From my understanding of projectile weapons, the energy of the projectile is what determines its penetration power, which would approximately be the bow's draw weight times its power stroke.
For some reason, the power stroke for crossbows seemed to be fixed throughout history, so people came up with all sorts of contraptions to give them mechanical leverage and allow them to construct even heavier crossbows. But why not construct a crossbow out of materials that allow for a longer acceleration time and heavier bolts, like a longbow?
The only example of a "crossbow" doing this i could find was the crossbows granddaddy, the Gastraphetes
To start with, this is only really a design feature of medieval European crossbows and weapons that probably derive from them. Crossbows from the Han Dynasty in China had their trigger mounted at the very end of the stock and used normal sized bows with long draw distances. I'm less familiar with later Chinese crossbows, but from what I know with the exception of the repeating crossbow which is it's own whole can of worms this remained fairly constant.
For European crossbows the short answer is that we don't know. No medieval crossbow maker, archer, or general wrote down an account of why crossbows were designed that way. We also don't know when or why Europe developed the rolling nut trigger instead of using the Chinese bronze trigger, for example. What we can do is look at the technology as we know it existed and make some educated guesses, but these are not definitive answers.
So, why have a shorter draw distance / power stroke? The best explanation I've seen is that a shorter power stroke allows for a shorter bow, but that then immediately begs the question of why would you want a shorter bow? There are fairly straightforward answers to that for composite and steel crossbows.
Steel is the most obvious: weight. Nobody wants a full longbow sized bow made of steel mounted on the end of their stock, a shorter thicker bow was a far better use of weight. Also you would have lost a lot of efficiency moving those huge steel limbs back when the crossbow was fired.
Composite crossbows were expensive and hard to make, making a shorter thicker and potentially more durable bow could have advantages over making a full sized composite bow and mounting it on a stock.
But with those explanations we're tackling this problem in reverse chronological order. From what we know about wooden crossbows (and, to be fair, it's not a whole lot) they also had quite short power strokes. The Berkhamsted Bow has been estimated as probably drawing between 8 and 10 inches, which is a lot longer than a fifteenth century steel crossbow but still much shorter than a yew longbow. To some extent weight and ease of use would have factored in to this, nobody was going to mount a six foot longbow to the end of a stock, but that doesn't fully explain why only 8 inches and not, say, 14-20 inches.
A problem here is we only have a handful of wooden crossbows to draw our conclusions from and two of them from the British Isles so it's hard to say if this was standard across Europe c.400 - c.1400 or if our limited evidence and the obvious later trend biases us to assume that this was a universal feature of European crossbows.
Another important factor to consider is the rolling nut trigger, briefly mentioned above. Because of the large Z shaped metal trigger used for the rolling nut it made the most sense to put the lock mechanism in the centre of the stock rather than at the end like the Chinese crossbows. Now we're getting into another can of worms, though. As mentioned, the Chinese trigger is older so why did Europe use the rolling nut? And did the rolling nut determine the draw distance or did the draw distance make the rolling nut appealing?
We have a paucity of evidence and a lot of room for speculation, but the short answer remains that we're not sure why European crossbows started being designed this way. It's possible that if the crossbow got greater scholarly attention this could be solved, but it's not a widely studied weapon (especially outside of German language publications) so for the time being it will likely remain an intriguing mystery.
Egon Harmuth's Die Armbrust has a good section on the mechanics of crossbows if you can read German (or are very patient and equipped with a dictionary and Google Translate).
Jens Sensfelder is another great author for all things crossbow design and mechanics. He mostly writes in German but he has published in English on occasion.
It doesn't answer the question, but a fascinating and relevant article is Erhard Franken-Stellamans' “A Mathematical Method for Determination of the Appropriate Draw Length for a Given Steel Bow”, Journal of the Society of Archer-Antiquaries 47 (2004). 92-97.