Popper’s writings cover a remarkably wide range of subjects. The spectrum runs from Plato’s theory of politics to the foundations of quantum mechanics. Yet even amidst this variety the philosophy of probability occupies a prominent place. David Miller once pointed out to me that more than half of Popper’s The Logic of Scientific Discovery is taken up with discussions of probability. I checked this claim using the 1972 6th revised impression of The Logic of Scientific Discovery, and found that of the approximately 450 pages of text, approximately 250 are to do with probability. Thus Miller’s claim is amply justified. It seems indeed that the philosophy of probability was one of Popper’s favourite subjects, and, as we shall see, Popper certainly enriched the field with several striking innovations. In this area, as in others, Popper held very definite views, and criticized his opponents in no uncertain terms. Popper was an objectivist and anti-Bayesian, and his criticisms were directed against subjectivism and Bayesianism.

INTRODUCTION

Keynes is most famous as an economist, but he was involved with probability from the very beginning of his research career, and in fact his first piece of academic research was in the philosophy of probability. Keynes's interest in probability began as an undergraduate. He started his degree in 1902, and in 1903 he was initiated into Cambridge's secret society of elite intellectuals, known as the Apostles. A paper that he read to this society in January 1904 contains his first discussion of the philosophy of probability (Skidelsky 1983: 152). In those days, the PhD degree did not exist in Cambridge, and bright graduate students were expected to submit dissertations in the hope of obtaining college Fellowships. In 1907, Keynes submitted a dissertation on probability for the prize competition at King's College Cambridge, but surprisingly he was unsuccessful. The college awarded Fellowships instead to two gentlemen by the names of Dobbs and Page. In 1909, however, Keynes did win a Fellowship at King's with a revised version of his dissertation on probability. Incidentally, Alan Turing, the computer pioneer, won a Fellowship to King's with a dissertation on probability on his first attempt in 1935. Keynes was on the committee which awarded Fellowships that year.

Introduction

Suppose we were to ask some students of philosophy to imagine a typical book of classical German philosophy and describe its general style and character, how might they reply? I suspect that they would answer somewhat as follows. The book would be long and heavy, it would be written in a complicated style which employed only very abstract terms, and it would be extremely difficult to understand. At all events a description of this kind does indeed fit many famous works of German philosophy. Let us take for example Hegel's Phenomenology of Spirit of 1807. The 1977 English translation published by Oxford has 591 pages, and as for style here is a typical passage:

Self-consciousness found the Thing to be like itself, and itself to be like a Thing; i.e., it is aware that it is in itself the objectively real world. It is no longer the immediate certainty of being all reality, but a certainty for which the immediate in general has the form of something superseded, so that the objectivity of the immediate still has only the value of something superficial, its inner being and essence being self-consciousness itself. The object, to which it is positively related, is therefore a self-consciousness.

Introduction

Popper's writings cover a remarkably wide range of subjects. The spectrum runs from Plato's theory of politics to the foundations of quantum mechanics. Yet even amidst this variety the philosophy of probability occupies a prominent place. David Miller once pointed out to me that more than half of Popper's The Logic of Scientific Discovery is taken up with discussions of probability. I checked this claim using the 1972 6th revised impression of The Logic of Scientific Discovery, and found that of the approximately 450 pages of text, approximately 250 are to do with probability. Thus Miller's claim is amply justified. It seems indeed that the philosophy of probability was one of Popper's favourite subjects, and, as we shall see, Popper certainly enriched the field with several striking innovations. In this area, as in others, Popper held very definite views, and criticized his opponents in no uncertain terms. Popper was an objectivist and anti-Bayesian, and his criticisms were directed against subjectivism and Bayesianism.

As well as carrying out his own research in the philosophy of probability, Popper stimulated interest in the subject among members of his department at the London School of Economics. When I arrived there as a graduate student in 1966, Imre Lakatos was editing a volume on one part of the subject (The Problem of Inductive Logic), and himself wrote a paper for the volume. Imre Lakatos did not return to the subject, but this paper is itself a notable contribution.

Public-enterprise finance, broadly defined, encompasses mechanisms for raising funds, distributing profits, and absorbing losses. Although these functions are shared with private finance, public-enterprise finance typically differs in one important respect. When private firms raise capital on a commercial basis, they incur explicit obligations to repay specific amounts or shares of their profits in the future. These obligations reflect the opportunity cost of funds for claims of similar riskiness. Thus, in raising funds, they also specify the distribution of future profits and losses. In contrast, when public enterprises raise funds, they often do so without incurring such explicit obligations. Even when public enterprises borrow from commercial sources, they often do so with governmental guarantees, implicit or otherwise. As a result, public enterprises incur explicit financing costs that do not reflect the private opportunity cost of funds employed in the enterprises, let alone their social opportunity costs.

In this chapter we argue that the typical pattern of public-enterprise finance yields a distorted picture of the opportunity cost of funds employed and, hence, of the profitability and efficiency of public enterprise. Although this should have no real effect in an idealized setting where all public-enterprise decisions are made in light of the full set of social costs and benefits – including the social-opportunity cost of funds employed – we believe that in practice the way public enterprises are financed often contributes to inappropriate investment and operating decisions. In our view, the structure of management incentives is the primary link between financing and the strategic and operating decisions of a public enterprise.