Yes, of course the 18.6 year nodal precession is mentioned. But this is in reference to a specific longitude. The Chandler wobble is independent of longitude and should respond to periods at which the moon crosses the equator through a complete ascending/descending cycle.
That cycle is the draconic month, which is 27.2122 days. The tropical month is 27.3216 days, which is the length of time the moon takes to appear at the same longitude. Those two are slightly different and the difference forms a beat cycle that determines how often the maximum declination is reached *for a particular longitude*. That is important for ocean tides as tides are really a localized phenomenon.
But the Chandler wobble is global and would only have triaxial components to second order. It doesn't really care where the moon is located in longitude when it reaches a maximum in declination excursion. Consider that at the poles, the longitude converges at a singularity. It's the physics of a spinning top under the influence of a gravity vector.
So the important point is that the moon is cycling a significant gravitational torque on the earth's axis every 1/2 a draconic period. When this is reinforced by the sun's biannual cycle, one gets the Chandler wobble period precisely.
Try Googling "Draconic AND Chandler wobble". You won't find anything. Sure you will find the 18.6 year cycle but that's because of an echo chamber of misleading information. This is important to understand, and if my own reasoning is faulty, this model I have been pushing for the Chandler wobble and QBO collapses. Those are both global behaviors. ENSO, on the other hand, has a longitudinally localized forcing and should show strong effects of the 18.6 year cycle ... and I have shown it does!