According to the keel effect page on Wikipedia it is the rolling moment created by the fuselage and vertical surfaces in a sideslip if the center of pressure is vertically offset from the center of gravity.
Functions of a keel
The keel of a ship (besides shifting the center of gravity below the metacenter) provides a stabilizing side force when the ship rolls. The same is achieved by a tail surface, but in aviation this is called roll damping. The much larger wing tends to dominate roll damping, so the effect of the vertical surface is small.
The next task of the keel is to balance the side force and rolling moment from the sails in crosswind conditions. It does this by letting the boat build up a sideslip condition until the side force from the keel and hull balances the side force from the sails, and the rolling moment from the side force and from heeling balances the rolling moment from the sails. The keel effect is this rolling moment due to a sideslip condition.
A few aircraft with very small wings and large vertical surfaces do indeed suffer from excessive moments created by side loads on those vertical surfaces. The best known of those is probably the Lockheed F-104 Starfighter. The massive tail on the upswept rear fuselage made 10° anhedral necessary to get the sideslip-induced rolling moment right.
Pendulum effect
The Wikipedia page also mentions the pendulum effect and calls both the fuselage contribution to the dihedral effect. Man, if there is ever a competition for the most misleading name of an effect, this would be the winning entry.
A pendulum is a mass mounted below the pivot or the metacenter, so it will stabilize in the down position. It seems the adherents of the pendulum effect belief believe that the plane is somehow hinged at the center of the wing. This is so wrong! A flying aircraft is not hinged, so all motion is around the center of gravity.
This is different with airships: As @JanHudec points out in the comments, the buoyancy of the lifting gas always works against gravity whereas the lift of a wing acts perpendicular to its span and the airspeed. This does produce a moment when the airship rolls, and the heavy gondola is pulled down by the same effect which pulls down a pendulum. Also, due to an airship's low density (by definition very close to the density of the air it displaces) the aerodynamic side forces on an airship are much larger compared to the inertial forces. This lets the airship effectively pivot around a point near the center of the envelope. Note that airships have no means of roll control - it is just gravity which ensures that the gondola stays on the lower side of the hull.
Now you can argue that the same is true for aircraft, only with much smaller aerodynamic forces. Yes, but those side forces are vanishingly small in comparison to the wing's aerodynamic forces which, again by definition, are close to the weight of the aircraft. A small change there will completely dominate the picture, and the small loads on the fuselage are just a rounding error. By moving the control surfaces, the pilot can shift the lift vector around almost as he/she likes, and the forces on the fuselage are insignificant.
Note that knife edge flight, a situation when the side force on the fuselage really is close to the aircraft's weight, needs lots of fuselage area and a powerful engine. Fuselage loads in normal flight are a fraction of those in knife edge flight.
