Money
Since Robertson tended to give much less weight than certain contemporaries to monetary considerations in the determination of cyclical fluctuations, it may be thought surprising that he took on the authorship, after returning from the First World War, of the 1922 Cambridge Economic Handbook on Money. Keynes might have initially intended to do it himself, but, being too busy to do so, may have put pressure on his erstwhile student, junior, and friend to write it in his stead.
Though the book soon became justly famous, this was in some large part because of the lucidity of its presentation and its wit and humour; space allows but one example: ‘The mere assurance of convertibility, it is thought, would have the same kind of soothing effect as the sound of church bells in the distance, and be equally unprovocative of action’ (Robertson 1948: 60). It was also the first occasion when Robertson used Alice quotations at the head of each chapter, henceforth one of his trademarks. (As macroeconomics has gained in mathematical rigour, so it has lost in style.) But, leaving style to one side, it is in some other respects, and with hindsight, slightly odd as a textbook on money. Although the value of money, the inverse of the price level, is emphatically stated to be the outcome of the interaction between the demand and supply of money—a quantity theory approach—there is relatively little formal discussion of the factors determining the demand, or the supply, of money.
The demand for money is taken to be conventionally related to incomes. Although ‘the demand for money, like that of the demand for bread, turns out to be the result of a process of individual weighing up of competing advantages at the margin (ibid.: 36; italics in original), the only example that Robertson gives of a serious shift in velocity, the Marshallian k, arises as a result of extreme inflationary (or deflationary) expectations, as in the German hyperinflation (ibid.: 117-119). For the remainder, Robertson tends to assume that velocity is generally stable, despite acknowledging that money is whatever may be currently acceptable in exchange (ibid.: 3-4). If the latter is true, might not financial innovations, and/or general reactions to monetary mismanagement, make both the definition of money and its velocity unstable? Again, there is remarkably little reference to the role of interest rates in equating the demand and supply of money. Nor is the supply of money treated with great care. The role of the central bank in controlling the supply of money does not appear until quite late in the book, in Chapter VIII on ‘The Question of the Cycle’.
Instead, Robertson focuses on three other questions. The first, which was a recurring theme for him, was whether it would be preferable to aim for a stable price level, a price level that declined slowly in line with productivity, or a gently rising price level, to provide encouragement for entrepreneurs. Whichever might be chosen, he was keen to add that its achievement should not be in the guise of a rigid rule, but a rough norm, or average, allowing some faster inflation in investment-led booms, balanced by some declines in prices either in depressions or when productivity gains were particularly strong. He was also prescient in his criticisms of the ‘real bills’ doctrine, which he termed the ‘Principle of Productive Credit’ (see Robertson 1928b [1940]).
The second main issue, as would be expected in a book first written in the early 1920s, concerns the exchange rate regime to be adopted. The discussion of this, in Chapter VII, is beautifully balanced. Moreover, in line with his quantity theory approach within a closed economy, Robertson develops the view that, in the longer run, the exchange rate between two economies is determined chiefly by relative national monetary growth operating through their respective price levels (see Humphrey 1992).
However, the core and focus of the book is in Chapter V on ‘Money and Saving’. The key quest of Cambridge macroeconomics in general, and of Robertson and Keynes in particular, in the interwar period was to analyse how the ex ante desire to invest and to save, which were driven by different forces and were often far from identical, could be adjusted so that at every point of time ex post investment and saving had to be exactly equal in a closed economy. While (real) interest rates, dependent on productivity and thrift, were a part of the story, Robertson does not put much weight on that equilibrating force, in either A Study of Industrial Fluctuation, Money, or Banking Policy and the Price Level.
Instead, the deus ex machina is forced saving via changes in the price level. Entrepreneurs generally have little wealth of their own, and so turn to banks for loans to provide ‘circulating capital’. This gets spent on labour, raw materials, and so on, and the money stock rises (loans create deposits). With the output available fixed from production the day before (Dt-1), the Robertsonian lag, and money velocity assumed constant (k is fixed), prices have to rise. But the rise in prices reduces real incomes, again with nominal incomes having been earned in the previous day, and thereby forces income recipients to spend less than they had intended, as already described in Section 2. That, however, is far from being the end of the story. The rising prices, especially, but not necessarily, if wage rates remain sticky, encourage all entrepreneurs to raise output in this same period. But what was central was that the adjustment process was via price and monetary changes, involving ‘forced saving’, or the various forms of ‘lacking’ already discussed.
This central plank of Robertson’s theory was extended and embellished in Banking Policy andthe Price Level, especially the key Chapters V and VI. What is remarkable, given what was to come later, was the recorded close involvement of Keynes in that analysis and his apparent commendation of it. In several ways, however, Banking Policy and the Price Level was something of a disaster. Robertson had a habit, shared with Lewis Carroll, of making up words; thus, in Money, he preferred to call bank deposits ‘chequeries’. Moreover, rather than pursue simplicity, he felt obliged to detail all possibilities. The result, of which Banking Policy and the Price Level is the main example, could be an almost impenetrable jungle of words.
Yet, the basic message of the price adjustment/forced saving approach is perfectly simple and intuitive. Moreover, Robertson was writing Money shortly after the 1919-1920 post-war cycle, largely driven by changes in prices of commodity inventories, and had the inflations in Central Europe to observe, so viewing the adjustment process through a flex-price lens was entirely understandable.
What happened between 1926 and 1936 was that Keynes moved on from a perfectly flex-price to a (temporarily at least) fixed-price model, so that adjustment took place primarily via output changes rather than through the price mechanism (see, for example, Skidelsky 2003: 481—482). Why he failed to take Robertson with him, or perhaps failed to try to do so, is not clear. Thus, rather than having initial injections of (investment) expenditures leading businessmen to rewrite price tags continuously, the expenditures (output being given) would result in (unplanned) reductions in inventories of finished goods. Robertson had been aware of the possibility of destocking, but had asserted that stocks of finished goods were relatively small. In practice, he then, post-1936, retreated quite a long way from his forced savings theory. Thus in his ‘A Survey of Modern Monetary Controversy’ (Robertson 1938 [1940]: 106), which was aimed to take some of the heat out of the disputes, he wrote, ‘The increase in expenditure tends to make prices rise, but the use of accumulated stocks, and the expansion of output, tend to prevent them from rising’. Note that in the above brief account, the Robertsonian lag, whereby everything available today for sale is produced in an earlier period, and period-by-period dynamics are maintained, the Keynesian analysis still holds. Nevertheless, Keynesians tended also to assume that output could be varied within the immediate short period to meet changing demand and, even in some extreme cases, that the full multiplier process could be completed within the immediate short period. Robertson had some fun at their expense, and reasonably enough preferred his own methodology, but it did not disguise the fact that in this key field of analysis, Robertson had lost the battle, and he knew it.
Robertson never again tried to develop a holistic account of macroeconomic adjustment, once ‘forced savings’ had been largely shot down. Instead, he tended to snipe at the errors and excesses of the Keynesian revolution. There were several of these targets for Robertson. In particular, the fact that most goods and services exhibit sticky prices in the short run does not mean that businessmen will not adjust prices in the medium and longer term when trends in input prices, competitive output prices, and demand for their own product became more clearly apparent. But Keynesians initially extended the stickiness of prices to continue indefinitely over time, until full employment was reached, a reverse-L-shaped supply curve. Moreover, there was a tendency from the late 1930s to the 1950s to lower the assumed level[1] of such full employment from perhaps 5% of the workforce unemployed in the 1930s to 1%, or even lower, in the 1940s and 1950s. Inflation occurring above this level of unemployment was then largely attributed to cost-push pressures, and the socially ‘correct’ remedy was incomes policies.
All this was anathema to Robertson. The heterogeneity of the real economy, bottlenecks, transitional costs, and so on, meant that there would be no kink at full employment. In a debate at the 1959 Conference of the International Economic Association on Inflation, Robertson referred to Phillips’s (1958) conclusion that a rate of unemployment of 2.5% should be able to preserve price level stability as ‘very favourable to the optimists’ (Robertson quoted in Hague 1962: 456). In Robertson’s opinion, the original element in Phillips’s article was not the negative correlation between the rate of growth in money wages and the rate of unemployment, but rather its contention that the relationship between these two variables was stable. Robertson noted that ‘one could not put much reliance on the results because it assumed there was a fixed psychological function relating the attitude of trade unions to the level of employment over a whole century’ (ibid.). Interestingly enough, as Blyth (1975: 306) notes, Robertson (while at the LSE) played a role in influencing Phillips to add the determination of prices and wages to the traditional Keynesian model of the early 1950s. Even more important, Robertson saw no proper empirical, or theoretical, basis for the progressive reductions in estimates of full employment, and believed that cost-push pressures only occurred because of the excessive level of demand, and that incomes policies could not work in such conditions. He had the courage to argue (in the Cohen Council 1958) that the equilibrium (NAIRU) level of unemployment was somewhat over 2%, not only higher than the then estimates of most Keynesians, but also higher than previous outcomes. For such counterrevolutionary views, he was excoriated by the bien pensant socialists of the day, including many in the Cambridge Faculty of Economics. Currently, however, when central banks worry about whether the NAIRU might be anywhere between, say, 4% and 6%, and any suggestion to adopt an incomes policy would be laughed out of court, it is patently obvious that Robertson won this larger battle. Mainstream macro is now Robertsonian, rather than Keynesian.
Robertson believed in the relative stability of the velocity of money, and was a sceptic about the stability of the consumption function. In what was known as the ‘battle of the airwaves’ between Friedman and Meiselman (FM) and Ando and Modigliani (AM), he would have been one of the few British economists to take the side of FM. Indeed, he adopted at one time, or another, almost all the analytical parts that lay behind Friedman’s monetarist counter-attack on the Keynesian position. Yet he never quite put it all together in that way, perhaps because he saw himself as Blondin, the tightrope walker, keeping a balance between different schools of thought, all of which had merit (Robertson 1938 [1940]).
A possible contributory factor might have been a slight hesitance, or uncertainty, about the determinants of the supply of money. Robertson is crystal clear that loans create deposits, and in some cases, he leaves it at that, for example, in his address on ‘Is There a Future for Banking?’ (Robertson 1952: Chapter 14). But then the question arises as to what limits the expansion of bank loans? In a slump, it often is the demand for such loans; analogies arise about taking horses to water. However, in a boom, Robertson sees the limiting factor as the availability of cash reserves, and he adopts, though mostly in words, the money multiplier analysis, which was becoming standard. At the same time, Robertson is aware of a problem: The banks can always get more cash by (forcing the discount market into) borrowing from the Bank of England, admittedly at a penalty rate, but one that is not much higher than the policy rate. Interest rates will then rise, but not by much unless the Bank is prepared to raise the whole structure of (short) rates significantly.[2] This problem became even worse after the Second World War when British banks were stuffed full of saleable short-term government debt, with the government being a largely interest-insensitive borrower.
Robertson was one of the few to recognise some of the flaws in monetary management by means of ratio controls. Thus:
If a proportion fixed by custom is arbitrary and misleading, a proportion fixed by law seems at first sight to be positively mischievous. An iron ration which you must not touch even in the throes of starvation is something of a mockery. Against such criticism it may be urged (though not too loudly) that in finance as in war rules are made to be broken on occasion, and that their object is not to ensure that certain things shall never be done, but that they shall not be done without good reason (Robertson 1948: 57).
And, ‘...it resembles the procedure of a certain municipality which tried to guard against a shortage of cabs by ordaining that there should always be at least one cab on the ranks’ (ibid.: 62). One of the authors of this paper has reused this analogy on numerous occasions, without having been fully aware of its provenance with Robertson.
Interest rate fluctuations play a relatively minor role in A Study of Industrial Fluctuation, Money, and Banking Policy and the Price Level; indeed, remarkably so in the latter two books. Whereas productivity and thrift play a fundamental role in determining the quasi-equilibrium longer-term real interest rate, in practice, short-term factors, such as gold flows, changes in bankers’ confidence, and shifting expectations of future inflation, could easily drive actual market rates away from the equilibrium level.
It is, therefore, somewhat ironic that Robertson’s best-known clash with Keynes was over interest rate determination, the loanable funds versus liquidity preference debate. As in the case of price stickiness, it was triggered by Keynes in his attempt to differentiate his product from what had gone beforehand, even though he himself was arguably going too far. First, Keynes argued that saving was, almost entirely, a function of income, and hardly, if at all, of interest rates, and that investment was (empirically) rather insensitive to interest rates. As a result, interest rates could not be determined by forces causing ex post I to equal ex post S at full equilibrium (or in other words, the Hicksian IS curve was very inelastic).
Second, Robertson, supported initially by Keynes, had argued that the demand and supply of money were equilibrated by changes in the general level of prices, the quantity theory of money. But now Keynes had shifted towards claiming price (or money-wage unit) stickiness so long as employment was below full employment. If so, then what would equate the demand and supply of money? Keynes’s answer, of course, became the liquidity preference theory. The extreme version of the liquidity preference theory could not, and did not, last, partly under fire from Robertson. Keynes was obliged to accept a ‘finance’ motive for holding money, and Hicks, with his IS-LM analysis, brought about a partial reconciliation.
Robertson was not fully reconciled with the IS-LM analysis, even though it did bring back productivity and thrift into the analysis of the determination of interest rates, from which Keynes had once hoped to banish it. Even so, it made the adjustment of the economic system to disturbances to supply-side monetary shocks occurring primarily via interest rates, whereas Robertson would argue that the money/bond margin was but one margin out of many; so supply-side monetary shocks could impinge the economy directly and not just via interest rates (see, for example, his paper on ‘What Has Happened to the Rate of Interest?’) (Robertson 1949 [1952]).
Another, more policy driven, concern of Robertson’s was that the Keynesian analytical apparatus was used, primarily after the Second World War, to justify using fiscal policy to maintain full employment, while holding interest rates down as low as possible (Dalton) in order to encourage investment, to lower the debt burden (on government), and, perhaps, to push the rentier into euthanasia. In part because of Robertson’s belief in the quantity theory, he believed that this line of policy would become inflationary, distortionary, and ineffective. His later, post-Second World War writings are full of pleas both for a higher general level of interest rates and for a much more aggressive counter-cyclical variation in them (see Dennison and Presley 1992).
To conclude this section and sum up, Robertson’s central analytical concept of forced saving following flex-price inflation was discredited when Keynes switched from a flex-price to a fixed-price adjustment mechanism. But wait long enough into the medium term and almost all prices regain a flex-price character. In this latter context, the balanced wisdom of Robertson/Blondin has eventually, albeit post-mortem, triumphed over the initial excesses of the Keynesians. Modern macro is primarily Robertsonian (rather than Keynesian), except that the IS-LM (now three equation) model still has money supply shocks channelled primarily via the interest rate, whereas Robertson would have preferred to revert to a more general quantity theory mechanism.
- [1] Skidelsky (2003: 711—712) has Beveridge giving a value of 8% to this and Keynes regarding 5% as‘normal’.
- [2] From Money (1948: 162—163): ‘But on these difficult matters the reader may well be excused from seeking further enlightenment in this little book, and the author from trying to impart it. It is enough torealise that by the use of their double weapon, Central Banks can up to a point check the expansion ofthe money-supply [sic]; and that while we cannot be sure that their power to do so is in all circumstancescomplete, there is good reason to believe that if it were used earlier and more resolutely than it has sometimes been in the past, many of the evil excrescences of a trade boom could be lopped away’.
