If you are comparing phosphorylated and dephosphorylated forms of the same protein, it is possible that a more negative charge induced by the presence of the phosphate group would cause a downward mobility shift (protein travels further down the band). However, this is not always the case; for example, enough occupied phosphorylation sites on a protein may increase the MW of the protein enough to cause an upward mobility shift, despite a more negative overall charge. I would look into phosphorylation sites on your protein of interest and further, see what other researchers have found when blotting for the same protein.
In Western Blot SDS used in SDS PAGE is an anionic detergent which linearizes proteins and coat proteins with uniform negative charge, as proteins can be positively, negatively or neutrally charged thereby imparting a similar charge-to-mass ratio to all proteins. So in case of phosphorylated proteins which will have a higher molecular weight than the non-phosphorylated ones will tend to run slower. It depends on many factors like the number of phosphorylation sites in the protein, percentage of the gel used and the resolving time. You have to look into the phosphorylation sites of your protein as mentioned by Jake. The molecular weight of proteins is considered for the mobility of proteins rather than the overall increase in the negative charge of the molecule.
Post-translational modifications does induce mobility shift of proteins and as mentioned by Jake in majority of the cases it results in upward shift although downward shift is also possible. In my experiments the modification usually causes upward shift of the proteins. To confirm whether your proteins are phosphorylated you can use phospho-specific antibodies.