Energy Economics I

Jeffrey Frankel

Harpel Professor, Harvard

University

ADA Summer School, Baku,

Azerbaijan
1

 (1)

Long-term trends:
Are oil prices fated
to rise as the world runs out?
(2)

Shorter-term
movements:
What causes swings
such as the 2008 price spike?
2

To think about oil prices,

a broad historical perspective is

essential

From the vantage point of 2006,

the last decade suggested
a permanent upward trend.

Now, 2001-2009 looks like

a classic bubble and crash.
3

Price of oil, 1995-2006

Permanent upward trend?

Price of oil, 2001-2009

Or was 2008 only a transitory

spike?

(1) Long-term world oil price

trend

(i) Determination of the price on world markets

(ii) The old structuralist school (Prebisch-Singer):

The hypothesis of a declining commodity price trend

(iii) Hypotheses of a rising price trend

Hotelling, non-renewable resources, & the interest rate

Malthusianism & the peak oil hypothesis

(iv) Empirical evidence

Statistical time series studies

Paul Ehrlich versus Julian Simon

(i) The determination of the

export price on world
markets

Developing countries tend to be smaller

economically than major industrialized
countries, and more likely to specialize in
the exports of basic commodities like oil.

As a result, they are more likely to fit

the small open economy model:

they can be regarded as price-takers,

not just for their import goods,

but for their export goods as well.

That is, the prices of their tradable goods are

generally taken as given on world markets.
7

The determination of the export price on world

markets
continued

The price-taking assumption requires 3 conditions:

low monopoly power,

low trade barriers,
and intrinsic perfect substitutability in the commodity as
between domestic and foreign producers

a condition usually met by primary products

and usually not met by manufactured goods & services).

To be literal,
not every barrel of oil is the same as every other
and not all are traded in competitive markets.
But the assumption that most oil producers
are price-takers holds relatively well.
8

A qualification: Monopoly
power

Saudi Arabia does not satisfy the 1st condition,

due to its large size in world oil markets.

If OPEC functioned effectively as a true cartel,

then it would possess even more monopoly
power in the aggregate.

However OPEC does not currently exercise

much monopoly power beyond that of Saudi
Arabia,

because so many non-members now produce oil

and
because even OPEC members usually do not feel
constrained to stay within assigned quotas.

The determination of the export price on world

markets
continued

To a first approximation, then,

the local price of oil =
($ price on world markets) x (the exchange
rate).

=> a devaluation should push up

the oil price quickly and in proportion

leaving aside pre-existing contracts

or export restrictions.

An upward revaluation of the currency

should push down the price in proportion.
10

(ii) The old structuralist

school
Raul Prebisch (1950) & Hans Singer

(1950)
The hypothesis: a declining long run trend
in prices of mineral & agricultural products

relative to the prices of manufactured goods.

The theoretical reasoning:

world demand for primary products is
inelastic with respect to world income.

That is, for every 1 % increase in income,

raw materials demand rises by less than 1%.
Engels Law, an (older) proposition:
households spend a lower fraction of their income
on basic necessities as they get richer.
Demand
=> P oil

11

Structuralists, continued

This hypothesis, if true, would imply that

specializing in natural resources was a bad deal.
Mere hewers of wood & drawers of water would
remain forever poor if they did not industrialize.
The policy implication of Prebisch:

developing countries should discourage international

trade with tariffs,
to allow their domestic manufacturing sector to
develop behind protective walls,
rather than exploiting their traditional comparative
advantage in natural resources

as the classic theories of free trade would have it.

12

Import Substitution
Industrialization policy (ISI)

was adopted in the 1950s, 60s and

70s

in most of Latin America and much of

the rest of the developing world.

The fashion reverted in subsequent

decades, however.
13

(iii) Hypotheses of rising

trends

Hotelling on depletable resources;

Malthus on geometric population
growth.

Persuasive theoretical arguments

that we should expect oil prices to show
an upward trend in the long run.
14

Assumptions for Hotelling

model

(1) Non-perishable non-renewable resources:

Deposits in the earths crust are fixed in total supply

and are gradually being depleted.

(2) Secure property rights:

Whoever currently has claim to the resource
can be confident that it will retain possession,

unless it sells to someone else,

who then has equally safe property rights.

This assumption excludes cases where warlords

compete over physical possession of the resource.
It also excludes cases where private oil companies fear
that their contracts might be abrogated
or their holdings nationalized.
15

If property rights are not

secure,

the current owner has a strong

incentive to pump the oil quickly,

because it might never benefit if the oil is left in the

ground.

That is one explanation for the sharp rise

in oil prices from 1973 to 1979:

Western oil companies in the 1960s had anticipated

that newly assertive developing countries would
eventually nationalize the reserves within their borders,
and thus had kept prices low by pumping oil more
quickly than if they been confident that their claims
would remain valid indefinitely
until they indeed lost control in 1973.
16

Price of oil, 1900-2006

Oil shocks

1 1973 Arab oil

embargo
2. 1979 fall of Shah of
Iran
3. 2008 spike

17

While we are on the subject

of the 1970s oil shocks

A more common explanation for the oil price

increases of 1973-74 and 1979-80 is
simply geopolitical disruptions:

Yom Kippur War and Arab Oil Embargo

Revolution in Iran and Fall of the Shah

Less common explanations:

Excessively easy monetary policy

coming from US Fed accommodation of Vietnam deficits

The world is running out of oil.

We consider both later.

18

One more assumption,

to keep the Hotelling model
simple:

(3) The fixed deposits are easily accessible:

the costs of exploration, development, &

pumping
are small compared to the value of the oil.

Hotelling (1931) deduced from these

assumptions the theoretical principle:
the price of oil in the long run should rise
at a rate equal to the interest rate.
19

King Abdullah of Saudi Arabia,

with interest rates close to zero,
apparently believes that the rate of
return on oil reserves is higher if he
doesn't pump than if he does:

"Let them remain in the ground for our

children and grandchildren..." (April 12,
2008)

20

The Hotelling logic:

The owner chooses how much oil to pump

Whatever is pumped can be sold at todays

price (price-taker assumption)

and how much to leave in the ground.

and the proceeds invested in bank deposits

or US Treasury bills, which earn the current interest
rate.

If the value of the oil in the ground is not

expected to rise in the future, then the owner
has an incentive to extract more of it today, so
that he earns interest on the proceeds.
21

The Hotelling logic,

continued:

As oil companies worldwide react in this way,

they drive down the price of oil today,

below its perceived long-run level.

When the current price is below its long-run level,

companies will expect the price to rise in the
future.

Only when the expectation of future appreciation

is sufficient to offset the interest rate will the oil
market be in equilibrium.

Only then will oil companies be close to

indifferent between pumping at a faster rate and
22
a slower rate.

Hotelling,

continued:

To say the oil price is expected to

increase at the interest rate means
that it should do so on average;

it does not mean that there wont be price

fluctuations above & below the trend.
The theory does imply that, averaging out
short-term unexpected fluctuations, oil
prices in the long term should rise at the
interest rate.
23

If there are costs of extraction &

storage?

-- non-negligible costs (but assume constant) ?

then the trend in prices will be lower

than the interest rate, by that amount.

If there is a constant convenience yield

from holding inventories?
then the trend in prices will be higher
than the interest rate, by that amount.

The arbitrage equilibrium equation:

E p oil = interest rate + costs convenience
yield
24

The upward trend idea is older than

Hotelling.

It goes back to Thomas Malthus (1798)

and the first fears of environmental scarcity:
Demand grows with population,
Supply does not.
What could be clearer in economics
than the prediction that price will rise?

25

Over the two centuries since Malthus,

or the 70 years since Hotelling,

exploration & new technologies have

increased the supply of oil at a pace
that has roughly counteracted the increase
in demand from growth
in population & incomes.[1]

[1] Krautkraemer (1998) and Wright & Czelusta (2003, 2004, 2006).
26

Hubberts
Peak U.S.

Just because supply has always increased

in the past does not necessarily mean it
always will in the future.

In 1956, M. King Hubbert, an oil engineer,

predicted that the flow supply of oil within
the US would peak in the late 1960s and
27

Hubberts
Peak U.S.

The prediction was based on a model

in which the fraction of the countrys reserves

that has been discovered rises through time,

and data on the rates of discovery versus
consumption are used to estimate the
parameters in the model.

Unlike myriad other pessimistic forecasts,

this one came true on schedule,

earning subsequent fame for its author:

28

Hubberts
Peak Global

Planet Earth is a much larger place

than the USA, but it too is finite.

Some analysts have extrapolated Hubberts words

& modeling approach to claim that the same pattern
will follow for extraction of the worlds oil reserves.
Some claim the 2000-2008 run-up in oil prices
confirmed a predicted global Hubberts Peak. [1]

It remains to be seen whether we are currently

witnessing a peak in world oil production,

notwithstanding that forecasts of such peaks

have proven erroneous in the past.

[1] E.g., Deffeyes (2005).

29

Hubberts Peak
global

HeatUSA.com blog

30

The complication: supply is not

fixed.

True, at any point in time there is a certain

stock of oil reserves that have been discovered.

But the historical pattern has long been that,

as that stock is depleted, new reserves are
found.

When the price goes up, it makes exploration &

development profitable for deposits farther
under the surface or underwater
or in other hard-to-reach locations.
especially as new technologies are
developed for exploration & extraction.

31

The empirical evidence

With strong theoretical arguments on both

sides, either for an upward trend or for a
downward trend, it is an empirical question.

Terms of trade for commodity producers had

a slight up trend from 1870 to World War I,

a down trend in the inter-war period,
up in the 1970s,
down in the 1980s and 1990s,
and up in the first decade of the 21st century.
32

What is the overall statistical trend

in commodity prices in the long
run?

Some authors find a slight upward trend,

some a slight downward trend. [1]
The answer seems to depend, more than anything else,
on the date of the end of the sample:

Studies written after the 1970s boom found an upward trend,

but those written after the 1980s found a downward trend,

even when both went back to the early 20th century.

[1] Cuddington (1992), Cuddington, Ludema & Jayasuriya (2007),

Cuddington & Urzua (1989), Grilli & Yang (1988), Pindyck (1999), Hadass
& Williamson (2003), Reinhart & Wickham (1994), Kellard & Wohar
(2005), Balagtas & Holt (2009) and Harvey, Kellard, Madsen & Wohar
(2010).
33

What is the trend in the price

of oil in particular?
1869-1969:

Downward

1970-2010:

Upward

The

The long run: Unclear

?
34

Price of oil, 1869-2009

35

Addendum 1:
Malthusians vs. Cornucopians
The wager of Paul Ehrlich against Julian Simon

Paul Ehrlich is a biologist, highly respected among

scientists but with a history of sensationalist doomsday
predictions regarding population, the environment, &
resource scarcity.

Julian Simon was a libertarian economist, frustrated by

the failure of the public to hold Malthusians accountable
for the poor track record of their predictions.

In 1980, Simon publicly bet Ehrlich $1000 that the prices

of 5 minerals would decline between then and 1990.
36

Malthusians vs. Cornucopians

The Ehrlich-Simon bet, continued

Ehrlichs logic was Malthusian:

because supplies were fixed while growth of

populations & economies would raise demand,
the resulting scarcity would drive up prices.
He mentally extrapolated into the indefinite future
what had been a strong rise in commodity prices
over the preceding decade.

Simons logic is called cornucopian:

Yes, the future would repeat the past.

The relevant pattern from the past was not the tenyear trend, however, but rather a century of cycles:

resource scarcity does indeed drive up prices,

whereupon supply, demand and, especially, technology,
respond with a lag, driving the prices back down.
37

Malthusians vs. Cornucopians

The Ehrlich-Simon bet, continued

Simon won the bet handily:

Every one of the 5 real mineral prices in

the basket declined over the next 10
years.
He was also right about the reasons:

In response to the high prices of 1980,

new technologies came into use, buyers
economized, and new producers entered the
38
market.

The Ehrlic-vs.-Simon bet

carries fascinating implications,

for Malthusians vs. Cornucopians,

environmentalists vs. economists,
extrapolationists versus contrarians,
and futurologists versus historians.

The main point:

Simple extrapolation of medium-term trends is

foolish.
One must take a longer-term perspective.
Statisticians need as long a time series as possible.
39

However, one should avoid falling prey to either of

two reductionist arguments at the philosophical
poles of Mathusianism & cornucopianism.

On the one hand, the fact that the supply of

minerals in the earths crust is a finite number,
does not in itself justify the apocalyptic conclusion
that we must necessarily run out.

As Sheik Yamani, the former Saudi oil minister,

said, "The Stone Age came to an end not for a lack of
stones and the oil age will end, but not for a lack of
oil.

Malthusians do not pay enough attention to the tendency

for technological progress to ride to the rescue.

40

On the other hand, the fact that the Malthusian

forecast has repeatedly been proven false in
the past does not imply that it will always be so
in the future.
One must seek, rather, a broad perspective,
in which all relevant reasoning & evidence
are brought to bear in the balance.

41

(2) Short-term
oil price volatility

and medium-term swings

Low short-run elasticities

Inventories moderate volatility
Cobweb cycle
Monetary influences
The 2008 spike
Speculators
42

Causes of Volatility

The market price of oil is volatile in the short run.

Because elasticities of supply & demand
with respect to price are low,
relatively small fluctuations in demand
(due, for example, to weather) or in supply
(due, for example, to disruptions) require a large
change in price to re-equilibrate supply and
demand.

Demand elasticities are low in the short run largely

because the capital stock is designed to operate
with a particular ratio of energy to output.

Supply elasticities are also often low in the short

43
run because it takes time to adjust output.

A given rise in demand causes

a small price rise
or
a big price
rise

Poil
D

with
High
elasticities

The increase
in demand drives up
the price

D'

Supply & demand for oil

with

Low elasticities
D

Poil

D' S

{
Supply & demand for oil
44

Volatility,

continued

Inventories can cushion the short run impact

of fluctuations, but they are limited in size.

There is a bit of scope to substitute across

different fuels, even in the short run.

But this just means that the prices of oil,

natural gas, and other fuels tend to experience
their big medium-term swings together.

45

Volatility,

continued

In the longer run, elasticities are far

higher, both on the demand side and the
supply side.

This dynamic was clearly at work in

the oil price shocks of the 1970s
quadrupling after the 1973 Arab oil embargo

doubling after the Iranian revolution of

1979,
which elicited relatively little consumer
conservation or new supply sources in the short
46
run, but a lot of both after a few years had

Volatility,

continued

In the medium run,

people started insulating their houses and
driving more fuel-efficient cars,
and oil deposits were discovered
& developed in new countries.

This is a major reason why the real price

of oil came back down in the 1980s-1990s.
47

Price of oil, 1970-2007

48

Volatility,

continued

In the medium term, oil may be subject to

a cob-web cycle, due to the lags in
response:

The initial market equilibrium is a high price;

1
the high price brings forth investment

and raises supply after some years,

which in turn leads to a new low price,

which discourages investment,

and thus reduces supply with a lag

and so on.

3
4

In theory, if people have rational expectations,

they should look ahead to the next price cycle before
making long-term investments in housing or drilling.
But the complete sequence of boom-bust-boom over the last
35 years looks suspiciously like a cobweb cycle nonetheless.

49

Monetary influences on oil

prices

The same arbitrage equation that implies a positive

long-run price trend also explains some shorter-run
price swings.

The real price of oil should be unusually high during

periods when real interest rates are low (e.g., due to
easy monetary policy),

so that a poor expected future return to leaving

the oil in the ground offsets the low interest rate.

By contrast, when real interest rates are high (e.g.,

due to tight monetary policy), current oil prices
should lie below their long-run equilibrium,

because an expected future rate of price increase

is needed in order to offset the high interest rate.

50

The mechanism?
High interest rates reduce the demand for oil,
or increase the supply, through 3 channels:
by increasing the incentive for extraction
today rather than tomorrow;
by decreasing firms' desire to carry
inventories (oil stocks held in tanks or tankers)
by encouraging speculators to shift out
of spot oil contracts, and into treasury bills.

All 3 mechanisms work to reduce the market price of oil,

as when real interest rates where high in the early
1980s.
A decrease in real interest rates has the opposite effect,
lowering the cost of carrying inventories, and raising oil
prices, as happened from Aug. 2007 to Sept. 2008.
51
Call it an example of the carry trade.

The monetary overshooting theory

summarized

Monetary contraction temporarily raises the real interest

rate

whether via rise in nominal interest rate, fall in expected inflation, or both.

Inventory demand falls. Oil prices fall.

How far?
Until oil is widely considered "undervalued"

-- so undervalued that there is an expectation of future appreciation

together with other advantage of holding inventories: the "convenience yield

that it is sufficient to offset the higher interest rate

and other costs of carrying inventories: storage costs plus any risk premium.

Only then are firms willing to hold the inventories

despite the high carrying cost.

In the long run, the general price level adjusts

to the change in the money supply.

As a result, the real money supply, real interest rate, and real commodity price eventually return to where they
were

Frankel "Expectations and Commodity Price Dynamics: The Overshooting Model," American J. of Ag. Economics

52

Monetary influences on oil prices,

continued

Very low US real interest rates boosted

commodity prices toward the end of the 1970s,

high US interest rates drove them down in the 1980s,

especially in $ terms;

especially in $.

In the years 2003-2010, low interest rates may again

have been a source of high commodity prices.

References by the author include Frankel, 1986, 2008a,b; Frankel &

Hardouvelis, 1985; Frankel & Rose, 2009; and "Real Interest Rates Cast a
Shadow Over Oil," FTimes, April 15, 2005.
Also Barsky & Summers, 1988; and Caballero, Farhi & Gourinchas, 2008 .
Barsky & Killian (2002) and Killian (2009) believe that many big oil price
shocks have in reality been endogenous with respect to monetary policy.

53

Influences on oil
inventories
Regressors

Real interest rate

coefficient

error

Standard
at 10%

Significant

1. Real rate

-5.96

0.29

2. Real rate
& linear trend

-0.69

0.35

The spot-futures spread, risk & Industrial Production also

appear significant in variants of the inventories
equation.
Source:
Table 4, Frankel, The Effect of Monetary Policy on Real Commodity
Prices, in Asset Prices and Monetary Policy, edited by John Campbell
(University of Chicago Press), 2008, pp.291-327
54

Influences on oil inventories,

Real interest
Lagged
rate
inventories

-0.394*
0.089

-0.056
0.032

-0.211*

Spot-

0.032
0.000

IP

Risk

risk

futures

-0.821* 0.397*

-0.002*

0.041

0.062

0.001

-0.079*

0.052*

0.000

0.013

0.020

0.000

-0.727*

0.085

-0.017

IP

continued

0.131

0.040

-0.071*
0.009

0.009

-0.005*
0.126

0.009
0.012

0.931*

0.001

0.000

0.937*

0.045
55

Reasons for the oil price rise,

culminating in the 2008 spike

As Jim Hamilton points out, it differed

from the previous big oil price
increases which arose in geopolitical
disruptions to supply.

Rather, there was a combination

of rapidly growing demand,
with stagnant supply.
56

Past oil price shocks arose in

geopolitical supply disruptions
Exogenous disruptions in world petroleum supply.

Date

Event

Drop in world
oil production
Nov.1956 Suez Crisis
10.1%
Nov.1973 Arab-Israel War
7.8%
Nov.1978 Iran Revolution
8.9%
Oct.1980 Iran-Iraq War 7.2%
Aug.1990 Persian Gulf War
8.8%

Drop in U.S
real GDP
-2.5%
-3.2%
-0.6%
-0.5%
-0.1%

Source: Hamilton (2003, p.11).

57

Price of oil, 1940-2008

March 2008

58

Reasons for the 2007-08 oil price spike,

continued

Instability amongoil producers; fear of US-Iran

conflict; misguided US ethanol subsidies; etc.,

But the solution must be macroeconomic:

Prices of other minerals & agricultural products

increased at the same time.

Many said speculators were driving up the price.

Two macroeconomic fundamentals could explain

the decades boom in commodity markets:

Low interest rates in the US;

rapidgrowth worldwide,
especially in China & India.

59

Reasons for the 2007-08 oil price

spike,

The rapid growth in world demand,

especially from Asia, is obviously part of it

continued

2003-2007
and now back again since mid-2009.

But what explains the accelerated

price rise in 2008?

All forecasts for growth in Asia & the world

had been downgraded by then.
Only the easing of US monetary policy fits.
60

Don Kohn and Paul Krugman: low interest rates or speculation

could not be the causes, because oil inventories were low.
It is true that low interest rates or speculation, other things equal,
should in theory increase firms desire to hold inventories.

US crude oil inventories did not appear

especially low or high in the graph above,
showing June 1998-June 2008 (from Bloomberg).
61

World markets are relatively integrated,

so it is world inventories that should matter most.
Oil inventories in developed countries were above
average
during most of the year, as the graph shows.
They rose in January 2008,
when the Fed aggressively cut interest rates.
These numbers are far from conclusive
Source: Oil Market Report:
International Energy Agency

62

Conclusion: Causes of 2008 Oil

Spike
Hamilton (2009)

Primary causes:
Booming demand and stagnant production

Possible secondary factors:

Low interest rates

Speculation

When the global recession is over,

the fundamental imbalance will return.

Thus the long-term and short-term may be linked:

Malthus/Hotelling/Hubbert prediction of global scarcity

The 2008 price spike
63

Addendum:
Are speculators bad?

Sure, speculators are important

in commodities markets.
The spot price of oil especially on a day-today basis, is determined in markets where
participants typically base their supply and
demand in part on their expectations of
future increases or decreases in the price.
That is speculation.
But it need not imply
bubbles or destabilizing behavior.

64

Are speculators bad?

Speculators often fulfill useful functions:

continued

If they know the price is temporarily high, they

sell short, thereby moderating todays high price.
If they have reason to think there will be a future
increase in demand, they go long,
thereby driving up todays low price and sending
the market signal needed to spur investment.
In these cases they are the messenger delivering
the news about economics fundamentals.

Admittedly, there are sometimes speculative

bubbles, a self-confirming movement of
the market price away from fundamentals.
65

A story from the 1955 movie East of

Eden:

The legendary James Dean plays Cal.

Cal goes long in the market for beans, anticipating an
increase in demand if the USA enters World War I.
Sure enough, the price of beans goes sky high,
Cal makes a bundle, and offers it to his father.
But the father is morally offended by Cals speculation,
not wanting to profit from others misfortunes, and angrily
tells him that he will have to give the money back.
Cal has been the agent of Adam Smiths famous invisible
hand: By betting on his hunch about the future, he has
helped raise the price of beans in the present, thereby
increasing the supply so that more is available
precisely when needed (by the British Army).
The movie even treats us to a scene
in which Cal watches the beans grow,
which real-life speculators seldom get to do.
66

In Giant, its oil that James

Dean invests in, before World
War II again raises demand

67

68