changes in the operational efficiency of national oil companies

21
Changes in the Operational Efficiency of National Oil Companies Peter Hartley George & Cynthia Mitchell Professor, Economics Department, Rice University, Rice Scholar in Energy Studies , James A Baker III Institute for Public Policy and Kenneth B. Medlock III James A Baker, III and Susan G Baker Fellow in Energy and Resource Economics & Deputy Director, Baker Institute Energy Forum and James A Baker III Institute for Public Policy Adjunct Assistant Professor, Economics Department, Rice University

Upload: mauli

Post on 24-Feb-2016

37 views

Category:

Documents


0 download

DESCRIPTION

Changes in the Operational Efficiency of National Oil Companies. Peter Hartley George & Cynthia Mitchell Professor, Economics Department, Rice University, Rice Scholar in Energy Studies , James A Baker III Institute for Public Policy and Kenneth B. Medlock III - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Changes in the Operational Efficiency of National Oil Companies

Changes in the Operational Efficiency of National Oil

CompaniesPeter Hartley

George & Cynthia Mitchell Professor, Economics Department, Rice University,

Rice Scholar in Energy Studies , James A Baker III Institute for Public Policy

and

Kenneth B. Medlock III

James A Baker, III and Susan G Baker Fellow in Energy and Resource Economics & Deputy Director, Baker Institute Energy Forum and

James A Baker III Institute for Public PolicyAdjunct Assistant Professor, Economics Department, Rice University

Page 2: Changes in the Operational Efficiency of National Oil Companies

Overview

We examine a sample of firms that includes fully government owned National Oil Companies (NOCs), partly government owned national oil companies (pNOCs) and shareholder owned oil companies (SOCs)

We reaffirm previous evidence that NOCs and pNOCS tend to be less revenue efficient than SOCs

Excessive employment and retail subsidies are again found to be significant causes of the reduced revenue efficiency of NOCs and pNOCs

We also find that partial privatization, and mergers and acquisitions, are likely to increase efficiency

Also, while oil and gas firms as a whole tended to become more efficient at producing revenue from 2001–09, NOCs and pNOCs on average improved more than SOCs

Page 3: Changes in the Operational Efficiency of National Oil Companies

Data

Primary source “Ranking the World’s Oil Companies” by Energy Intelligence

Company annual reports also used to check and revise, and provide missing, data

We began with almost 150 firms but the methods require a balanced panel

This constrained both the number of years and the number of firms The inputs and revenues of merger partners were combined in years prior

to a merger Nippon Oil was dropped as an outlier in specialization in downstream

activities Ultimately, the data set covered 61 firms for the period 2001-2009

We added data on oil and natural gas prices from the EIA and average retail fuel prices from the Metschies surveys

Page 4: Changes in the Operational Efficiency of National Oil Companies

Model

We focus on revenue technical efficiency for several reasons: The different products sold are most naturally aggregated using prices Political pressure is likely to force a NOC to subsidize sales to domestic

consumers Revenue is a key objective for both public and private firms For many firms, revenue figures are more readily available than physical

outputs of different commodities Crude or natural gas output: Q = F(L)×Rsv×G(E), where E =

cumulative output Marketed products: R = H(K,L,Q) Revenue is then p(1–s)R for a vector of product prices p and

corresponding percentage subsidies s for each component of R

Page 5: Changes in the Operational Efficiency of National Oil Companies

Variables used in the analysis Rev = Revenue in $ million Emp = Number of employees OilRsv = Oil reserves in millions of barrels NGRsv = Natural gas reserves in billions of cubic feet Refcap = Refining capacity in thousands of barrels per day Oilp = Average US import oil price, average OPEC and non-OPEC oil prices (in $/barrel) NGp = Henry Hub, Japan LNG import prices, EU pipeline and LNG import prices year = 1, 2, … 9 for 2001–2009 respectively GovSh = Government ownership share; GovSh+ = I(GovSh > 0) VertInt = Vertical integration measure (product sales/liquids production both in

thousands of barrels per day) Premerge = 1 in the years before firms merged, 0 in years following a merger; 0 for non-

merging firms RetSubs = percentage deviation of gasoline and diesel retail prices in the headquarter

country below the average US retail prices; 0 if average retail prices are greater than or equal to US retail prices

Page 6: Changes in the Operational Efficiency of National Oil Companies

Methods overview

We use two methods to calculate revenue efficiency: data envelopment analysis (DEA) and stochastic frontier analysis (SFA)

DEA is non-parametric but does not allow for measurement error or other sources of variation across firms unrelated to inputs or relative efficiency

SFA is parametric, permitting a structural interpretation of the results

However, if the assumptions, including those relating to the structure of the error terms, are invalid the interpretation may be misleading

SFA also allows for other sources of random variation and tests of goodness of fit

The aim is not to compare the methods but to check robustness of the results

Page 7: Changes in the Operational Efficiency of National Oil Companies

Annual DEA efficiency scores

Page 8: Changes in the Operational Efficiency of National Oil Companies

Some observations The efficiency scores for most of the firms increased over the nine-

year period Notable exceptions among SOCs: Occidental, Chesapeake, BG, EOG, CNR,

Devon, Talisman, Noble and Plains PDV from Venezuela followed a rising then falling pattern ending with a

lower score in 2009 than any earlier year Partially privatized PDO from Oman also follows a rising then falling

pattern TNK starts with a score of 0.077, jumps to 0.78 when TNK-BP is formed in

2003, stays on the frontier until 2007, but then drops back to 0.73 in 2009 Like the major IOCs, the most efficient pNOCs – Statoil-Hydro,

Sinopec and PTT – are on the frontier every year ENI, CNOOC and Petrobras generally improve with ENI and CNOOC

attaining the frontier Saudi Aramco is on the frontier from 2005–09, Sonangol from 2004–

07, QP in 2009 and Kuwait’s KPC in 2007

Page 9: Changes in the Operational Efficiency of National Oil Companies

Average DEA scores by category

Trend in SOC scores is not statistically significantly different from zero

Both NOC and pNOC scores trend up at the same rate of around 0.015/year

NOC and pNOC scores vary more year to year, partly because a larger fraction of SOCs have a maximum score of 1 each year

Page 10: Changes in the Operational Efficiency of National Oil Companies

Panel Tobit models of DEA scores

Government ownership and time:

After allowing for firm-specific effects (> 80% of residual variance), partial government ownership has the same effect as full ownership

NOCs and pNOCs start with 30% lower DEA scores in 2001, but by 2009 their scores are only 8% lower

Allowing for retail subsidies, efficiency changes from mergers, and an effect of vertical integration we find:

Page 11: Changes in the Operational Efficiency of National Oil Companies

Technical change measures

Page 12: Changes in the Operational Efficiency of National Oil Companies

Comments on technical change measures

The average of all technical change measures across years and firms is 1.129, implying that on average the frontier expanded over the decade

The average measure across firms exceeded 1 for all pairs of years except 2001–2002

Years 2006–07 and 2008–09 also showed somewhat weaker technical progress

2002–03 had highest average measure, although only in 2004–05 were all measures greater than 1

Across all years, the average technical change measure for SOCs is 1.141 compared to 1.123 for NOCs and 1.111 for pNOCs

With the exceptions of CNOOC and PDO in 2008–09, all the large technical change expansions, and the majority of the frontier contractions, occur for SOCs

Page 13: Changes in the Operational Efficiency of National Oil Companies

Comments on technical change measures

Measures for SOCs vary more than measures for pNOCs and NOCs Among the five SOCs that remain on the frontier every year:

ExxonMobil stands out with technical change measures very close to 1 in all years, although the measures for BP are only slightly more dispersed

Wintershall and BHPBilliton have the largest dispersions in technical change

The dispersion in technical change measures for Marathon is intermediate Among the pNOCs on the frontier every year (StatoilHydro, Sinopec

and PTT) Dispersions in StatoilHydro and PTT measures are quite similar to those of

Marathon Except for 2006-07, the measures for Sinopec are about as dispersed as

those for ExxonMobil, but their average is substantially above 1, implying that Sinopec had to make changes to remain on the frontier each year

Page 14: Changes in the Operational Efficiency of National Oil Companies

Panel regression on Malmquist total productivity change measures

Mergers and a reduction in retail fuel subsidies both raise total productivity

The insignificance of ΔVertInt and the GovSh measures is consistent with the DEA results

Page 15: Changes in the Operational Efficiency of National Oil Companies

Stochastic frontier model 1

For a Cobb-Douglas production function, log of revenue will depend linearly on the logs of the input variables and the log of prices

Error terms:

vit and ui are distributed independently of each other and the regressors with estimated values μ = 1.506 (0.2657), η = 0.0291 (0.0047), (0.0888) and (0.0031) so more than 89% of the variation in the composite error term is due to the one-sided systematic efficiency differences or other sources of firm heterogeneity

Page 16: Changes in the Operational Efficiency of National Oil Companies

Efficiency measures compared

Panel Tobit regression:

Page 17: Changes in the Operational Efficiency of National Oil Companies

SFA model 2 We would expect ln(1–s) to appear in the revenue function Relative inefficiency of NOCs is also likely to be manifest in reduced

labor productivity, or an interaction between GovSh and Emp in the revenue function

We also allowed for a structural model of the one-sided inefficiency error term using the specification of Battese and Coelli (1995) whereby the mean of the firm-specific inefficiency measures uit depends on firm-specific covariates zlit

where wit is a truncated normal with mean zero and variance such that the point of truncation is wit ≥ –zitδ

We allowed GovSh, Premerge and VertInt as potential z variables The non-efficiency error component vit is iid and independently

distributed from the uit, which also are independently distributed from each other

Page 18: Changes in the Operational Efficiency of National Oil Companies

SFA model 2 estimates

The best estimated model

with estimated error structure

and with

Page 19: Changes in the Operational Efficiency of National Oil Companies

Conclusions

The two inefficiency measures were highly positively correlated Retail subsidies were a major source of reduced efficiency for many NOCs and

pNOCs Government ownership also tends to produce a lower productivity of labor

Partial privatization reduced this effect on average, but it did not decline over time There was also an additional residual negative impact of government ownership

that was not affected by partial privatization and disappeared over the decade The DEA and Malmquist analyses found evidence that mergers tended to raise

the efficiency of the merging firms, but we did not find a similar effect in the SFA SFA, but not DEA, found that higher final product sales/liquids production was a

significant source of firm heterogeneity More variable technical change measures for SOCs may indicate more

innovation occurs in such firms The DEA and Malmquist analyses provided some interesting details on the

relative efficiencies and efficiency changes of particular firms over the decade

Page 20: Changes in the Operational Efficiency of National Oil Companies

Additional slides

Page 21: Changes in the Operational Efficiency of National Oil Companies

Contributions to dominating composite firms