This is not my chart. This is Statistics Canada. If they produced false data, then I guess we should ask them.
You keep saying this is a direct relationship between unemployment and crime. What the chart is showing is that differences in unemployment seem to correlate very strongly with differences in robbery. These are the two parallels. The point of the chart was to show that variances in social conditions can explain variances in crime quite effectively, while there is no similar chart that anyone can produce that would show similar variations in sentencing affecting variations in crime.
My point was a very simple one, and this is only illustrative of one of many factors that could be contributing to crime. If we want to reduce crime and are very serious about it, are we going to put all our attention and effort into measures for which there is no evidence of their having any impact, or are we going to concentrate our efforts on those factors that appear to have a strong correlation?