The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
CONFIDENTIAL AMERICAN PROSPERITY AND THE BRITISH BALANCE-OFfPAYMENTS PROBLEM Randall Hinshaw Division of Research and Statisticsf Baard of Governors of the Federal Reserve System. September$ 1944 .FOREWORD I t is widely recognized that Britain, at the end of the war, will in a l l likelihood be confronted with a serious balance-of-payments problem. Loss of foreign-investment income and the emergence of an immense external debt have resulted in a situation in which, at the pre-war volume of exports and imports, Britain would be faced with an annual balance*-of~payments deficit that might easily amount to 400 million pounds. Unless Britain is able to meet this deficit by borrowing from abroad, or unless the external sterling debt is drastically scaled down, Britain will be forced either to undergo a marked reduction in imports or to achieve a very substantial increase in exports. The latter alternative will for Britain be much the less painful, particularly if there is a thriving demand for British wares• In the absence of such a demand, Britons may feel compelled to resort to bilateral trading arrangements and to demand from British sources of supply preferential treatment for British exports. To what extent can the United States assist in the solution of this British problem by maintaining a high level of domestic prosperity (and thus of imports)? The ensuing report is an attempt to answer this question. Estimates are based on somewhat elaborate s t a t i s t i c a l analysis, and are made on the assumptions (l) that the United States does not raise tariffs above present levels and (2) that international trade is conducted on a multilateral basis. On these assumptions, i t is estimated that at a peace-time national income of $140,000,000,000 per year (somewhat below the present level), and at a pricG level for American imports 30 per cent greater than the 1937 level, American imports from Britain would amount to $493,000,000 per year, and imports from gritain attributable to American imports from third countries would amount annually to $916,000,000* Adding these two figures, we obtain a sum of $1,409,000,000, representing the total estimated imports from Britain directly and indirectly attributable to the American economy at this level of national income (and of import prices). Since the corresponding e s t i mate for 1937 (at 1937 import prices) is $567,000,000, the estimated increase in British exports resulting from the increase in American national income (and import prices) over the 1937 level amounts to $842,000,000 (&2Q9,OOO,OOO)--a very substantial contribution to the narrowing of the British bal&nce-of-payments deficit. Estimates for other levels of national income and of import prices are available in Tables I-V. AMERICAN PROSPERITY AND THE BRITISH BALANCE-OF-PAYMENTS PROBLEM1 Randall Hinshaw, Division of Research and Statistics, Board of Governors of the Federal Reserve System* It is generally agreed that by maintaining a high level of national income (and thus a high level of imports), the United States can diminish the grave balance~of~payments difficulties that are expected to confront Britain at the end of the war* No attempt, however, appears to have been made to estimate the extent to which the United States can aid Britain in such fashion* This is hardly surprising, since even the most careful guess would have to rest, in some degree, on dubious assumptions and precarious extrapolations* Nevertheless, the problem is both important and intriguing, and in the ensuing discussion an effort will be made to estiirate the extent to which the level of British exports would be affected by a level of American income after the war well above the level during any pre-war year. To make such a guessf it will not, of course, be sufficient to estimate merely the increase in United States imports from Groat Britain* This is only part of the problem* Of possibly equal (or even greater) importance, so far as the level of British exports is concerned, are the reverberations of an increase in American national income via third countries* The total effect on British exports of an increase in American income would comprise (\) the increase in American imports from Britain plus (2) any increase in imports from Britain by the re6t of the world that was attributable tc the increase in total American imports. Effect of Changes in National Income on American Imports The starting point of this study will bo an investigation of the manner in which, in the past, American imports have fluctuated with national income. The period covered comprises tho years 1922 through 1937* Chart I is ex scatter-di&gram in whiph annual figures for the two variables are correlated* Tho national-income figures used are the annual estimates prepared by the Department of Commerce; the import figures are from the official series referred to as Unitod States The writer is indebted to Messrs* J* H* Adlert Howard S* Ellis, Walter Gardner, Alexander Gerschenkronf and Lloyd Metzler for cooperation and advice, *The British balance-of-payments problem is briefly analyzed in an appendix to this report*/ "imports entered for consumption/1 In order to adjust the value of imports to reflect more accurately the cost of imports to American consumers, customs duties (which during this period varied from between 13,2 and 19#8 per cent of total imports) were added to the import figuresf* As would be expected, there is a positive correlation between imports end national income. Moreover, the correlation is high, the linear correlation coefficient being +t95. Indeed, the correlation is so high that one is tempted to conclude that national income is the only important variable affecting the level of imports and to assume that the level of imports at high levels of income can be reliably estimated by extrapolating the regression line. This, however, would imply that changes in the level of import prices have little effect on the level of imports, and can be disregarded in making predictions.^ Actually, certain irregularities in Chart I suggest that national income is not the only important variable affecting the value of imports. By connecting, in chronological order, the points on the diagram, it becomes clear that the value of imports was high in relation to national income during the period 1922-27 and low in relation to national income during the period 1934-37, Is it possible that this pattern was in some way related to the level of import prices? To answer this question, an index of import prices was correlated with national income, and the points in the scatter^diagram were connected in chronological order,4 The result, shown in Chart II, is most striking. Import prices, it will be seen, were high in relation to national income during tbse years (1922-27) when tb value of imports was high in relation to national income, and were low relative to national income when the reverse was true (1934-37)• This at once suggests that the price elasticity of the American demand for imports is less than unity, since, at a given level of incomef the value of imports is high when import prices are high and is low when import prices are low, In Chart III, the "quantity" of imports (i.e., the value of imports at 1937 prices) is correlated with national income.& As would be expected, "'"Figures for both imports end customs duties wero taken from the Statistical Abstract of the United States, 1941, p» 527. ^ h e regression line in Charts I, Xl$ III, and V was fitted to the data, by tho method of least squares f *Only if the price elasticity of demand for imports were unity would the value of imports be unaffected by changes in the level of import prices. 4 The import price index her® used i§ that found in the Statistical Abstract of the United States, }94}, p* 5$2 (figures for the ye^rs preceding~"T930 are taken iVom the Department of Commerce publication, Foreign Trade of the United states, 1 W 4 , pf 38)* For our purpose, an adjustment in the index was made to a^low for changes in customs duties. This was accomplished by multiplying the index by the quotient of value of imports plus customs duties, divided by value of imports* The base of the index was then changed to the year 1937* Figures for quantity of imports were derived by multiplying, for each year, an index of quantity (the base of which had been changed to 1937) by the value of imports tor 1937 • Tfte import quantity index was obtained from the same sources (same p&gep) &s the import price index* the pattern is the reverse of that in Chart II; when import prices are high in relation to national income (as during most of the twenties), the quantity of imports is low relative to national income, and the reverse is true when import prices are low relative to income (as during the middle thirties). Effect of Changes in Import Prices on American Imports In order to obtain a clearer idea of the influence of price on imports, it will be helpful to abstract, statistically, from the influence of national income. One way of accomplishing this is the method of multiple correlation. We shall employ this method shortly, but before doing so will resort to another method that may have already suggested itself to the reader. If quantity of import9 and import prices are plotted on the same diagram, it will be found that a positive, rather than a negative, correlation between the two variables resultsf This is due to the preponderant influence of changes in national income. Both quantity of imports and import prices tend to be high in periods whon national income is high and low when national income is low* If, however, national income were to remain unchanged, one would expect a negative correlation between quantity of imports end the level of import prices. One way of testing this presumption is to correlate the deviations from the regression line in Chart II with the deviations from the regression line in Chart III. This is done in Chart IV. The result, it will be noted, is a striking negative correlation^ revealing unmistakably that, besides being sensitive to changes in* national income, the quantity of imports is also sensitive to changes in price. Is it possible, from the deviation figures in Chart IV, to construct a demand ourve for African ImportsT Such an attempt is made in Chart V, in which absolute figures are substituted for the deviation values. The major problem in constructing this curve was to derive two figures—one for quantity and one for price—-to which to add (algebraically) the deviations plotted in Chart IV. These figures were obtained as follows: In the original data, the year 1937 was used as a base for both quantity and price; the quantity of imports was defined as the value of imports at 1937 prices, and the price index number for 1937 was set at 100, Thus in Chart V, the price and quantity figures for 1937 are simply the original figures (3,01 billion dollars for volyjne and 100 for price). By subtracting (algebraically) the 1937 volume deviation from the 1937 volume figure and the 1937 price deviation from the 1937 price figure, we obtain two figures (2#55 billion dollars and the index number 137.6), to the first of which can be added the quantity deviations and to the second the price deviations^ Plotting the resulting sums, we obtain the curve found in Chart K& This curve may be interpreted as a demand curve for American imports on the assumption that national income is at the 1937 level. The negative correlation between price and quantity is high (-#92), and the mean elasticity of the curve within the range of observations is well under unity (~57) A more elegant, and in some respects more reliable, method of ascertaining the manner in which the quantity of imports reacts to changes in national income end in the import price level is the method of multiple correlation. By this method can also be determined the degree to which changes in the volume of imports are explained by changes in national income and in import prices* That changes in the latter two variables account very largely for changes in the quantity of imports is revealed by the high "coefficient of multiple correlation" obtained for the three series ever the period covered* Variations in the volumo of imports would be completely accounted for by changes in national income end in import prices if this coefficient wore l c 0| actually, tho coefficient is #97. The degree to which chsngee in the quantity of imports are explained (l) by changes in the import prico level when national income remains unchangod or (2} by changes in national income when import prices remain unchanged can be expressed statistically by "coefficients of partial correlation*" These coefficients, for the period covered in this study, are both very High* The partial coefficient showing ths correlation between quantity of imports and nation©! income when tho import price level is held constant is +#97, and the coefficient showing the correlation between quantity of imports and import prices v/hen national income is held constant is <~*92. As would be expected, the latter correlation is negative, indicating that, at a given national income, the quantity of imports varies inversely with price. Estimate jof African Imports at High Levels of National Income In multiple correlation, an equation is obtained in which, by inserting any given values for the independent variables (in this case national income and the import price level), the value of the dependent variable (in this case quantity of imports) can bo estimated,* If the coefficients of multiple and partial correlation are high, as they are in this case, one can bo confident that the actual value of the dependent variable will not be far from the estimated value, provided th© values for the independept variables are within the range of the original data. When the values of tho independent variables are outside the range of observed cases, one cannot, of course, rest assured that the estimated value of the dependent variable will approximate the actual value, but even in this situation the estimating equation may provide the basis for a good guess* the and the the The ensuing estimates rest on this presumption^ It is assumed that war has not substantially altered tho relationship of national income import prices to the quantity of imports, and it is also assumed that same relationship holds true for levels of national income well above level of any pre-war year. The equation obtained in this case may bo expressed as follows; X^ * ,8759 + f0408X2 - cOO9OXg, whore Xj represents volume of imports, Xg national income, and X 3 the level of import prices. Employing the estimating equation obtained from multiple correla* tionf we can construct a table showing the value of imports to be expected at various levels of national indome and at various levels of import prices. This is done in Table l^ TABLE I ESTIMATED UNITED STATES IMPORTS (Billions of Dollars) National Income 70 90 i$e 110 120 130 140 150 of I m ]s o r t (1937 • 100) t • • Le V 0 1 i _i00 110 120 130 2ce 3.0 3.5 3,9 4.4 4.8 5.3 5.7 6f2 6,6 3.2 3.7 4.2 4,6 3,3 3,9 4.4 4.9 5.4 6.0 6.5 7,0 7.6 3,2 3.6 4.1 4,5 4.9 5.3 5.7 6.1 5.1 5.6 6,1 6T6 7.1 P r i C 6 6 140 150 3.5 3.6 4.0 4,6 5,2 4.2 5.7 6.3 6.9 7.5 8,0 4.8 5.4 6.0 6.6 7.2 7.9 8.5 Since the price elasticity of the American demand for imports is less than unity, the value of imports rises, not only with national income, but with the level of import prices. As tho above table shows, the extent of the increase in value of imports as the other two variables rise is rather spectacular. At a national income of 70 billion dollars (slightly below the 1937 level of 71,5 billions), end at the 1937 level of import prices, annual imports amount to 2.8 billion dollars. At a national income of 120 billions, end at a level of import prices 30 per cent higher than that of 1937, the estimated level of imports is exactly double the 3,937 level of 3.0 billion dollars. Effect o£ High American National Income on British Exports HOY/ shall we estimate the post-war level of American imports from Britain? The simplest and perhaps the best procedure is to apply to the estimates in Table I the pre-war ratio of American imports from Britain to total American importsf Actually, this ratio varied considerably during the inter-war period^ During the years 1923-29, imports from Britain amounted to 9,1 per cent of total American imports; during the years 1933-^38, to 7*1 per centf Annual figures for the period 1923^38 ranged from as high as 10f8 per cent (1923) to as low as 5*6 per cent (1932)» The average for the £ixt<jen years was 8f2 per cent, which coincides with tho figure for 1936* Table IJ contains estimates, at various levels of income and of import prices, of American imports from Britain on the conservative assumption that such imports amount to 7 per cent of total American imports as estimated in Table I. TABU! II ESTIMATED UNITED STATES IMPORTS FROM BRITAIN (Millions of Dollars) National Income 70 80 90 100 110 120 130 140 150 s : ; : L e v e l of I m p o r t (1937 - 100) P r i c e s 100 110 120 JL30 140 _150 198 227 255 284 312 341 369 398 426 211 242 274 305 336 368 399 431 462 223 257 291 325 360 394 428 462 233 270 307 344 381 418 456 493 530 242 282 322 362 402 442 482 522 562 250 293 336 378 497 421 464 507 549 592 a Billione of Dollars The figures in Table II represent an estimate of what we may call the direot effect of an increase in American national income on British exports * As already pointed out, there is also an indirect effectr-i»e f f the effect of an increase in American imports on British exports via third countries* An increase in American imports provides the rest of the world with exchange resources, a part of which may be expected to be converted into sterling and used to purchase British exports« While the extent of this indirect effect is a matter of conjecture, i t would appear reasonable to assume that the fraction of en addition to the rest of the world* s exchange resources that would be spent on British exports would approximate the pre-existing fraction of the rest of the world1s imports that come from Britain, The lattor fraction can bo readily estimated from figures provided by the League of Nations in i t s publication, The Network of World Trade• Figures are available for the years 1928, 1935, and 1938f The numerator of the fraction is simply British exports minus exports to the United States, end the denominator is world imports minus total American and British imports. The fraction appears to be rather stable; expressed as a percentage, i t is 14 per cent for each of the three years. By "rest of the world" is here meant the world minus the United States and Britain* Table III contains estimates, for various levels of American income and of import prices, of the amount of imports purchased from Britain by countries other than the United States as a result of American purchases from these countries, The estimates are obtained by taking 14 per cent of estimated total American imports other than from Britain (i«e*, 14 per cent of the difference between the figures in Table I and the figures in Table II), TABLE III ESTIMATED IMPORTS FROM BRITAIN ATTRIBUTABLE TO UNITED STATES IMPORTS FROM THIRD COUNTRIES (Millions of Dollars) Lev e 1 National Inoom© 70 80 90 XOO 110 120 150 140 15GI a of I inp o r t (1937 •> 100) P r ic e s • 110 120 130 140 150 368 393 421 475 451 433 502 571 640 709 778 847 916 450 525 599 673 747 822 896 970 465 545 624 985 100 528 581 634 687 740 509 568 626 684 743 801 414 478 542 605 669 732 796 860 793 859 923 704 783 863 943 1,022 1,045 1,102 Billions of dollars* In comparing Table III with Table I I , i t is interesting to observe that the estimated indirect effect on British exports of an increase in American national incoiEe is much greater than the estimated direct effect* By adding the two together, we secure an estimate of the total amount of imports from Britain that can be traced, directly or indirectly, to the American economy* Sums of the two sets of figures (the figures of Table II plus the figures of Table III) cure shown in Table IV, Table V shows the estimated total increase in imports from Britain resulting from an increase in American national income above the level of 70 billion dollar* (approximately the level of 1937). In order to present the matter from the British point of view, the figures arc converted into pounds sterling at the current official exchange rate of 4*035 dollars per pound• The figures in Table V strongly support the British contention that the United States can contribute greatly to the solution of the TABLE IV ESTIMATED IMPORTS FROM BRITAIN DIRECTLY AND INDIRECTLY ATTRIBUTABLE TO THE AMERICAN ECONOMY (Millions of Dollars) L e v e l National Income 70 80 90 100 110 120 130 140 150 of Import (1937 - 100) P r i c os 100 110 120 130 140 150 567 648 730 811 893 975 604 693 783 873 962 637 735 833 930 666 772 878 985 692 807 921 715 837 960 1,028 1,126 1,224 1,322 1,420 1,091 1,197 1,303 1,409 1,515 1,035 1,149 1,264 1,378 1,492 1,606 1,082 1,204 1,327 1,449 1,572 1-694 1 ,056 1 ,138 1 ,219 1,052 1,142 1,232 1,321 'Billions of dollars. TABLE V ESTIMATED INCREASE IN IMPORTS FROM BRITAIN RESULTING FROM INCREASE IN AMERICAN NATIONAL INCOME ABOVE 1937 LEVEL (Millions of Pounds) i : National Income : •• 70 80 90 100 110 120 130 140 150 a Lo Vo 1 of Im p o r t (1937 * 100) P r i c e 6 100 110 120 130 140 150 0 20 40 61 81 101 121 142 162 9 31 54 76 98 120 143 165 187 17 42 66 90 114 139 163 187 212 25 51 77 104 130 156 182 209 235 31 60 88 116 144 173 201 229 258 37 67 97 128 158 188 219 249 279 Billions of dollars. British balance-cf-payments problem simply by maintaining domestic prosperity. At a national income of 140 billion dollars (about 12 per cent below the current level), and at an import price level 30 per cent greater than in 1937, the increase in imports from Britain -would probably be sufficient to cover well over half the annual British deficit. These estimates, however, rest on two assumptions• It is assumed that the United States does not raise tariffs above present levels, and it is assumed that international trade is conducted on a multilateral basis* The indirect effect (the effect via third countries) of an increase in American imports depends, of course, on the second assumption, and it should be remembered that, quantitatively, the indirect effect appears to be considerably more important than the direct effect (ife», the increase in American imports from Britain). But both indirect sztd direct effects depend on the first assumption that the United States does not raise tariff barriers. APPENDIX THE BRITISH BALANCE-OF-PAYMENTS PROBLEM The average annual British balance of payments for the three-year period 1936-38 was as follows: BRITISH BALANCE , 1936-38 (Millions of Pounds) Credits Debits Importsa Government payments 950 7 95? Exportsa Investment income Shipping income Other income Defioit 563 203 105 43 43 957 a Includes silver bullion and specie* Exports include re-exports* How has the war changed this picture? It is estimated that income from foreign investments has fallen by about half* The post-war shipping situation is decidedly obscure, but there are reasons to hope that British shipping income will remain about the same as before the war* Let us assume that this is the casef and let us make the same assumption in regard to "other income*tf To estimate the deficit which, in the absence of borrowing, will have to bo met either by a reduction in imports or by en increase in exports, let us assume that the post-war volume of exports end of imports remains at the 1936*38 level* Even though volume remains unchanged, the value of exports and imports will, for a period at least, be higher as a result of higher pricey* At the present tin£, British wholesale food prices are more than 60 per cent higher than in 1938, and the same is true of British wholesale prices as a whole* Let us make the conservative assumption that both import and export prices in the immediate post-war period are 30 per cent higher than in 1936^38* (This assumption is favorable in that it implies no deterioration in British terms of trade,) One further adjustment remains to be made* Britain during the course of the war has incurred a huge external debt in the form of blocked sterling balances* The exact amount of thie debt is a matter of conjecture, but estimates (as of the end of 1944) run as high as 3 billion pounds** Lot -**Lord Keynes, in the New York Times for July 7, 1944, was quoted as s&ying that blocked sterling balances will amount to $12,000,000,000 (t3,000,000,000) by the end of this year* 11 us assume that, as a result of agreement, the sterling debt is retired at the rate of £100,000,000 a year* Our balance now appears as follows: ESTIMATED POST-KHAR BRITISH BALANCE OF PAYMENTS (Millions of Pounds) Debits Credits Imports 1,235 Government payments 113 Exports Investment income Shipping income Other income Deficit 732 100 105 43 368 1,348 On the above as* rtions, we derive an annual balance-of--payments deficit of ovor throo affd million poundsf In principle, this deficit co^ld be met by a reduction in imports, and it is possible that substantial curtailment of import? may occur, It should be pointed out, however, that the British demand for imports appears to be highly inelastic, both with respect to price end with respect to national income, ais the following table shows; INDEX OF PER CAPITA VOLUME OF BRITISH IMPORTS (1929 • 100) 1924 1925 1926 1927 1928 89 92 96 98 95 1929 1930 1931 1932 1933 100 97 100 87 88 1934 1935 1936 1937 1938 93 94 100 105 100 It will be seen that, despite wide variations in prices and in national income during the above years, the por capita volume of British imports varied but little* It is interesting to observe that per capita real imports wore the smno in such widoly divergent years as 1929, 1931, 1936, and 1938* Thus it would seem that, in the absonce of borrowing, the British deficit will have to be metf in the main, by a marked expansion of exports• u 8 fc- I UNITED STATES IMPORTS (Custaiis Duties Added) CHART I I 200 40 60 80 100 BfcXXUttl XHCQB (Billion* of DeOLUzv) 120 CHART III Billions of Dollars 5 * - - o / A f 2 _i 1 1 1 1 o ca L 8 taj Mr r\ n I M r M i^f 1 m\ t iy t A 'S r •1/ I f1 [ ( / r t 0 H / r M j i r J i L Mi 1 / Li r s I; i Ua A fw I T \ 1 % w T \j f / f 1 1 / s 2 i 1 20 _L 60 80 100 XaTIOBaL DC016 (Billions of Dollars) 120 CHART IV o I n - + / SI / +pU / A/ S 1 • ?! > 1 . 1 s +fcU \ 1 LL2 - |1 s ^^-^ CO I I \ V\ \ \ > © w It K 1 V / LJ • 1 L :3 K 3 PRICE pd i 2 V > V s o t 1 \s s > \ If \I ^ t \ w \\\ 1 I 1 \\ 1 11 1 \ 1 IN CO N CO N s ^ ^ 6 _J rS i r27 o -600 -I4OO -200 0 +200 *liO0 IMPORT VOLUME DEVIATIONS (Millions of Dollars) +600 _ CHART T 200 % \ \ \1 \ 180 x\ \M \\\\ - ps to 11 » 1 160- \\A \i in] - k H4O JHr 8 ••ill* 1 £120 }a 1 I Il\ Ji\ _^\ 100 \\ 1 \T^> \ \ \ \ \ H \ \ \ \| \ \ 80 60 \ \\ \ \ \ \| \ \ V \\ \\\ \ k M\ \\ I \ \ rs o 01 i 20 \ \ \ \ \ \ \/\ \\ \\ \ \ \\ 1 2 3 4 5 6 VOLUME OF IMPORTS (Value at 1937 Prices in Billions of Dollars)
@FedFRASER