r = r r| r|
≤ C
1ς
C
γ w
h|1 − h h
′
|
γ c
w
c
b
−c
w
c
b
|v|
γ
dv = C
1ς
C
γ w
h |b
′
− bb
′
|
γ c
w
c
b
−c
w
c
b
|v|
γ
dv ≤ C
2ς
h|b
′
− b|
γ
, where C
sς
is a finite constant that depends on ς
i
; for example, C
1ς
≡ sup
u
c
∈U
c
u
d i
∈U
d
m
ς
u
c
+ hv, u
d i
f u
c
+ hv, u
d i
dv ∞. Similarly, K
2
=
u
d i
∈U
d
,u
d i
=u
d
,, ,
,w u
c i
− u
c
h ,,
, ,
γ
|λ
′
− λ|
γ
m
ς
u
c i
, u
d i
× f u
c i
, u
d i
du
c i
= h|λ
′
− λ|
γ u
d i
∈U
d
c
w
c
b
−c
w
c
b
w v
γ
m
ς
u
c
+ hv, u
d i
× f u
c
+ hv, u
d i
dv ≤ C
3ς
h|λ
′
− λ|
γ
≤ C
3ς
hn
−γ σ
|r
′
− r|
γ
. It follows that
E [|K
b
′
r
′
,iu
− K
br,iu
|
γ
ς
i
] ≤ c
γ
C
2ς
∨ C
3ς
h|b
′
− b|
γ
+ |r
′
− r|
γ
, A.3
where a ∨ b = max a, b . Then by the C
r
inequality E
[|K
b
′
r
′
,iu
|
γ
ς
i
] ≤ c
γ
E [|h
′
K
h
′
λ
′
,iu
− hK
hλ,iu
|
γ
ς
i
] + c
γ
E [|hK
hλ,iu
|
γ
ς
i
] ≤ c
γ
C
2ς
∨ C
3ς
h|b
′
− b|
γ
+ |λ
′
− λ|
γ
+ C
4ς
h ≤ C
5ς
h. A.4
Note that |h
′ α
− h
α
| = |b
′ α
− b
α
|n
−αδ
= αn
−αδ
b
∗ α−1
|b
′
− b| ≤ C
6
h
α
|b
′
− b|, A.5
where the second equality follows from the intermediate value theory and b
∗
lies between b
′
and b. Then by the C
r
inequality, and
A.3 –
A.5 , we have that for any α 0 and γ 0,
E [|h
′ −α
K
b
′
r
′
,iu
− h
α
K
br,iu
|
γ
ς
i
] ≤ c
γ
E [|h
−α
K
b
′
r
′
,iu
− K
br,iu
|
γ
ς
i
] + c
γ
E [|[h
′ −α
− h
−α
]K
b
′
r
′
,iu
|
γ
ς
i
] ≤ c
γ
h
−αγ
E [|K
b
′
r
′
,iu
− K
br,iu
|
γ
ς
i
] + c
γ
|h
′ −α
− h
−α
|
γ
E [|K
b
′
r
′
,iu
|
γ
ς
i
] ≤ c
2 γ
h
1−αγ
|b
′
− b|
γ
+|λ
′
− λ|
γ
+ c
γ
C
6
h
−αγ
|b
′
− b|C
5ς
h ≤ C
7ς
h
1−αγ
|b
′
− b|
γ
+ |λ
′
− λ|
γ
. In the general case where p
c
1 or p
d
1, with a lit- tle bit abuse of notation we use row vectors to denote
the bandwidth parameters. We can write h
s
= b
s
n
−δ
s
and λ
t
= r
t
n
−σ
t
for some b
s
, r
t
, δ
s
, σ
t
0, s = 1, . . . , p
c
, and
t = 1, . . . , p
d
. Let b = b
1
, . . . , b
p
c
and r = r
1
, . . . , r
p
d
.
Similarly, h
′
, λ
′
= h
′ 1
, . . . , h
′ p
c
, λ
′ 1
, . . . , λ
′ p
d
and b
′
, r
′
= b
′ 1
, . . . , b
′ p
c
, r
′ 1
, . . . , r
′ p
d
are connected through h
′ s
= b
′ s
n
−δ
s
and λ
′ t
= r
′ t
n
−σ
t
for b
′ s
, r
′ t
0. Then using the fact that our multivariate kernel function is a product of univariate kernel
functions, we can follow the above arguments and readily show that
E [|h
′ −α
K
b
′
r
′
,iu
− h
α
K
br,iu
|
γ
ς
i
] ≤ C
8ς
h
1−αγ
b
′
− b
γ
+ λ
′
− λ
γ
, A.6
where C
8ς
is a finite constant depending on ς
i
. Below we use
C to denote a generic constant that can vary from equation to
equation. Now we prove i. Let J
1n,st