Why are so many companies like harbeth making old speakers designs and charging thousends?

I have kinda stayed away from this thread,  but as I think it through,  hopefully my input will hold some merit....
"Why are so many companies like harbeth making old speakers designs and charging thousends?"  Thousands....
Really,  in the big picture,  Every speaker design can be called Old. 
What changes speaker design is that a company or designer will like the look, spec, sound of a new driver and develop a new speaker around that driver or drivers...
In the end,  most designer prefer a sound of a certain type of crossover slope and will normally gravitate to using those slopes when possible.
A 6/12 slope has a certain sound,  all 6db slopes another,  12/18 or all 24 db slopes, etc, etc, etc.  Each type of crossover has a flavor that each person prefers.
Harbeth in particular has their own midrange driver,  Unless they contract a new part or parts,  I have to believe that they will continue to use the parts that they feel is their best foot forward.  Maybe each company needs to try different parts, as in a ribbon or a horn load, or exotic material drivers, but once you get a potion for success and you make good money doing it,  it gets hard to rock the boat.
I know there is a huge following out there for all 6db per octave speakers,  I personally like Odd order slopes, properly aligned and phase compensated. 
So in the end,  true NEW parts being introduced doesn't happen everyday and right off the top of my head, I can think of maybe 6 or 8 or so different crossover topologies.  The 2 that I use most are most likely butterworth or Linkwitz Riley.  New crossover technology hasn't happened in decades.

I've avoided this one, too, but...

Active digital crossovers (and room correction) are a more recent innovation, so there's been some recent technology developments out there, but Tim's point is mostly taken.

IMO, there are few innovations that justify the price increases we've seen over the past 25 years.  Active speakers (particularly if multiple channels of amplification are included) might be a partial exception, but overall - I've got to believe that it's largely margin.  I find it next to impossible to attribute the cost of most contemporary  loudspeakers bearing a five (or six) digit price tag to parts, or manufacturing overhead, or R & D, or inflation.  Gross margins on those products have to be very high on average.  

That doesn't mean the manufacturer is getting rich on them, either. If he sells three pairs a year, the fact that each pair bears a healthy mark-up isn't going to buy that mansion on Maui.  Small volume companies need higher margins to survive, but it doesn't change the fact that margins are high.  The specific question from the OP is legit:  They may not be the priciest speakers out there, but  I'm pretty confident that Harbeth makes fat margins on each sale.

BTW, that doesn't mean people shouldn't purchase a pair if they like the sound and/or look of them. That's also true of a six figure loudspeaker.  Porsche has very high margins for an auto manufacturer and not too many folks belabor that point.  Designer furniture (and that's a big part of what a six figure loudspeaker is) also routinely commands prices that you don't want to justify on a cost of manufacture basis.

People enjoy all kinds of high margin purchases.  I just think they shouldn't try to justify the purchase by wishing away the mark-up.

I think its just the market is limited these days for many niche high end audio products so profit margins per unit must be higher to stay in business   Factor in high overhead costs to market and distribute internationally and such and there you go.

Have the prices of the speakers (and everything else) gone up ... or has the value of the money gone down?  

Hi Marty,
     "Active digital crossovers (and room correction) are a more recent innovation, so there's been some recent technology developments out there, but Tim's point is mostly taken" 
Back in 1980 we were experimenting with the stk086 chip used individually on drivers to actively cross, bi/tri/quad amp etc. (we actually made a few fabulous sounding amps with this simple chip)  We could not do the hard core analytics for room correction then that a laptop program can provide now,  but even then it was quite simple to graph and eq a room, I suspect that room correction as we think of it today has been around longer than we realize.