The thing is that a wizard's earnings should be based on the opportunity cost of not paying a wizard to cast some particular spell. If a wizard can produce some service that people would want to buy for less than other providers, then people seeking that service are going to bid up the wizard's payments until the cost matches the other providers. Since many things that a wizard can theoretically provide have no easy to price equivalent service (what's the value of a Teleport Other spell?), this exercise can't necessarily determine how much a wizard should get paid, but it can put a lower bound on a wizard's wages. If a wizard can maintain a Wealthy lifestyle by casting Create Food, there's little reason to think that he's going to work for Average wages by enchanting objects.
The Opportunity Costs of Candles and Continual LightAs mentioned, a lot of spells provide services that don't have an easy to price equivalent. But some spells do. For example, Continual Light provides enough light to eliminate darkness penalties in a 1, 2, or 3 yard radius for an average of 7 days. $7 of tallow candles provides 1 yard of light for 7 days; a $20 lantern and $14 of lantern oil provides 2 yards of light for 7 days. A wizard who can make an acceptable living by casting Continual Light for $7 or less can undercut candlemakers and oil salesman; a wizard who needs more money than that might have a niche market (underwater lighting, for instance) or need to find other spells to earn money.
Continual Light is an interesting spell for this example because it is easy to learn: it requires Magery 0 and a single prerequisite within the same college. Even an IQ11, Magery 0 spellcaster can learn with not much effort: learning two hard skills, one to a professional level, is equivalent to getting a BS with a major and a minor. It has a linear energy cost of 2 per yard of light produced, so 2 energy for a 1 yard light and a 4 energy for a 2 yard light, nicely matching the doubled cost of moving from candles to lantern oil.
Using the standard magic system, a hypothetical IQ11, Magery 0 spellcaster with Light-9 and Continual Light-13 can recover 6 FP per hour, succeeds in casting Continual Light 86% of time, and can expect to produce 20 castings of a 1 yard Continual Light per 8 hour working day. Selling these spells for a mere $6 each and leaving some leeway for disastrous critical failures (roughly twice a year, our wizard forgets how to cast Continual Light for an average of 2 weeks and every other year he summons a demon) gives an income that is high end Comfortable, if not Wealthy, in GURPS terms. A delving or enchanter grade wizard, with IQ15, Magery 3+, and Recover Energy-15 puts out 95 castings a day and is Very Wealthy if not Filthy Rich.
In a Threshold system, a wizard's spellcasting frequency is limited by the lower of his Threshold or his Recovery Rate. Thaumatology suggests a Recovery Rate of 8, which puts the sample minimal spellcaster at a low Average wage, and a delving grade wizard (without an increased Recovery Rate) at Average or possibly Comfortable. My College Ritual Book system suggests that Recovery Rate should be purchased as an advantage. A low end spellworker with Recovery Rate 10 makes at least Average wage, and a delving wizard can earn Comfortable to Wealth wages, depending on his exact Recovery Rate.
So what does this mean?This is a pretty simple thought experiment, but it strongly suggests that most lighting in a world of magic should either be Continual Light spells, or that wizards should be paid more than the standard GURPS model. The standard model assumes that wizards get paid Average wages and that spells cost about $1/FP to cast. More realistically, wizards should get paid Wealthy to Filthy Rich wages and spells should cost $5/FP to cast (or probably more, since this analysis doesn't begin to factor in the rarity costs of being the only IQ16 wizard in 200 miles who knows how to cast Teleport Other).
Which isn't to say that all PC wizards should be rich and well-employed, casting Continual Light all day. PC wizards tend to have all kinds of issues that make them unsuitable for employment, such as Obsessions with becoming liches or an Elder Thing ancestry. It just means that NPC wizards who lack those issues should be well paid, and willing to charge people a pretty penny for spell-casting services.
Aside: What About Enchantment?At the start of this discussion, I noted that the GURPS Magic treatment of the enchantment of magic items assumes certain wages and works out the cost of magical items from there. It does have a brief section on changing those assumptions. The cost of Quick and Dirty enchantment goes from $1/point to $25/point (Filthy Rich enchanters with Enchantment-20 could easily have Continual Light-20 and be raking it in on the lighting market, so they don't come cheap) while Slow and Sure enchantment goes from $33/point to $666/point.
Alternately, people can ditch the GURPS Magic enchantment rules, which don't produce very fun or interesting results anyway. In Dungeon Fantasy, PCs can't become enchanters, and the exact details of the enchanting process are unknown. It's entirely possible that Q&D items are still $1/point because it doesn't take a master enchanter and a circle of five enchanter-grade wizards costing $6600/day to produce 4 Q&D items/day. Maybe a single enchanter can pump out 6 of them a day.