See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/237799633

On the Second Borel-Cantelli Lemma for Strongly Mixing

Sequences of Events

Article  in  Journal of Applied Probability · June 1997

DOI: 10.1017/S0021900200101020

CITATIONS READS

12 351

1 author:

Dirk Tasche
Swiss Financial Market Supervisory Authority FINMA
66 PUBLICATIONS   3,480 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Calibration of posterior probabilities View project

Proving Prediction Prudence View project

All content following this page was uploaded by Dirk Tasche on 04 November 2017.

The user has requested enhancement of the downloaded file.

 J. Appl. Prob. 34, 381-394 (1997)

Printed in Israel

C Applied Probability Trust 1997

ON THE SECOND BOREL-CANTELLI LEMMA

FOR STRONGLY MIXING SEQUENCES OF EVENTS

DIRK TASCHE,* Technische Universitdt Berlin

Abstract

Assume a given sequence of events to be strongly mixing at a polynomial or exponential

rate. We show that the conclusion of the second Borel-Cantelli lemma holds if the series

of the probabilities of the events diverges at a certain rate depending on the mixing rate

of the events. An application to necessary moment conditions for the strong law of large

numbers is given.

BOREL-CANTELLI LEMMA; STRONG MIXING; STRONG LAW OF LARGE NUMBERS

AMS 1991 SUBJECT CLASSIFICATION: PRIMARY 60F20

SECONDARY 60F15

1. Introduction

If (A,),~ is a sequence of independent events, then the relation

(1) IP(A,)=co => P UAm = 1

n=l n=1 m=n

holds. This is the assertion of the second Borel-Cantelli lemma. If the assumption of

independence is dropped, (1) fails in general.

However, (1) remains true if independence is replaced by pairwise independence

(Durrett (1991), Theorem (6.6) ch. 1) or by uniform mixing (Iosifescu and Theodorescu

(1969), Lemma 1.1.2'). Rieders (1993) showed by an example that strong mixing without

any further assumption does not guarantee the validity of (1). Hence, the fact that

Yoshihara's (1979, Theorem 1) Borel-Cantelli lemma for strongly mixing events requires

a condition on the size of the probabilities is not surprising.

This article is intended to give refinements of Rieders' and Yoshihara's statements.

Intuitively, one might expect that the less dependent the events are the better relation

(1) will fit. Our results confirm this intuition in the following manner.

For a fixed rate of convergence of the strong mixing coefficients of the events A,, we

determine how fast the rate of divergence of the series C P(A,) has to be in order to

make (1) valid. In the case of decreasing probabilities (Theorem 2.2) this rate is smaller

than in the general case (Theorem 4.1 and Proposition 4.3). By Theorem 2.2 we deduce

Received 20 June 1995; revision received 9 October 1995.

* Postal address: Ingeborgstrasse 62, 81825 Mtinchen, Germany.

381

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 382 DIRK TASCHE

necessary moment conditions (Corollary 2.3) for the strong law of large numbers in the

case of a strongly mixing stationary sequence. These conditions are useful in Section 3

to show by an example (Proposition 3.1) that the conclusion of Theorem 2.2 is essentially

sharp in the case of a polynomial mixing rate.

Throughout this paper we will denote by [x] the integer part sup{z z E : z ? x} of the

real number x, by I(A) the indicator function of the event A and by log+ the function

log+(x):= log(max(x, e)), where log denotes the natural logarithm.

2. Decreasing probabilities

First let us specify the mixing coefficients we are dealing with in this and the subsequent

sections.

Definition 2.1. Let (Zk),~N be any sequence of random variables on a common prob-

ability space.

(a) The numbers az(n), defined for all n E N by

az(n) = sup{IP(A n B)-P(A)P(B)I: kE f-,

AEa(Zj :1 <j =k), BEa(Z : k+n<j< o)},

are called strong mixing coefficients of (Zk). The sequence (Zk) is called strongly mixing

if lim,n + z (n) = 0.

(b) The numbers az(n), defined for all n E =- by

az(n) := sup{ IP(A n B)-P(A)P(B)I : k E N, A E a(Zi : 1 _j < k), B E a(Z,+k)},

are called a-mixing coefficients of (Zk). The sequence (Zk) is called !-mixing if

limn - az (n)= 0.

(az(n)) and (Iz(n)) are obviously non-increasing sequences with az(n) ? az(n) for all

n. Thus a-mixing is weaker than strong mixing. We will omit the index Z if there is no

danger of confusion.

For a sequence (Ak)k ~N of events the corresponding mixing coefficients and mixing

properties are defined by considering the random variables Zk := I(Ak). With these

notations the main result of this section is as follows.

Theorem 2.2. Let (Ak)k E N be a sequence of events such that k 4 P(Ak) is non-increas-

ing. Then P(n,"= _ LU"M, Am,)= 1 is implied by either of the following two conditions.

(i) There exists some rE [- 1,co[ such that

(2) n-/(r'+2)P(A,)=xo and n'4a(n)< o.

n=l n=l

(ii) There exists some bE ]0, 1[ such that

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
On the second Borel-Cantelli lemma for strongly mixing sequences of events 383

SP(A,)

(3) ---n= 00 and a(n)= O(b").

n= log1(n)

We will prove Theorem 2.2 using some lemmas. But first let us state the announced

corollary whose proof will be deferred to the end of the section. Compare the corollary

to the strong law of large numbers for i.i.d. real random variables (X,), , I which can be

formulated as

in

'E(IX, )< oo if and only if lim - Xj exists and is finite a.s.'

n-x? n j=1

Corollary 2.3. Let (Xn,),, be a sequence of identically distributed real random

variables such that limn.n- I.n'=, X exists and is finite a.s. Then we get the following

implications.

(i) XC0"= 1 x (n)/n < co o E(log+(IX, I))< o.

(ii) IX, n"5x(n)< o for some r E ]- 1,cj[ = E(XI i(r+)/(r +2)< cc.

(iii) 5x(n)= O(b") for some b E]0, 1[ = E(IXlllog+(IX,1))< 00.

In Section 3 we will see by an example that implication (ii) cannot be improved

essentially.

The proof of Theorem 2.2 is based on the following simple inequality, which is closely

related to inequality (6) by Tasche (1995).

Lemma 2.4. Let (A,)n ,, be a sequence of events with corresponding &-mixing

coefficients (5(m)),, , and let k ,---, k, be any positive integers. Define so:= 1 and

s, := 1 +i=, k fobr j 1 ,- ,.-, n. Then for all events B1, B2, B,B with Bj E a(Aj) for

j = 1, 2,..., s,, -1, s,, we have

(4) P2 B, 6 P(B,,) + I (kj) H P(BS,).

i=1 j=o j=1 I=j+

Proof. (By induction on n.) For n= 0 there is nothing to show. Suppose that (4)

holds for some n > 0. By the definition of a(k, +,) and the induction hypothesis we obtain

then

P B) <P( B BBl , S+ < P( B, P(B,~)+ +(k? , )

n+l n n+ l

< H P(BP)+ L (kj) H P(Bs,)+a?(k,,,),

j=O0 j=1 =j+l

which equals the right-hand side of (4) with n replaced by n + 1.

The next lemma supplies us with the probabilistic part of the proof of Theorem 2.2.

Lemma 2.5. Let (k,), , be a sequence of positive integers and let (A,) and the numbers

s be as in Lemma 2.4. Suppose that X&1~ a(k,)< c and that Xf=0 P(A, ,)= ct for every

ne R. Then P(nQ _ U=, A,,)= 1 holds.

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 384 DIRK TASCHE

Proof. Denote by A the complement of the event A. Thus we have to show that for

fixed n eN

(5) lim P ( Am = 0.

N-+ oo m=n

Since P(Qm=, Am) is non-increasing in N, we only need to prove (5) for an appropriate

subsequence (N,), N. Choose N,:= n + s, for all v E EN.

Now fix an e > 0. By assumption there is an integer M such that I_ =M+ I 1(km)< e.

The divergence of the series 7~_ P(A,+j) implies

0 = lim 7 (1 -P(A,+)) = lim Hi P(A,+sj).

v_0oo j=M+1 v+0oo j=M+1

Therefore we can deduce by Lemma 2.4

lim sup P lim sup P(An+sj)

Vm=n -V_00 j=0

M 00_

+ E ~(km) + E (k,) lim sup H P(A,+j) ? e.

m=M+1 i=1 vXo0 j=i+l

This yields (5), because e can be chosen arbitrarily small.

For the analytical part of Theorem 2.2 we need the following change-of-variable

formula for series, which can be proven by means of the integral test. Its proof will be

omitted.

Lemma 2.6. Let ? be a positive number and let 4): [, oo[ -+ ]0, co[ be a function with

the following properties.

(i) 4 is continuous with limx_.oo (x)= o.

(ii) 4 is continuously differentiable in ], co[. 4)' is positive and non-increasing or non-

decreasing. limx,4 0'(x) 40 exists and is finite.

(iii) There exist constants m, M > 0 such that m ? 4)'(x + 1)/1'(x) ? Mfor allx E [~-, o[.

Denote by y the inverse function of 4 and set : = 4(?). Then for every positive and non-

increasing sequence (a(n)),1N the series E,+i ,1 a([y(n)]) converges if and only if the series

n=[4]+, a(n)4'(n) converges.

Proof of Theorem 2.2. A straightforward application of Lemma 2.6 yields the follow-

ing implications.

For r > -1:

(6) Z n(n)<cx = a([n1/+l1)])< oo.

n=l n=l

For r= -1:

Sd(n)

n=l n2 n=l

and

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 On the second Borel-Cantelli lemma for strongly mixing sequences of events 385

(7) a(n)= O(b") with bE ]O, 1[ =~ ( ) < 00for all a ]0, -log b[.

These observations suggest how to choose kn in Lemma 2.5. First, suppose r > - 1 and

let k:'= [j"r+ 1)]. Then

(8) s,, = kj jI L+<f xl/(r+)dxl = ((n+ 1)(r+2)/(r1)- 1)

Use Lemma 2.6 with

r+l

(x) : = ((x + 1)(i +2)/(W + 1)

r+2

and = y-' to deduce that, for all nE N,

Z P(An,+[lm)])= 00 z m-1(r+2)P(An+m)=co

S= 1 m = 1

(9)

9 mZ-1/(r+2)p(Am)= co.

As (P(Am)),N is non-increasing, by (8) for all m, nE N the inequality P(An+[Am)I) <

P(A,,+,,) holds. Hence, under condition (i), Lemma 2.5 together with (6) and (9) implies

P(-nQ:I U",, Am)= 1. The proof for the case r= -1 is obvious.

Now, let us consider the situation under condition (ii). Fix an arbitrary number

a E ]0, log b[ and let k := [log+(i)/a]. Then the same reasoning as in (8) leads to

[e-1 1 n+l

(10) sn + - (n + 1)log .

aae

Thus, this time we have to choose

e-1 1 x+l

(x) := - + - (x + 1)log

aae

and again b := -'. Observe that limy,. /'(y+ 1)1'(y)= 1. By Lemma 2.6 we get

therefore for all n E

SP(A+fmli+[))= e z P(A,+ )/'(m)=c0,

m= 1 mf=1I

and by (10), (7) and Lemma 2.5 it only remains to prove

" P(A,) o

(11) o =o 0 Z P(A,+m,,)'(m)=co forallnEJN.

., = log +(m) m=]

By the definition of y the inequality y(x) ? x/a holds for x ? 3. Hence, 4(y) ? ay for

y > 4(3). This in turn implies

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 386 DIRK TASCHE

aa

log(4(y)+ 1) = log(ay + 1)

for y > 0(3), and thereby also (11).

Proof of Corollary 2.3. Because the arguments in the other cases are very similar,

we confine ourselves to considering the case of condition (ii). Since the random variables

X, are identically distributed, the following implication is easily deduced by means of

the integral test for every r >- 1:

(12) n-fi(r+2)P(IXI >cn)< oo for some c > 0 E(IX,I(r+')/(r+2))< C

n=l

Fix any r> -I and assume E(IX,I(r+'1)/(r+2))=) . Then (12) implies that the

series 1~ n- 1(r+2)P(IX,I >cn) diverges for all c> 0. By Theorem 2.2 it follows that

P(sup, IX,/n = oo)= 1. But sup, IX,/nl I 2 sup, In-' ` =1, XI, and thus n-'C= X

cannot converge almost surely. This contradiction implies the assertion.

3. An example

We will show that Rieders' (1993) example in a slightly modified form can also serve

as a counterexample in the situation of Corollary 2.3.

Proposition 3.1. For every r > - 1 there exists a sequence (X,),,n of identically dis-

tributed real random variables with the following properties.

(i) E , n'a(n) < 0o.

(ii) E(IX,I(r+1)/(r+2) log+(IXlI))= 00.

(iii) limn- n-1 j=1 Xi exists and is finite almost surely.

Remark. Unfortunately, the example does not seem to work in the case of an exponen-

tial mixing rate. So the question is still open whether exponential mixing implies the

second Borel-Cantelli lemma without any specified rate of divergence of I P(A,).

Proof of Proposition 3.1. Fix r > - 1 and define a probability distribution (Pk)k EN

on the positive integers by setting

Pk := ck-(3 +r)(log(k))-2,

where c is a constant ensuring ECk~ Pk = 1. We define the transition probabilities of the

Markov chain (Y,)n,=N by

p , if i=1,

P(Y,+I=j I Y,=i):= 1, if i=j?+1,

0, otherwise.

Then (Y,) has a unique invariant distribution u = (Uk)k EN given by j, =( ]jpj)-' and

Uk = L =1 jk pj for k 2. Choosing o( Y,)= j, we make (Y,) stationary. Note that (Y,)

is Rieders' example in time-reversed representation.

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 On the second Borel-Cantelli lemma for strongly mixing sequences of events 387

By Corollary 1 of Kesten and O'Brien (1976), we know that the strong mixing co-

efficients cy (n) satisfy

(13) n nWa,(n)< oo.

n=l

Let g: N -+ ]0, oo[ be a non-decreasing function (which will be specified later) and

let X,:=(-j1)'g(Y,,) for all nE N. Then (13) implies (i). Define To:

inf{n > 1: Y,= 1}, Tk:= inf{n > Tk- : Y,= 1}, Rk:= Tk- Tk-l and N(k):=

I{1 ?j < k: Y= 1 }-1 for k ? 1. Then on the event {To<n} the following identity

holds a.s.

1 1To 1 N(n) 1

S- X= X+ E x+x

1j= j = 1 n i=l j=Tij+l + j=TN(n)+l

(14) 1 I N) R n

1 N~n)1

(140 ( 1).g(j) + - (Y-11) g(j) + - I Xi -

/j=l 1 "= /j1 j j=TN(,)+l

Since g is positive and non-decreasing we have I|=I (- 1)jg(j)l ?g(k) for all k E N.

Therefore we get the inequalities I ,1 (-1) g(j)l ?g(R,) and

1 1 "(")+'1 TN("n)

xiz A?- I x? +-zX

Sj = TN(,,)+ I n1 = TN(n) +1 l j=n+1

1 R,(n)+l 1 TN(n)+L-n

- (-01)+g(j) - Z (-1)ig(j)

n j=1 n j=1

< - g(RN(,n)+1).

Observe that (Rk)kE, is i.i.d. with P(R, =j)=pj for all jE N. Hence, if we assume

1~, g(j)pj to be finite, by the strong law there exists a number hE CR such that

1 m Ri

lim - I (- 1) g(j) = h a.s.

m--c0 m i= 1 j=l

By the same argument limm,, Rm/m = 0 holds. By the ergodic theorem limn,, N(n)/n=

y, > 0 a.s., and by (14), this implies the almost sure convergence of n-' C_, X,. So all

we have to do in order to obtain (ii) and (iii) is to choose g(k) such that

E", g(k)p, < 0o and that

E(lX,1(r+l)/(r+2) log+(lX, )) = Z (g(k))(r+)/(r?+2) IOg(g(k))k= *c.

k=l

This is achieved by the choice g(k): = kr+2

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 388 DIRK TASCHE

4. General probabilities

Among the assumptions of Theorem 2.2 the monotonicity of the probabilities appears

unnatural when compared to the assumptions of the second Borel-Cantelli lemma under

independence. Durrett's (1991, Theorem (6.6) ch. 1) version of the lemma suggests a

way to avoid this taint and even to get a stronger conclusion. The price to be paid for

the improvement is higher rates of divergence of E P(A,) under comparable mixing

conditions.

Notation. For any sequence (A,),, Nof events in a common probability space and all

nE- let

(n) := sup{IP(A.+k I Ak)-P(An+k)I : kE -N, P(Ak) > 0}

and

y(n) := sup{ P(An+k n Ak)-P(An+k)P(Ak)I : k E }.

These coefficients are two-dimensional analogues of the well-known strong mixing

and uniform mixing coefficients and are obviously dominated by them. The idea of the

proof of parts (b) and (c) of the following theorem has been suggested by Rio's (1993)

proof of his Theorem 1.1 (a).

Theorem 4.1. Let (A)., E be a sequence of events in a common probability space.

Define S,: = ~, I(A,) and c. := 1 P(A,) for all n E N.

(a) If liM , c, = oo and E~kl r(k)< oo, then lim ,o SI/c, = 1 almost surely.

(b) If lim,, n-1+(t+2)cn= OO and EX=I k'y(k)< oo for some t O, then limn , SI/c = 1

in probability.

(c) If limn-a c,(log+(n))-'='0 and y(n)=O(b") for some b ]O, 1[, then

limn, S,/c, = 1 in probability.

Remarks. (1) Part (b) of the theorem can be regarded as a revision of Yoshihara's

(1979) Theorem 1. He stated his result under the assumptions kl=1 (a(k))" '+') < co for

some t > 0 and lim inf,,, n"P(A,) > 0 for some a E ]0, (t + 1)/(t + 2)[ on the mixing

coefficients and the probabilities resp. of the events. These conditions are slightly

stronger than our conditions. Concerning the probabilities this can be shown by a short

calculation. For the comparison of the conditions on the mixing coefficients, see Rio

(1993, pp. 592/593).

(2) If limn, S,/c, = 1 in probability, then limk,, Sn/lCnk = 1 a.s. for some subsequence

and therefore P(on, Un,_ Am)= 1. Nevertheless, the question is interesting whether

the conclusion in (b) and (c) can be sharpened to almost sure convergence. Also, we

have to ask whether the rate condition on the mixing coefficients in (a) is really necessary.

In Proposition 4.2 and Proposition 4.3 we provide partial answers to these questions.

(3) Concerning the rate of divergence of I P(A,), the conditions in (b) and (c) are

obviously stronger than the corresponding conditions in Theorem 2.2. To get an intuition

for the gap, observe that in the case of non-increasing probabilities the condition

C",l n-li(t'+)P(A,)= c0 for some e > 0 implies lim sup._, n-1/' 1, P(Ak)= c00.

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 On the second Borel-Cantelli lemma for strongly mixing sequences of events 389

(4) It would be nice to find a sequence of strongly mixing events Am such that

P(O,I lU,, Am) equals 1 but S,/c, does not converge in probability.

Proof of Theorem 4.1. We follow as far as possible Durrett's (1991, p. 44) proof. For

all nE GN and all 6 > 0 Chebyshev's inequality implies

var(S,,)

P(IS,- c, I > 6cn)(,,)

(6c)2

(15) = (cn) ( P(A)-P(Ak))+2 1 (P(Ak A-P(Ak)P(A))

5 -+2(6c,,)-2 IP(Ak n A)-P(Ak)P(A)I .

Cn k=1 j=k+I

(a) Let M:= I' rI (k)< oo. Then (15) implies

1 n n l+2M

P(I S,-cn > 6c,) <$ +2(6c,)-2 ql(j-k)P(Ak) <1 ?2M

62Cn k= 1j=k+ +k 62Cjn

and the rest of the proof is the same as in Durrett's book.

(b) By (15), to get the desired convergence in probability we have to show

(16) lim c2 Z Z IP(Ak Ai)-P(Ak)P(A j)I =0.

n - k=1 j=k+1

For k ? 1, let m,: = [k'/('+2)] and observe

n-In

k= j=kIP(AkA)-P(A)P(A)I

n-1 nik+k n-1 n

(17) <n - min(P(A ),P(Aj)) + (j - k)

k=1 j=k+l k=l j=mk+k+1

n-I n-I n-k

<I kll(t+2)P(Ak)+ (J).

k=l k=1 j=mk+l

The first term is well-mannered, since

n-1

(18) c;2 1 kll(t+2)P(Ak) ? n!!2/2)Cn 0 as n -* c.

k=l

Concerning the second term on the right-hand side of (17), note that

mk +1 j l n-k [kl/('+2)] Ij- 1 and k n n-j

k ? n and kl/"~' +2 j

kk ? min(n,j+2).

This implies

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 390 DIRK TASCHE

n-1 n-kn

f E y(j) ? Ey(j)min(n,jt+2)

k=1 j=mk+l j=1

(19) "m.

j- '+(j)+n 2 (j).

j=1 j=mn+1

Since Y~=. n'y(n) < oo, we have

n' E y(k)= 1: k'y(k) -O as n --oo.

k=n k=n

By (19), this yields

n-1 n-k

k=1 j=mk+1

< C(m)2 E J'T(J)+cnn2n2/(t+2) (n 1/(t+2) ()

j=1 j=mn+l

? (c'nl/(t+2))2 J t(J)+(m+ 1)t (Jj) -+ 0 as n oo.

j=1 j= mn +

Together with (17) and (18) this implies (16).

(c) Again we have to prove (16). Choose M > 0 such that y(n) Mb" for all n E N.

Let mk := [1ogl/b k] for k ? 1. Then

Z Z IP(Ak nA)-P(Ak)P(A)I

k=1 j=k+l

n-1 k+mk n

(20) - < P(Ak)+M Y, b' -k

k=1 j==k1 j=k+lmk+

n-1 n

( mkP(Ak)+M Z bj-)

k=1 j=k+mk+l

The definition of mk leads to

n-1 n - n-1

k=1 j=k+mk+l -b k=l

(21) 1 n 1 1+log n

1-b k=1k 1-b

Combining (21) with (20) implies

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 On the second Borel-Cantelli lemma for strongly mixing sequences of events 391

cn2 Z Z IP(Ak nAj)-P(Ak)P(Aj)I

k=1 j=k+l

log (n) 1 " logj 1 +log n

EC P(A)+

c, (-log b)c, j=1 log+(n) (1 -b)c2

which tends to 0 as n -* cG, since limn,- (log n)c, 0=O by assumption. This completes

the proof.

Up to now, we have proved our results by elementary computations. In order to give

an impression of what can be reached by invoking general limit theorems, we finish our

considerations with two further results on the a.s. convergence of S,/c,. The first proposi-

tion is based on ideas from Blum et al. (1963) and their concept of *-mixing.

Proposition 4.2. Let (A,)n,, be a sequence of events such that 1 I1 P(A,) = co. For

every n E N let

P(A)P(B)
(n):= sup -(A)P(B) - 1 : AE a(A,,..., Ak),

BE a(An+k), P(A)P(B)> 0, k E

and suppose limn., ,(n) = 0. Then

. I= I I(Ai)

lim =1 a.s.

Proof. Let X,:= I(A,)-P(A,), ~ "= a(A1,..., A,) and b,:= Y= P(A,). Without

loss of generality assume bn > 0 for all n E N. Our mixing assumption implies that

equation (2.22) from Hall and Heyde (1980, p. 40) is satisfied (with f(n) := f(n)). Hence,

by Theorem 2.20 of Hall and Heyde the assertion will follow if we show that

supn b1' = E(IXX I)< oo and 1"~ b-2E(X,2)< Oc. The first relation is obvious. For the

second we compute

N N P(A,)

Z b <2E(X2) b2

,n=1 =,n b 2

1 u bn-b_ 2 1 2

P(A - n=2 bnb, _ P(A1) bu = P(A )

Letting N --+ o yields the desired inequality.

Our last proposition is based on a recent result of Rio (1995) on a.s. convergence of

strongly mixing sequences of square integrable random variables.

Proposition 4.3. Let (A,),~ be a sequence of events such that C., P(A,) = c~ and

( 2P (A )t/(t+1)
(22) n < co

for some t ? 0. Suppose that the c-mixing coefficients of (A,) (cf. Definition 2.1) satisfy

Xl1,L n'a(n) < c0. Then

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
392 DIRK TASCHE

Z.X= I(A,)

lim =1 a.s.

n -o X1=1 P(Ai)

Remarks. (1) In the proof of Proposition 4.2 we had to show

0 P(A,) <

n=1 (i P(A,))2

For large t, condition (22) is very similar to this inequality, but it does not follow from

I" I P(A,)= 00o.

(2) There is no clear hierarchy between Theorem 4.1(b) and Proposition 4.3. For

t=O the condition lim,, n-1/(t+2)c, = oo is implied by (22). However, for t E N choose

1, if n= k'+2 for some k E N

P(A,) =

0, otherwise

to see that this implication fails in general. Conversely, the example

P(A) log n (t +1)/(t+2)

n

shows that (22) does not follow from lim,_ n-U1(t+2)cn= 00.

(3) If the sequence (P(A,)),nA is non-increasing, for every t ? 0 condition (2) of

Theorem 2.2 is implied by the assumptions of Proposition 4.3. Indeed, by Kronecker's

lemma condition (22) implies

=l_ 1P(A,)t/(t+l)

lim = 0.

n -+ (n=1 P (A,))2

Since

P(An,) < P(An,)t(t+1) P(A,)(t+2)/(t+1)

i=1 i=1 i=1

for every n E ~, it follows that ',I P(A,)(t+2)/"(t+1)=o, which in turn implies that

-nO= n-1/(t+2)P(A,)= o00.

Proof of Proposition 4.3. Define the random variables X, and the numbers b, for

n E N as in the proof of Proposition 4.2. The assertion will be implied by Kronecker's

lemma, if we prove a.s. convergence of the series X",1 X,/b,. To get this convergence we

will examine the assumptions of Rio's (1995) Corollary 1. In his paper, Rio only uses

the notion of strong mixing. However, a careful inspection of his Lemma 1 reveals that

in Corollary 1 the assumption on the strong mixing coefficients can be replaced by the

same assumptions on the -mixing coefficients. Thus we have to show

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 On the second Borel-Cantelli lemma for strongly mixing sequences of events 393

Sb'2 a(x/2)(Qi(x))2dx < oo,

i=1

where the functions Qi for i N and a are defined by

Qi(u):= sup{t O : P(jIXA > t) > u} and

a(u):= sup{k+ I : kE CNo, (k)> u} resp.

(with 5(0) := and sup 0 := 0).

First, consider the case t > 0. Then, by Hilder's inequality, for every i e N with the

property P(A,) $ we have

a(x/2)(Qi (x))2dx

fg 1 )1/(t+ 1) 01 t/(t+1)

0( a(x/2)t+ldx j0(Qi (x))2(t + )/tdx)

= 2 (n +t- 1)'(a(n) - 5(n + 1))

1=0

S[(1 -P(Ai))P(Ai)2(t + 1)/'t + P(Ai)(1 - P(Ai))2(t + 1)/t ]t/(t + 1)

00/(t + 1)

< 2 E(n+ 1)'l(a(n) - a(n+ 1)) 2P(A,)''' +1).

n=0

If P(A,)> this inequality also holds true. Thus, since by assumption

0i-1 P(Ai)"(t +l)'b~< 00 it remains to show Y"0 (n + 1)'+'(a(n) - (n+ 1))< c. This

follows from

I (n+ 1)"'(a(n)- (n+ 1))

n=0

_ (0) + (t + 1) (n + 1)'i(n), N

n=l1

for all N E N. Concerning the case t= 0, observe that Q, (x) ? 1 and therefore

Sb 2 a(x/2)(Q(x))2dx ? (n) b2<.

i=_ n=O i=1

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
All use subject to JSTOR Terms and Conditions
 394 DIRK TASCHE

References

BLUM, J. R., HANSON, D. L. AND KOOPMANS, L. H. (1963) On the strong law of large numbers for a class

of stochastic processes. Z. Wahrscheinlichkeitsth. 2, 1-11.

DURRETT, R. (1991) Probability: Theory and Examples. Wadsworth, Belmont, CA.

HALL, P. AND HEYDE, C. C. (1980) Martingale Limit Theory and its Application. Academic Press, Boston.

IOSIFESCU, M. AND THEODORESCU, R. (1969) Random Processes and Learning Springer, Berlin.

KESTEN, H. AND O'BRIEN, G. L. (1976) Examples of mixing sequences. Duke Math. J1 43, 405-415.

RIEDERS, E. (1993) The size of the averages of strongly mixing random variables. Statist. Prob. Lett. 18,

57-64.

RIo, E. (1993) Covariance inequalities for strongly mixing processes. Ann. Inst. H. Poincard 29, 587-597.

RIo, E. (1995) A maximal inequality and dependent Marcinkiewicz-Zygmund strong laws. Ann. Prob.

23, 918-937.

TASCHE, D. (1995) First-passage time moments of strongly mixing stationary sequences. Preprint 452.

Technical University of Berlin.

YOSHIHARA, K. (1979) The Borel-Cantelli lemma for strong mixing sequences of events and their applica-

tions to LIL. Kodai Math. J 2, 148-157.

This content downloaded from 155.198.30.43 on Thu, 17 Mar 2016 22:46:52 UTC
View publication stats
All use subject to JSTOR Terms and Conditions