A complete seismic inversion would solve for the physical parameters in the scattering potential. Most imaging algorithms stop short of that, solving instead for a reflectivity function. To produce the physical model behind the reflectivity, further analysis is required. Similarly, to synthesize seismic data one often starts from a reflectivity rather than a scattering potential.

Inversion for or modeling from a reflectivity model is legitimate, if incomplete, though one does have to ask, just what is reflectivity? The reflection coefficient is a well defined concept for plane waves impinging upon an abrupt boundary of no curvature. Asymptotically, the concept extends to waves far from source or receiver impinging upon boundaries of reasonable curvature. Consequently, when we think about reflectivity we are usually thinking about discrete specular reflections produced at reflecting boundary surfaces. However, changes in real-Earth properties may or may not be abrupt, so confining reflectivity to a boundary surface is at best an idealization, at worst wrong. Moreover, imaging algorithms normally produce images of volumes, not surfaces. That is, they produce a value at every point within a volume representing the Earth's subsurface, not just along selected surfaces within the volume. Imaging algorithms solve for a pointwise reflectivity function, not a specular reflectivity function.

This is as it should be. Specular reflectivity is the integrated response of a boundary surface, not the response of a single point on that surface.

Conventions

Vector conventions Vector quantities appear in bold-face type (e.g. x). Two-dimensional projections of threedimensional vectors are given a ∼ overhead, as in x = (x, y), as a projection of x = (x, y, z). Matrices are represented as bold-face capitals.

Operator argument conventions

Some quantities can be viewed as operations which map a set of input coordinates onto a set of output coordinates. We represent such quantities with the form D (xg|xs;t). The xs is placed right of the vertical bar to indicate that it is an input coordinate, while the xg is placed left of the vertical bar to indicate that it is an output coordinate. The t is placed on the far right to separate it from the input and output coordinates. We could define a source time as an input coordinate and a measurement time as an output coordinate but, since we consider only the difference between these two times to be important, have chosen not to. Our notation is analogous to a generalized matrix notation in which the receiver locations xg indicate rows and the source locations xs indicate columns. Were we to employ Dirac notation (commonly used in physics), D (xg|xs; t) would become 〈xg|D(t)|xs). Though better in many respects, Dirac notation is not in common use within the geophysics community, so we refrain from using it in this volume.

Explanation of performance ranking

In this appendix we classify the overall performance of the fifteen NOCs in our sample into four categories – high, upper middle, lower middle, and low. The assessments are based on the detailed case studies presented in Chapters 5–19. This classification is then used in other parts of this book, such as the Chapter 3, on how national governments attempt to govern NOCs, as well as the Conclusion that summarizes what we have learned about the factors that explain performance (Chapter 20). As we describe below, characterizing performance at the level of the individual firm – especially for firms that are often secretive and designed to perform many functions – is a necessarily qualitative and subjective process. Nonetheless, any study that aims to explain performance must make an effort to measure its quarry. This Appendix explains what we mean by “performance” and explains our scoring for each case in the spirit of transparency and debate.

For the purposes of comparing NOCs (and NOCs with IOCs), we initially sought quantitative metrics. The Introduction (Chapter 1) summarizes some of the existing quantitative research. Within the realm of what can be measured quantitatively, existing studies show that state ownership does in fact have a noticeable effect on standard performance metrics such as output efficiency, per-barrel operating cost, investment cost, and net present asset values. However, this kind of research is difficult to implement because there are no universally applicable benchmarks for performance comparison between individual companies; the variables that are most interesting to measure are usually the hardest to quantify.

Introduction

With more than 70 percent of the world’s oil reserves, national oil companies (NOCs) dominate world oil markets (see Chapter 1). Fourteen of the top twenty oil companies in the world are NOCs or newly privatized NOCs. Moreover, NOCs appear to be increasing their control over global oil reserves (Kretzschmar et al. 2009). Due to their increasing pre-eminence in world energy markets, NOCs have recently been the focus of a cottage industry of academic studies. A number of studies have found profound differences between the performance of NOCs and independent, private oil companies. Eller et al. (2007), Victor (2007), and Wolf (2009) all find that NOCs are dramatically less efficient than their private counterparts. Moreover, Wolf and Pollitt (2008) find that NOC performance improves dramatically following privatization. However, NOC performance is far from monolithic. Some NOCs are able to perform as well as major private companies, while others fall significantly short (Victor 2007). Indeed, this entire volume explores how NOCs perform and how they differ from private enterprise.

These findings raise two interlocking puzzles about the structure of the global petroleum industry. First, if private oil companies usually perform so much better than nationalized companies, why have so many countries nationalized private oil companies’ assets and established state-owned petroleum companies? Second, why do so many countries continue to maintain nationalized oil companies in the face of strong evidence of weak performance?

Introduction

In this chapter, I analyze how states govern national oil companies (NOCs) – typically their chief sources of revenue and technocratic expertise. Some governance systems set clearly defined lines of authority, separating policymaker, regulator, and operator responsibilities, whereas others blur those lines almost completely. States differ in their control over NOC budgets, appointments, investments, and contracting arrangements with international oil companies (IOCs).

Despite these differences, a common thread exists: Most NOC governance systems are a hybrid of corporate governance, public administration, and regulation. Many NOCs’ founding laws and organizational by-laws mimic private sector corporate governance models. NOCs typically have a board of directors, hold annual shareholder meetings (with the state as the only or more important shareholder), and pay dividends. Many states also subject their NOCs to the same public administration tools that states use to control government agencies, like legislative budget approvals. And several states use regulatory powers to indirectly control their NOCs, with some also commandeering their NOCs as energy regulators. NOC governance reflects this hybrid blend because of the special challenges posed by managing oil companies. NOCs are almost invariably the leading state-owned enterprise (SOE) and often the largest domestic company, public or private, in the domestic economy. And unlike the perennial problem in many SOEs, which lose money, NOCs usually produce copious rents.