A History of the Federal Reserve, Volume 2 (70 page)

The London gold market became the principal market for sales by gold producers, especially Soviet and South African sales. During the first half of 1960, the London price remained between $35.08 and $35.12 with little trading (Coombs, 1976, 48). That changed as the 1960 U.S. election approached. Rumors spread that, as president, Kennedy would follow Roosevelt in 1933–34 by devaluing the dollar. Purchases by central banks and others increased, raising the gold price. An opportunity opened for central banks to buy gold at the U.S. Treasury and resell it at a profit in the London market.

The Bank of England had a limited role as residual buyer or seller. Increased activity by other central banks as buyers left the Bank as residual supplier. It could replenish its gold holdings profitably by selling dollars to the U.S. Treasury at the official price, but it was reluctant to profit from the arbitrage. More worrisome was the market’s ability to circumvent U.S. policy of selling gold only to governments (ibid., 51)

The solution in December 1961 established a cooperative arrangement under which the principal European central banks shared part of the responsibility for the London gold market. Seven central banks agreed to commit half of the $270 million subscription to the gold pool. The United States was responsible for the other half.
¹⁵⁶
The Bank of England managed intervention, using its own gold. Participating central banks repaid the Bank at the end of each month in proportion to their participation. The aim was to keep the gold price below $35.20, the delivered price of gold purchased in New York and sold in London (ibid., 62). Participating central banks agreed to make gold purchases from the U.S. Treasury or other central banks, not from the market.

Although the agreement fixed shares in the reimbursement, it left the final distribution to the individual central banks. They retained their right to demand gold from the Treasury. Coombs (ibid., 63) wrote: “We hoped that such conversions would not take place immediately after the receipt of dollar proceeds. But it would be entirely within the option of the central bank whether to convert in one week, one month, several months, or not at all.” Some held additional dollars. Others converted to gold, so the United States paid for more than half but less than all the gold sold by the pool.

In March 1962, pool members agreed to share the proceeds of purchases as well as sales. The banks relinquished their right to compete
for gold. Thereafter, the principal producers, the Soviet Union and South Africa, faced a single buyer, the Bank of England, acting as agent for the group of central banks. The pool set a maximum price of $35.08 that it would pay. This price was the London price of New York gold after the president agreed to waive the 0.25 percentage point charge on direct sales (Dillon papers, Box 33, May 26, 1961)

156. The commitments (in millions) were: West Germany, $30; Britain, France, and Italy, $25 each; Switzerland, Netherlands, and Belgium, $10 each (Coombs, 1976, 62).

The gold pool distributed $650 million in gold to the Treasury in its first twenty-one months of operation. The main reason for the large purchases was a bad Soviet harvest in 1963–64 that required the Soviet Union to sell gold to buy grain abroad. The pool lasted until 1968. While it lasted, it prevented central banks from competing for gold, lowered the price paid to producers, and kept the gold price close to $35. It may have reduced demands on the U.S. gold stock by encouraging countries to hold additional dollars. Like many other measures, the pool may have delayed the breakdown of the Bretton Woods arrangements. It could not prevent it.

During the years that the pool operated, member countries sold gold worth (net) $2.5 billion on the London market. The U.S. share was $1.6 billion (Schwartz, 1987b, 342). During approximately the same period, 1961–67, the U.S. gold reserve declined more than $5 billion, the difference reflecting direct sales from the U.S. gold stock. In the same period, the industrial countries in G-10, especially France, added more than 100 million ounces, $3.5 billion, to their official gold reserves. The purchases by the G-10 equal 97 percent of U.S. sales outside the gold pool. These data suggest that, despite Coombs’ claims and efforts, the pool did not function as intended; the G-10 replaced most of their sales from U.S. stocks.

What
was
achieved?
By August 1971, when the gold window closed, Federal Reserve currency swap lines reached $11.7 billion, a long way from the modest experiment, limited to $500 million, only nine years earlier (Hetzel, 1996, 39). Coombs (1976) reports the details of these operations and gives a sense of his excitement and his belief that he prevented many crises. Coombs’ book fell short, however, on the analytic side; he did not explain what the Federal Reserve achieved by active intervention beyond possible initial effects. In this, he reflected the almost complete absence of an explanation in the 1961–62 discussions of what Martin and the FOMC expected to achieve by intervention. There are many statements recognizing that currency market intervention could not reduce the payments imbalance. At various times, Martin and his colleagues cited increased productivity, reduced government spending abroad, foreign aid, military assistance, and other real factors as the principal causes of the problem or the sources of potential improvement. And they recognized that currency market intervention was just another type of open market operation.
The only advantage was that the swap arrangement delayed a reduction in the gold stock and reduced dollars held by foreigners. The offset was an obligation to repay the borrowed foreign currency in ninety days, with one possible renewal (Later extended). It differed mainly because it gave a temporary stay to gold sales.

There is a glaring omission from the 1961–62 discussions. The terms “sterilized” and “unsterilized intervention” never appear. In practice, the Federal Reserve sterilized its intervention. It had shifted to an interest rate target some of the time and soon after, most of the time.
¹⁵⁷
The special manager withdrew reserves when he sold foreign exchange for dollars. This action increased the market interest rate (and reduced free reserves), inducing the domestic desk to purchase government securities, thereby supplying reserves and reducing the interest rate back to the target. The net effect was to change the mix of assets held by the Federal Reserve and the market. Purchases of foreign exchange shifted the mix in private portfolios also. The market faced smaller exchange risk and greater interest rate risk. Given the large size of the stocks of government securities and foreign assets and the relatively small size of foreign exchange operations, this effect was trivial. Whatever announcement effect it might have on impact, the conclusion of years of research is that there is no lasting effect of sterilized intervention.

Purchases of domestic securities increased reserves and reduced the United States’ interest rates relative to rates abroad. The difference in interest rates induced capital to move to the countries with higher relative rates. To effect the transaction, holders sold dollars, bought foreign exchange, and purchased foreign assets. The difference in interest rates just equaled the expected depreciation (or appreciation) of the exchange rate over the term of the securities. All of this worked in reverse following an open market sale of domestic securities.

Federal Reserve staff and officials believed that interest rate differences induced currency movements. They discussed this arbitrage operation many times. No one related interest arbitrage to the role of the special manager or to the expected effect of his operations. To the contrary, no one mentioned that he would not achieve much.

Discussions concentrated on the many legal issues, as we have seen. Concerns for independence from the Treasury surfaced occasionally, but no one recognized that sterilization removed this problem. Neither Ralph Young and the Board’s staff nor the staff at New York and the other banks
prepared an economic analysis. Except for George Mitchell, none of the FOMC members had worked as economists, so no one was disposed to ask how the operation would work and what it might achieve. As in the 1920s, efforts to get central bank cooperation replaced careful analysis of what cooperation could achieve. No one asked why it was desirable to delay foreign demand for gold by swaps with three- or six-month duration if the problem arose from continued military or foreign aid that exceeded the private sector’s surplus, as many at the FOMC believed.

157. A free reserve target was an imperfect substitute. The short-term interest rate moved randomly around the interest rate consistent with the free reserve target.

The payments deficit declined very little if at all during 1961 to 1963. Knipe (1965, 165) suggests that “the full size of the deficits was partially concealed during 1961, 1962, and 1963. This was done . . . by regarding prepayments to the U.S. on loans, prepayments to the U.S. on military exports, and the issuance by the U.S. Treasury of new-type securities to foreign nations, as export-type items.” Adjusted for these items, current account deficits were higher by $0.7 to $1.5 billion in each year. Official figures later reported the data with and without the adjustments.

At the Treasury, Robert Roosa worked actively to develop and negotiate lending and borrowing arrangements and the gold pool. The Treasury issued some “Roosa bonds” denominated in foreign currency to reduce dollars held abroad. The bonds had longer duration than swaps, but like swaps they were not a solution. The arrangements may have delayed adjustment. They could not prevent it. Delay was useful if conceived as part of a strategy that permitted lower inflation in the United States to adjust the real exchange rate. By late 1963, this adjustment began to occur. The payments balance fell to $0.5 billion, and the adjustment continued.

What
was
believed?
The problem the System faced in the 1960s was similar in some respects to those in the late 1920s after restoration of widespread convertibility into gold. In both periods, central banks or governments tried to maintain fixed nominal exchange rates by cooperating and used cooperation in a vain attempt to avoid adjustment of misaligned real exchange rates. In both periods, France cooperated reluctantly or not at all.
¹⁵⁸
And in both periods, the main weakness in the exchange rate system arose because any central bank holding a large stock of dollars could precipitate a crisis by demanding gold. The rules provided no means to adjust to such an attack.

While these flaws in the gold exchange system were common to the two periods, there were significant differences. The United States sterilized
gold inflows to avoi
d inflation in the 1920s. It lost gold because it created inflation in the 1950s and again after 1964. The policies of the 1960s aimed at creating cooperative arrangements in which other countries shared responsibility for maintaining the system. The flaw in any system of this kind was a failure to recognize that cooperative arrangements work in the long-term only if countries have common objectives or incentives to cooperate. The incentives were not absent, but they were often weak.
¹⁵⁹
At one point, President Kennedy threatened to take U.S. forces out of Europe to remind France and Germany of their dependence. Later, the Vietnam War muted interest in cooperation because the stock of dollars continued to increase, and the war was unpopular abroad. European governments became reluctant to finance it in any case but especially because financing the U.S. deficit promoted domestic inflation that was also unpopular.

158. In Meltzer (2003, 210), I suggest that the main problem in the 1920s was incompatible objectives, not failure to cooperate. Cooperation lessened the short-term problem but did nothing to resolve the longer-term problems—misaligned real exchange rates, (ibid., 178).

In principle the United States could have adjusted its real exchange rate either by deflating relative to its trading partners or by revaluing gold. The first option worked for a time by creating more inflation abroad than at home. Financing the higher budget deficits used to finance the Vietnam War and the Great Society and slower productivity growth narrowed the difference in inflation rates and reversed the improvement in the U.S. trade balance. Principal policymakers dismissed the second option, either devaluation or floating, as unworthy of serious consideration.

Treasury Undersecretary Robert Roosa set the basic policy early in the Kennedy administration. Roosa (1965, 27) believed not only that fluctuating rates were destabilizing but also that “public authorities then come under pressure to manipulate the rates, and usually do. This leads to competitive devaluation, and on to trade and exchange restrictions.” Further, floating exchange rate degenerate into “disorderly chaos if they do not have some fixed point of reference” (ibid., 27). This created a “sense of rubbery unreality concerning the validity of all prices” (ibid., 28).
¹⁶⁰

Under Bretton Woods rules, central banks agreed to keep exchange rates within one percent of their fixed parities. Roosa rejected wider bands to increase flexibility, for example during non-synchronized recessions or
expansions. These could become a type of “de facto devaluation” (ibid., 31). He regarded convertibility of dollars into a gold as a
“privilege”
that foreign governments retained because they accepted
“responsibility
. . . for the continued smooth functioning of the system” (ibid., 33; emphases in the original).

159. In a review of Coombs’ book, Jordan (1978, 416) noted its significant flaw: “After relating all his frustrations with uncooperative foreign officials and U.S. Treasury officials, Coombs does not reach the conclusion that . . . a system that was so dependent on the personalities of such a large number of individuals representing such a diverse array of interests was fundamentally flawed.”

160. Roosa (1965, 30) did not dismiss fluctuations in forward exchange rates. Instead, he welcomed those fluctuations as a market mechanism that substituted for fluctuations in reserve positions. These fluctuations were limited, however, by the spot rate. If the forward rate rose or fell too far from the spot market rate, it began to influence the spot rate to move toward its
upper or lower band.