Munich Personal RePEc Archive
The long-run relationship between
money and prices in Mexico: 1969-2010.
Gomez-Ruano, Gerardo
Universidad Iberoamericana
2014
Online at https://mpra.ub.uni-muenchen.de/93647/
MPRA Paper No. 93647, posted 07 May 2019 13:18 UTC
1
! "
" ! # "
! "
" # $%&% '($( "
!
2
!
) * + , "
" " $%%-
# " * $%./ 0 1 $%-(
2 " 1
3 "
4 " 0 4 4 '($$ "
" " 0 5 '((% "
"
! 4
" !
" $%&% '($(
4 ## "
6 ### 6 #*
6 *
3 6
! " # $
7
8 8 "
8 ! 8
8 ! 8 0 "
!
!
! "
# #
9 "
" "
: "{ }"
: " "X
(∙)
: " "
= = ( )
1 This is an informal presentation, which follows Priestley (1981), Fuller (1996), Brockwell and Davis (1991), and Box et al. (2008).
4
" ( , )"
( , ) = ( − )( − ) = ( − )( − ) , ( , )
, (∙,∙) 3 ( , )
: " " "
{ }" = " ( , ) = ( , ) , !
#
"
) 7
: 2
; {" } !
7
" =
#$(" , " ) = % 0, ≠ 0
( , = 0
< "
; $" 0 ; $
; {) } $
) = #) + "
# {" }
# |#| < 1 " " = 0" {) } 7
5
) = . # "
/0
) = 1. # "
/0
2
= . # "
/0
= . # ∙ 0
/0
= 0
#$() , ) ) = ) , )
= 1. . # #3" " 3
3/0 /0
2
: # #3" " 3 −
4 ≠ + − $ #$(" , " ) = 0 ≠ 0 , "
− 4 = + − $" ! 4 = $ −
: 4 $ 9 "
0 6 0
# ≤ 0
= −| | 4 4 4 = $ −
7 1. . # #3" " 3
3/0 /0
2 = 1. #3 | |#3" | | 3" 3
3/0
2
= 1. # 3 | |" | | 3
3/0
2
6
= ( . # 3 | |
3/0
= ( #| | 1 − #
# 7 0" 6
#
4 , "
!
: 0 ! " 0
!
# "
"sin( ) cos( ) :
0
Fig. 1. Sine and cosine functions
7
; "
/= 45° " > ?
2A ! @&(
"
!
1 7
B = C cos( ) + sin( )
C ( " "
{B }
B = C cos( ) + sin( )
= C cos( ) + sin( )
= 0 ∙ cos( ) + 0 ∙ sin( )
= 0
" ;
cos( − D) = cos( ) cos(D) + sin( ) sin(D) " C = 0 "
B B = (C sin( ) + cos( ))(C sin( + ) + cos( + ))
= C sin( ) sin( + )
+ C (sin( ) cos( + ) + cos( ) sin( + )) + cos( ) cos( + )
= C sin( ) sin( + ) + C (sin( ) cos( + ) + cos( ) sin( + )) + cos( ) cos( + )
= ( sin( ) sin( + ) + ( cos( ) cos( + )
8
= ( cos( )
# " 2 = 0
B" ( 2( sin ( ) + cos ( ) = 1
" " 0
!
# " E" E ≡ 2A
7
B = C cos(E ) + sin(E )
= C cos(2A ) + sin(2A )
# B ! "
!
1 {EG}G/ ! " {#G}G/ "{HG}G/
" " (G "
3 "
I = .(#Gcos(EG ) + HGsin(EG ))
G/
" {I }
# "
' " !
7
2 That is, J = # + KH = $(cos L + K sin L) = $MNO, where $ = |J| = √# + H , cos L =Q3 , and sin L =R3 .
9
I = . S MN TG( )U|V|
/
S = WX Y XZ 1
2 (# − KH ), ! ≥ 1 0, ! = 0 1
2 \#| |+ KH| |], ! ≤ −1
K K = −1" 4^_( ) "
{S } /
! {EG}G/ {#G}G/ {HG}G/
` = # + KH a = ` = # + K H
: #$(`) = (` − a)(`∗− a) " `∗ 3
`6 " ` = # + KH `∗ = # − KH
# {S } /
#$(S ) = (| | ! ≠ 0 " "
! EG #$(SG) + #$(S G) = (G
!
; "
" I !
! 7
#$(I ) = #$(I0) = #$ c . S
/
d = #$ c. #G
G/
d = . (G
G/
2
= 0 4 sin(0) = 0"
10
; "
I
$% % # "
: "
!
0 > ? ;
7 ;
" "
A " "
! " !
; 9 '7
Fig. 2. An example of the alias effect
# " 0
$ '" 0 #
" ! $ ! '
11
# 0
" ! E∗ #
" "
! E∗ = A # "
0 @
B " ! "
! ! 7 #
' / √10 C
# " ! E (0, A
" "
!
! "
;
4 "
I = . S MN TG( )U|V|
/
: {S } / "
! E (0, A 0
# "
3 This maximum observable frequency is also known as the Nyquist frequency.
12
= eMN U `(E)
e
: `(E) `∗(E) =
`(−E)" `(E) = 0 2 E < 0 "
`" 3
I
6
`(E) 0 SG"
– A A
D I" 7
#$( ) = #$( 0) = #$ g e `(E)
e h = e | `(E)|
e
+
i(E) = | `(E)|
i(E) ! E # i(E) "
i(E) = ℎ(E) E" ℎ(E) > ? "ℎ(E0) E0 /
) 7
{" } > ?
! "
4 Recall that if a function is differentiable, it means its derivative is continuous. And continuity of a function (∙) implies that for every k 7 0 and 0, there is a neighborhood of 0, such that
| ( ) − ( 0)| < k for all in that neighborhood.
13
ℎ(E) =lem" −A ≤ E ≤ A" ( "
) ! E " 3 3
!
9 ( = 1"
> ? " !
Fig. 3. The power of a white noise process
$
; |#| < 1 " = 0 "
; $ # ! E7
ℎ(E) = e( Q nop(U) Qlm m)" −A ≤ E ≤ A" ( "
: # 7 0" !
! 6 # < 0 9
( = 1 # = 0.5 = 0.25
#$() ) = lmQm =rs
14
Fig. 4. An example of the power for an AR(1) process
4 > EE" > ?
" 3
! ; " "
"
!
9 " "
" ! "
& % "
4 "
4 { , } { , }6 "
"
, = eMN U ` (E)
e
15
, = eMN U ` (E)
e
(E) ! E { , } { , }6
(E) = t \ ` (E), ` (E)]
#$( ` (E)) #$( ` (E)) u t F
" " (E)
" 0 "
{ , } { , }" ! E # "
! 0 ≤ (E) ≤ 1 # " 0 "
(E) { , } { , } ! E
(E) = t ( ` (E), ` (E))
#$( ` (E)) t
B " " "
(E) " 0 " !
` (E) ` (E)
A 0 " " (E) ≠ (E)
%
:
$%-( = !
5 At the mexican central bank’s website.
16
0 $
/ 5 > ? $
$ " ' 0
0 G " @ 0 G " /
/ $ $
" ' " @
G 5 $%&(
!
< H # <H#
& $%&%" !
, $%&% '($( .
4 <H# " $%&%
'($(
6 This bureau is called the Instituto Nacional de Geografía y Estadística (INEGI).
7 The CPI is known in Mexico as the Indice Nacional de Precios al Consumidor (INPC).
17
Fig. 5. CPI monthly log-change
" " 5 5 '($( " I '($(
; " "
$%&% $%.'"
0 G
0 0
$% ' % (
5 "
18
Fig. 6. M1a monthly log-change
$% ' % )
/ > ? " " 0
0 "
19
Fig. 7. M4a monthly log-change
# " / $ 2 "
!
& '
9 0 " #* *
> ? > ?
; " 0 "
: ! 7
" $ " / " " $
" /
20
"
$% '
: " 0
* $% ' 7
! >
? $%&&
Fig. 8. Log of the power estimate for prices
0 " &
'((J @
21
* $% ' ( $
7
Fig. 9. Log of the power estimate for M1a
B " > ? " <
&" @" '
* $% ' ) 9 " /
$ 7 > ? " 3 0 "
&" @" ' /" ' -
8 A seasonality around 2.4 months may seem awkward. It is actually not because 2.4 months are one fifth of a year. In other words, this is a harmonic frequency of the yearly frequency. Together, the cycles of length 6, 3, 2.4, and 2 months, help describe common yet not uniform seasonalities.
22
Fig. 10. Log of the power estimate for M4a
A 0 " /
@ /
$% % "
$ " / " D 0
" 7 $
$ , "
0
* & % " ' (
! " ! "
!
23
Fig. 11. Coherency estimate for (M1a, prices)
: $
! '
* & % " ' ) 9 / " "
' 7
( - "
& " $
24
Fig. 12. Coherency estimate for (M4a, prices)
B
$%
; " 0
#
"
$% ( ; "
$ 7
25
Fig. 13. Gain estimate of prices over M1a
9 ' " > ? ( '6
& " > ? ( -=
$
$% ) # / " > ? ( '
$ 6 > ? $
'
26
Fig. 14. Gain estimate of prices over M4a
2 / $ " >
? # " $
/ "
5
5 '((%
& $ " ' #
# "
5 B)<+
4
27
" ! " 3 0 5
; $= " !
$% % " % # "
; " !
> ? !
! "
* $% % " ' ( $
! ! /
@
28
Fig. 15. Coherency at frequency zero for (M1a, prices), rolling sample
$%&% $%.& $%%=
$%.& $%-. "
$%-- $%%/ %
* $% % " ' ) /
+ 0 $%-- "
9 Strictly speaking, it was a crawling-peg regime. For detailed accounts of mexican economic history see, for example, Moreno-Brid and Bosch (2010), or Kuntz (2010).
29
Fig. 16. Coherency at frequency zero for (M4a, prices), rolling sample
7
"
$%-- 0
$% # "
D 0 ! " !
# " 0 > ?
* $% ( $
7 "
0
30
Fig. 17. Gain at frequency zero from prices over M1a, rolling sample
2 " "
* $% ) /
$ # " 7
0 $%-- 0" 0
31
Fig. 18. Gain at frequency zero from prices over M4a, rolling sample
# "
& %
: " 3 "
5 4 4 '($$
A4
; "
5
! $(
10 Benati employs 25-year-width windows. We chose 15-year-width windows, given the smaller span of our sample. Our results are fairly comparable though.
32
# 0 0
!
7 A4;" AI" < " D " 2 " )
K " 0 2 "
A4;" AI" 2
( %
; " 5 "
! ( .=7 3 0
$=
4 4 '($$ " " "
0 " A4;
& (
4 #* $%&% '($(
< "
33
9 $ " $J ( -J
$$B " $J /
$ $J
4 * $=
; " " > ?
> ? " $ # G
; 5
0 A4 AI > ?
9 " / "
Acknowledgements
# 0 < < L < H H
References
; " + : I $%%$ " >, 0 ; < <
) ?" " =%" @" -$. -=-
5 " 1 '((% " >1 ) # ?"+ , -
! " $('." ) < 5 0
11 By a “lasting deviation” we mean what happens at the zero frequency or the “long-run”. Recall that the zero frequency measure tells us what happens for arbitrarily long cycles.
34
5 " ) H " D 0 " < '((- "$ ! . "
& " " D : 4
5 0 " H D ; + $%%$ "$ ! $% " % " "
4
9 " : ; $%%& " ! $ ! " " D :
4
" < : D $%&& " > 4 4 ) * ?"
" @/" $" $=( $&$
I " 4 9 '($( "/ 0 1 &
2 " ) < M 4 N ,
1 " ) $%-( " > # O ?".
3 '" .(" =" $((= $($/
5 " D < D 5 '($( " " 2
" 9 < ) P
H " 5 $%-$ "! . " $ ! " ; H
" 1 $%%- ". / " $% %" H A H
4 " D H 4 '($$ " > # O 7
5 0 ?". 3 '" $($" $" $(% $'-
Appendix
; H $%-$ ℎ(E) { }
ℎ(E)v = 12A . w(4)x(4)v cos(4E)
(y ) / (y )
35
z "w(4) " x(4)v
4 "
! "
x(4)v = 1z . ( − ̅)( |3|− ̅)
y |3|
/
̅ =y∑y/ " { } " " { }
!
" } = 12$'
w(4) =~rm•p€•(~)~ − cos(‚)ƒ" ‚ =„e| |…† =e| |0
# { , } { , }"
(E) = ‡( (E)) + (ˆ (E)) ℎ (E)ℎ (E) ‰ u
(E) " ˆ (E) ! { , }
{ , } ! E" ℎNN(E) K = 1,2
! E
"
12 See Priestley (1981) and Andrews (1991).
36
v = 1\ (E)(E) v ] + \ˆ (E)v ] ℎ (E)v ℎ (E)v 2
u
ℎvŠŠ(E) "
!
v =(E) se∑(y )/ (y )w(4)‹x (4)v + x (−4)v Œcos(4E)"
ˆ (E)v =se∑(y )/ (y )w(4)‹x (4)v − x (−4)v Œsin(4E)"
x (4)v =y∑ \ , − •••]\ , − •••]
# #$\ v ] = }(E) 2z (1 − (E) )
9 " { , } { , }"
(E) = ( (E)) + (ˆ (E)) u ℎ (E)
v =(E) Ž\ (E)v ] + \ˆ (E)v ] • u ℎ (E)v
" ! " "
37
# #$\ v ] = }(E) 2z (E) •1 + 1 (E) ‘