Spain’s Statistical Office (Instituto Nacional de Estadística , INE) provides yearly series of ‘resident’ population from 1971. INE also presents annual series of ‘de facto’ population for 1900–1991, in which figures for census benchmark years are linearly interpolated. Roser Nicolau (2005) collected and completed the series back to 1858. More recently, Jordi Maluquer de Motes (2008) has constructed yearly estimates of ‘de facto’ population for 1850–1991 and spliced them with ‘resident’ population for 2001. In order to do so, Maluquer de Motes started from census figures at the beginning of each census year adding up annually the natural increase in population (that is, births less deaths) plus net migration (namely immigrants less emigrants ). I have followed Maluquer de Motes’s approach with some modifications. Thus, I have accepted census benchmark years’ figures and Gustav Sündbarg (1908) estimate for 1850 and obtained the natural increase in population with Nicolau (2005) figures for births and deaths from 1858 onwards, completed for 1850–1857 with Sündbarg (1908) net estimates at decadal averages equally distributed. 1 My main departure from Maluquer de Motes approach has been with regards to net migration for which I have accepted Blanca Sánchez-Alonso (1995) estimates for 1882–1930, completed back to 1850 and forth to 1935 with statistical evidence from Spanish and main destination countries’ sources (see Sect. 7.3.4). For the years of the Civil War (1936–1939) and its aftermath (1940–1944), I have accepted José Antonio Ortega and Javier Silvestre (2006) gross emigration estimates for 1936–1939, assuming no immigration during the war years, and distributing evenly an upward revision of their return migration estimates for 1940–1944, while assuming no gross emigration during World War II . 2 In order to obtain a consistent series for 1850–1970, I have spliced population estimates linearly by distributing the difference between the estimated population obtained by forward projection of the initial census benchmark figure for the year of the next census benchmark and the observed figure at the new census using expression (16). Lastly, I have linked the linearly interpolated series for ‘de facto’ population for 1850–1970 with the ‘resident’ population series from 1971 onwards to get a single series. 3 Fortunately, the difference between the ‘de facto’ and ‘resident’ series over 1971–1991 is negligible. 4

FormalPara Notes
  1. 1.

    Sündbarg (1908) estimates are reproduced in Maluquer de Motes (2008: 145). I have used the average birth and death rates in 1858–1860 for the years 1850–1857, except in the case of 1855–1856 for which the death rate (45 per 1000) estimated for 1855 as a consequence of cholera epidemics by Pérez Moreda (1980: 398) has been used. I have also used the average of birth and death rates in 1870 and 1878–1880 for the years 1871–1877 in which data on total births and deaths are missing.

  2. 2.

    Ortega and Silvestre (2006) consider the 162,000 net migration figure during 1940–1944 grossly underestimated. Pérez Moreda (1988: 418) reckoned a maximum permanent exile of no more than 190,000 people, a figure below the 200,000 provided by Tusell (1999) and much lower than a post-Civil War exile estimate (300,000) (Tamames 1973). I have accepted Pérez Moreda’s conjecture.

  3. 3.

    Choosing ‘resident’ over ‘de facto’ population allows me to keep consistency with Spanish official national accounts, which employ ‘resident’ population.

  4. 4.

    The average ratio between the resident and de facto population over 1971–1991 is 0.9956 with a coefficient of variation of 0.0048.