When I was a kid, my father used to teach me a couple simple "rule of thumb" formulae of iron-core transformer design. 45/s turns per Volt, and P = s^2, where s = square of the EI iron core's central column in cm^2. I.e. the available electric power grows with the fourth power of the core's "thickness" :-) And, note that tiny wattages have you end up with extremely high turn counts of pretty fine wire... lots of series resistance.
My naive understanding is, that in order for the transformer not to suck/waste power when unloaded / open-ended, you need the transformer to have a certain high-enough inductance for a given primary voltage (measured with an open-ended secondary). Again for small cores, this translates into high turn counts of fine wire.
If tiny iron-core transformers were more efficient than similar-wattage switchers, legal regulations on stand-by power consumption in consumer electronics would make the iron-core stuff omnipresent. The reality is, that only switchers can be made with a good enough efficiency in the tiny form factors (with next to zero idle power consumption).
This is not to say that all switchers are elegant, stabilized and efficient. For instance, some of the super-sexy tiny and light Nokia chargers (ex works standard accessory) contained the most minimalistic switcher I've ever seen: a self-oscillating three-legged "transistor" in the primary, cheap elyts, no filtering, no feedback! Unsurprisingly, those chargers are energetically inefficient, regulation under load is very poor (compared to a similar wattage tiny iron-core trafo) and I didn't really check the EMI.
Some cheap slightly bigger wall warts (switchers) already have a proper feedback (via an optocoupler) and some minimal filtering, but are pretty inefficient (get hot). Same thing with aftermarket noname notebook adaptors.
I've recently been amazed at the current-generation Lenovo notebook adaptors (with the rectangular yellow plug). This small "plastic brick" adaptor runs perfectly cool! Amazing, compared to some notebook adaptors I had in the past. I would hazard a guess at modern TrenchFET/FinFET transistors and sync rectification at the secondary side.