Knowledge that Transforms

Cover title

Robert Triffin, Herbert Grubel, The Adjustment Mechanism to Differential Rates of Monetary Expansion among the Countries of the European Economic Community, The Review of Economics and Statistics, Vol. 44, No. 4 (Nov., 1962), pp. 486-491

This paper considers some of the shortcomings of an income per I capita measure of and suggests an alternative capitalized value of expected future income per capita. Part I considers briefly some of the variables that a generally acceptable empirical welfare function should include and defines the capital value measure; Part II compares this measure of with the commonly used measure, income per capita; Part III presents and analyzes the results of computations of economic well-being for a number of United States cities, by both measures; Part IV considers some implications of a present value measure for interregional migration.

As technological developments have altered production techniques, types of mechanical equipment, and varieties of outputs, society has begun to recognize that economic progress involves not only changes in machinery but also in men – not only expenditures on equipment but also on people. Investment in people makes it possible to take advantage of technical progress as well as to continue that progress. Improvements in health make investment in education more rewarding by extending life expectancy. Investment in education expands and extends knowledge, leading to advances which raise productivity and improve health. With investment in human capital and non-human capital both contributing to economic growth and welfare and in what is probably an interdependent manner, more attention should be paid to the adequacy of the level of expenditures on people.

To illustrate the alternative implications for banks' capital adequacy of sales of short- and long-term securities, Luckett employs a formula based on the “Form for Analyzing Bank Capital.”22 Ibid., p. 55. However, implicit in the illustration is the assumption that the bank in question has only a bare sufficiency of capital for its asset portfolio before the loan, and the treatment seems to ignore the fact that the capital funds of banks are increasing over time. Insured commercial banks have increased their capital substantially since December, 1952. Between December, 1952, and each of five other dates arbitrarily selected as turning points from the chart showing the ratio of seasonally adjusted loans to demand deposits for all commercial banks,33 Ibid., p. 57. The turning points are December, 1953; December, 1955; December, 1957; June, 1959; and December, 1960. banks' capital accounts have increased by more than the 16 1 2 per cent of increased loan volume required by an expansion of loans achieved by a reduction of cash holdings.44 For example, between December, 1952, and December, 1960, banks increased capital accounts by $8.75 billion and valuation reserves on loans by $1.45 billion (a total of $10.2 billion). Loans were increased by $55.15 billion. If the increase in loans had been made by a reduction of cash holdings, the maintenance of capital adequacy would have required additional capital of $9.1 billion (0.165 × $55.15 billion). Had securities been sold, the required increase in capital would have been less. Valuation reserves on loans are included in the computation of capital for measuring adequacy. The analysis assumes that the capital requirements of fixed assets, etc., have not changed significantly. Data were taken from the Annual Reports of the FDIC, 1952, 1953, 1957, 1959, and 1960. These data do not, of course, prove that Luckett's reasoning is not valid for an individual bank. It should also be noted that the capital loss incurred by the sale of a security impairs the capital position as well as reducing the capital requirement. The difference in the capital requirement between short- and long-term securities is 3 1 2 per cent. If the sale of the longer security results in a capital loss of more than 3 1 2 per cent in excess of that which would have been incurred on the sale of a shortterm security, no net capital saving will have been achieved—unless the liquidity premium of 6 1 2 per cent is relevant because of low primary and secondary reserves. The theory that banks will be concerned less with capital losses than with adequate liquidity, while undoubtedly true in ordinary situations, is based on a misinterpretation of Warren Smith's analysis.55 Warren G. Smith, “On the Effectiveness of Monetary Policy,” American Economic Review, XLVI (September, 1956), 590–91. The suggestion that a capital loss can be made up by only a “nominally wider spread” in yields is based on a formula dealing with a switch from Treasury bonds into corporate securities of the same maturity.66 Luckett, op. cit., n. 9, p. 57, and Smith, op. cit., nn. 8 and 10, pp. 591–92. The yield differential between a bond and a shortterm loan, needed to offset any capital loss incurred, will be considerably greater than Smith's figures show. Furthermore, the costs of administering a loan exceed those of investment portfolio supervision, so that any differential in gross yield should be enhanced by any cost difference, as well as by any requisite risk premium.77 The “Functional Cost Analysis” carried out by the Federal Reserve Bank of Boston shows this to be true for the average, though not necessarily for the marginal, cost—the latter being the relevant cost in this computation. Another reason suggested for a phlegmatic attitude toward deliberately incurred capital losses is that “the tax-option on capital gains and losses enjoyed peculiarly by commercial banks would tend to undercut the desire to avoid taking capital losses.” Permission to treat securities as stock-in-trade for tax purposes does induce banks to engage in tax-swaps which require banks deliberately to incur capital losses. However, the two-way option need have little or no influence on the choice of the security sold to finance an increase in loan volume. A bank can incur a loss by means of a tax-swap and can immediately replace the bond sold with a non-similar security. “Non-similar” would seem to offer a very wide range of permissible substitutes, so that no significant alteration of the portfolio is necessary.88 See Sam B. Chase, Jr., “The Lock-in Effect: Bank Reactions to Security Losses,” Monthly Review, Federal Reserve Bank of Kansas City, June, 1960, pp. 9–16. Reprinted in L. S. Ritter, Money and Economic Activity (2d ed.; Boston, 1961), pp. 171–77, n. 1, p. 10, for an example of the latitude of interpretation of the ruling “not identicalor similar.” The capital loss can therefore be incurred without any necessity to shorten the average maturity of the investment portfolio: short-term holdings can be reduced to increase loan volume at the same time as the capital loss is registered through a tax-swap. The shortening of the security portfolios of commercial banks which has taken place during the last ten years may well have been caused by factors other than the need to offset the illiquidity inherent in greater loan volume. Any securities sold to increase loan volume must be a part of the residual loanable funds, i.e., those funds which a bank, with its given liability structure, would be happy to lend if sufficient credit-worthy loan demand appeared. Prior to the 1950's, commercial banks had experienced a protracted period of monetary ease during which they were seldom, if ever, forced to sell securities at a loss in order to expand loan volume. Additionally, yields were very low, so that the income produced by any residual loanable funds was of considerable importance. It is therefore reasonable to suppose that the reintroduction of active monetary policy found the commercial banks with bond portfolios less than perfectly suited to the new conditions. The reintroduction of considerable flexibility in bond prices, coupled with a secular increase in yields, may well have shortened the optimum maturity of the portfolio. Those securities which constitute the residual of loanable funds should fulfil no particular liquidity service99 Roger A. Lyon, Investment Portfolio Management in the Commercial Bank (New Brunswick, N.J., 1960), p. 139. and are invested to secure an adequate return during periods of insufficient loan demand. The selection of the type of security that will offer the highest rate of return on these funds obviously involves many estimates of future behavior—the timing of the eventual sale, future bond prices, changes in the shape of the yield curve, etc.1010 The explanation is carried out in terms of a single security for the purpose of simplicity. The portfolio manager is concerned with the real yield of the security after taxes; this fact in itself is likely to shorten the optimum mean maturity of the portfolio. A further minor question, if Luckett's theory is not altogether correct, is why banks should have chosen to unload their long-term securities mainly when expanding their loan volume rather than shifting into shorter securities throughout the whole period. One explanation of the cyclical quality of the unloading is that, although commercial banks have considerable scope to disguise the actual level of published profits through the use of valuation reserve accounts and fast write-down of security book values, banks always wish to show quite adequate profit figures to their stockholders and to the public. It therefore seems rational to take the capital losses forced on them by the secular increase in yields on those securities which they do not propose to hold to maturity, at times of relatively high net current earnings. High net current earnings per unit of assets and per units of capital are achieved during periods of high interest rates, tight money, and growing loan demand. The author suggests that this behavior on the part of commercial banks may have important policy consequences. It is not certain that the banks have not sold their longer-term securities to holders of idle cash balances and thereby increased velocity directly. Certainly, life insurance companies and mutual savings banks do not seem to have acquired the securities sold by commercial banks, since the total government security holdings of these two types of institutions have been decreasing secularly since the early 1950's. Savings and loan associations have increased their total holdings of government securities secularly, but no definite cyclical pattern of maturities is apparent.1111 U.S. Savings and Loan League, Savings and Loan Fact Book (Washington, 1961), pp. 86 and 93. In view of the high yield—certainly compared with those of the fairly recent past—at which these bonds were sold, the possibility of finding holders in the public and non-financial sectors does not seem too remote. Surely, however, the crucial point is not whether dumping leads directly or indirectly to higher velocity but rather what will happen when banks have succeeded in shortening their portfolios to some sort of satisfactory mean maturity. Luckett suggests that the absence of longer-term securities will impede rises in velocity caused by portfolio switches because banks will come up against a ceiling because of their liquidity loss. Portfolio switches will be on a smaller scale, and therefore the rise in velocity will be reduced unless other institutions take over the intermediating functions now performed by commercial banks. But what matters in an increase in velocity is the volume of unused loanable funds, not the maturity of the security in which these funds happen to repose. The normal reserves should be sufficient to preserve liquidity, and the cyclical shifts in loan demand will always enable banks to increase velocity through portfolio switches, especially if banks are deliberately made more liquid during recessions as a stimulatory move by the Federal Reserve System. What Professor Luckett has demonstrated is yet another weakness of the locking-in effect as a deterrent to velocity-increasing portfolio switches.

Gray first takes me to task for having misapplied Warren Smith's analysis dealing with the wider spread needed to recoup capital losses when a bank switches securities at depressed prices.33 Warren L. M. Smith, “On the Effectiveness of Monetary Policy,” American Economic Review, XLVI (September, 1956),590–93. Gray correctly observes that, in applying Smith's analysis to the spread between Treasury bonds and customer loans, I have implicitly carried over Smith's assumptions that the two securities are of identical maturity. While this assumption is unobjectionable in the Smith context, it is clearly out of place in my own. Gray's point is well taken. Nevertheless, I wonder if the error is quite so glaring as Gray seems to feel. In the first place, the increased loan demand is likely to be considerably more permanent than any particular loan. Thus, while the bank may be making only a 3-month loan, it will generally expect to be able to relend the funds to another customer when the 3 months are up. Moreover, Silverberg's analysis indicates that the typical response of banks to an increased loan demand is to sell off their intermediate-term bonds (1–5 years) and allow their longer-term bonds (over 5 years) to become intermediate-term through the passage-of-time effect. Given these two conditions, the difference in maturity may not be so important as would at first appear.44 Gray takes a step in his analysis at this point which I find puzzling. On p. 647 he says: “Furthermore, the costs of administering a loan exceed those of investment supervision, so that any differential in gross yield should be enhanced by any cost difference, as well as by any requisite risk premium.” But, since the raison d'être for the spread is risk and administrative costs, I fail to see why the differential should increase when a bank increases its loans. As an alternate explanation of the shortening of the security portfolios of commercial banks, Gray suggests that this may have been a secular adjustment occasioned by the reintroduction of flexible monetary policy in 1951. If I understand him correctly, Gray would argue that the shortening would have occurred even in the absence of an expanded loan demand because the “reintroduction of considerable flexibility in bond prices, coupled with a secular increase in yields may well have shortened the optimum maturity of the portfolio” (p. 648). Gray himself raises the major objection to this interpretation of the data: Why should a secular adjustment occur in a cyclical fashion? His explanation is that banks took their capital losses in times of high net current earnings in order to be able to “show quite adequate profit figures to their stockholders and to the public.” While I cannot, of course, disprove this theory, it does seem to me to be stretching things to assert that banks have deliberately engaged in maximizing their capital losses for purposes of window dressing. Moreover, even if one were to accept this explanation, it still does not account for the cyclical upswings in the long-term bond holdings of banks; we are still left with the question of why, if banks were engaged in the strictly secular action of reducing their portfolio of long-term governments, they should have acquired nearly $15 billion worth in 1953 and over $7 billion worth in 1958. In my paper I left it an open question whether the banking system had disposed of its long-term governments through open-market sales or through the passage-of-time effect.77 Op. cit., p. 58, n. 12. Silverberg's comment closes this gap and, in so doing, makes a significant contribution to the discussion.