Basically, there is nothing one can do, short of a VPI PLC, or newer SDS.
(or other similar speed controllers)
I'll explain later in my post what can help alleviate it some.
The major reason the error is there, is no pulley, nor platter can ever be perfectly machined to be a dead on 33-1/3, or 45 rpm. Ever so slight diameter errors cause this to happen.
My own HW-19, either in its original jr form, or with the upgraded MK-IV Platter ran slightly slow until I got the SDS.
Some tips that might lessen these errors.
One, yes do buy a new Belt, and to help make that belt last as long as possible, it makes sense to assist the Platter upon start up to grasp the Record Clamp after it is properly clamped to the record, and assist the platter to come up to speed. It only takes a second to do.
I'm lately using one of the newfangled black Belts for the HW-19 bought from LP Gear. So far, it's been about a year, and the Belt is holding up very well.
Platter, and Motor Lubrication.
Per the VPI manual, the Hurst Motor can be lubed with one small drop of 40 weight oil.
I would suspect an oil such as Mobil 1 5-30, or 10-30 should also work fine as well. Insure you place that small drop of oil right at the lowest point of the Motor Shaft to insure it goes where it needs to.
The Platter bearing is something that needs to be always periodically lubed, and again, the Mobil 1 5-30 or 10-30 will work very well for this application.
If this Bearing Shaft, it's Bearing Ball, and Thrust Plate that is at the bottom of the well has been neglected, the surface contact areas of both can increase in size, and I suppose this could cause a slight increase in drag, as well as increase bearing noise, and decease speed stability.
The .250" Bearing Ball can be replaced if any signs of wear are present, either with a stock Metal Ball from VPI, or a harder, smoother Ceramic Bearing from someone like Boca Bearing.
Delving further, the Thrust Plate may also be still available from VPI? Of this I am unsure? These other items to look at would be cheaper fixes than the PC, or SDS, and perhaps should be looked into first, and see what the results are? I hope this helps. Mark