The chase by an ‘economic hound’ ^{Footnote 6}

Thus began Phillips, in his first published article, with that characteristic humility for which he was known and which is repeatedly – and justly – emphasised by Dr Bollard, in Hares. ^{Footnote 8}

Dr Bollard ‘confesses’, ‘I have largely ^{Footnote 9} avoided graphs and avoided equations’ (p. vi). This is a laudable aim, especially if the ‘economic times’ of this humble genius are described with reasonable fidelity, and provided that a good list of references is given, so that one may proceed to tackle the more technical papers dealing with the kinds of issues that were broached by Phillips. I think Dr Bollard succeeds, ‘largely’, in achieving this aim, but not ‘entirely’ satisfactorily, especially not in part II.

Approximately a third (74 out of 232 pages of the main text) of Hares tells the story, in that outmoded grammatical mood of the subjunctive, of the first 23 years of Phillips’ pre-economic life of 60 years. The first four chapters are liberally sprinkled with phrases such as ‘would have’, ‘would be’, ‘would also have’, ‘could have’, ‘may have’, ‘we can only speculate’ and ‘suggest’. This is a plausible reconstruction of that part of Phillips’ life, spent largely in the Antipodes, East and South East Asia. The author’s knowledge of the New Zealand landscape is superbly invoked in describing the kind of life that may have moulded this quiet, increasingly self-sufficient man who refrained from any kind of complaining of whatever hardship he, or his family, had to endure – and there were copious instances of them, as poignantly narrated in the engaging first part of Hares.

Apart from the time Phillips spent in a Japanese Prisoner of War camp, and the so-called British Institute of Technology courses, ^{Footnote 10} I have no especial knowledge or expertise on any of the topics discussed and narrated in part I. I shall, therefore, simply note some of the salient points of that part of Bill Phillips’ life before his arrival in London, ‘nearly’ around ‘Christmas 1937’ (Hares, p. 43).

The way the independent spirit of a New Zealander may have been inspired by the natural setting of the country is evoked, almost sotto voce, in Hares:

And,

This landscape may well be a testimony to the irrelevancy of nature’s bountiful endowments to the fostering of intellectual greatness.

The early and growing life of Phillips, spent in the almost idyllic natural surroundings of the spectacular North Island of New Zealand, cannot be more different from a POW camp in Java or ‘digs’ in London! But these ‘idyllic surroundings’ were the backdrop of the harshness of life, the overcoming of which may have cultivated the resilience, the almost single-minded determination to resolve any problem with which he was confronted, a hallmark of the remarkable achievements of Phillips.

Three examples of the innovative ways Bill Phillips came to terms with, and overcame, the difficulties of the limited resources of a rural life – however idyllic the environment was – deserve mention, because they reflect his attitude and determination, in the face of adversity.

First, like many budding geniuses, his thirst for knowledge, growing in a rural community, where his daily journeys to school was nothing short of an ordeal, in terms of the time and the various modes of transport he had to resort to – cycling, walking and a goods train ride, capped by another ‘brisk walk uphill three-quarters of a mile to the high school’ (Hares, p. 25) – to go from home, and return home, early and late, respectively, led to innovative solutions to a problem. As described in Hares, Phillips found,

And, then,

These noble characteristics – mastering the mechanism of a useful machine, teaching himself a skill, whether cerebral or technological, breaking inconvenient rules and making sure the circumvention was done in a way that made it possible for the ‘enforcers’ of an inconvenient law to ignore the offender, all at age ‘about 14’ – stood him in good stead in the years to come, in difficult and better times.

Bill Phillips’ father Harold (‘known by his contemporaries as Housego, his mother’s maiden name’, later ‘inherited’ by Bill Phillips, Hares, p. 7) is said to have been very adept at mathematics and reading. Reading Shakespeare would become a life-long interest. Harold Housego was also adept at adapting technology to make electricity available at the home, was innovative in introducing modern livestock breeding methods and was an enthusiast of outdoor sporting activities. Many of these traits passed on to Bill Phillips, except ‘reading Shakespeare’ and enthusiasm for sporting activities.

His personal exploits and feats, under the difficult captivity by victorious Japanese forces in South East and East Asia, in the Island of Java, have become legendary. They are too well known for me to comment on Bollard’s sympathetic description of the ordeals and scars that Bill Phillips underwent during the years as a POW.

The negative remarks on Laurens Van der Post in Hares, and this South African’s pitiless way of disregarding the important role played by Phillips, in those dark, final, days at the POW camp, receive the kind of opprobrium that they deserve: ‘Even though he obviously knew his name, the egotist van der Post may have deliberately excised Bill’s name from his account to keep the focus on himself’ (Hares, Note 15, p. 240). ^{Footnote 11}

Part II is divided into seven chapters, Chapters 6–10 largely spanned by the eight path-breaking contributions by Phillips, from 1949 to 1962 – the exceptions being the 5th and the 11th chapters of the book, ‘Post-war London’ and ‘China and an ending’, respectively. There is also a brief, 13-point, summary Annex and a 10-page Endnote section on each of the chapters of the book. With the possible exception of point 8, all other claims in the Annex are demonstrably false.

Chapter 5 describes the 2 years of Bill Phillips’ life, between finding ‘himself back in London in early 1946’ and the end of the glorious summer of 1947 – a period of momentous changes in society, and in the way Phillips himself began to fashion an interesting academic life, although with serious ‘birth pangs’. It is a well-written chapter, with the general social, political and economic background to the world in which Bill Phillips found himself, so that this qualified and experienced electrical engineer’s decision to pursue a sociology degree at the London School of Economics (LSE), with increasing interest in economics, might make some sense.

The chapter also documents the linguistic abilities of Bill Phillips – fluent or proficient (at least with reading knowledge), apart from English, in Chinese, Russian, French, Dutch, Malay and German.

Two points raised in this chapter deserve attention. One is the contradictory interpretation by Leeson and Sleeman of the importance – or not – of the wartime experience on his choice of Sociology (Hares, Note 5 to Chapter 5, p. 242). The other is the unfortunate penchant of the author of Hares to ‘indulge’ in unnecessary and inaccurate ‘Cambridge bashing’ (despite his own Cambridge ‘pedigree’, Hares, Note 1, p. 240).

First, ‘unlike the fellow travelers of Cambridge University’, the author writes, ‘the conversation around the LSE corridors would have reflected all shades of views’ (Hares, p. 92, italics added). Second, on the same page of Hares, one reads, ‘Newspaper headlines trumpeted the arrest of spy scientist Klaus Fuchs, and the defection of spy diplomats Guy Burgess and Donald Maclean, rocking the Cambridge establishment’ (p. 92).

The LSE economics department was not the bastion reflecting ‘all shades of views’, nor is there anywhere an identifiable ‘Cambridge establishment’ that was rocked. Moreover, the Fuchs, Burgess and Maclean ‘affairs’ did not become ‘newspaper headlines’ till the 1950s, well outside the interval of time discussed in this chapter.

Chapter 6, titled ‘Hydraulic machine’, is the story of the MONIAC, how and why it was conceived, constructed, implemented and displayed, with mathematical and theoretical foundations in mechanics, hydraulics and macroeconomic theory. It is an important story and I doubt anyone else than the author of Hares can tell it better or more forcefully. ^{Footnote 12} That, of course, does not mean the chapter’s contents are without blemish.

There is no evidence whatsoever that Phillips, or the British Institute of Technology course in electrical engineering that he took, was aware of the burgeoning developments in non-linear (electrical) circuit theory in the 1930s. Nor did any economist at the LSE – at that time – have any knowledge of non-linear modelling of trade/business cycle theories. Without acknowledging this technical ‘gap’ in Phillips’ knowledge base, there is a risk of untenable claims and assertions, as – indeed – there are, in Hares, on his contributions to macroeconomic dynamics and stabilisation policy.

This chapter is the longest in the book, and justifiably so. In a way, it is the beginning and the foundations upon which Phillips’ remarkable life as an economist developed. There is, however, nothing in the way of MONIAC’s relationship to the differential analyser, although in sporadic ways it is pointed out that it was designed to ‘solve’ differential equations (essentially the initial value problem of such equations), assuming, implicitly, the Peano existence theorem and some kind of Lipschitz continuity. ^{Footnote 13}

It is in this context that Robertson’s perceptive query should be understood:

There is no indication in Hares of how Phillips clarified ‘his argument’, in the ‘letter of September 1950’. Did Phillips acknowledge that the MONIAC was a special differential analyser, which could handle only differential equations – it would not be able to ‘handle difference equations’, without ad hockeries?

Yet, a persistent reader can easily understand how Bill Phillips went about constructing the MONIAC, how he underpinned the mechanics of the working machine in the macroeconomic theory of the times (with the help of Walter Newlyn) and how his background in electrical engineering would have helped him in putting together the electro-mechanical dynamics of an essentially hydraulic machine – all this by reading this chapter carefully, weeding out the infelicities.

But ‘weeding out the infelicities’ will not be easy – for the chapter is interspersed with technically incorrect claims, infelicities in references of classics and a complete misunderstanding of the meaning of optimal control theory, ^{Footnote 14} all amidst the generous portrayal of the genius that Phillips was, in conceiving the MONIAC.

Several of the assertions about Irving Fisher’s Hydraulic Machine are wrong – about when the initial machine was built (p. 96), erroneous statement that the cisterns were connected by ‘levers’ (p. 96), incomplete and incorrect reference to Fischer’s 1982 thesis in the reference list and failure to note that Fisher’s mechanism for the hydraulic determination of equilibrium prices was underpinned by an appeal to Pascal’s Principle – and that this is a fundamental difference with the mechanism of the MONIAC.

The claims for Leontief’s assertions, in Hares, pp. 96–97, are wrong; ^{Footnote 15} Leontief discusses ‘the fundamental difference’ between the flow of water in a river and the flow of goods in an economy, as well as the drawing of ‘erroneous conclusion’ from a ‘false analogy’ (between the two flows).

In Note 5, p. 242, the reference to the title of the second edition of Allen’s Mathematical Economics is incorrect. The composition of the ‘Cambridge Circus’ group (p. 104) did not include Keynes, but did include Meade. The claim that Hayek’s The Road to Serfdom ‘would later help win him the Nobel Prize’ (p. 105) cannot be substantiated. Both statements about the two modern classics by Hicks and Samuelson, ^{Footnote 16} on pp. 113–114, are hopelessly wrong. There was no ‘third edition of Foundations’ (p. 114) in which ‘Samuelson synthesized a quasi-Keynesian Hicksian framework’. ^{Footnote 17} Phillips could not ‘have been exposed to the intellectualism and policy interest of … Kaldor’, at the LSE (p. 114) – simply because he (Kaldor) had left by the time Phillips arrived there.

It is claimed in Hares that this reviewer pointed out that the MONIAC ‘would not pass the Turing test for being an analogue of the real world’ (p. 116)! This is an incorrect claim, theoretically and factually; theoretically, because ‘the Turing test’ is the basis for a kind of test for what has come to be known as Artificial Intelligence; factually, because I never made such a preposterous statement.

The Skidelsky story (p. 117) about Marty Feldstein who ‘used the machine to explain the Keynesian circular flow of income’, although amusing, should never have been rehashed the way it is done in Hares. Skidelsky adds a qualifying remark (italics added), that Feldstein’s explanation was ‘based on an actual machine Alec Cairncross had caused to be constructed at the LSE’! How could ‘Ramsay McDonald and the other politicians around at this time’ – that is, 1929–1931 – understand ‘the principles of [MONIAC]’, when Keynes had only just begun writing the General Theory! Machlup was never at Harvard (p. 122).

Finally, there is no clear indication of the importance of a flexible, non-linear, accelerator and a dynamic multiplier, jointly generating and maintaining aggregate fluctuations; instead, almost all the discussion is against the backdrop of a linear accelerator. I believe this was partly due to Newlyn’s influence on Phillips; Newlyn adds, incorrectly, that Goodwin was inspired to use the (non-linear) accelerator as a result of its (linear) incorporation in the Mark II MONIAC (Reference LeesonLeeson, 2000, editor: Chapter 8). It was, in fact, the other way about!

Chapter 7, on ‘Stabilization and computing’, is only a page shorter in length than the previous one, but is as informative and interesting, although far less marred by the kind of infelicities of Chapter 6 and its notes. It is about the ‘interregnum’ professional years between the construction of the MONIAC, the successful beginnings of an academic economist, at an outstanding Institution, the enrichment of a personal life via a romance which led to a marriage and a lifelong loving and devoted partnership, and the next stage in a research career that blossomed into work that became pioneering, with many ramifications in the theoretical, empirical and applied worlds of macroeconometrics, dynamic macroeconomics and continuous-time estimation.

The two key contributions by Phillips, for Chapter 6, are the classic Economic Journal articles of 1954 and 1957, but also his reviews of Kalecki’s book of 1954 ^{Footnote 18} and Tustin’s, also of 1954, belong to this period, ^{Footnote 19} as should Phillips’ Zurich cybernetic paper of 1958. Two important events of this period – one academic and the other personal – need to be mentioned (Hares, pp. 138–139 and pp. 130–131 ^{Footnote 20} ). As described,

Then, as if to celebrate these honours,

The subsection on ‘Personal and academic life’ (pp. 150–155) is a relaxed and generous description of the domestic life of Bill and Valda Phillips and the duties and burdens of an increasingly successful academic economic career of Bill Phillips, culminating in his appointment as the Tooke Professor of Economics ^{Footnote 22} at the LSE.

To return to the five classic contributions mentioned above, by Phillips, two preliminary comments need to be made. In referring to Phillips’ Zurich paper, ^{Footnote 23} Bollard writes,

I doubt the author of Hares has read this article by Phillips carefully or that he understands the difference between classical and optimal control theory; none of the claims about optimal control and optimal control theory can be substantiated. Phillips’ Zurich Conference paper remains a good contribution – not a pioneering one – to the literature on Cybernetics, which was taking a different direction from Optimal Control Theory.

Incidentally, the implicit criticism of the National Physical Laboratory’s focus being ‘too theoretical’, implying thereby also a critique of the work being done by Phillips ‘on the analogue computer’ (p. 150), is, at best, amusing. Also, the ‘radically advanced new digital computer, the Automatic Computer [sic!] ^{Footnote 24} Engine (ACE)’ (p. 146) (designed by Alan Turing).

Before moving on to the ‘stabilisation’ problem, supposedly pioneered by Phillips, within a servomechanism, or negative feedback framework, it will be useful to remember, first, that the following assertion in Hares is meaningless:

The opposite is the fact (Hares, pp. 145–146) that analogue computers were – and are – utilised in continuous-time problems, discrete time or discrete data problems are natural repositories of digital computers. Phillips, late in his economic modelling life, struggled to find a technical solution to the problem of continuous-time estimation on digital computers. It is about reconciling scientific computing and computability ^{Footnote 25} theory.

Second, both Herbert Simon and Richard Goodwin pioneered the use of servomechanism theory in economics, long before Bill Phillips. Moreover, Bill Phillips was comprehensively wrong ^{Footnote 26} to state the following in what he wrote to Charles Holt in October 1956:

Third, it was Goodwin, in the late 1940s, who pioneered the evaluation of the dynamics of policy stabilisation, using short-run aggregative models of non-linear fluctuations, in the Anglo-Saxon literature.

I wish it had been emphasised, in Hares, that Phillips’ elementary conclusions on policy dynamics were based on an ultra-simple, linear model of the economy. Whether he was motivated by his electrical engineering training or his nascent interests in servomechanism theory, Phillips was ill-informed about the non-elementary conclusions on dynamic stabilisation policy intrinsic to non-linear circuit theory and non-linear dynamics, both of which had been copiously used in mathematical macrodynamics.

‘Ultimately’, claims Dr Bollard, ‘the Phillips curve article is about a statistical relationship’ (Hares, p. 170). Yet, the previous section ^{Footnote 27} (the second of this chapter), which begins with the Reference LaidlerLaidler (2002) quote that ‘Every economist has heard of Bill Phillips, most of them for the wrong reason’ (p. 223), is supposed to have drawn ‘primarily on Lipsey (2000)’, as the author states in Endnote 2 to this chapter (Hares, p. 245). I am afraid the contents of this chapter show no evidence whatsoever that either Reference LaidlerLaidler (2002) or Lipsey (2000) has been properly read.

First, had Dr Bollard read the Laidler piece carefully, most of the erroneous claims on optimal control theory could have been avoided – but not all, because Reference LaidlerLaidler (2002) himself adds another dangerous comment on the topic:

The second sentence in the above quote is not correct. ^{Footnote 28}

Second, Reference LaidlerLaidler’s (2002) observation that ‘the idea of the curve came from Phillips’ reading of Bent Hansen’s (1951) Theory of Inflation …’ (p. 228) leads on to Lipsey’s ‘three considerations’ ^{Footnote 29} on Phillips’ own view of what the curve that bears his name should mean, where the most important of these is that it depicted the economy’s lack of self-adjustment capabilities, and therefore, the curve represented a disequilibrium phenomenon, as in Bent Hansen, in the tradition of Lindahl. ^{Footnote 30} To the credit of the author of Hares (pp. 169–170), all three points Phillips raised with Lipsey are handsomely acknowledged – although without mentioning the underpinning in Bent Hansen’s work.

The other important point, emphasised in Hares, in connection with the geometric form of the relation that came to be called the Phillips Curve, is the following: ‘[I]t was a distinct improvement on the old textbook treatment, with its sharp dichotomy between stable prices and inflation’ (p. 169; italics added). This statement should be coupled to Lipsey’s ‘historical’ observation ^{Footnote 31} that removing the classical dichotomy was important in useful models of stabilisation policy – but ‘useful’ also in the sense of an engineering mechanism that could encapsulate reasonably smooth flows.

To the idea that the three articles by Phillips of 1954, 1957 and 1958 ‘need to be seen as a unit’, I would add Phillips (1950) – and the part played by his background in electrical engineering in the construction of the MONIAC. ^{Footnote 32}

The ‘Friedman connection’ is pointed out in this chapter because it was, in many ways, the launching pad for what came to be the dominant kind of macroeconomics at the frontiers today. Moreover, Dr Bollard, in Hares, was perceptive in reporting that ‘Friedman argued [that] the [Phillips] curve should have been stated in real not nominal wages, and Friedman’s recollection is that Bill was persuaded of this’ (p. 178; italics added).

Coupled to the long-run nature of the Bent Hansen model, a ‘complete’ Phillips model of the dynamics of an advanced industrial economy would incorporate a consistent interaction between the short-run dynamics of aggregate fluctuations and long-run growth. This was provided by Reference Goodwin and FeinsteinGoodwin (1967), in his Dobb Festschrift contribution, where the one non-linearity that generated the dynamics between wage shares and unemployment was a real Phillips curve.

Finally, much is made in this chapter of Friedman’s indebtedness to Bill Phillips for the adaptive expectation hypothesis, which underpinned his expectations-augmented Phillips curve in that famous 1967 address. Not the author of Hares, nor Bill Phillips, and not even Friedman seem to have read the famous Muth article of Reference Muth1961 carefully. ^{Footnote 33} Had they done so, they would have realised that the adaptive expectation hypothesis was intrinsic to the cobweb model of the cycle.

This chapter, too, is marred by infelicities. Tinbergen (1952), on p. 160, should be Tinbergen (1951); everything about Samuelson’s Foundations of Economic Analysis – on p. 171 of Hares – is incorrect; ^{Footnote 34} ‘Fabo Schaparelli’ refers, most likely, to Fabio Schiantarelli (Endnote 5 to this chapter, p. 245); Endnote 3 to the final chapter gives the wrong reference (Leeson, 1997, rather than Reference Dorrance and LeesonDorrance and Leeson, 1997) and is incorrect about the claim on Meade (Hares, p. 246):

No such substantiation for the claim can be found in Dorrance and Leeson, for one thing, and what does Bollard mean with ‘destructive arguments’?

Chapters 9, 10 and 11 are descriptions of Phillips’ life and work during the last decade and a half, of times spent in London, Australia (Canberra) and, finally, back in New Zealand (Auckland). The return to the Antipodes, now committed to developing his almost lifelong interest in China, thwarted by an unkindly shortened life, first by illness, then by death, is narrated with compassion and sadness by Dr Bollard.

Incidentally, there is a fine analysis of the formalism of Phillips (1961) in Reference AllenAllen (1967: Chapter 20, especially § 20.2). Allen subjects it to clear local stability analysis.

I must, however, point out that the claims of priority or originality for the work by Bill Phillips on growth cycles and continuous-time estimation are slightly exaggerated by the author of Hares. The chapters, at least 10 and 11, are sprinkled with the familiar infelicities – research in mathematical growth theory, in the US, was being led by Evsey Domar; bracketing Mundell with the other distinguished Professors of Massachusetts Institute of Technology (MIT) is a typical mistake (Hares, p. 188); Endnote 1 to Chapter 10 is puzzling in a book otherwise devoid of mathematical ‘hieroglyphics’ – especially since the reader was forewarned already on p. vi!

Alas, the index is inadequate and the reference list woefully proof-read.

Reflections and ruminations

This pithy characterisation is, in this reviewer’s opinion, a sad reflection of the preoccupations and ideological stances of the ‘profession’.

When Arthur Porter built a differential analyser with Meccano Set parts, for Douglas Hartree, ^{Footnote 35} it ‘proved unexpectedly successful’ and did what the MONIAC did to demystify some of the macroeconomic controversies of the times. Above all, both Phillips and Irving Fisher stressed the importance of transparency and teaching economic principles, (monetary) macroeconomic fundamentals (and utility-based microeconomics).

Unfortunately, Phillips’ noble name will forever be associated with the Phillips Curve. Many will also consider the ostensibly original introductory contributions on proportional, derivative and integral policies for stabilising an unstable macroeconomy, ^{Footnote 36} developed by means of simulation in models that were underpinned by servomechanism theory, as pioneering – as, indeed, they were, to some extent. Still others may make sporadic references to Phillips on growth theory and dynamic continuous-time estimation.

But a unified consideration of the MONIAC and the four macroeconomic classics of Phillips (1950, 1954, 1957, 1958), with some consideration of Phillips (1961, 1962), is the way to pay homage to this humble giant of a theorist. Even within these, it is the construction of the MONIAC that should be considered the jewel in the crown of the contributions by Bill Phillips.

It gave me particular pleasure to read this book by Dr Bollard, ‘warts and all’, simply because – in the end – I believe that is the main message of this book, too.