Tens of thousands of years ago, "productivity" was pretty much a function of weather, and "business cycles", insofar as the term can be applied, followed productivity very closely, both in time and in linear correlation. In the last couple centuries of the second millennium, a drop in expected future demand led to a pullback in investments, both in capital and in new labor relationships, resulting in unemployment, resulting in a drop in demand. That we aren't living in caves eating berries owes much to our stockpile of capital, both physical and intellectual, but the greater the share of our economy that is devoted to producing durable goods (including capital), the more sensitive our near-term production is to expectations of slightly less near-term production, leading both to a greater variation in production resulting from a particular exogenous fluctuation in productivity and to a greater variation in production resulting from nothing at all.
I think I've mentioned a couple of times that I believe the most underappreciated macroeconomic stylized fact of my lifetime is the doubling of the portion of GDP going to depreciation, and I've even suggested (I believe) that this is part of why the 1990 and 2000 recessions were long and shallow rather than sharp and steep; perhaps without the shift away from producing long-duration capital, the 2008-2009 collapse would have been closer to the scale of 1929-1933, better response by the Federal Reserve notwithstanding. Perhaps this is related to the general deceleration in growth over that time; perhaps it will also, though, improve our ability to make forecasts going forward, insofar as they depend more on "fundamentals" and less on self-fulfilling expectations.