Parallel resisitors can work very well as a dummy load so long as you work within their limitations. I have made a 100W load from 2W resistors. It got rarther large and the strays limited the frequency at which the match was good, but it was adequate for my needs and a lot cheaper than a ready made one. You can tune out the inductance to get a good match at a spot frequency if you ned to. Don't forget to screen it well, both from the RF leakage point of view and for safety reasons.
Peter
You're apparently assuming constant ESL independent of resistance values. That's highly unlikely because resistors differ both by wire gauge and number of turns. Another point to consider is skin effect of resistance wire. For an exact calibrated dummy load, you'll hardly avoid to measure the actual impedance at 13.56 MHz with a network analyzer or RF impedance bridge.
If your application is 13.56 MHz with small bandwidth, compensation of ESL shouldn't be a problem. But in general, low inductance resistor technologies with thin resistor elements (less skin effect) would be preferrred.
Yes, though it may be easier to use a shunt capacitor. As FvM says you will need to make a measurement and see what you actually get at 13MHz, it may not be what you expect from a measurement on a single device.
I I were making a load for use at 13MHz I would use TO220 resistors, unless of coarse I had the vitreaous enamel resistors already.
Peter