So I found this
There seems to be a pretty tight correlation between retail sales as a fraction of disposable personal income (left axis) and the average weekly hours worked (right axis).
In the long run they are both declining, of course, but its not immediately clear why they should be declining at proportional rates. Over the business cycle it makes sense if there is causation running from consumption to hours work. People try to save more and additional savings leads to less work.
It makes less sense if the correlation is running from optimal hours worked to spending. In that case we would expect consumption to fall during downturns, people are choosing less work which leads to less income which leads to less consumption. However, consumption smoothing would suggest that consumption as a fraction of income should rise. In this graph our proxies for consumption spending as a fraction of income falls.
Now it could be that consumption is smoothed through non-retail channels: health care, service streams from existing housing and durables, etc. I am not sure, however, that when all the i’s are dotted and t’s crossed that this is theoretically consistent.
Here are the correlations
and the first difference