And some days pass on by
I'll be working here forever
At least until I die
--Huey Lewis and the News
Study of productivity using GDP per capita data rather than the standard BLS productivity series in part to review levels and trends over a longer period of time (BLS productivity tracking began in the late 1940s).
The most common measure of productivity is a ratio of output to labor resources that produce output (a.k.a. partial factor productivity with respect to labor). In the numerator is a measure of output (such as GDP). In the denominator is a measure of labor.
In the BLS series, an estimate of hours worked is used. This comes with obvious estimation challenges as many if not most American workers do not report their hours worked. An advantage of a per capita productivity measure is that focuses on straight head count of the citizenry--presumed to be less subject to error.
However, head count measures of productivity have their own limitations. Viewing output relative to the entire population ignores the fact that only a fraction of all citizens are working to produce that output, and that fraction might change over time. Indeed, the workforce participation rate has been declining for decades. Moreover, headcount measures fail to account for the possibility that workers may work more or less time to produce a given amount of output.
Nonetheless, output per capita provides a useful lens for evaluating productivity over time. It reflects an 'effective productivity' of sorts because it reflects output per all consumers given extant social constraints on production.
This is wholly consistent with capital theory. Saving has been declining. Fewer and fewer resources have been set aside for productivity improvement projects.
Stated differently, as capital is consumed, productivity will decline.