The Rising Influence of Rising Affluence

[Democritus_of_Abdera]

Re: CLF Annual Meeting...

The failure of CLF's current Boad of Directors to schedule the annual meeting is a travesty. It solidifies my decision to not cast any votes for Kirsh when I get an opportunity.

Equally important, in my opinion, is when CLF sets the date of record regarding which shareholders will be eligible to vote at the annual meeting. If the date of record is set close to the date it is announced, institutional investors might have difficulty recovering the shares they lent to shorts in time to vote.

In this regard, it should be noted that Casablanca's original request to delay the record date was due to the original record date of March 7 being only four days away from when CLF approached Casablanca to discuss a settlement (see Casablanca's SEC 13D/A filing today (April 21); http://www.sec.gov/Archives/edgar/data/764065/000090266414002196/p14-1086exhibit11.htm ).

We believe the Board has confused shareholders in its press releases dated March 7, 2014 and related public statements by conflating two very different topics: the record date for the shareholder meeting (which Cliffs and Casablanca discussed), and the actual annual meeting date (which was never discussed) . A record date merely establishes the date on which a shareholder must own shares if they are to be eligible to vote in the annual meeting. Cliffs’ record date was originally set for March 7, 2014—a deadline that was only four days away when the Company contacted Casablanca on March 3, 2014, and indicated its desire to discuss a settlement. In this highly-preliminary discussion, Casablanca suggested that delaying the impending record date might be a productive way to lay the groundwork for meaningful settlement discussions. Cliffs responded by indefinitely suspending its annual meeting date—stating it was doing so “in order to accommodate this request.” Casablanca only suggested that the record date be pushed back, but did not at any time request that the annual meeting itself be delayed.