<p>^ I read it. Four times now. As I see it you are saying (good ol’ numbered points again):</p>
<p>1: The segment of the population that has insurance has a certain rate of ER visits.</p>
<p>2: The segment with no insurance has a higher rate of ER visits than segment #1.</p>
<p>3: If segment #2 were to get insurance, their rate of ER visits would drop.</p>
<p>4: It would drop by more than 4.75%, the number I said was the max in post #253.</p>
<p>For this to be true, the (now insured) uninsured would have to have a much lower rate of ER visits than the (still insured) insured people had before we changed anything.</p>
<p>Going back to my example, for the rate of ER visits to drop by 10%:</p>
<p>Before law: 1000 people, of whom 160 are uninsured. 100 go to the ER, of whom 20 are uninsured.</p>
<p>So, just like before:</p>
<p>840/80 = 1 in 10.5 insured people.</p>
<p>160/20 = 1 in 8 uninsured people.</p>
<p>Now, suppose we think making the uninsured people get insurance will cause the number of ER visits to drop to 90. The insured people haven’t had anything changed, so we will still have 1/10.5, or 80, of them. That leaves 10 that can come from our formerly uninsured group, which numbers 160. So…</p>
<p>160/10 = 1 in 16 people from the formerly uninsured group go to the ER.</p>
<p>But 1 in 10.5 of the people who had insurance before go to the ER! So this must mean that the formerly uninsured people are only 2/3 as likely to go to the ER as the rest of the population! This is absurd, unless you think having lacked insurance in the past improves your health once you get it.</p>