An increate in government borrowing G - T raises the interest rate and thus has the potential to lead to less private investment. Government borrowing can then be said to "crowd out" private borrowing.
The change in G, which is the horizontal movement of the I+G-T curve, can be decomposed into an increase in S and a decrease in I.
It is important to note that, from the point of view of short-run policy implications, crowding out has no effect because capital is taken as fixed. Over the longer run, however, crowding out could certainly lead to a smaller capital stock than would otherwise exist, lowering the potential for production.
How important could this effect be? Consider a worst case scenario. If the deficit is $250 billion and the labor force is 125 million, then the deficit is $2,000 per worker. If unemployment is 5% of the labor force, then the deficit is $40,000 per unemployed worker. The question becomes, could $40,000 per unemployed worker be better invested in education and retraining or in whatever the deficit is financing? Keep in mind that the $40,000 per unemployed worker is only one year's deficit.
Crowding out is clearly much more of a concern when the economy is operating close to capacity and increased resources absorbed by the government have to displace resources that might have otherwise been allocated to the private sector. If there is a great deal of slack in the economy, as there was in the Great Depression, then the increase in government spending might well generate a sufficient increase in output that a deficit would have little effect on private investment. In terms of the present diagram, which shows a static analysis, the change in G could shift the savings curve.
Crowding Out (economics), Wikipedia.