* stata_codebook.do - attach long-form notes to the .dta files (run once in Stata).
* Generated by build_data_dictionary.py - do not edit by hand.

* ---- source_data.dta ----
use "source_data.dta", clear
label data "Basque regional panel: GDP + 13 synthetic-control predictors"
note _dta: The basque dataset bundled with the Synth R package (Abadie, Diamond & Hainmueller 2011); originally compiled by Abadie & Gardeazabal (2003). 18 units x 43 years = 774 rows, 1955-1997. Region 1 is the Spain national aggregate (dropped); region 17 is the treated Basque Country (terrorism onset 1970). gdpcap is observed every year; the predictor columns are pooled period averages / snapshots and are therefore sparse. Sectoral and school variable names were dotted in the original (sec.agriculture) and are renamed to underscores here for Stata validity; values unchanged.
note regionno: Integer identifier for the regional unit; 1 is the Spain national aggregate, 2-18 are the 17 autonomous communities (17 = Basque Country, treated).. Construction: From the bundled basque dataset.. Units: 1-18. Source: Synth package (basque)
note regionname: Name of the regional unit (e.g. 'Basque Country (Pais Vasco)', 'Cataluna', 'Madrid (Comunidad De)').. Construction: From the bundled basque dataset (some names stored without diacritics).. Units: string. Source: Synth package (basque)
note year: Annual time index, 1955-1997. Terrorism onset (treatment) is dated to 1970.. Construction: From the bundled basque dataset.. Units: year. Source: Synth package (basque)
note gdpcap: Outcome variable: real GDP per capita in thousands of 1986 US dollars.. Construction: Observed regional GDP per capita; this is the outcome the synthetic control matches and predicts.. Units: 1986 thousand US$. Source: Abadie & Gardeazabal (2003) via Synth
note sec_agriculture: Share of regional production in agriculture (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.agriculture.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note sec_energy: Share of regional production in energy (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.energy.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note sec_industry: Share of regional production in industry (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.industry.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note sec_construction: Share of regional production in construction (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.construction.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note sec_services_venta: Share of regional production in marketable ('venta') services (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.services.venta.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note sec_services_nonventa: Share of regional production in non-marketable ('nonventa') services (sectoral predictor).. Construction: Pooled 1961-1969 sectoral average; one value per region. Originally sec.services.nonventa.. Units: % of production. Source: Abadie & Gardeazabal (2003) via Synth
note school_illit: Number of people with no schooling / illiterate (education predictor, raw count).. Construction: Education-level figure; converted to a within-region percentage share before matching. Originally school.illit.. Units: count. Source: Abadie & Gardeazabal (2003) via Synth
note school_prim: Number of people with primary education (education predictor, raw count).. Construction: Education-level figure; converted to a within-region percentage share before matching. Originally school.prim.. Units: count. Source: Abadie & Gardeazabal (2003) via Synth
note school_med: Number of people with intermediate/secondary education (education predictor, raw count).. Construction: Education-level figure; converted to a within-region percentage share before matching. Originally school.med.. Units: count. Source: Abadie & Gardeazabal (2003) via Synth
note school_high: Number of people with higher education (education predictor, raw count).. Construction: Education-level figure; collapsed with school_post_high and converted to a share before matching. Originally school.high.. Units: count. Source: Abadie & Gardeazabal (2003) via Synth
note school_post_high: Number of people with post-higher education (education predictor, raw count).. Construction: Education-level figure; collapsed into school_high and converted to a share before matching. Originally school.post.high.. Units: count. Source: Abadie & Gardeazabal (2003) via Synth
note popdens: Population density at the time of treatment (1969 cross-section; static control predictor).. Construction: Single 1969 value per region. Originally popdens.. Units: people / km^2. Source: Abadie & Gardeazabal (2003) via Synth
note invest: Investment as a share of GDP (predictor).. Construction: Annual investment-to-GDP ratio; pre-treatment values enter the predictor set.. Units: % of GDP. Source: Abadie & Gardeazabal (2003) via Synth
save "source_data.dta", replace