Need a meaningful way to describe heavily-skewed numbers for 50 states

[dunnsidhe]

Hi there,
I apologize if this is an insultingly stupid question, but I managed to enroll in a political research class that involved very little actual research until this end of the semester. My problem is this. I'm attempting to describe a change dollars per state in a meaningful way, and the numbers are very skewed (not sure if that matters just yet, but I'll go on).
For example, here's are two variable, described in Stata:

dollars08                                                          (unlabeled)
-------------------------------------------------------------------------------

                  type:  numeric (double)

                 range:  [0,16081.83]                 units:  .01
         unique values:  49                       missing .:  0/50

                  mean:   593.904
              std. dev:   2292.47

           percentiles:        10%       25%       50%       75%       90%
                              1.57     13.95     54.47    309.69   1209.55

dollars07                                                         (unlabeled)
-------------------------------------------------------------------------------

                  type:  numeric (float)

                 range:  [.34,2801.83]                units:  .01
         unique values:  50                       missing .:  0/50

                  mean:   245.761
              std. dev:   471.463

           percentiles:        10%       25%       50%       75%       90%
                             1.865     10.38    65.525    234.95    770.86

From these, I generate a variable describing the difference:

dollars_diff                                                            (unlabeled)
-------------------------------------------------------------------------------

                  type:  numeric (float)

                 range:  [-2518.97,15509.74]          units:  1.000e-08
         unique values:  50                       missing .:  0/50

                  mean:   348.143
              std. dev:   2282.93

           percentiles:        10%       25%       50%       75%       90%
                          -364.915    -94.97       .02     62.06    892.27

If I transform dollars_diff into 4 ordinals along the percentiles displayed above, is that a meaningful way to look at the data? I'm sorry if I'm not providing enough information. Let me know what else is required and I'll immediately follow up. I genuinely appreciate the help anyone here can offer.
Thanks!
D