The Human Fertility Database
Explanatory notes
The HFD is entirely based on one and the same type of initial data  officially registered birth counts by calendar year, mother's age (and/or cohort) and (whenever possible) biological birth order. These data, together with total female population exposure from the Human Mortality Database (www.mortality.org) and parityspecific female population exposure from selected population censuses, population registers, or largescale surveys, are further processed using a uniform set of methods. The major HFD output includes detailed data on births, unconditional and conditional fertility rates, cohort and period fertility tables as well as selected aggregate indicators such as total fertility rates, mean ages at childbearing, and parity progression ratios.
For each country, there are four blocks of data provided:
Summary Indicators
This data block consists of two subblocks distinguishing summary indicators of period and cohort fertility.
Period summary indicators include the following data:
 Total number of live births.
 Crude birth rates.
 Total fertility rates (including total fertility rates by age 40).
 Tempoadjusted total fertility rates, using Bongaarts and Feeney's method.
 Mean ages at birth (including mean ages at birth by age 40).
 Standard deviation in the mean age at birth (including the standard deviation in the mean age at birth by age 40).
Cohort summary indicators are as follows:
 Completed cohort fertility (including the completed cohort fertility by age 40).
 Parity progression ratios.
 Mean ages at birth (including mean ages at birth by age 40).
 Standard deviation in the mean age at birth (including the standard deviation in the mean age at birth by age 40).
These summary indicators are, where appropriate, calculated on the basis of fertility rates by age; and, when available, by birth order. In the case of period summary indicators, these rates are sorted by Lexis squares, and for producing cohort summary indicators, horizontal parallelograms are used.
Detailed explanations of the Lexis diagram and the Lexis shapes as well as a description of the calculation methods behind the displayed values are given in the HFD Methods Protocol.
AgeSpecific Data
The agespecific data are also presented in two subblocks, separating the period and the cohort dimension. Both subblocks cover the same array of indicators. Period data are provided by Lexis triangles, squares, and vertical parallelograms, while cohort data are organised by horizontal parallelograms.
The fertility indicators provided in this block are as follows:
 Live birth counts for all birth orders combined and (when available) by birth order by all Lexis shapes (Lexis triangles, squares, vertical and horizontal parallelograms).
 Female population exposures by all types of Lexis shapes (Lexis triangles, squares, vertical and horizontal parallelograms).
 Unconditional agespecific fertility rates for all birth orders combined and (when possible) by birth order (by Lexis triangles, squares, vertical and horizontal parallelograms).
 Cumulative fertility rates (by Lexis squares, vertical and horizontal parallelograms).
Fertility Tables
This block contains two subblocks: "Period fertility tables" and "Cohort fertility tables".
The subblock "Period fertility tables" consists of period fertility tables as well census or registerbased fertility tables, specific for age and parity, and provides the following data:
 Fertility tables.
 Female population exposures by age and parity.
 Conditional period fertility rates, controlling for age and parity. These data are extracted from period fertility tables and/or census or registerbased fertility tables featured in this subblock.
 Period table summary indicators, which include parity and ageadjusted total fertility rates (PATFR) and table mean ages at birth.
The subblock "Cohort fertility tables" show the cohort fertility tables.
The period fertility tables are built on the basis of data by Lexis squares, while the construction of the cohort fertility tables involves data by horizontal parallelograms. Period data are indexed by calendar year, whereas cohort data are indexed by year of birth.
Period fertility tables in the HFD include the following variables:
Year 
Calendar year of occurrence of birth 

x 
Age at childbearing in completed years 

w_{i1}(x) 
Weight of parity i1 in the total female population exposure in age interval [x, x+1) 

m_{i}(x) 
Conditional agespecific fertility rates (occurrence/exposure rates) in age interval [x, x+1) 

q_{i}(x) 
Probability of giving an ith birth in age interval [x, x+1) 

l_{i1}(x) 
Table population (table cohort size) of parity i1 by age x 

b_{i}(x) 
Table number of births of order i in age interval [x, x+1) 

L_{i1}(x) 
Table population exposure of women of parity i1 in age interval [x, x+1) 

Sb_{i}(x) 
Cumulative (in respect to age) births of order i by age x 
Population weights of parity i1, w_{i1}(x) are assessed in two ways depending on data availability.
 For countries that have sufficiently long time series of orderspecific period data on births, weights w_{i1}(x) are reconstructed from the time series of period data by cumulating fertility of birth cohorts over their reproductive life span.
 For countries where orderspecific data on births are available for a short period only, population census or population register data are employed to derive the initial (for one starting year, 'base year') ageparity distribution of women. The selected census or register from which data are used is called the 'golden' census. In exceptional cases, data from a largescale survey can be considered for the 'golden census'.
Census or registerbased period fertility tables include the same life table functions as the period fertility tables. In these life tables, however, the weights w_{i1}(x) are based on selfreported or officially estimated information on parity of women originating from censuses or population registers.
The census or registerbased period fertility tables are calculated for each census year for which suitable data are available. The population register data are treated the same way  the period tables are computed for all years covered by the register.
Estimation of the population weights and period fertility tables are described in the Methods Protocol.
Cohort fertility tables comprise the following columns (functions):
Cohort 
Mother's year of birth 

x 
Age at childbearing in completed years 

i 
Birth order of child (15+) relates to parity i1 (04+) 

b_{i}(x) 
Table number of births of order i in age interval [x, x+1) 

l_{i1}(x) 
Table population (table cohort size) of parity i1 by age x 

q_{i}(x) 
Probability of having an ith birth in age interval [x, x+1) 

m_{i}(x) 
Conditional agespecific fertility rates (occurrence/exposure rates) in age interval [x, x+1) 

Sb_{i}(x) 
Cumulative (in respect to age) births of order i by age x 

chi(x) 
Average number of children by age x in the highest parity category i_{4+} 
Note that i is the birth order varying from 1 to 5+ that relates to parity i1 varying from 0 to 4+.
Calculation of cohort fertility tables is described in the Methods Protocol.
To facilitate rapid downloads of large amounts of (output) data, the Human Fertility Database offers two series of zipped data files: "By data type" and "By country".
Input data
All input data used for the computation of output fertility indicators are provided on the country pages. These data include the raw data originally collected for each country, and which have been converted into a standardised format (described in the Data formats document). For every country included in the HFD, there are four files, which contain data on:
 Births by calendar year, the age of the mother, (when available) the mother's year of birth, and (if available) birth order.
 The age and parityspecific distribution of the female population (from population censuses, population registers, or largescale surveys).
 Births by calendar year and month.
 Female population sizes and death counts by calendar year and age (for countries and periods not available in the HMD).
In addition to the data files, the Input Database includes two documents in PDF format. The first one contains Notes regarding specific points or parts of the input data. The second one contains References to sources of the input data.
Information about formats of the respective input and output data files can be found in the Data formats document.