Column efficiency is inversely proportional to particle diameter. Efficiency is a function of "N" or the number of theoretical plates on a column. A theoretical plate is one instance of an analyte entering and then leaving the stationary phase. Plate height (HETP) is the average distance a solute will travel into a particle before re-entering the mobile phase. So to put it all together, a lower plate height means more theoretical plates, which in turn means higher efficiency.
Plate height is directly proportional to particle size, since it takes an analyte less time to enter and then leave a smaller particle than it does a bigger one. So, all other things being equal, a smaller particle size means sharper peaks and more resolution.
There are some disadvantages to using a smaller particle size. The obvious drawback is backpressure. Operating pressure and particle size have an inverse square relationship. A decrease in particle size by a factor of 2 would yield a pressure increase by a factor of 4. Or more practically, going from a 5um to a 3um particle would increase your pressure by a factor of 2.78.
A more subtle disadvantage of smaller particle columns is their vulnerability to clogging. Smaller particle size means the space between the particles, or interstitial space, is also smaller. The interstitial space is where large impurities can become stuck and cause column clogging. Every column has a frit on the front of the column as a last-resort filter to keep impurities out. When transitioning to a smaller particle size column it's a good idea to find out the pore size of this frit and filter your samples with an equal or smaller pore size filter. This isn't as much of a headache when going from 5um to 3um, but becomes a huge issue when using sub-2um columns.
Finally, you might consider using a larger particle size column because of price. Typically, a smaller particle size will cost you more with all other things staying the same.
