The take-away: the case for downsizing CLE isn't nearly as obvious as it is for CVG or MEM, but is still pretty compelling.

Domestic O/D passengers, in millions (rank among US airports) {O/D pax per average departure}
LAX 31.1 (2) {120}
ORD 28.8 (3) {68}
DEN 22.7 (7) {75}
EWR 19.8 (12) {119}
SFO 16.5 (16) {110}
IAH 13.5 (23) {54}
IAD 10.3 (28) {76}
CLE 7.4 (40) {66}
CVG 4.3 (61) {27}
MEM 4.2 (62) {42}

The {last figure} can be thought of as "how many local pax are on the average flight?" -- except that you should assume it's actually half that figure, since the divisor counts only departures and not arrivals. It's a wonky number, but it makes all the difference; not long before AA shut its hub at RDU, I remember that AA officials told the local newsmedia, "just two more local passengers on every flight and we'll keep it open." Well, in the end, they left, going from 200+ to 45 flights a day. They're back at about 70/day now.

As you can see, CVG is pretty much the closest thing in America to a pure transfer station (perhaps excluding some FedEx/UPS sites), whereas growth at LAX and EWR can be supported by O/D pax alone.

For comparison, some O/D counts for downsized AA or US hubs:
STL 10.2 (29)
SJC 9.5 (32)
RDU 8.3 (37)
BNA 7.8 (38)
PIT 7.5 (39)

We all agree that DEN and IAD have additional capacity -- plus stronger O/D, stronger local economies, and better international connections. Even IAD, the next smallest hub on the CO-UA list, has 40% more O/D than CLE (and could potentially grab more of DCA's 13.6M O/D pax), while DEN's local market is 300% larger. BTS gives us a look at the 2007 passenger traffic ex-CLE. By far the most popular destination for Clevelanders, with nearly 1/8 of departing passengers? Chicago. CLE accounts for just under 5% of CO mainline passengers, a share that would shrink to under 2% in CO-UA. It's less than half the size of UA's mainline operation at IAD.

Data are from Eclat Consulting and BTS; passenger counts are year ending 31 Mar 2007, average departure counts are for Jul 2007.