Because it's a really good coincident indicator. Let's look at industrial production (the most visible manifestation of manufacturing activity and also a stat for which we have a long series of historical data) and compare it to GDP. The data series below have been normalized so that both real GDP and IP are based on 100. I've also made the IP data quarterly by summing the monthly data.
The above chart shows the relationship between industrial production and real GDP. While correlation does not equal causation, these two data series have a similar trajectory for most of the other's existence.
The above three scatter plots show the relationship between the year over year percentage change in industrial production with the year over year percentage change in real GDP. As above, the IP information was placed into quarters by summing each three month series and both data series were "normalized."
As with the first chart, notice the clear positive relationship between industrial production and real GDP; when IP increases, do does GDP while the reverse is also true.
