The 2012 and 2013 columns are empty, not zero

What do an export's blank last years do to a chart? Each row is one of twelve countries in one year from 1960 to 2013, with its total fertility rate or a blank. The file's headers run to 2013, but no country has a value after 2011. A chart that reads blanks as zero shows fertility collapsing in a year; one that carries 2011 forward invents two years of data. This export keeps the blanks blank.

Interactive chart available when JavaScript loads.

Every path stops at 2011

Each path is one country's last twelve years, and every one ends at the same point. The axis runs to 2013 because the file does, and the shaded band is what the file holds there: nothing. In the 2012 rows, 100.0% of values are missing. In the 2011 rows, 0.0% are.

Interactive chart available when JavaScript loads.

Two blank columns in a full grid

Each cell is one country-year, dark where the file holds a value. Fifty-two columns are full and the last two are empty for every country, so the gap is the export's, not any one country's. 624 of the 648 country-years hold a value, and the 24 that do not all sit in 2012 and 2013.

Interactive chart available when JavaScript loads.

Twelve, twelve, twelve, zero, zero

The last five years of the file, counted by whether each country-year holds a value. Twelve countries report through 2011; then the count drops to zero, not because fertility did anything but because the export ends. Overall, 3.70% of the file's cells are blank, every one of them in these two columns.

Interactive chart available when JavaScript loads.

Takeaway

The file ends in 2011; the headers end in 2013. Read the blanks as zero and twelve countries' fertility collapses in a year; carry 2011 forward and two years of data appear from nowhere. Leaving the cells empty is the only honest option, and it costs two columns of white space.

Limits

This is a snapshot of a historical export, so the blanks say the values were unavailable in this file, not that they are unavailable now. Twelve countries chosen for contrast, not a sample.