Commodity market outlook - Crude oil focus - Summer 2020 August 2020 - Numera Analytics
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Commodity market outlook
Crude oil focus – Summer 2020
August 2020
Numera Analytics |Investment recommendations
• All commodities – Overweight. Raw material prices should recover
further on improved global mobility, stronger industrial activity in China
and a weaker US dollar. We maintain absolute and relative overweight
positions for one-year holdings.
• Crude oil – Overweight. OPEC production cuts have helped restore the
market balance. A recovery in road traffic and stronger industrial activity
should lift consumption to ~ 98M / bbl by 2021. We now expect Brent to
exceed $60 / bbl by mid-2021.
Numera Analytics | 21. All commodities
Numera Analytics |Broad commodity outlook
• Demand for industrial commodities has improved markedly in June
and July, fueled by stronger cyclical conditions in China. This also
helps explain differences in performance across commodities.
• A pick-up in global mobility and stronger industrial activity in China
should support commodity prices over the next 12 months.
• The dollar is weakening on stronger risk sentiment and monetary
stimulus. USD depreciation is now the most likely outcome.
• Both factors should lift overall commodity prices. There is a 76%
chance that the S&P GSCI benchmark will strengthen 12M out.
Numera Analytics | 4Commodities demand is on the rise
Commodities activity index
Deviations from trend
4
Inflationary (↑)
2
Standard deviations
0
-2
Deflationary (↓)
-4
-6
90 93 96 99 02 05 08 11 14 17 20
Note: Index isolates common fluctuations across a panel of 18 industrial commodities, thereby abstracting
from market-specific (idiosyncratic) supply shocks. Index adjusts for USD moves.
Source: Numera Analytics Numera Analytics | 5Fueled by stronger activity in China
China real activity index
Deviations from trend
4
0
Standard deviations
-4
-8
03 05 07 09 11 13 15 17 19
Note: Index isolates common fluctuations across a panel of 20 CN economic indicators, thus
abstracting from industry-specific (idiosyncratic) supply shocks.
Source: Numera Analytics Numera Analytics | 6Which explains differences in performance
S&P commodity price index
Deviations from trend
Current Trough
40%
24%
5.8%
0%
-16% -10%
-26%
-40%
-80%
S&P GSCI Energy Agriculture Industrial Precious
metals metals
Note: Energy represents 62% of the S&P GSCI index (of which crude oil is 59%). Agriculture
and metals are roughly 15% each.
Source: S&P; Numera calculations Numera Analytics | 7Given China’s greater influence on metals
China raw materials demand
% of global consumption
2019 2000
54%
Base metals
51%
Iron ore
28%
Gold
28%
Grains
15%
Crude oil
Note: Base (non-ferrous) metals are the sum of aluminum, copper, nickel,
lead and zinc. Grains include corn, wheat, soybean and rice.
Source: WBMS, World Steel, USDA, BP; Numera calculations Numera Analytics | 8The US dollar continues to weaken
USD during deep recessions
Broad dollar appreciation
16%
15%
2008/09
11%
Dollar has depreciated
6% 4% since mid-May on
improved risk sentiment
COVID-19
1% 1%
-4%
0 15 30 45 60 75 90 105
Days following peak
Note: Change in US dollar against currencies of major trading partners. Grey line shows change
versus mid-February. Green line shows evolution following Lehman bankruptcy.
Source: JP Morgan; Numera calculations Numera Analytics | 9And should depreciate further 12M out
USD appreciation probabilities
12-months ahead (%)
March July
80%
63% 61%
60% US dollar appreciation
much less likely than
41% 42%
40%
early in the pandemic
34%
30%
20%
0%
CAD vs. USD EUR vs. USD GBP vs. USD
Note: Likelihood of USD strengthening against the CAD, EUR and GBP 12M out. Probabilities
above 50% indicate USD appreciation is the most likely outcome.
Source: Numera Analytics Numera Analytics | 10Overall commodity prices should recover
S&P commodity price index
2010 = 100
180
135 95% P.I.
90
45
0
00 02 04 06 08 10 12 14 16 18 20 22
Note: Fan chart shows 50 / 80 / 95% prediction intervals.
Source: History: S&P; Forecast: Numera Analytics Numera Analytics | 11High likelihood of positive returns 12M out
S&P commodity price index
12M probability forecast (%)
22%
12M predictive distribution
Very high chance of
positive returns on
Positive 12M holdings
returns
76%
Left-tail risk:
-63%
-100% -70% -30% 0% 40% 70% 110% 140%
Note: Expected returns are the mean of the predictive distribution. Black dot is the 5% conditional value-
at-risk, expected future losses below the 5% quantile. Shaded area denotes likelihood of positive returns.
Source: Numera Analytics Numera Analytics | 122. Crude oil
Numera Analytics |Market review
Numera Analytics |Oil market review: Demand
• Global oil consumption has increased sharply since bottoming out in
April, but remains ~7M bbl / day (6.7%) below pre-COVID levels.
• The pick up in demand is directly related to stronger global mobility, as
transportation accounts for 60% of worldwide oil usage.
• Within transportation, the bulk of the improvement reflects a recovery
in road travel, although traffic levels are far from normalizing. Jet fuel
demand is exceptionally low due to a collapse in air travel.
• Global industrial activity (16% of demand) also bottomed out in April,
improving in May and June as production in China normalized.
Numera Analytics | 15Demand still far from pre-COVID levels
Change in global oil consumption (SA)
Mbbl / day and % (vs. 01/20)
0
-5.5%
-5.9%
-2.3%
-2 -12% -11%
-10%
Million bbl / day
-4
-6 In April, consumption
down 19 Mbbl (-18%)
-6.7%
-8
World US EZ JP Other DM CN Other EM
Note: Chart breaks down change in global oil consumption (SA) between January and July 2020.
Bars measure change in million bbl / day.
Source: Energy Information Agency (EIA); Numera calculations Numera Analytics | 16Evolution mirrors global mobility trends
Mobility and COVID deaths
World ex. Latin America
0% 1
New COVID
-16% 0.75
deaths (R)
-32% 0.5
-48% 0.25
Away-from-
home traffic (L)
-64% 0
07/03 04/04 02/05 30/05 27/06 25/07
Note: Green line shows changes in retail and recreation traffic versus Jan. 2020. Red line shows new
COVID deaths per million. Figures correspond to global aggregate excluding Latin America.
Source: ECDC, Google Mobility; Numera calculations Numera Analytics | 17As 60% of demand tied to transportation
Global oil demand by end-use
% of total, 2019
Buildings
Other
10% 15%
Industry
16%
Transportation
98M bbl/d
Gasoline
CN
/ diesel
51%
60%
8%
Jet fuel
Source: BP Statistical Energy Review; Numera calculations Numera Analytics | 18Road traffic has yet to fully recover
Traffic congestion in metropolis
Year-on-year (%)
Latest (02/08) Trough
0%
-12%
-25% -17%
-24%
-36% -34%
-50%
-52%
-57% -57%
-75%
-100%
S. Paulo CDMX NYC Paris N. Delhi London Shanghai Tokyo
Source: TomTom Traffic Index Numera Analytics | 19While dismal auto sales restrict growth
Vehicle sales by country
6M year-to-date, SA (%)
UK -36%
BR -34%
MX -33%
IN -29%
EZ -25%
DM -25%
US -21%
World -18%
RU -17%
EM -10%
CN -6%
JP -5%
Note: Bars compare seasonally adjusted change in light vehicle (car + truck) sales
by country between 12/19 and 06/20.
Source: Government agencies, Numera Analytics Numera Analytics | 20Jet fuel is the hardest hit end-use
COVID impact on global air traffic
Change in seating capacity (%)
Domestic International
0%
-25%
-50%
-75%
-90% -91% -91% -89%
-100%
World Asia Europe N. America
Note: Bars compares seating capacity in international and domestic flights in June 2020
versus ‘business as usual’ (the ICAO baseline).
Source: International Civil Aviation Organization (ICAO) Numera Analytics | 21Industrial activity is slowly recovering
COVID impact on industrial activity
Global output losses (%)
0%
Baseline
-5% Great
Recession
-10%
-15% At the peak, COVID-19
losses of 15%
-20%
0 2 4 6 8 10 12 14
Months after initial shock
Note: Chart shows the realized performance of global industrial production during
deep recessions. Output losses measured as deviations from a no-shock baseline.
Source: Numera Analytics Numera Analytics | 22Oil market review: Supply
• Oil production has fallen sharply since the OPEC+ April agreements.
OPEC output is now roughly 9M / bbl below capacity.
• In combination with stronger consumption, pledged and voluntary
production cuts have now restored the market balance.
• Despite tighter market conditions, above-ground inventories remain
exceptionally high, continuing to weigh on oil prices.
• Shale production in North America is down 2M bbl / day since March,
reflecting a sharp decline in active drilling rigs. With WTI in the low
$40s, shale producers are still at a high risk of bankruptcy.
Numera Analytics | 23Production down sharply since April
Change in global oil production
April – July (million bbl/day)
0
OPEC production
down 8.4M bbl /
day (-24%)
-5
Oil production 12%
lower than in April
-8.4
-10
+0.4
-2.4
-0.8 -0.5 -11.9
-15
OPEC RU US CA RoW World
Source: EIA; Numera calculations Numera Analytics | 24OPEC operating well below capacity
OPEC crude oil supply
Million bbl / day
36 Great Recession Oil crisis S. Aramco attack
Capacity
32
28 Surplus capacity at an all-
Production time high of 9M bbl / day
due to OPEC+ agreement
24
20
06 08 10 12 14 16 18 20
Source: EIA Numera Analytics | 25Supply cuts have helped restore balance
Global crude oil balance
Millions of bbl / day
9 Excess demand (↑)
0
Balance restored
on rising demand,
-9
Excess supply (↓) falling supply
-18 In April, excess oil
supply of 21M bbl
-27
97 99 01 03 05 07 09 11 13 15 17 19
The crude oil balance is computed as the difference between global consumption and production.
Positive figures indicate excess demand.
Source: Energy Information Administration (EIA) Numera Analytics | 26Inventories remain exceptionally high
OECD crude oil inventory
Millions of barrels, SA
2000
Other OECD
1600
1200 United States
800
97 99 01 03 05 07 09 11 13 15 17 19
Note: Commercial inventories only (excludes US Strategic Petroleum Reserves).
Source: EIA Numera Analytics | 27Although days of supply now diminishing
US crude oil inventories
Weekly days of supply
45
40
2020
35
Much higher stock levels
than during 2015 oil crisis
30
2015
25
2010-19
20
J F M A M J J A S O N D
Note: Commercial inventories only (excludes US Strategic Petroleum Reserves).
Source: EIA; Numera calculations Numera Analytics | 28NA supply impacted by low oil prices
NA oil rigs and WTI spot
Rig count (SA) and USD / bbl
2000 120
1500 90
WTI spot (R)
1000 60
500 Active oil 30
rigs (L)
0 0
11 12 13 14 15 16 17 18 19 20
Source: Baker Hughes, World Bank; Numera calculations Numera Analytics | 29Shale production has been hard-hit
US oil production by region
March - July (millions bbl/day)
March-May May-July
0.8
5%
0.0
-18% -19%
-24%
-0.8 -14%
-1.6 WTI fell below break-even
-11%
on existing wells ($25-30),
prompting shut-downs
-2.4
US Bakken Eagle Ford Permian Other shale Non-shale
Source: EIA; Numera calculations Numera Analytics | 30Bankruptcy risk is still very high
US high yield bond spreads
Basis points (bps)
2000
1500
High yield
Energy Very high premium, which
1000 high yield given ample Fed liquidity
implies high default risk
500
0
01 03 05 07 09 11 13 15 17 19
Note: Spreads computed against the Moody’s Aaa investment-grade benchmark.
Source: ICE BofAML, Moody’s; Numera calculations Numera Analytics | 31Oil market review: Prices
• Oil prices have risen $25 since late April as improved mobility and
OPEC production cuts have helped restore the market balance.
• Trader sentiment remains bearish (as indicated by e.g. net short
positions), but is less of a drag on prices than in March and April.
• Weak economic activity and ample inventories explain most of the
drop in real oil prices year-to-date (-37% through 6M).
• We estimate Brent is currently trading at a ~ $10 discount versus its
‘fair’ value. The wedge is explained by bearish sentiment and low
precautionary purchases by end-users, given high stock volumes.
Numera Analytics | 32Oil prices up $25 since late April
Oil market events timeline
Brent spot price (USD / bbl)
80 2. Iranian General
Soleimani killed
3. WHO first reports
COVID transmission
60
1. Drone strike
Saudi Aramco
40
4. OPEC talks fail 5. OPEC+
agreement
7. Peak COVID
20 deaths globally
6. Negative
WTI futures
0
28/08 27/09 28/19 26/11 26/12 27/1 25/2 25/3 27/4 5/28 25/6 7/27
Source: FRED, Numera Analytics Numera Analytics | 33Market volatility has diminished
Oil volatility and economic uncertainty
OVX index and World EPU
180 600
135 450
Economic
Oil market Uncertainty (R)
90 volatility (L) 300
45 150
0 0
07 09 11 13 15 17 19
Note: Red line is the CBOEs crude oil ETF volatility index (OVX). Green line shows the GDP-weighted
global economic policy uncertainty index of Baker, Bloom and Davis.
Source: CBOE; Baker, Bloom & Davis Numera Analytics | 34But trader sentiment remains ‘bearish’
Net long positions – Crude oil
Thousands of contracts
30
COVID-19
0
Net short positions suggest
bearish sentiment, but less
-30 so than early in the pandemic
-60
02 04 06 08 10 12 14 16 18 20
Note: Net long positions are the difference between long and short crude oil positions by traders
in the New York Mercantile Exchange (NYMEX).
Source: Commodity Futures Trading Commission (CFTC) Numera Analytics | 35OPEC cuts are helping stabilize prices
Drivers decomposition – Oil prices
Year-to-date (%)
Demand USD Production Stocks Speculation / prec.
20%
Cuts helping
0% stabilize prices
-20% -14%
-24%
Sentiment less of a drag
-40%
than in March and April
-37%
-60% -43%
-61% -60%
-80%
01/20 02/20 03/20 04/20 05/20 06/20
Note: Contribution of market drivers to change in real Brent prices vs. December 2019. ‘Demand’ captures changes
in broad economic activity and oil-specific demand shocks. Speculation is proxied by net long NYMEX oil positions.
Source: Numera Analytics Numera Analytics | 36Low demand remains the biggest drag
Drivers decomposition – Oil prices
6M year-to-date (%)
12% 10%
0%
-2%
-12% -10%
-15%
-24% -20%
Weak demand still
-36% the biggest drag
-37%
-48%
Real oil Demand US dollar Production Stocks Speculation /
prices precautionary
Note: Contribution of market drivers to change in real Brent prices 6M YTD (right-hand bar in previous chart).
Source: Numera Analytics Numera Analytics | 37Oil prices trading at a ~ $10 discount
Brent spot vs. ‘fair’ value
Risk premium – USD / bbl
50
Premium (> 0)
25
USD / bbl
0 Brent trading at
a ~ $10 discount
-25
Discount (< 0)
-50
03 05 07 09 11 13 15 17 19
Note: The oil risk premium is obtained by comparing the Brent spot price to its modelled ‘fair’ (or equilibrium) value. A
positive gap indicates oil is trading at a premium, e.g. due to geopolitical risk. A negative gap signals bearish sentiment.
Source: Numera Analytics Numera Analytics | 38Market outlook
Numera Analytics |Oil market outlook
• We expect global oil consumption to recover further, averaging ~ 98M
bbl / day in 2021. Given depressed air travel, however, it is unlikely
that oil will exceed 100M bbl (its Q4/19 level) before late 2021.
• Financial markets expect oil prices to rise 6% ($3 / bbl) by mid-2021.
In our view, prices have greater upside. There is a 78% chance that
Brent will exceed the July 12M futures contract ($46 / bbl). This
improves its relative appeal versus other risky assets.
• The high probability reflects further improvements in transportation
demand, alongside limited oil investments and a weaker USD outlook.
The fact that oil is still trading at a discount also improves its upside.
Numera Analytics | 40Demand factors dominate at a 12M horizon
12M variance decomposition – Brent
Share of variability, 1986-20 (%)
Demand US dollar
17%
39%
11% Production
4%
Stocks
12%
16%
Speculation Precautionary
Note: Pie chart breaks down the contribution of market drivers to 12M (YoY)
fluctuations in real Brent prices. ‘Sentiment’ encompasses both precautionary
purchases by consumers (e.g. refineries) and speculative moves by traders.
Source: Numera Analytics Numera Analytics | 41Oil demand should continue recovering
Global oil consumption
2020-21 outlook (Mbbl / day)
92
Predictive distribtuion
Demand should recover
further, but will likely fall
short of 100M bbl
2020
98
> 100M bbl
2021
28%
80 84 88 92 96 101 105 109
Million bbl / day
Note: Chart shows probability distributions for average oil consumption worldwide in 2020 and 2021. Base scenario is
the mean of each distribution. Shaded area denotes the likelihood of oil demand exceeding 100M bbl / day in 2021.
Source: Numera Analytics Numera Analytics | 42Aided by stronger economic activity
DM GDP outlook
Trillions of (2010) USD
65
95% P.I.
55
45
35
06 08 10 12 14 16 18 20f 22f
Note: Fan chart corresponds to 70% and 95% prediction intervals. Grey bars are OECD recession dates.
Source: Numera Analytics Numera Analytics | 43Air travel is the weakest link
Global air traffic seating capacity
Deviations from baseline (%)
20%
Projection
0%
ICAO expects air travel to
Domestic be down 25% 12M out, a
-20% loss of ~ 2M bbl / day
-40%
-60% International
-80%
-100%
1/20 4/20 7/20 10/20 1/21
Note: Chart compares global air traffic seating capacity relative to ‘business as usual’.
Source: ICAO Numera Analytics | 44Inventory overhang should diminish
OECD crude oil inventories
Monthly days of supply
100
90
80 Expect an inventory
drawdown as demand
strengthens
70
Interquartile range
60
50
40
00 02 04 06 08 10 12 14 16 18 20
Note: Commercial inventories only (excludes US Strategic Petroleum Reserves).
Source: EIA Short-term Energy Outlook (STEO) Numera Analytics | 45Low CAPEX should keep NA supply tight
Largest declines in US oil investment
Quarter-on-quarter, SAAR (%)
Oil-specific Recession
0%
-25%
-50%
-58% -56% -55%
-61%
-75% -66% -66%
-78%
-84%
-100%
Q4/86 Q2/20 Q4/09 Q1/16 Q1/02 Q2/15 Q1/81 Q3/91
Note: Bars show the eight largest quarterly declines in US oil mining investment since 1970. Green
bars represent oil market-specific events. Blue bars correspond to declines during recessions.
Source: US BEA; Numera calculations Numera Analytics | 46Markets expect Brent to hit $46 12M out
Brent NYMEX futures
USD / bbl
50
July 2020 Markets expect Brent
46 to rise by 6% to $46 /
bbl one-year out
42
40 May 2020
39
USD / bbl
30 March 2020
20
1 3 5 7 9 11
Contract maturity (months)
Source: CME Group Numera Analytics | 47But market view is quite conservative
Crude oil price outlook
Baseline and probabilities
Brent Outlook Baseline Probabilities (%)
Horizon (USD / bbl) > July ($43) > Futures < $30 > $70
3M ahead $49 66% 64% 9% 5%
6M ahead $55 78% 75% 11% 16%
12M ahead $61 81% 78% 9% 35%
Note: Baseline forecast is the mean of the predictive distribution. Probabilities show the likelihood of exceeding specific thresholds over
a given horizon. These include the July month-end price (in red) and the futures price for a specific maturity.
Source: Numera Analytics Numera Analytics | 48High probability of oil price inflation
Likelihood of higher Brent prices
12M ahead probability (%)
90%
81%
60%
50%
30%
0%
05 07 09 11 13 15 17 19
Note: Chart tracks the probability of Brent spot prices rising over a one-year period. Probabilities
above 50% (dashed red line) indicate oil price inflation is the most likely outcome.
Source: Numera Analytics Numera Analytics | 49Improving the upside for oil investments
Crude oil vs. DM stocks
Excess returns probability
60% 60%
12M outperformance probability
50% 50%
40%
30%
Q1-17 Q3-17 Q1-18 Q3-18 Q1-19 Q3-19 Q1-20
Note: Chart plots likelihood of crude oil investments outperforming DM stocks over a 12M
holding period. Probabilities above 50% indicate most likely outcome.
Source: Numera Analytics Numera Analytics | 50Favourable risk-reward trade-off 12M out
Risk-reward comparison
12M investments
21%
Better risk-reward 1 Oil Despite being a very risky
investment, oil offers higher
risk-adjusted expected returns
5
Real expected returns
14%
EU stocks
7% 4
5 US stocks
EM bonds 6 EM stocks
3
2 US bonds Worse risk-reward
0%
0% 3% 6% 9% 12%
Downside risk
Note: Numbers in red rank assets by their Omega ratio, a measure of investment quality
that compares expected returns to downside risk (defined as average expected losses).
Source: Numera Analytics Numera Analytics | 51You can also read
