I'll deal with your first example first: standing on a table in that situation is not particularly an advantage. It's also, therefore, very unlikely behaviour for a martial artist. (Not counting Feng Shui players). Standing slightly up from your opponent, on the other hand, is an advantage for many of the reasons below. It brings your kicks to better target points with less effort, while costing you nothing.
Historically, however, standing on tables or fighting indoors is an unlikely scenario. Let's discuss the more common swordsman (or in your example, armoured knight) fighting uphill against similar opponent. In this case the higher-ground advantages are huge.
(To help make examples clearer: our swordsmen are on a hillside. Tom is higher up, Lowry is lower down and facing uphill.)
The assumption that Tom needs to 'bend over' to block is false. The leg guards work perfectly well while standing in a normal fighting crouch, regardless of the origin height of the attack - it has to be at Tom's leg height regardless. At worst Tom ducks slightly.
Lowry, on the other hand, gains defence for his legs but loses it at the head, and loses access to attack Tom's head. Is this a fair exchange for Lowry? No! (Hint... which would you rather lose, leg or head?)
Endurance matters! The defence advantage for Tom is large. Tom needs to guard his legs and abdomen more... some of the least tiring guards, because your arm drops low. Lowry needs to guard his head more... needing him to lift the weight of his weapon more often. And he has to block against greater force, because Tom has gravity aiding in the downswing.
On this theme, gravity is Tom's friend. Most styles use more cuts down than up, for a reason - those cuts have more power with less muscular effort. (And for the beginner, cuts down are also easier to execute and to feint with.) Lowry also has to lift the blade more to perform equivalent cuts and thrusts.
Footwork matters! In a duel both fighters move in all directions. Stepping backwards uphill leaves you less likely to trip than stepping backwards downhill. And more likely to catch yourself before dying. (In a massed formation during battle the rules are different, but that applies to most things.)
Balance matters! If both are competent, neither fighter will particularly lean forward or back... or they'll lose. (Tom ducks down, not leans forward, to strike the head.)
Reach really matters... but only to the key target points. All else being equal, on a mutual thrust, Tom hits the head while Lowry hits the calf. Guess who's limping away from that fight?
For armoured knights with swords the situation is even worse, because the legs and chest are very hard to injure on a man in harness. (That's what the armour is for, after all.) The good target points on a suit of plate armour are mostly upper body. (Although this match is historically unlikely outside competitions; poleaxes would make more sense, and are slightly more even because of the leverage advantages they grant.)
Vision matters! As wraith808 correctly says, looking up is harder than looking down. And it's a lot harder if you're wearing a helmet. Lowry has to work harder to keep his eyes on Tom. And can more easily be distracted by attacks coming at the head.
Speed matters! When Tom attacks he's closing the distance downhill, with gravity aiding his strike. Lowry is climbing to do the same.
All this only applies if the height difference is a couple of feet, or a not-too-steep-hill. At more than that this isn't a swordfight as such; Tom is hacking down at a climbing opponent.
This is a summary of a complex issue... but the purpose of game mechanics is to summarise complex issues. A flat bonus is not an unreasonable representation of this effect.