Well the first year of the Iraq War cost the US $54 billion, according to congress's budget[0]. This doesn't include the total cost of the supporting infrastructure need to be able to deploy troops in Iraq quickly, but we can estimate that using the increase in defence budget from 2002-3, or $94 billion ($132B in 2020)[1].
According to Wikipedia, Minuteman III ICBMs have a 2020 unit cost of $20 million[2], so for the cost of an Iraq invasion, the US could have fired about 6600 missiles. Considering the invasion toppled the Iraqi government, it's pretty unlikely that firing 6600 missiles with conventional payloads would have been anywhere near as effective.
The comparison we're making is whether precision attacks, presumably on roughly building-sized targets, would be cheaper to do from long range via ICBMs (with conventional warheads), or via much cheaper but shorter-range missiles. My guess is that neither ICBMs nor shorter-range missiles could have accomplished what the U.S. military accomplished in Iraq. Presumably missiles alone were responsible for a small portion of that $54 billion.
If I can trust https://en.wikipedia.org/wiki/LGM-30_Minuteman a Minuteman III (which is the current ICBM design used by the US) will land within 800ft (240m) of its intended target 50% of the time. And outside that circle the other 50%.
In other words, you can't really target a "building-size" target with these (with maybe exceptions like the Pentagon).
For nuclear payloads, a few hundred meters of error is much less of an issue, of course.