Exhibit 2

TBA	Price	CC PSA			FC PSA			OAD			OAC
		old	new	diff	old	new	diff	old	new	diff	old	new	diff
FNCL4.5	95.64	146.5	185.7	39.1	137.0	156.2	19.2	5.00	5.02	0.01	-0.99	-0.91	0.09
FNCL5.0	98.17	200.0	224.8	24.8	156.7	191.4	34.7	4.19	4.26	0.06	-1.43	-1.37	0.06
FNCL5.5	100.23	296.2	266.8	-29.4	229.2	233.6	4.4	3.18	3.28	0.10	-2.05	-1.86	0.19
FNCL6.0	101.86	368.3	440.3	72.1	333.8	345.9	12.1	2.18	2.10	-0.08	-1.98	-2.00	-0.01
FNCL6.5	103.11	401.8	460.7	58.9	326.0	401.5	75.5	2.24	2.00	-0.25	-1.22	-1.27	-0.05
FNCL7.0	104.72	432.7	449.2	16.4	375.8	404.5	28.7	2.28	2.14	-0.14	-0.45	-0.66	-0.20
FNCL7.5	105.81	519.3	531.6	12.3	488.0	496.0	8.1	2.17	1.96	-0.21	0.23	-0.04	-0.27
FNCL8.0	106.88	519.3	507.2	-12.1	517.3	495.6	-21.8	2.15	2.09	-0.06	0.19	-0.02	-0.22

As we could have expected, the premiums got shorter in the OAD sense and more negatively convex.

Conventional MBS versus Government MBS

Every MBS analyst is aware that, to the contrary of historical facts and perceptions, recently originated Ginnies were prepaying faster than Fannies and Freddies in 2004 – 2005. Both old and new prepay models reflect this fact, which is not a single-month surprise. Our government prepay model is faster – according to the recent trend. Old government pools remain slower than conventionals; for details, see reports.

Fannies versus Freddies

Freddie MBS look cheaper than Fannies in all coupons, but we are aware of some stunning distortions in the assumptions. We use the Bloomberg median levels for the inputs and those differ widely between FNCL6.5 (WAM of 337 resulting in an OAS of 16 bps) and FGLMC6.5 (WAM of 325 resulting in an OAS of 40 bps). A detailed look at the VALL screen suggests that 9 brokers out of 11 contribute similar WAM assumptions for these two TBAs, whereas 2 others differ considerably. One of them actually determines the median point. Since Fannies and Freddies should not differ drastically, this is a good argument

>>>