Well the game has been under Beta a little over a year now, and with the data gathered although it is balance short-term we can't be so certain about long-term as the game can progress for many years (it's a browser football manager game - footballstrategy dot org)

D you have a list/spreadsheet/... of the economy ?

And especially a list of how/where new money is generated(you know, out of thin air) and where/how it is removed.

Actually, there are job titles for it. The first is "designer", the second is "quality assurance".

As mentioned, you start with your data tables or designs or tunable variable data, whatever you have for your game.

Then you take either a spreadsheet or Ye Olde Fashioned Paper, and make a lot of notes. Record the funds you started with, how much it should have grown, how much it actually grew. Note what is fun and what is not. Note what could be adjusted, and adjust it.

Sometimes problems are more obvious than others. In one game I worked on, the end-level summary listed starting funds, money from foo, money from bar, money from baz, and final funds. Someone discovered the sum didn't work out. Pulling an old version, they discovered the sum hadn't worked out for over four months and nobody on QA had caught it. We were late in development and couldn't hunt down where it came from, so the fix was to add an "other changes" line, do the math, and any difference between the numbers was shown there.

Finding exploits of the economy is also the job of both the designer and QA. Both should work together in evaluating the designs to figure out any loopholes or exploitable flaws. After implementation, QA needs to spend time working with everything money-related to shake out exploits. Common exploits are to start and action involving money or items, and then cancel it. Sometimes it will consume items, other times it will give free resources. It needs to be tested for everything that produces or consumes the resources, at every step of the process. For example, if a spell causes you to pick three ingredients, try picking one then cancelling, or picking two and then cancelling, or picking two then backing up then going forward again, and so on, through as many variations as you can imagine. Patience, precise notes, and careful math are all mandatory, meaning few testers are good at it.