive rise in 1930–33, partial recovery thereafter) is robust, precise percentages can vary by a few points depending on the reconstruction [4] [6].
4. The most contested years and why 1932–1933 matter
The early 1930s are where reconstructions diverge most—1933 is commonly cited as the high‑water mark, with the BLS reconstruction peaking at 24.9% in 1933, a figure that appears in many secondary summaries and databases [7] [1]. Disagreements reflect methodological choices: whether to count marginally attached workers, how to treat farm employment and seasonal patterns, and differing numerator/denominator labor‑force definitions drawn from incomplete administrative records and state UI data prior to nationwide reporting standards [1] [5].
5. How to read these numbers in historical context
The reconstructed BLS percentages should be read as best‑effort, conceptually harmonized estimates—useful for broad historical comparison and narrative but not precise down to tenths of a percentage point; the BLS and historians alike encourage use of multiple series (BLS official reconstruction, Lebergott, Darby, and NBER compilations) to bracket uncertainty and to understand how definitional choices move the estimates [2] [3] [4]. For practical purposes, the BLS reconstruction (3.2% in 1929 → near‑25% peak in 1933 → falling to 16.9% by 1936) provides the canonical chronology most cited in policy and secondary accounts [1] [4].
6. Limits of available reporting and where to go next
Primary documentation of the BLS methodology is the 1948 Monthly Labor Review technical note and the BLS historical pages; researchers seeking variant runs should consult Lebergott , Darby , NBER data tables and FRED’s historical series, all of which are cited and linked in the archival literature and modern databases [1] [3] [4]. This account sticks to published reconstructions and scholarly series; if a reader needs month‑by‑month series or alternative labor‑force denominators, those require consulting the specific NBER or academic tables that re‑weight or redefine unemployment for different conceptual exercises [4] [3].