820 CHAPTER 30. CONTINUOUS STOCHASTIC PROCESSES

Proof: Suppose Q ∈ [0,T ]×Ω is progressively measurable. This means for each t,

Q∩ [0, t]×Ω ∈B ([0, t])×Ft

What about (s,ω) ∈ [0, t]×Ω, (s,ω)→∫ s

0 XQdr? Is that function on the right B ([0, t])×Ft measurable? We know that Q∩ [0,s]×Ω is B ([0,s])×Fs measurable and henceB ([0, t])×Ft measurable. When you integrate a product measurable function, you doget one which is product measurable. Therefore, this function must be B ([0, t])×Ftmeasurable. This shows that (t,ω)→

∫ t0 XQ (s,ω)ds is progressively measurable. Here is

a claim which was just used.Claim: If Q is B ([0, t])×Ft measurable, then (s,ω)→

∫ s0 XQdr is also B ([0, t])×Ft

measurable.Proof of claim: First consider A×B where A ∈B ([0, t]) and B ∈Ft . Then∫ s

0XA×Bdr =

∫ s

0XAXBdr = XB (ω)

∫ s

0XA (s)dr

This is clearly B ([0, t])×Ft measurable. It is the product of a continuous function of swith the indicator function of a set in Ft . Now let

G ≡{

Q ∈B ([0, t])×Ft : (s,ω)→∫ s

0XQ (r,ω)dr is B ([0, t])×Ft measurable

}Then it was just shown that G contains the measurable rectangles. It is also clear thatG is closed with respect to countable disjoint unions and complements. Therefore, G ⊇σ (Kt) = B ([0, t])×Ft where Kt denotes the measurable rectangles A×B where B ∈Ftand A ∈B ([0, t]) = B ([0,T ])∩ [0, t]. This proves the claim.

Thus if Q is progressively measurable, (s,ω)→∫ s

0 XQ (r,ω)dr ≡ f (s,ω) is progres-sively measurable because for (s,ω) ∈ [0, t]×Ω,(s,ω)→ f (s,ω) is B ([0, t])×Ft mea-surable. This is what was to be proved in this special case.

Now consider the conclusion of the proposition. By considering the positive and neg-ative parts of φ (X) for φ ∈ E ′, and using Pettis theorem, it suffices to consider the casewhere X ≥ 0. Then there exists an increasing sequence of progressively measurable simplefunctions {Xn} converging pointwise to X . From what was just shown,

(t,ω)→∫ t

0Xnds

is progressively measurable. Hence, by the monotone convergence theorem, (t,ω) →∫ t0 Xds is also progressively measurable. ■

What else can you do to something which is progressively measurable and obtain some-thing which is progressively measurable? It turns out that shifts in time can preserve pro-gressive measurability. Let Ft be a filtration on [0,T ] and extend the filtration to be equalto F0 and FT for t < 0 and t > T , respectively. Recall the following definition of progres-sively measurable sets.

Definition 30.3.6 Denote by P those sets Q in FT ×B ([0,T ]) such that for t ∈[−∞,T ]

Ω× (−∞, t]∩Q ∈Ft ×B ((−∞, t]) .

820 CHAPTER 30. CONTINUOUS STOCHASTIC PROCESSESProof: Suppose Q € [0,7] x Q is progressively measurable. This means for each f,00.1] x QE B((0,t)) x FWhat about (s, @) € [0,t] x Q, (s,@) + Jo Zodr? Is that function on the right F ((0,1]) xF, measurable? We know that QM [0,5] x Q is A([0,s]) x A; measurable and hence&((0,t]) x FY; measurable. When you integrate a product measurable function, you doget one which is product measurable. Therefore, this function must be 4([0,t]) x F;measurable. This shows that (¢,@) + {j 2 (s,@)ds is progressively measurable. Here isa claim which was just used.Claim: If Q is 4 ([0,t]) x YF; measurable, then (s,@) > Jy Zodr is also B ((0,t]) x F,measurable.Proof of claim: First consider A x B where A € A((0,t]) and B € .F;. Then[ Paxear= | Fy Zndr = Xy\ 0) | Xx (s)dr0 JO JOThis is clearly @([0,t]) x A, measurable. It is the product of a continuous function of swith the indicator function of a set in .¥;. Now letG= {0 € B((0,t]) x F : (s,@) > [ Xo (r,@)dr is B((0,t]) x F; measurable}Then it was just shown that Y contains the measurable rectangles. It is also clear thatG is closed with respect to countable disjoint unions and complements. Therefore, Y D0 (.4;) = B((0,t]) x F, where .% denotes the measurable rectangles A x B where B € F,;and A € &([0,4]) = A([0,T]) 1 [0,t]. This proves the claim.Thus if Q is progressively measurable, (s,@) > Jj 20 (r,@) dr = f (s,@) is progres-sively measurable because for (s,@) € [0,t] x Q,(s,@) > f (s,@) is A([0,t]) x F, mea-surable. This is what was to be proved in this special case.Now consider the conclusion of the proposition. By considering the positive and neg-ative parts of @ (X) for @ € E’, and using Pettis theorem, it suffices to consider the casewhere X > 0. Then there exists an increasing sequence of progressively measurable simplefunctions {X,} converging pointwise to X. From what was just shown,t(t,@) > i X,dsis progressively measurable. Hence, by the monotone convergence theorem, (t,@) >Jo Xds is also progressively measurable.What else can you do to something which is progressively measurable and obtain some-thing which is progressively measurable? It turns out that shifts in time can preserve pro-gressive measurability. Let .F; be a filtration on [0,7] and extend the filtration to be equalto Yo and ¥r fort <0 andt > T, respectively. Recall the following definition of progres-sively measurable sets.Definition 30.3.6 Denote by P those sets Q in Fr x B([0,T}) such that for t €[—2, T|Q x (—-,1]NO € F; x B((—~,t]).