There are a couple reasons for this misconception. Historians used to use door heights and bed lengths as indicators of height. The problem is that short doorways and beds aren’t indicative of short people. Low doorways are common for heat retention in Northern European houses. And short beds are indicative of how people used to sleep; it was thought that laying flat was unhealthy so people slept partially sitting up, bolstered by pillows. Therefore beds were shorter.
In some periods of history, people were indeed shorter. In the Victorian period (1837-1901), the movement from life on the farms to life in the newly industrial cities led to malnutrition, vitamin deficiencies, and respiratory problems on a scale previously unknown. Measurements of skeletons from cities in this time period are significantly smaller than skeletons from the same cities in the period before and after the Victorian period. Interestingly, the skeletons of country people from the same time period are larger than those of Victorian city people.
Let’s look at something a little more modern. Would it shock you to know that the base measurements in pattern books from 1916 are 34-26-42? (That's six inches larger in the hips than the modern analogous size!) Tailor's books and patterns from the late teens and early 1920s don’t go below a size 32" bust. And that's considered a "teen" size. The modern size charts of one of the "Big Three" pattern companies starts adult female sizes at a bust 29.5"!
But let's look at sizes on early 1920s patterns. Here's a chart:
|20s bust||20s hip||now bust||now hip|
That's a three to six inch difference (depending on bust size) between bust and hip measurements as opposed to a mere two inch difference between bust and hip (regardless of size) nowadays.
So they weren’t smaller “back then”. But maybe their sizing was a little more realistic than what we’re used to today.
Werner, A. London Bodies: the changing shape of Londoners from prehistoric times to the present day. 1998: Museum of London, London.