Slides - FGV/EPGE

Report
Variable Long-Term Trends
in 100+ Mineral Prices
John T. Cuddington
William J. Coulter Professor of Mineral Economics
Colorado School of Mines
August 16-17, 2012
Rio de Janeiro, Brazil Conference on
“The Economics and Econometrics of Commodity Prices”
sponsored by the Getulio Vargas Foundation and VALE
My Home:
Colorado School of Mines
Division of Economics and Business
www.econbus.mines.edu
• CSM is the oldest university in the CO state system
(1874-)
• CSM is a small, elite university focusing on engineering
and applied science
• CSM’s Division of Economics and Business Programs
o BS - Economics
o MS – Engineering and Technology Mgt (ETM)
o MS, PhD – Mineral and Energy Economics
• I’d show you pictures, but we all can’t live there!
2
Long-Run Trends in Mineral
Prices: Overview
• Motivation: policy, theory, empirics
• Objective: to explore the use of band-pass
filters for extracting LR trends
• Empirical results for some long-span data
• Conclusions
• Extensions: Super Cycles (20-70 years)
3
Motivation - Policy
• Policy makers – keen interest during periods
of sharply rising resource prices, perceived
‘shortages’ or geo-political threats to
availability
• Will we run out of various nonrenewable
resources? (Limit to Growth debate)
• Will they be exhausted before they become
economically obsolete, or vice versa?
• Real prices are a key measure of economic
scarcity; long-span mineral price data is
readily available
4
Tilton (2003) RFF Book: On
Borrowed Time? Assessing the Threat
of Mineral Depletion
• “Mining and the consumption of nonrenewable mineral
resources date back to the Bronze Age, indeed even the
Stone Age…(p.1)
• “What is new is the pace of exploitation. Humankind
has consumed more aluminum, copper, iron and steel,
phosphate rock, diamonds, sulfur, coal, oil, natural gas,
and even sand and gravel during the past century than
all earlier centuries together. (p.1)
5
Causes of Explosion in Mineral Use
o Advances in technology allow [exploration and] extraction…at lower and
lower cost. [Shifts mineral supply curves down]
o Advances in technology also permit new and better mineral commodities
serving a range of needs.[Shifts mineral demand curves out/up]
o Rapidly rising living standards in many parts of the globe are increasing
demand across the board for goods and services, including many that use
mineral commodities intensively in their production [Shifts the derived
demand for minerals out/up]
o Surge in world population means more and more people with needs to
satisfy. [Shift the derived demand for mineral in or out depending on the
relative mineral intensity of various goods.
Source: (Tilton 2003, p.1)
6
Hotelling Theory of
Nonrenewable Resources
• Hotelling’s (1931) ‘benchmark’ theory of nonrenewable
resources
o Shadow price of resource stock (in the ground) = price
– marginal extraction and production cost
o Hotelling model implies the r percent rule: shadow
price should rise at a rate equal to the interest rate
o Hotelling also predicted that resource consumption
would decline monotonically over time.
o The competitive market outcome was Pareto efficient:
Don’t worry everything will work out fine!
7
Extensions of the Hotelling Model:
Getting the theory to match the fact!
• See Gaudet (2007) and Slade and Thille (2009) for recent
discussions
• Declining resource quality (Ore grade, accessibility)
• Exploration for additional reserves
• Recycling – in effect, adds to reserves
• Technological advances that impact demand or supply of
nonrenewables
• Theoretical models developed by Pindyck (1978), Heal (1981),
and Slade (1982) predict a U-shaped time pattern for prices
with technological advance initially dominating, but
ultimately being overpowered by depletion.
8
Empirical Evidence on Long-Term
Price Trends
o The ‘game’ is to get the longest data span possible and apply the most
robust univariate time series techniques. For some nonrenewables,
data go back to the mid 1800s
o Much of the literature focuses on estimating either TS or DS
specifications in order to estimate the constant long-term trend (albeit
it with the possible search for occasional breaks).
o TS Model
ln Pt = a + bt + e t
o DS Model
D(ln Pt ) = b + ut
9
U-Shaped Price Paths
• Margaret Slade (1982 JEEM) fit (deterministic) linear and
quadratic trend models for eleven nonrenewables from
1870 through 1978 [Aluminum, Copper, Iron, Lead,
Nickel, Silver, Tin, Zinc, Coal, Natural Gas, Petroleum].
• Quadratic trend model is flexible enough to allow for up
to one change in direction of the time trend line,
including the U-shape behavior
• Concerns:
o Linear and (presumably) quadratic trend model are subject
to spurious regression issues in the presence of unit roots.
10
Overall conclusions from review
of empirical work
• Conclusions on the significance of the time
trend depend critically on presence/absence
of unit roots and/or breaks
• Any trend is small and difficult to estimate
precisely, given the huge year-to-year
volatility in the price series.
11
Continued…
•
Tilton (2003, p. 54) summarizes his survey of literature on long-term price
trends this way:
•
“History also strongly suggests that the long-run trends in mineral
prices…are not fixed. Rather they shift from time to time in response to
changes in the pace at which new technology is introduced, in the rate of
world economic growth, and in the other underlying determinants of
mineral supply and demand.
•
“This not only complicates the task of identifying the long-run trends that
have prevailed in the past, but cautions against using those trends to
predict the future. Because the trends have changed in the past, they
presumably can do so as well in the future.”
•
Empirics should allow for variable trends – that is, the gradual evolution in
LT trends without constraining the trends to be constant (or u-shaped) over
time.
Band-pass filters provide one way of doing this if our objective is data
description and historical analysis, rather than hypothesis testing.
•
•
12
Our departure point:
Variable Long-run Trends
• Nonrenewable prices in the long run
will reflect the tug-of-war between
exploration, depletion and
technological change.
• There is no reason to expect that
balance among these forces should
remain constant over the longest
available data span.
13
Band-Pass Filters
•
“When confronting data, empirical economists must somehow isolate features of
interest and eliminate elements that are a nuisance from the point of view of the
theoretical models they are studying. Data filters are sometimes used to do that.”
(Cogley, 2008, p. 68)
•
Explaining how data filters work, Cogley (2008, p.70) notes: “The starting point is the
Cramer representation theorem,… which provides a basis for decomposing xt and its
variance by frequency. It is perfectly sensible to speak of long- and short-run
variation by identifying the long run with low-frequency components and the short
run with high-frequency oscillations.”
•
“Many economists are more comfortable working in the time domain, and for
purposes it is helpful to express the cyclical component as a two-sided moving
average [with infinitely many leads and lags].” (Cogley, 2008, p.71)
•
Although the ‘ideal’ filters have infinitely many leads and lags, actual filters
necessarily involve lead/lag truncation. There are different methods for doing this
(e.g., Baxter-King, Christiano-Fitzgerald)
Actual filters may be symmetric (centered) or asymmetric (uncentered).
o Symmetric – no phase shift
o Asymmetric - allow the filtered series to be calculated all the way to the ends of the data
set
•
14
Applications
• Band-pass (BP) filters allows us to:
o Extract cyclical components within a specified range of periods
(or frequencies) from an economic time series.
o Decompose any time series into a set of mutually exclusive and
completely exhaustive cyclical components that sum to the series
itself.
• Note: The highest-frequency (or shortest period) cycle that can be
identified equals 2 times the data frequency
• Initial application: Baxter and King define ‘business cycle
fluctuations’ as lying in a ‘period window’ between 6 and 32
months.
• Comin-Gertler (2006) Medium-Term Macroeconomic Cycles
• Cuddington and coauthors: super cycles in mineral prices
15
Our Definition of the
‘Long Run’
Pt º Pt (2, 70) + Pt (70,¥)
Pt (2, 70) = 'aggregate'cyclical component
Pt (70,¥) = long - term trend component
16
Economist Commodity Price Index
US dollar terms, in logs
Preliminary Look at
The Economist
Industrials
Commodity Index
5.0
4.5
4.0
3.5
3.0
1875
1900
1925
1950
1975
2000
Economist Commodity Price Index
US dollar terms, log-difference
.6
•
.4
Apparent downward trend after
early 1920s
.2
.0
•
-.2
from -40% to +40%
-.4
1875
1900
1925
1950
1975
2000
Economist Commodity Price Index
US dollar terms, second log-difference
Annual percentage changes range
•
Increase in volatility after early
1920s
.8
•
.4
.0
Average annual growth rate is not
statistically different from zero
-.4
-.8
1875
1900
1925
1950
1975
2000
17
30-Year Moving Average:
Centered vs. Trailing
5.00
4.75
4.50
4.25
4.00
3.75
3.50
3.25
3.00
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
Economist Commodity Price Index (US dollar terms, in logs)
Centered 30-Year Moving Average
Trailing 30-Year Moving Average -- Note severe phase shift!
18
Economist Industrial Commodity Index (EICI):
Annual Growth Rates
.5
.4
.3
.2
.1
.0
-.1
-.2
-.3
-.4
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
Economist Commodity Price Index (US dollar terms, log-difference)
Centered 30-Year Moving Average of Annual Growth Rates
19
EICI:
Centered 30-year Moving Average Growth Rate
.02
.01
.00
-.01
-.02
-.03
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
20
ACF-Band-Pass
Filter Results
on Long-run
Trend
5.00
4.75
4.50
4.25
4.00
3.75
3.50
3.25
3.00
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
Economist Commodity Price Index (US dollar terms, in logs)
Trend Component = ACF-BP(>70)
•
Long-run Trend in EICI is
negative until mid-1980s,
RP_NC_DUM
then turns upward
2
•
One one change in direction
•
Not the classic U-shape that
Pindyck-Heal-Slade would
1
predict
•
Remember: EICI contains
both renewable and
0
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
nonrenewable resources
21
10.0
Long-run
Trends in
LME6:
9.2
9.5
8.8
9.0
8.4
8.5
8.0
8.0
7.6
7.5
7.0
1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
2010
7.2
1900
1910
1920
1930
1940
Aluminum (Natural Logs)
AL_L_2_70_NC
1950
1960
1970
1980
1990
2000
Aluminum, Copper
Nickel, Lead
Tin, Zinc
2010
Copper (Natural Logs)
CU_L_2_70_NC
10.4
8.0
7.8
10.0
7.6
•
Wide variety of price
paths
•
Some have more than
one change in
direction
•
Can we tell metal
specific stories about
the roles of
exploration/discovery,
depletion, and
technological change?
7.4
9.6
7.2
9.2
7.0
6.8
8.8
6.6
8.4
1900
1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
2010
6.4
1900
1910
1920
Nickel (Natural Logs)
NI_L_2_70_NC
8.8
10.4
8.4
10.0
8.0
9.6
7.6
9.2
7.2
8.8
6.8
1910
1920
1930
1940
1950
Tin (Natural Logs)
1960
1970
1940
1950
1960
Lead (Natural Logs)
10.8
8.4
1900
1930
1980
SN_L_2_70_NC
1990
2000
2010
6.4
1900
1910
1920
1930
1940
1950
Zinc (Natural Logs)
1970
1980
1990
2000
2010
PB_L_2_70_NC
1960
1970
1980
1990
2000
2010
ZN_L_2_70_NC
22
Long-Run Variable Trend Rates
for LME6
.02
.01
.00
-.01
-.02
-.03
-.04
00
10
20
30
40
AL_NC_D
PB_NC_D
50
60
CU_NC_D
SN_NC_D
70
80
90
00
10
NI_NC_D
ZN_NC_D
23
.12
Variable Trend
RATES in the
USGS 101
Minerals
.08
.04
.00
-.04
-.08
-.12
00
10
ABM_NC_D
AL_NC_D
ASB_NC_D
BARITE_NC_D
BI_NC_D
CLAY_NC_D
CU_NC_D
FELDS_NC_D
FESLA_NC_D
GA_NC_D
GRAPH_NC_D
HG_NC_D
KYAN_NC_D
MGMTL_NC_D
MO_NC_D
NAS_NC_D
PEAT_NC_D
PRL_NC_D
RE_NC_D
SCAB_NC_D
SN_NC_D
STEEL_NC_D
TALC_NC_D
TISCP_NC_D
VRM_NC_D
ZR_NC_D
20
30
40
50
ABN_NC_D
ALOX_NC_D
AU_NC_D
BAUXI_NC_D
BR_NC_D
CO_NC_D
DIAM_NC_D
FEORE_NC_D
FESTE_NC_D
GAR_NC_D
GYP_NC_D
I_NC_D
LI_NC_D
MICAS_NC_D
MSCLAY_NC_D
NB_NC_D
PGM_NC_D
PUM_NC_D
S_NC_D
SDASH_NC_D
SNDGRC_NC_D
STNC_NC_D
TE_NC_D
TL_NC_D
W_NC_D
60
70
80
ABNSS_NC_D
ALUM_NC_D
B_NC_D
BE_NC_D
CD_NC_D
CR_NC_D
DIATO_NC_D
FEPIG_NC_D
FLUOR_NC_D
GE_NC_D
HE_NC_D
IN_NC_D
LIME_NC_D
MICASP_NC_D
MTLAB_NC_D
NI_NC_D
PHS_NC_D
QTZ_NC_D
SALT_NC_D
SE_NC_D
SNDGRI_NC_D
STND_NC_D
TH_NC_D
TRIP_NC_D
WLA_NC_D
90
00
10
AG_NC_D
AS_NC_D
BALL_NC_D
BENT_NC_D
CEM_NC_D
CS_NC_D
FCLAY_NC_D
FESCR_NC_D
FULE_NC_D
GEM_NC_D
HF_NC_D
KAO_NC_D
MGCOM_NC_D
MN_NC_D
N_NC_D
PB_NC_D
POT_NC_D
RAREARTH_NC_D
SB_NC_D
SI_NC_D
SR_NC_D
TA_NC_D
TI_NC_D
V_NC_D
ZN_NC_D
Hmmm?
What am I supposed to learn
from this?
24
7. 6
7. 2
6. 2
8. 5
6. 0
8. 0
5. 8
7. 5
5. 6
7. 0
5. 4
6. 5
1 4. 5
1 3. 0
5. 2
6. 0
1 2. 5
5. 0
5. 5
4. 8
5. 0
4. 6
4. 5
4. 4
4. 0
1 0. 0
1 4. 0
8. 0
6. 2
7. 5
9. 5
1 3. 5
6. 8
6. 0
7. 0
9. 0
8. 5
6. 4
5. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Abr asiv es ( m anuf act ur ed) ( ni ogs
l )
ABM _ L _ 2 _ 7 0 _ NC
5. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Abr asiv es ( nat ur al) ( ni ol gs )
ABN_ L _ 2 _ 7 0 _ NC
5. 2
Abr asiv e Specal
i Si
cil a ( ni ol gs)
ABNSS_L_2_70_NC
4. 8
4. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Alum num
i
( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Alum na
i ( ni ol gs)
11
3. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
2. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Clay - Ballcay
l ( ni ol gs)
BALL_L_2_70_NC
8. 0
7. 5
Bar ti e ( ni ol gs)
BARI TE_L_2_70_NC
2. 0
8. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Bauxit e ( ni ol gs)
BAUXI _L_2_70_NC
18
9. 2
17
8. 8
16
8. 4
15
8. 0
Ber yum
ill ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
5. 8
Br om ne
i ( ni ol gs)
13
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Chr om um
i
(ni ogs)
l
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
CR_L_2_70_NC
Cesum
i
( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
CS_L_2_70_NC
Copper ( ni ol gs)
6. 4
6. 0
5. 4
6. 2
5. 8
5. 2
5. 6
6. 0
12
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
4. 4
9. 5
3. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Cem ent ( ni ogs)
l
9. 0
8. 5
8. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
CEM _L_2_70_NC
7. 6
Clays ( ni ol gs)
6. 0
CO _L_2_70_NC
2. 8
5. 6
2. 6
5. 2
6. 4
2. 4
4. 8
6. 0
3. 0
3. 6
2. 8
2. 2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
4. 4
5. 6
5. 2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Felds par ( ni ol gs)
FELDS_L_2_70_NC
2. 0
4. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
I r on or e ( ni ol gs)
FEO RE_L_2_70_NC
1. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
I r on oxide pigm ent s ( ni ol gs)
FEPI G _L_2_70_NC
1 5. 5
23
8. 5
5. 5
1 5. 0
22
8. 0
5. 0
1 4. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
I r on and s t eel s cr ap ( ni ol gs)
FESCR_ L _ 2 _ 7 0 _ NC
1 0. 8
I r on and st eelsag
l ( ni ol gs)
FESLA_L_2_70_NC
1 4. 0
1 0. 6
1 3. 6
21
7. 5
4. 5
20
7. 0
4. 0
1 0. 0
6. 5
3. 5
9. 8
6. 0
3. 0
5. 5
2. 5
1 0. 2
1 3. 2
1 4. 0
1 2. 8
19
1 3. 5
9. 6
18
5. 6
5. 2
1 2. 4
9. 4
1 3. 0
17
12
Cobalt ( ni ol gs)
3. 0
3. 2
Cla y Fir e cay
l ( ni ol gs)
FCLAY_L_2_70_NC
6. 0
13
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
CLAY_L_2_70_NC
6. 8
3. 4
6. 4
4. 0
1 0. 0
3. 2
1 0. 4
14
4. 2
1 0. 5
3. 4
CD_L_2_70_NC
3. 8
3. 7
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Diat om ti e ( ni ol gs)
DI ATO _L_2_70_NC
15
4. 6
Cadm um
i ( ni ol gs)
3. 6
4. 4
4. 8
5. 0
2. 6
16
4. 6
5. 0
5. 2
4. 4
6. 8
4. 8
5. 2
5. 6
3. 6
B_L_2_70_NC
4. 0
3. 9
5. 0
5. 4
5. 8
5. 4
2. 8
7. 2
17
1 1. 0
4. 0
Diam ond ( ndust
i
r ai )l ( ni ol gs)
DI AM _L_2_70_NC
18
3. 8
4. 1
4. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
CU_L_2_70_NC
1 1. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
4. 2
3. 8
7. 2
4. 0
2. 6
4. 4
4. 8
5. 0
Bor on ( ni ol gs)
4. 7
4. 1
4. 4
4. 2
3. 0
14
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1 2. 0
5
4. 5
3. 2
5. 0
7. 6
5. 6
AU_L_2_70_NC
4. 2
2. 8
BR_L_2_70_NC
4. 3
3. 4
6. 0
G old ( ni ol gs)
4. 8
7. 2
5. 2
16
14
ASB_L_2_70_NC
4. 2
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
5. 4
18
5. 5
Asbest os ( ni ol gs)
6
3. 6
20
6. 0
AS_L_2_70_NC
4. 4
BI _L_2_70_NC
3. 8
5. 6
22
6. 5
Ar senic ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
4. 5
6
Bism ut h ( ni ol gs)
6. 8
6. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
4. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
7. 2
1 5. 2
4. 3
Clay- Bent onit e ( ni ol gs )
BENT_L_2_70_NC
BE_L_2_70_NC
24
7. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
8
7
1 1. 5
7. 6
1 5. 6
7
9. 0
8. 4
8. 0
8
2. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
5. 0
9. 5
1 2. 0
8. 8
1 7. 2
9
1 0. 0
3. 2
3. 2
5. 5
9
1 0. 5
3. 0
1 2. 5
3. 2
5. 5
1 1. 0
3. 5
1 3. 0
4. 0
3. 6
6. 0
10
10
1 3. 5
4. 4
3. 6
6. 0
11
4. 0
4. 0
4. 0
6. 5
ALUM _L_2_70_NC
1 1. 5
1 4. 0
4. 8
4. 4
6. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Alum num
i
O xide ( ni ol gs)
ALO X_L_2_70_NC
1 2. 0
4. 5
4. 4
5. 0
AL_L_2_70_NC
5. 0
1 4. 5
5. 2
4. 5
AG _L_2_70_NC
1 5. 0
7. 0
1 7. 6
1 6. 0
5. 2
5. 0
7. 0
Sivl er ( ni ol gs)
7. 0
5. 4
7. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
7. 5
7. 5
1 6. 4
5. 5
1 1. 0
8. 0
8. 0
5. 6
8. 0
1 1. 5
8. 5
8. 5
1 6. 8
6. 0
1 2. 0
6. 0
9. 0
5. 8
6. 5
1 2. 5
5. 0
1 2. 0
9. 2
2. 0
9. 0
1 1. 6
16
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
I r on and s t eel ( ni ol gs)
FESTE_L_2_70_NC
1 1. 5
Clay- Fuler
l s ear t h ( ni ol gs )
FUL E_ L _ 2 _ 7 0 _ NC
Fluor spar ( ni ol gs)
FLUO R_L_2_70_NC
1 1. 5
17
1 1. 0
G A_L_2_70_NC
5. 0
G ar net ( ndust
i
r ali ) ( ni ol gs )
G AR_ L _ 2 _ 7 0 _ NC
6. 6
G er m anium (ni ogs
l )
G E_L_2_70_NC
4. 7
4. 6
9. 5
4. 8
6. 2
1 0. 5
15
1 0. 5
1 0. 0
8. 0
8. 0
9. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M er c ur y ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
HG _L_2_70_NC
8. 0
I odine ( ni ol gs)
1 2. 0
4. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
I _L_2_70_NC
I ndium ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
3. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Ky anit e ( ni ol gs)
4. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
KYAN_L_2_70_NC
6. 8
7. 2
1 1. 2
12
6. 4
12
11
6. 0
11
10
5. 6
10
9
5. 2
9
8
4. 8
8
7
4. 4
7
6
Lit hium ( ni ol gs)
LI _L_2_70_NC
5. 2
6. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Lim e ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
LI M E_L_2_70_NC
1 0. 8
1 0. 8
3. 2
6. 4
6. 0
1 0. 4
4. 8
1 0. 4
5. 6
5. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
8. 0
5. 2
9. 6
1 7. 2
7. 6
6. 8
6. 6
6. 4
2. 4
6. 0
4. 4
5. 6
4. 0
5. 2
3. 6
1 0. 0
9. 6
9. 6
9. 2
1 6. 8
4. 4
7. 4
1 6. 4
7. 2
4. 0
5. 5
2. 0
8. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M anganese ( ni ogs)
l
1. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M N_L_2_70_NC
5. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M olybdenum ( ni ogs)
l
M O _L_2_70_NC
4. 8
9
4. 4
8
4. 0
7
3. 6
6
Clay- M si c elaneous
l
clay and shale ( ni ol gs)
M SCL AY_ L _ 2 _ 7 0 _ NC
4. 2
4. 8
3. 2
4. 4
2. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M et alcil Abr asvi es ( ni ol gs)
M TLAB_L_2_70_NC
4. 5
Nit r ogen ( ni ol gs)
9. 2
13
4. 0
8. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1 6. 5
11
1 6. 0
8
10
7
4
8
3. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Phos phat e r ock ( ni ol gs)
PHS_ L _ 2 _ 7 0 _ NC
6. 0
Pot as h ( ni ol gs)
2. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
PO T_L_2_70_NC
Pertil e ( ni ol gs)
1 2. 0
8. 0
1 1. 5
7. 8
1 1. 0
7. 6
1 0. 5
7. 4
1 0. 0
7. 2
Pum ci e and pum ci ti e ( ni ol gs)
PUM _ L _ 2 _ 7 0 _ NC
1 0. 8
2. 8
3. 4
2. 6
1 0. 4
4. 0
5. 6
Rhenium ( ni ol gs)
2. 4
8. 5
6. 6
8. 4
1. 4
2. 2
Sicil on ( ni ol gs)
1 2. 0
1 2. 4
1 1. 6
1 2. 0
1 1. 2
1 1. 6
1 0. 8
1 1. 2
1 0. 4
1 0. 8
SI _L_2_70_NC
Tin ( ni ol gs)
1 1. 5
S_L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Salt ( ni ol gs)
5. 3
2. 8
5. 2
2. 6
6. 2
6. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
SALT_L_2_70_NC
Ant m
i ony ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
SB_L_2_70_NC
6. 0
Sicil on Car bide ( ni ol gs)
SCAB_L_2_70_NC
1 3. 6
1 3. 2
5. 6
2. 4
1 2. 8
5. 2
2. 2
1 2. 4
4. 9
2. 0
1. 8
1 1. 6
4
4. 4
3
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Sand and gr avel ( const r uc t on)
i ( ni ol gs )
SNDG RC_ L _ 2 _ 7 0 _ NC
1 0. 0
16
9. 6
15
9. 2
14
8. 8
13
8. 4
12
4. 6
1. 6
4. 5
1. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Sand and gr avel ( ndus
i
t r al
i ) ( ni ol gs)
SNDG RI _ L _ 2 _ 7 0 _ NC
1 2. 0
4. 8
4. 8
4. 7
SN_L_2_70_NC
1 2. 0
5. 4
Sulf ur ( ni ol gs)
6. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
5. 0
1. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
3. 1
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
5. 1
1. 6
6. 4
7. 2
7
5
8. 8
SE_L_2_70_NC
6. 8
RE_L_2_70_NC
6
6. 8
Selenium ( ni ol gs)
8. 0
6. 6
2. 6
7. 0
Soda ash ( s odium car bonat e) (ni ogs
l )
SDASH_ L _ 2 _ 7 0 _ NC
8. 4
3. 5
3. 4
0
2. 8
2. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
7. 2
7. 0
3. 6
3. 2
3. 0
2. 2
9. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
7. 4
3. 3
2. 4
9. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
9. 2
3. 7
1
8
3. 2
9. 2
4. 4
7. 6
7. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Rar e ear t hs ( ni ol gs)
RAREARTH_L_2_70_NC
Plat num
i
- gr oup m et als ( ni ogs)
l
PG M _ L _ 2 _ 7 0 _ NC
9. 6
3
1 2. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
PEAT_L_2_70_NC
8. 8
4
1 3. 0
Q uar t z c r yst al ( ndus
i
t r ai)l ( ni ol gs)
Q TZ_ L _ 2 _ 7 0 _ NC
9. 6
4. 8
Peat ( ni ol gs)
3. 8
1 3. 5
3
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1 0. 0
5. 2
4. 0
2
4
7
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
PRL_L_2_70_NC
5. 6
1 4. 8
PB_L_2_70_NC
3. 9
1 5. 0
5
2. 5
3. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Lead ( ni ol gs)
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1 4. 0
3. 6
2. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
NI _L_2_70_NC
1 5. 2
2. 8
5
6
9
5
Nic kel ( ni ol gs)
6
1 4. 5
3. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1 5. 5
9
11
3. 8
3. 7
3. 2
6. 4
Niobium ( Colum bium ) ( ni ol gs)
NB_ L _ 2 _ 7 0 _ NC
12
12
3. 5
8. 4
Sodium s ulf at e ( ni ol gs)
NAS_L_2_70_NC
10
4. 0
3. 9
1 5. 6
8. 8
6. 6
N_L_2_70_NC
14
4. 1
3. 2
1 6. 0
3. 6
6. 8
6. 0
8. 8
5. 0
7. 0
6. 2
9. 2
M ci a ( scr ap and f ak
l e) ( ni ol gs)
M I CASP_ L _ 2 _ 7 0 _ NC
1 7. 6
4. 8
1 0. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M ci a ( sheet ) ( ni ol gs)
M I CAS_L_2_70_NC
1 0. 0
6. 0
HF_L_2_70_NC
6. 8
6. 2
M agnesium m et al ( ni ol gs)
M G M TL_L_2_70_NC
7. 8
1 0. 4
2. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
M agnesium com pounds (ni ogs
l )
M G CO M _L_2_70_NC
7. 0
6. 5
Haf nium ( ni ol gs)
7. 0
5. 8
4. 1
7. 0
Clay - Kaolni ( ni ol gs)
KAO _L_2_70_NC
7. 4
1 1. 6
7. 0
5. 0
I N_L_2_70_NC
4. 0
7. 5
7. 5
5. 2
11
13
4. 2
5. 4
8. 5
6. 8
HE_L_2_70_NC
6. 6
8. 5
5. 6
4. 2
12
Helui m ( ni ol gs)
4. 3
4. 4
13
9. 0
9. 5
G YP_L_2_70_NC
4. 4
5. 8
1 0. 0
G ypsum (ni ogs)
l
4. 5
9. 0
6. 0
4. 6
14
9. 5
G r aphit e ( nat ur al) ( ni ol gs)
G RAPH_L_2_70_NC
G em st ones ( ni ogs
l )
G EM _L_2_70_NC
1 0. 0
6. 4
16
1 1. 0
G al
um
il ( ni ol gs)
St r ont um
i
( ni ogs)
l
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
SR_L_2_70_NC
St eel ( ni ol gs)
5. 6
1 1. 2
5. 4
1 0. 8
5. 2
1 0. 4
5. 0
1 0. 0
4. 8
9. 6
4. 6
9. 2
11
4. 4
8. 8
10
4. 2
1 1. 2
4. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
STEEL_L_2_70_NC
1 0. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
St one ( cr ushed) ( ni ol gs)
STNC_L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
St one ( dim ension) ( ni ol gs)
STND_ L _ 2 _ 7 0 _ NC
6. 4
1 1. 5
6. 0
1 1. 0
5. 6
1 0. 5
5. 2
1 0. 0
4. 8
9. 5
4. 4
9. 0
Tant alum ( ni ol gs)
TA_L_2_70_NC
5. 5
5. 4
5. 2
1 1. 0
5. 0
1 0. 5
4. 8
4. 6
1 0. 0
4. 4
9. 5
5. 3
5. 2
5. 1
5. 0
4. 2
1 0. 0
1 0. 4
9. 0
9. 6
1 0. 0
8. 5
4. 0
3. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Talc and py r ophyltil e ( ni ol gs)
TAL C_ L _ 2 _ 7 0 _ NC
8. 8
Telul r um
i
( ni ol gs)
TE_L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Thor um
i
( ni ol gs)
TH_L_2_70_NC
8. 0
7. 6
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Tit anium m et al( ni ol gs)
TI _ L _ 2 _ 7 0 _ NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Tit anium s cr ap ( ni ol gs)
TI SCP_L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Thalum
il
( ni ol gs)
TL_L_2_70_NC
8. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Tr pol
i i ( Nat ur al Abr asiv e) ( ni ol gs)
TRI P_ L _ 2 _ 7 0 _ NC
4. 0
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Vanadium ( ni ol gs)
V_L_2_70_NC
4. 9
8. 5
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Ver m ci ulti e ( ni ol gs)
VRM _L_2_70_NC
4. 8
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Tungst en ( ni ol gs)
W _L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
W olal st onit e ( ni ol gs)
W LA_L_2_70_NC
7. 0
6. 8
8. 4
6. 6
8. 0
6. 4
7. 6
6. 2
6. 0
7. 2
5. 8
6. 8
5. 6
6. 4
5. 4
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Zinc ( ni ol gs)
ZN_L_2_70_NC
1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010
Zir c onium m ni er al concent r at es ( ni ol gs )
ZR_ L _ 2 _ 7 0 _ NC
25
Conclusions
• The extreme volatility of mineral prices (even w annual
frequency data) makes it very difficult to say anything
definitive about long-term trends
• Our band-pass filter analysis suggests that long-term
trends vary widely over time, often changing direction
more than once rather than following the U-shaped
pattern suggest by (some) theory
• Studying aggregate commodity indexes is a dubious
activity, given variety of underlying price behaviors
26
Extensions:
Bass-pass Filter
Analysis of Super
Cycles
(20-70 Years)
Economist Commodity Price Index:
Super-Cycle Component
.3
Difficult to
interpret
.2
.1
.0
-.1
-.2
-.3
-.4
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
•
Cuddington-Jerrett (2008)
on LME6
Indicator of SC Expansion
•
2
on Steel, Pig iron, and
1961-1977
1879-1918??
Molybdenum
2000-ongoing?
1934-47
Jerrett-Cuddington (2008)
1
•
Zellou-Cuddington (2012)
on crude oil and coal
0
70
80
90
00
10
20
30
40
50
60
70
80
90
00
10
27
Appendix: USGS Data
•
The USGS website has annual data for 101 non-energy minerals from 1900 (in many
cases) through 2010. Both nominal unit values and real unit values, using the U.S.
CPI as the deflator, are available. This allows for a rather exhaustive coverage of the
mineral commodities.
•
Source: http://minerals.usgs.gov/ds/2005/140/#data
•
“The U.S. Geological Survey (USGS) provides information to the public and to
policy-makers concerning the current use and flow of minerals and materials in the
United States economy. The USGS collects, analyzes, and disseminates minerals
information on most nonfuel mineral commodities.
•
“This USGS digital database is an online compilation of historical U.S. statistics on
mineral and material commodities. The database contains information on
approximately 90 mineral commodities, including production, imports, exports, and
stocks; reported and apparent consumption; and unit value (the real and nominal
price in U.S. dollars of a metric ton of apparent consumption). For many of the
commodities, data are reported as far back as 1900. Each commodity file includes a
document that describes the units of measure, defines terms, and lists USGS contacts
for additional information. [Accessed August 2, 2012]
•
Insert List and years covered for each (to do) ***
28
References (in progress)
Benati, L. 2001. “Band-Pass Filtering, Cointegration, and Business Cycle Analysis,” Working Paper No 142. Bank of England.
Cristiano, L. and T. Fitzgerald. 2003. “The Band Pass Filter,” International Economic Review 44, 435-65.
Cogley, Timothy. 2008. “Data Filters,” in Steven N. Durlauf and Lawrence E. Blume (eds.) The New Palgrave Dictionary of
Economics, 2nd Edition in Eight Volumes, Palgrave MacMillan.
Cogley, T. and J. Nason. 1995. “Effects of the Hodrick-Prescott Filter on Trend and Difference Stationary Time Series: Implications for
Business Cycle Research,” Journal of Economic Dynamics and Control 19, 253-78.
Comin, Diego, and Mark Gertler. “Medium-Term Business Cycles.” American Economic Review 96, no. 3 (June 2006): 523–551.
Cuddington, John T., Rodney Ludema and Shamila Jayasuriya. 2007. “Prebisch-Singer Redux,” in Daniel Lederman and William F.
Maloney (eds.), Natural Resources and Development: Are They a Curse? Are They Destiny? World Bank/Stanford University Press.
Cuddington, John T and Daniel Jerrett. 2008. “Super Cycles in Metals Prices?” IMF Staff Papers 55, 4 (December), 541-565.
Gaudet, G. 2007. “Natural Resource Economics Under the Rule of Hotelling,” Canadian Journal of Economics 40: 1033–59.
Heap, Alan. 1995. CitiGroup
Hotelling, Harold. “The Economics of Exhaustible Resources.” Journal of Political Economy 39, no. 2 (April 1, 1931): 137–175.
Murray, C. 2003. “Cyclical Properties of Baxter-King Filtered Time Series,” Review of Economics and Statistics 85, 472-76.
Osborn, D. 1995. “Moving Average Detrending and the Analysis of Business Cycles,” Oxford Bulletin of Economics and Statistics 57,
547-58.
Slade, Margaret. 1982. “Trends in Natural-Resource Commodity Prices: An Analysis of the Time Domain,” Journal of Environmental
Economics and Management 9, 122-137.
Slade, Margaret and Henry Thille. 2009. “Whither Hotelling: Tests of the Theory of Exhaustible Resources,” Annual Review of Resource
Economics 1, pp. 239-260.
Tilton, John E. On Borrowed Time? Assessing the Threat of Mineral Depletion. Washington, D.C.: Resources for the Future, 2003.
Zellou, Abdel and John T Cuddington. 2012. “Is There Evidence of Super Cycles in Crude Oil Prices?” SPE Economics and
Management (forthcoming).
29
Thank You!
Comments welcome
My e-mail: [email protected]
Many thanks to the Getulio Vargas Foundation and
VALE for sponsoring and hosting this conference
30

similar documents