More to the point, why should players want to have your new item rather than one of the existing items? At the same time, why should your players ever want any of the old items again if your new item is clearly better?
If you don't address this question of power balancing items, you'll wind up with players all making the same choices, leading to homogeneity of appearance and action that devalues the time and money you spent developing other content.
"But how can this balance be achieved, Master?" So glad you asked, Grasshopper. ;-)
The key is to explicitly and consciously design every usable item in the game to have both advantages and disadvantages. For every characteristic that offers something you want, there should be some other undesirable characteristic that goes along with the good stuff -- for every benefit, there should be an unavoidable cost.
The point of this is to insure that there's not just one item that everyone takes because it's the "best," but that different items will be "better" as environments and circumstances change. As long as you're careful to insure that the alternatives are balanced -- that one class of objects is never always clearly better than another class -- then you're creating a world of "interesting choices."
I should add that while it's possible to have something of this effect using just benefits, the effect is easier to produce when you can use both benefits and costs. Having both of these increases the number of interactions that objects have with their owner and the environment, which are what item usage decisions are based on.
The next step is to apply this approach to the characteristics of both loot/reward items and crafted items. Then when players create or receive an item, then can decide which characteristic they care more about maximizing.
I know this must seem pretty abstract, so let's look at an example.
Let's say you've won a "Developer For A Day" contest, and you've decided you want to offer a new class of crafting tools that have more effects on crafting.
The first step (from a functional point of view) is to think about what effects the object will have. Just for the sake of discussion, let's say you decide that the new crafting tools will have the following characteristics:
influences assembly result
influences experimentation results
influences the number of items manufacturable in 1 run
influences speed of producing prototypes or manf. items
Each of these characteristics would be a 0-100 result that modifies the baseline result. That is, if a tool has an Assembly Effectiveness of 50, it doesn't do anything to the basic assembly calculation; if it had an AE of 100, it would add some nice amount to your chances of getting a great assembly result; if it had an AE of less than 50, it would actually reduce your chances of getting a desirable assembly result. (Note that you could use some other scale if you wanted; "0-100" isn't the thing to focus on here.) (Also note that you don't actually have to make "undesirable" mean "worse effects" -- just preventing you from getting the best effects could seem like enough of a disadvantage.)
So if 50 is the "no effect" point, that allows you to assign "desirable" (above 50) and "undesirable" (below 50) values to each characteristic.
Next, you decide how you want to group benefits and costs. In this example, for crafted tools I'd suggest giving Assembly Effectiveness and Experimentation Effectiveness an inverse relationship, increasing Manufacturing Output should significantly reduce Construction Speed, and increasing Construction Speed should decrease the typical values of all three of the other characteristics.
By setting up these inverse relationships, you create a kind of item for which there is no "best" form -- each player will decide which is best for his or her personal needs.
And it doesn't stop there. Now you can also allow for a new kind of crafting tool that is dropped as loot or as a quest reward. All the characteristics still apply; what changes are the benefit/cost relationships. Maybe for a looted crafting tool, you can get a great Experimentation Effectiveness but it always comes with a very slow Construction Speed. Maybe a crafting tool presented as a quest reward lets you crank out many more units than usual in Manufacturing Output (and at an acceptable rate), but crafting items with that tool will suffer significant penalties to Assembly Effectiveness and Experimentation Effectiveness.
By specifying that usable items will have multiple characteristics that have meaningful in-game effects, and by further specifying that high values of one characteristic will always be balanced by low values in some other characteristic(s), you can balance the perceived value of items no matter how they're provided.
This seems like it might help to address the "crafted vs. looted/quested" item quality concern in a way that doesn't undercut anyone. Is there a problem with it that I'm not seeing?