I think you already know the answer: people reacting to events that increase uncertainty, and in environments with “low visibility”, game theory dictates that the blind follow the blind.
From another excellent paper by Bradford Cornell,

On a day to day basis, however, only a small minority of largest moves are associated with what appears to be unexpected fundamental information regarding the economy. The mystery is only deepened by the offsetting nature of many of the largest moves.
The results reported here largely mirror those published more than two decades ago by Cutler, Poterba and Summers. Despite the explosion in information technology, enhanced market regulation, innovation in stock trading and the introduction of new equity related financial products, large movements in the market remain as common and mysterious as ever. Only a minority of the 50 largest moves in the last 25 years can be tied to fundamental economic information that could have had a pronounced impact on cash flow forecasts or discount rates. If anything, the mystery has deepened because the size of the unexplained market movements has grown.
The most significant difference between the current results and those of Cutler, Poterba and Summers is that the large changes are more bunched in the current sample. More than half of the 50 largest moves occur in 2008 and 2009. Furthermore, 21 of the 50 largest moves occur on what are basically strings of consecutive days. This bunching further deepens the mystery because, with one exception, when large moves occur on consecutive days, the signs oscillate, often for no apparent reason. It is as if investors significantly alter their long-run views regarding cash flows and discount rates only to change their minds the next day.

(Nothing displayed below?)
[pdf width=”100%” height=”1100px”]https://s3.amazonaws.com/2013-wealthcop-PDF/Cornell.2012-What-Moves-Stock-Prices.pdf[/pdf]