What bothers me about 5.d5 is that it seems to offer targets for counterplay at an early stage.  If this is a problem, 5.Qa4 may cut out several options for black and white could possibly play d5 at a more favourable time.  I am not aware of any published analysis on either of these lines; these are only my observations.

What do think these games from the Argentinian GM Campora?  His plan for black seems promising.  Here are two recent games of his:

[Event "ESP-chT"]
[Site "Lanzarote"]
[Date "2003.05.06"]
[Round "4"]
[White "Granados Gomez,Manuel"]
[Black "Campora,Daniel Hugo"]
[Result "0-1"]
[Eco "A20"]
1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.d5 Bb4+ 6.Bd2 Qe7 7.Nc3 0-0 8.a3 Bxc3 
9.Bxc3 d6 10.e3 Nbd7 11.Ne2 c5 12.0-0 Ne5 13.Bxe5 Qxe5 14.Nc3 Bg4 15.Qc2 Rfe8 16.Rfe1 Bf3 
17.Nb5 Re7 18.Qc3 Qh5 19.Nxd6 Ng4 20.h3 Bxg2 21.Kxg2 Nxf2 22.Rg1 Rd8 23.Kxf2 Rxd6 24.Kg2 Qe2+ 
25.Kh1 Rh6 26.h4 Rf6 27.Rg2 Qg4 28.Kh2 h5 29.b4 Qc8 30.bxc5 Qxc5 31.Qd4 Qxd4 32.exd4 Rd6 
33.Rb1 Rc7 34.g4 Rxc4 35.Rxb7 Rxd4 36.Rxa7 R6xd5 37.g5 e3 38.g6 Rxh4+ 39.Kg3 Rg4+ 40.Kf3 Rxg6 
41.Rxg6 fxg6 42.Kxe3 Rf5 43.a4 Kh7 44.a5 g5 45.a6 Rf6 46.Ra8 Kg6 47.a7 Rf7 48.Ke4 g4 
49.Ke3 Kg5 50.Ke2 Kh4 51.Ke3 g5 0-1

[Event "Asuncion zt 2.5"]
[Site "Asuncion"]
[Date "2003.09.14"]
[Round "9"]
[White "Giardelli,Sergio Carlos"]
[Black "Campora,Daniel Hugo"]
[Result "0-1"]
[Eco "A20"]
1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.d5 Bb4+ 6.Bd2 Qe7 7.Nc3 0-0 8.e3 Re8 
9.Nge2 d6 10.dxc6 Nxc6 11.a3 Bxc3 12.Bxc3 Bg4 13.h3 Bxe2 14.Qxe2 Rac8 15.0-0 Ne5 16.Rad1 Nxc4 
17.Bxf6 Qxf6 18.Rd4 d5 19.Rxd5 Nxb2 20.Rb1 Na4 21.Qg4 Nc3 22.Rf5 Qe7 23.Rb4 g6 24.Ra5 b6 
25.Rh5 Kh8 26.Rh6 Rc6 27.Bf1 f5 28.Qh4 Qxh4 29.Rxh4 Rd8 30.g4 Rd1 31.Kg2 f4 32.exf4 e3 
33.fxe3 Nd5 34.Rb2 Nxe3+ 35.Kf2 Rxf1+ 36.Kxe3 Re6+ 37.Kd4 Rxf4+ 38.Kd5 Re7 39.Rb4 Ref7 40.Kc6 Rxb4 
41.axb4 Rf4 42.b5 Kg7 43.Kb7 Ra4 44.Kb8 h5 45.Kb7 Kf8 46.Kb8 Ke7 47.Kc7 Rc4+ 48.Kb7 Kd6 
49.Ka8 Kc7 50.g5 Rxh4 0-1

I would like to draw attention to the line 1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4, which is not extremely popular, but has had some recent successes at a high level.  For example:

[Event "Bled ol (Men)"]
[Site "Bled"]
[Date "2002.10.26"]
[Round "10"]
[White "Rowson,Jonathan"]
[Black "Bologan,Viktor"]
[Result "0-1"]
[Eco "A20"]
1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.Nc3 d5 6.Bg5 Bb4 7.Qb3 Bxc3+ 8.Qxc3 0-0 
9.e3 Nbd7 10.cxd5 cxd5 11.Rc1 b6 12.Qa3 a5 13.Ne2 Ba6 14.Bh3 Rc8 15.Kd2 Rxc1 16.Rxc1 h6 
17.Bf4 Qa8 18.Rc7 Rc8 19.Qc3 g5 20.Bd6 Rxc7 21.Qxc7 Bb5 22.Bxd7 Nxd7 23.Nc3 Bc6 24.a3 Kh7 
25.b4 axb4 26.axb4 b5 27.Qa5 Qe8 28.Qc7 Nf6 29.h3 Kg6 30.Be7 Qd7 31.Qxd7 Nxd7 32.h4 f6 
33.hxg5 fxg5 34.Ke2 Kf5 35.Nb1 Kg4 36.Nd2 Nb6 37.Bd8 Nc4 38.Nb3 Kh3 39.Be7 Nb2 40.Nc5 Kg4 
41.Ne6 Nd3 42.Bf8 h5 43.Be7 h4 44.gxh4 gxh4 45.Bd6 h3 46.Kf1 Bd7 47.Nc5 Be8 48.Nb3 Nb2 
49.Ke2 Bh5 50.Nc5 h2 51.Bxh2 Kh3+ 0-1

I believe black has good chances for equality in many lines when he plays ...d5, while at the same time having good winning chances.  When analyzing this line, I found it difficult to achieve an advantage with white, since in many lines where black achieved ...d5, the fianchetto seemed to make white development's awkward in this French-like structure.  In comparison to a game like Korchnoi-Bacrot (which was 2...c6), White's play is slower in the French structure, since he has already committed to Bg2.  I'll provide the game; it's pretty nice:

[Event "Albert m"]
[Site "Albert"]
[Date "1997.09.10"]
[Round "4"]
[White "Kortschnoj,Viktor"]
[Black "Bacrot,Etienne"]
[Result "1-0"]
[Eco "A20"]
1.c4 e5 2.g3 c6 3.d4 e4 4.Nc3 d5 5.cxd5 cxd5 6.Qb3 Ne7 7.Bf4 a6 8.f3 exf3 
9.Nxf3 Nbc6 10.e4 Be6 11.Rd1 dxe4 12.d5 Nxd5 13.Rxd5 Bxd5 14.Nxd5 Bd6 15.Qxb7 0-0 16.Bxd6 Qxd6 
17.Qxc6 Qxc6 18.Ne7+ Kh8 19.Nxc6 exf3 20.Kf2 Rac8 21.Nb4 Rc1 22.Nd3 Ra1 23.Bg2 Rxa2 24.Bxf3 Rd8 
25.Rc1 g6 26.Ke3 Kg7 27.Rc7 Ra5 28.Ra7 Rd6 29.g4 Rb5 30.b4 h5 31.gxh5 gxh5 32.Bb7 Rbb6 
33.Bc8 Kf6 34.Nc5 Kg6 35.Rxa6 f5 36.h4 f4+ 37.Kf3 Rf6 38.Rxb6 Rxb6 39.Nd3 Rc6 40.Bb7 Rc3 
41.Kxf4 Kf6 42.Be4 Rb3 43.Ke3 Rb1 44.Kd4 Rg1 45.b5 Ke6 46.b6  1-0

By the way, the line given in Kosten's Dynamic English is the game Ragozin-Povah:

[Event "Lloyds Bank op"]
[Site "London"]
[Date "1994.08.??"]
[Round "2"]
[White "Ragozin,Evgeni"]
[Black "Povah,Nigel E"]
[Result "1-0"]
[Eco "A20"]
1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.Nc3 d5 6.cxd5 cxd5 7.Bg5 Bb4 8.Rc1 0-0 
9.Nh3 Nbd7 10.Qb3 Bxc3+ 11.Rxc3 h6 12.Bf4 Qb6 13.Qxb6 Nxb6 14.b3 Bd7 15.Kd2 Rfc8 16.Rc5 Bc6 
17.f3 exf3 18.exf3 Re8 19.Rhc1 Re6 20.g4 g5 21.Bg3 Nbd7 22.R5c3 Bb5 23.Re1 Rxe1 24.Bxe1 Re8 
25.Re3 Re6 26.Rxe6 fxe6 27.Bg3 Kf7 28.Nf2 Ng8 29.h4 Ne7 30.h5 Nc6 31.Kc3 Be2 32.Nd3 Kf6 
33.Bd6 b6 34.Nc1 Bb5 35.f4 Nd8 36.fxg5+ hxg5 37.Bf3 Nf7 38.Bc7 Bf1 39.a4 a5 40.Kc2 Ke7 
41.Kd2 Nh6 42.Nd3 Nf7 43.Nf2 Kf6 44.Ke1 Ba6 45.Bd1 Kg7 46.Bc2 Nh6 47.Bg6 b5 48.Bd3 bxa4 
49.Bxa6 axb3 50.Kd2 Nf6 51.Be2 a4 52.Kc3 Nf7 53.Bb5 Ne4+ 54.Nxe4 dxe4 55.Bxa4 Nh6 56.Bxb3 Kf7 
57.Bd1  1-0

White got a slight edge, though I think black can improve.  Anyway, the "Keres French structure" after 1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.Nc3 d5 seems like a tough nut to crack, even in the positions where white should be slightly better.  When analyzing positions in this line, it seemed like black had to time to consolidate, and thereafter it was hard for white to make progress.

Well, I thought about this line for while, since this seems like a very natural continuation for black.  Then I found an interesting plan that was tested in a correspondence game.  The idea is simple: play Qa4 to prevent ...d5, and play d5 later on.  Here's the game:

[Event "DDR corr"]
[Site ""]
[Date "1987.??.??"]
[Round "0"]
[White "Schmeisser,Peter"]
[Black "Van Semmern,Frank"]
[Result "1-0"]
[Eco "A20"]
1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 5.Qa4 d6 6.Nc3 Bf5 7.e3 Be7 8.Nge2 0-0 
9.d5 Na6 10.b4 cxd5 11.cxd5 Nc7 12.Qb3 Qc8 13.Bb2 a5 14.b5 a4 15.Nxa4 Ncxd5 16.0-0 Qc7 
17.Nd4 Bg6 18.Nc3 Nxc3 19.Bxc3 Rfc8 20.Rfc1 Qd7 21.a4 d5 22.a5 Ne8 23.b6 Bf6 24.Qb5 Bf5 
25.Qxd7 Bxd7 26.Bb4 Bd8 27.Bf1 f6 28.Bc5 Rcb8 29.Rcb1 Rc8 30.a6 Rxc5 31.axb7 Rb8 32.Ra8 Rxb7 
33.Rxd8  1-0

5.Qa4 following 1.c4 e5 2.g3 Nf6 3.Bg2 c6 4.d4 e4 seems very logical to me.  I don't see any problems with it yet, and it seems to cause a lot of problems for black!  I wonder why this idea has had such little attention.  It's probably because 4...e4 is relatively rare, and I guess 5.Qa4 is a fairly nonstandard idea.  Maybe this could be a recommendation in The Dynamic English 2!