The architects of the early postwar international monetary standard, the Bretton Woods system, believed that the failure of surplus countries to adjust was one of two major flaws in the interwar gold standard of the 1920s. The other was competitive devaluation, or beggar-thy-neighbor policies. The Bretton Woods Agreement established the International Monetary Fund as a public intermediary in the international monetary system. The fund’s key features were (1) an agreement to lend and borrow to adjust “temporary” imbalances in international payments and (2) a structural adjustment arrangement to correct “permanent” imbalances by changing exchange rates while preventing competitive devaluations.
Countries with a “temporary” current account deficit could use the fund to borrow from countries in surplus. This provision sought to avoid the problem that the United States and France created by failing to expand and inflate in response to gold inflows at the end of the 1920s. Their decisions forced deficit countries to contract without triggering an equilibrating expansion in the surplus countries. Under Bretton Woods rules, deficit countries did not have to contract. They could borrow the funds accumulated by the surplus countries.
The structural adjustment provisions permitted countries to correct persistent or permanent imbalances by adjusting exchange rates. A major problem with this provision was that central banks and governments could not distinguish temporary from persistent imbalances ex ante or even for some time after deficits appeared. A related problem was that fund rules did not make it clear what should happen when the principal reserve currency country—the United States—ran persistent trade or current account deficits.
Reliance on gold as a principal reserve asset of the fund and the member countries gave the appearance of a gold-based system. This appearance probably strengthened the belief that inflation would remain modest and thus contributed to the slow adjustment of inflationary anticipations in the 1960s. In practice the system was based mainly on the dollar, and there proved to be no binding restrictions on the supply of dollars under the Bretton Woods system.
5. I return to this discussion, and proposals for change, in volume 2.
The principal designers of the International Monetary Fund were John Maynard Keynes of Great Britain and Harry Dexter White of the United States. Keynes spent the war years, until his death in 1946, at the British Treasury. White was an economist at the United States Treasury. In contrast to the 1920s, when Governors Benjamin Strong and Montagu Norman were the principal architects of the postwar international monetary arrangements, power and influence over international monetary arrangements rested firmly in the two treasuries. Here, too, central banks had a subsidiary role.
At the New York bank, John H. Williams became one of the principal opponents, so he was kept from membership on the United States delegation. The Federal Reserve never formally considered the Bretton Woods Agreement and was not asked to do so. As the system developed, however, Williams’s proposal for an international system, based on the dollar, soon supplanted many of the features of the Keynes-White plan.
THE ADMINISTRATION’S WARTIME PROGRAM
There are both similarities and differences in the financing programs for the two world wars. Table 7.1 shows that interest rates remained lower and rose less in World War II, and the measured rate of inflation was lower also. Price controls distort the timing of price changes for the period. When controls were removed, in third quarter 1946, the deflator rose at a 45 percent annual rate, releasing most of the changes suppressed by wartime controls.
The first observation for each war is for the quarter in which the United States entered the war—second quarter 1917 and fourth quarter 1941. Second is the observation for the quarter in which the war ended—fourth quarter 1918 and third quarter 1945. Third is the observation for the postwar quarter in which wartime inflationary pressures began to recede, as measured by the rate of growth of the monetary base. Annualized rates of change for money and prices are computed from the first to the third date shown in the table.
Financing World War II was a much larger task. The cost of the war was substantially larger both absolutely and relative to GNP.6 Real GNP was approximately two and a half times greater in the later war, and the level of the deflator was similar in both periods, but government debt increased nearly ten times as much, as the table shows. The larger increase in debt occurred despite the larger share of taxes and faster growth of base money in World War II. Also, the Federal Reserve chose a different method of supplying reserves and supporting the Treasury market. In World War I, the Federal Reserve System did not have an open market policy. Banks obtained reserves by borrowing at the discount window using Treasury securities as collateral. In World War II, the System supplied reserves principally by open market purchases. Since the Federal Reserve supported a pattern of rates, it became the residual buyer. This left control of reserve changes to the banks’ decisions, much the same as in World War I.
6. Feinstein, Temin, and Toniolo (1997) put the cost of the two wars at 13 percent and 45 percent of United States GNP at the time. For Germany, they estimated the costs as 53 percent and 76 percent.
With long- and short-term interest rates comparatively lower in the 1940s, the demand for real money balances was higher. In World War I, base money, money, and prices rose at about the same rate, 10 to 12 percent. Real balances declined slightly. In World War II, base money and money rose at about the same rate (16 percent), but prices rose at less than half that rate, reflecting the rising demand for cash balances. The rise in real cash balances financed spending and inflation at the end of the war and therefore became a cause for concern.
Beginning in 1942, the government severely curtailed automobile production and took all residual production. Production of other durables was curtailed also; spending declined and saving increased. Part of the saving was held as money because higher mobility of the population increased the demand for currency (Cagan 1965).
Chart 7.2 shows the relation of base velocity to a long-term interest rate and highlights quarterly data from 1942 to first quarter 1951. The chart suggests that much of the quarterly movement in wartime and postwar velocity (the reciprocal of average cash balances) is consistent with the long-term relationship. Velocity was historically low, and average cash balances were correspondingly high, principally because long-term interest rates remained close to the 2.5 percent maximum.
Chart 7.3 looks at the war and postwar period on a finer scale. The positive relation remains, but the effect of the 2.5 percent interest rate ceiling is now visible. Observations at the ceiling rate, mainly in 1943 and 1944, suggest that the ceiling was binding in these years. Extrapolating from the linear relationship, the data suggest that without the ceiling, interest rates and velocity would have been higher and average cash balances correspondingly lower during part of the war years. For much of the period, however, the ceiling rate seems not to have affected money holding.7
The opposite side of the much larger rise in cash balances was the much smaller increase in the public’s share of the debt. Morgenthau’s Treasury urged individuals to purchase debt, but he was unwilling to pay them to do so. The Treasury issued series E war bonds at prices as low as $18.75 per bond and war savings stamps for as little as 10 cents, which could cumulate to a bond purchase. The Treasury encouraged corporations, schools, and other institutions to sell bonds and stamps through payroll deduction and appeals to patriotism. These actions were not enough to offset the low interest rates paid on the debt. The nonbank public acquired a smaller portion of the debt in World War II than in World War I. Commercial banks acquired 40 percent of debt held outside the government and the reserve banks. Although many citizens and corporations pledged to buy bonds during bond drives, they sold many of the bonds to banks after the bond drive ended.
7. Base velocity is computed as the ratio of GNP from Balke and Gordon 1986 to high-powered money from Anderson and Rasche 1999.
Secretary Morgenthau set three major objectives for war finance
(Blum 1967, 14–15). He wanted to finance 50 percent of the war by direct taxation, to finance most of the rest by voluntary purchases of bonds, and to maintain low interest rates. He believed that low interest rates would minimize the cost of the war. He succeeded in his third objective, came close to his first, and managed to avoid most of the pressures from Congress and other parts of the administration calling for compulsory bond purchases.8
For calendar years 1942–45, total government spending was $306 billion, revenues were $138 billion, and GNP was $740 billion. These periods correspond to the war years, with a few additional months of demobilization and reconversion to peacetime resource use at the end. Based on these data, tax collections were 45 percent of spending, only $15 billion short of Morgenthau’s goal.9
8. J. M. Keynes advocated a compulsory saving scheme as part of a British plan for war finance. Many of Keynes’s followers in the United States wanted to adopt Keynes’s program. Morgenthau opposed it, in part because the United States economy started the war with output far below capacity, in part because he was able to market the debt at historically low interest rates (Blum 1965, 297, 299). In World War I, the Treasury assigned a quota for banks’ purchases. The quota was a minimum subscription for each bank (Sproul Papers, Monetary Policy, 1940–41).
9. Eccles (1951, 381) includes the prewar defense spending in his calculation. For July 1940 to December 1945 he reports spending as $380 billion financed by $153 billion of taxes (40 percent), and $228 billion of borrowing and money creation. Nonbank investors acquired about $130 billion but sold some of their bonds to commercial banks after the bond drives. Elsewhere, Eccles gives contemporary data for June 1940 to June 1946, the period including prewar preparation. Total cost was $398 billion, 44 percent paid by taxation, 56 percent by borrowing and money creation (Board Minutes, November 26, 1947, 4).
Tax Policy
Morgenthau had little success getting Congress to approve his tax policy. Despite a Democratic majority in both houses, he did not fully meet his revenue goal or get his preferred tax policy. By 1944, relations between Congress and the administration became so strained that, with large majorities, both houses of Congress overrode the president’s veto of a tax bill for the first time in United States history. The administration did not try again to change tax rates during the war.
The main sources of conflict were the level of rates and the distribution of the tax burden. Many congressmen favored a sales tax. Morgenthau opposed on equity grounds; the sales tax would put more of the burden on low-income earners, a group he tried to shelter. At the opposite end of the income distribution, Roosevelt favored a limit of $25,000 on individual after-tax income, $50,000 for families. This proposal had so little appeal that Congress did not consider it seriously.
The 1943 tax bill made a lasting change in the tax system by introducing withholding at the source. Before 1943, taxpayers paid taxes in March on the previous year’s income. Withholding shifted most tax collection to the current year, a pay-as-you-go system for wage earners and some others. Withholding greatly simplified enforcement, as the number of taxpayers expanded to include 40 million to 50 million returns on incomes as low as $600 a year.10
Morgenthau at first opposed the withholding plan because Congress proposed to forgive all 1942 tax liabilities (due in March 1943) when withholding began. His main objection was that, with progressive taxation and high wartime rates, high-income taxpayers (and wartime profiteers) would benefit most. He was able to limit tax forgiveness and introduce some progressivity. The bill forgave $50 or 75 percent of the lower of 1942 or 1943 tax liabilities. Withholding began on July 1, 1943.
Morgenthau recognized inflation as a tax on households. He claimed he preferred direct taxation to inflation, but he would not allow interest rates to rise.11 However, he proposed some fiscal changes to reduce household income. One of his proposals would have raised the Social Security tax on labor income during the war, with the proceeds returned after the war, if needed, as unemployment compensation (Blum 1965, 313). Perhaps without fully recognizing the change, Morgenthau had become a proponent of countercyclical fiscal policy.
10. The proposal was advocated in 1941 by Beardsley Ruml, head of R. H. Macy’s department store and chairman of the New York Federal Reserve bank. Hence it was often referred to as the Ruml plan.
11. Using the average values of the monetary base and GNP for the war years, the government taxed nearly 10 percent of the base through inflation. The base was about 5 percent of GNP and 15 percent of government spending, so the inflation tax on the base is about 0.5 percent of GNP and 1.5 percent of government spending.
At the Federal Reserve Marriner Eccles saw the war as a major shift in demand that had to be met by substantial tax increases. In 1940–41 he agreed with the Keynesians who argued that, given the high unemployment at the start of the war, the country could increase both “guns and butter.” By 1942 he was concerned that the administration and Congress would be slow to recognize that the problem was no longer an excess supply of goods. There was an excess supply of money and excess demand for goods (Eccles 1951, 346–47).12
Debt Finance
Morgenthau foresaw that the war would require an unprecedented volume of borrowing. The Treasury and the Federal Reserve agreed on the desirability of ceiling rates of interest, high tax rates, and selling bonds mainly to the nonbank public. Morgenthau described relations with the Federal Reserve as “more harmonious during the war than they had ever been during the years of the New Deal” Blum (1967, 15). Board members shared this view.
Differences about substance remained, however (Board Minutes, April 9, 1942, 8). Eccles and some others preferred a mandated program—forced saving—to Morgenthau’s mainly voluntary bond purchase program. The Board offered proposals in each of the eight bond drives intended to increase sales to nonbanks, restrict speculation in bonds, and limit the role of banks to short maturities. The Treasury accepted few of these suggestions.
There were other differences about debt finance. Although the Treasury agreed on the aim of selling as many bonds as possible to nonbank investors during bond drives, it was less concerned than the Federal Reserve about whether the purchasers held the bonds after the drive. Getting the bonds sold at prevailing rates was its overriding interest.
Three main problems arose. First, with interest rates lower on short-term than on long-term debt, the Treasury faced an upward-sloping yield curve. Bank and nonbank holders sold shorter-term securities and reinvested in longer-term bonds. Second, as in World War I, the Treasury permitted a “borrow and buy” policy. To ensure that bond drives were successful, banks lent money to finance bond purchases at interest rates below the bonds’ yield. Many banks agreed to buy the bonds from their customers after the drive. Since the buyers could profit by buying the bonds, they oversubscribed the new issue. This gave the appearance of public subscription but depended on bank financing. Third, Treasury certificates with one year or less to maturity were troublesome throughout. The Treasury first offered certificates in 1942 at a yield of 0.8 percent. The rate was above the rate required by the market, so prices rose to a premium. As the certificates approached maturity, they sold at a premium over Treasury bills. Banks sold them to the Federal Reserve at a profit. The Federal Reserve tried repeatedly to get the yield reduced to 0.75 percent on new issues or to shorten the maturity and lower the rate, but the Treasury would not change (Minutes, FOMC, March 1, 1944, at 11:40, 1).
12. Eccles cites a conversation with Roosevelt in December 1940, just after Roosevelt had announced the lend-lease program to help Britain. Roosevelt understood that technically it made no difference to the economy whether we lent the British money or lent them goods. But he believed the public would favor lending goods but oppose lending money. “If we made a dollar loan to the British, it would seem to our people that we were giving the British money, of which we were short, instead of goods which were in surplus” (quoted in Eccles 1951, 348).
Officially, the Tr
easury opposed the borrow and buy policy. In practice, it did little to prevent it (Eccles 1951, 361). As a result, nonbank purchasers acquired $147 billion of government securities (including nonmarketable war bonds) but held only $93 billion. Corporations subscribed to about $60 billion in bond drives but increased their holdings only $19 billion.
Commercial banks financed bond purchases by selling Treasury bills and other low-yielding securities to the Federal Reserve. With bill rates pegged and ceiling rates set on all other Treasury securities, the banks moved to the higher end of the yield curve. To limit bank purchases of long-term debt, many of the bonds were made “bank restricted.” Small and medium-sized banks complained that mutual savings banks and savings and loans could buy the restricted bonds and thus were able to offer higher returns to savers. In 1944 the rules changed to permit commercial banks to purchase restricted securities during bond drives up to 10 percent of their savings deposits (Board Minutes, December 7, 1943, 2–4). Overall, bank purchases were limited to $10 billion during all bond drives. Bank holdings increased by $57 billion, however (Eccles 1951, 362).13
As in World War I, debt finance was much less successful than claimed after the war bond drives. The monetary base doubled in the four years ending fourth quarter 1945, an 18 percent compound average annual rate of increase. Purchases of Treasury securities account for almost all of the $18 billion increase in the base.14
A History of the Federal Reserve, Volume 1 Page 85
