Most recessions are preceded by several months in which unemployment is rising significantly from the relative low. Its probably the most reliable recession predictor (visually, looking at a graph, the oscillations between the closely-spaced 1980 & 1981 recessions seems to be the only case of a recession not having that signal, and there are very few cases of a clear multimonth rise after the (usually shortly post-recession) recession-associated peak that do not signal another recession.
Most recessions are preceded by several months in which unemployment is rising significantly from the relative low. Its probably the most reliable recession predictor (visually, looking at a graph, the oscillations between the closely-spaced 1980 & 1981 recessions seems to be the only case of a recession not having that signal, and there are very few cases of a clear multimonth rise after the (usually shortly post-recession) recession-associated peak that do not signal another recession.