I am measuring the impact of offshoring to China on domestic relative labour demand. 

For this I constructed two measures of offshoring with data from a world input-output table. They are both broad measures, not narrow, which means that it includes all offshored goods form industry i to all industries except i in China. The first broad offshoring measure of industry i in country c contains all imported intermediate inputs of industry i in country c from industry j (i≠j), summed over all j in China. The second measure of offshoring of industry i in country c contains all exported intermediate inputs from industry i in China to industry j (i≠j) in country c, summed over all j. This second measure captures the effects of offshoring on labour when the offshored good was first outsourced domestically. 

The two measures are thus constructed of exactly the same data. My question is, will including them both in a regression lead to sort of double counting? I think that not including them both might also lead to biased results as one measure can pick up the effect of the other.